The Variation ofAnimals and Plants

under DomesticationVolume II

Charles Darwin

Wor

k re

prod

uced

with

no

edito

rial

resp

onsi

bilit

y

Notice by Luarna Ediciones

This book is in the public domain becausethe copyrights have expired under Spanish law.

Luarna presents it here as a gift to its cus-tomers, while clarifying the following:

1) Because this edition has not been super-vised by our editorial deparment, wedisclaim responsibility for the fidelity ofits content.

2) Luarna has only adapted the work tomake it easily viewable on common six-inch readers.

3) To all effects, this book must not be con-sidered to have been published byLuarna.

www.luarna.com

VOLUME II.

CHAPTER 2.XIII.

INHERITANCE continued—REVERSION ORATAVISM.

DIFFERENT FORMS OF REVERSION. INPURE OR UNCROSSED BREEDS, AS IN PI-GEONS, FOWLS, HORNLESS CATTLE ANDSHEEP, IN CULTIVATED PLANTS. REVER-SION IN FERAL ANIMALS AND PLANTS.REVERSION IN CROSSED VARIETIES ANDSPECIES. REVERSION THROUGH BUD-PROPAGATION, AND BY SEGMENTS INTHE SAME FLOWER OR FRUIT. IN DIFFER-ENT PARTS OF THE BODY IN THE SAMEANIMAL. THE ACT OF CROSSING A DIRECTCAUSE OF REVERSION, VARIOUS CASESOF, WITH INSTINCTS. OTHER PROXIMATECAUSES OF REVERSION. LATENT CHAR-ACTERS. SECONDARY SEXUAL CHARAC-TERS. UNEQUAL DEVELOPMENT OF THE

TWO SIDES OF THE BODY. APPEARANCEWITH ADVANCING AGE OF CHARACTERSDERIVED FROM A CROSS. THE GERM,WITH ALL ITS LATENT CHARACTERS, AWONDERFUL OBJECT. MONSTROSITIES.PELORIC FLOWERS DUE IN SOME CASESTO REVERSION.

The great principle of inheritance to be dis-cussed in this chapter has been recognised byagriculturists and authors of various nations, asshown by the scientific term ATAVISM, de-rived from atavus, an ancestor; by the Englishterms of REVERSION, or THROWING-BACK;by the French PAS-EN-ARRIERE; and by theGerman RUCKSCHLAG, or RUCKSCHRITT.When the child resembles either grandparentmore closely than its immediate parents, ourattention is not much arrested, though in truththe fact is highly remarkable; but when thechild resembles some remote ancestor or somedistant member in a collateral line,—and in thelast case we must attribute this to the descent of

all the members from a common progenitor,—we feel a just degree of astonishment. Whenone parent alone displays some newly-acquiredand generally inheritable character, and theoffspring do not inherit it, the cause may lie inthe other parent having the power of prepotenttransmission. But when both parents are simi-larly characterised, and the child does not,whatever the cause may be, inherit the charac-ter in question, but resembles its grandparents,we have one of the simplest cases of reversion.We continually see another and even moresimple case of atavism, though not generallyincluded under this head, namely, when theson more closely resembles his maternal thanhis paternal grand-sire in some male attribute,as in any peculiarity in the beard of man, thehorns of the bull, the hackles or comb of thecock, or, as in certain diseases necessarily con-fined to the male sex; for as the mother cannotpossess or exhibit such male attributes, the

child must inherit them, through her blood,from his maternal grandsire.

The cases of reversion may be divided into twomain classes which, however, in some in-stances, blend into one another; namely, first,those occurring in a variety or race which hasnot been crossed, but has lost by variation somecharacter that it formerly possessed, and whichafterwards reappears. The second class in-cludes all cases in which an individual withsome distinguishable character, a race, or spe-cies, has at some former period been crossed,and a character derived from this cross, afterhaving disappeared during one or several gen-erations, suddenly reappears. A third class,differing only in the manner of reproduction,might be formed to include all cases of rever-sion effected by means of buds, and thereforeindependent of true or seminal generation.Perhaps even a fourth class might be instituted,to include reversions by segments in the same

individual flower or fruit, and in different partsof the body in the same individual animal as itgrows old. But the two first main classes will besufficient for our purpose.

REVERSION TO LOST CHARACTERS BYPURE OR UNCROSSED FORMS.

Striking instances of this first class of caseswere given in the sixth chapter, namely, of theoccasional reappearance, in variously-colouredbreeds of the pigeon, of blue birds with all themarks characteristic of the wild Columba livia.Similar cases were given in the case of the fowl.With the common ass, as the legs of the wildprogenitor are almost always striped, we mayfeel assured that the occasional appearance ofsuch stripes in the domestic animal is a case ofsimple reversion. But I shall be compelled torefer again to these cases, and therefore herepass them over.

The aboriginal species from which our domes-ticated cattle and sheep are descended, nodoubt possessed horns; but several hornlessbreeds are now well established. Yet in these—for instance, in Southdown sheep—"it is notunusual to find among the male lambs somewith small horns." The horns, which thus occa-sionally reappear in other polled breeds, either"grow to the full size," or are curiously attachedto the skin alone and hang "loosely down, ordrop off." (13/1. 'Youatt on Sheep' pages 20,234. The same fact of loose horns occasionallyappearing in hornless breeds has been ob-served in Germany; Bechstein 'Naturgesch.Deutschlands.' b. 1 s. 362.) The Galloways andSuffolk cattle have been hornless for the last100 or 150 years, but a horned calf, with thehorn often loosely attached, is occasionallyproduced. (13/2. 'Youatt on Cattle' pages 155,174.)

There is reason to believe that sheep in theirearly domesticated condition were "brown ordingy black;" but even in the time of Davidcertain flocks were spoken of as white as snow.During the classical period the sheep of Spainare described by several ancient authors as be-ing black, red, or tawny. (13/3. 'Youatt onSheep' 1838 pages 17, 145.) At the present day,notwithstanding the great care which is takento prevent it, particoloured lambs and someentirely black are occasionally, or even fre-quently, dropped by our most highly improvedand valued breeds, such as the Southdowns.Since the time of the famous Bakewell, duringthe last century, the Leicester sheep have beenbred with the most scrupulous care; yet occa-sionally grey-faced, or black-spotted, or whollyblack lambs appear. (13/4. I have been in-formed of this fact through the Rev. W.D. Foxon the excellent authority of Mr. Wilmot: seealso remarks on this subject in an article in the'Quarterly Review' 1849 page 395.) This occurs

still more frequently with the less improvedbreeds, such as the Norfolks. (13/5. Youattpages 19, 234.) As bearing on this tendency insheep to revert to dark colours, I may state(though in doing so I trench on the reversion ofcrossed breeds, and likewise on the subject ofprepotency) that the Rev. W.D. Fox was in-formed that seven white Southdown ewes wereput to a so-called Spanish ram, which had twosmall black spots on his sides, and they pro-duced thirteen lambs, all perfectly black. Mr.Fox believes that this ram belonged to a breedwhich he has himself kept, and which is alwaysspotted with black and white; and he finds thatLeicester sheep crossed by rams of this breedalways produce black lambs: he has gone onrecrossing these crossed sheep with pure whiteLeicesters during three successive generations,but always with the same result. Mr. Fox wasalso told by the friend from whom the spottedbreed was procured, that he likewise had goneon for six or seven generations crossing with

white sheep, but still black lambs were invaria-bly produced.

Similar facts could be given with respect totailless breeds of various animals. For instance,Mr. Hewitt (13/6. 'The Poultry Book' by Mr.Tegetmeier 1866 page 231.) states that chickensbred from some rumpless fowls, which werereckoned so good that they won a prize at anexhibition, "in a considerable number of in-stances were furnished with fully developedtail-feathers." On inquiry, the original breederof these fowls stated that, from the time whenhe had first kept them, they had often pro-duced fowls furnished with tails; but that theselatter would again reproduce rumpless chick-ens.

Analogous cases of reversion occur in the vege-table kingdom; thus "from seeds gathered fromthe finest cultivated varieties of Heartsease (Vi-ola tricolor), plants perfectly wild both in theirfoliage and their flowers are frequently pro-

duced;" (13/7. Loudon's 'Gardener's Mag.' vol-ume 10 1834 page 396: a nurseryman, withmuch experience on this subject, has likewiseassured me that this sometimes occurs.) but thereversion in this instance is not to a very an-cient period, for the best existing varieties ofthe heartsease are of comparatively modernorigin. With most of our cultivated vegetablesthere is some tendency to reversion to what isknown to be, or may be presumed to be, theiraboriginal state; and this would be more evi-dent if gardeners did not generally look overtheir beds of seedlings, and pull up the falseplants or "rogues" as they are called. It has al-ready been remarked, that some few seedlingapples and pears generally resemble, but ap-parently are not identical with, the wild treesfrom which they are descended. In our turnip(13/8. 'Gardener's Chronicle' 1855 page 777.)and carrot-beds a few plants often "break "—that is, flower too soon; and their roots are gen-erally hard and stringy, as in the parent-

species. By the aid of a little selection, carriedon during a few generations, most of our culti-vated plants could probably be brought back,without any great change in their conditions oflife, to a wild or nearly wild condition: Mr.Buckman has effected this with the parsnip(13/9. Ibid 1862 page 721.); and Mr. Hewett C.Watson, as he informs me, selected, duringthree generations, "the most diverging plants ofScotch kail, perhaps one of the least modifiedvarieties of the cabbage; and in the third gen-eration some of the plants came very close tothe forms now established in England aboutold castle-walls, and called indigenous."

REVERSION IN ANIMALS AND PLANTSWHICH HAVE RUN WILD.

In the cases hitherto considered, the revertinganimals and plants have not been exposed toany great or abrupt change in their conditionsof life which could have induced this tendency;

but it is very different with animals and plantswhich have become feral or run wild. It hasbeen repeatedly asserted in the most positivemanner by various authors, that feral animalsand plants invariably return to their primitivespecific type. It is curious on what little evi-dence this belief rests. Many of our domesti-cated animals could not subsist in a wild state;thus, the more highly improved breeds of thepigeon will not "field" or search for their ownfood. Sheep have never become feral, andwould be destroyed by almost every beast ofprey. (13/10. Mr. Boner speaks ('Chamois-hunting' 2nd edition 1860 page 92) of sheepoften running wild in the Bavarian Alps; but,on making further inquiries at my request, hefound that they are not able to establish them-selves; they generally perish from the frozensnow clinging to their wool, and they have lostthe skill necessary to pass over steep icy slopes.On one occasion two ewes survived the winter,but their lambs perished.) In several cases we

do not know the aboriginal parent- species, andcannot possibly tell whether or not there hasbeen any close degree of reversion. It is notknown in any instance what variety was firstturned out; several varieties have probably insome cases run wild, and their crossing alonewould tend to obliterate their proper character.Our domesticated animals and plants, whenthey run wild, must always be exposed to newconditions of life, for, as Mr. Wallace (13/11.See some excellent remarks on this subject byMr. Wallace 'Journal Proc. Linn. Soc.' 1858 vol-ume 3 page 60.) has well remarked, they haveto obtain their own food, and are exposed tocompetition with the native productions. Un-der these circumstances, if our domesticatedanimals did not undergo change of some kind,the result would be quite opposed to the con-clusions arrived at in this work. Nevertheless, Ido not doubt that the simple fact of animalsand plants becoming feral, does cause sometendency to reversion to the primitive state;

though this tendency has been much exagger-ated by some authors.

[I will briefly run through the recorded cases.With neither horses nor cattle is the primitivestock known; and it has been shown in formerchapters that they have assumed different col-ours in different countries. Thus the horseswhich have run wild in South America are gen-erally brownish-bay, and in the East dun-coloured; their heads have become larger andcoarser, and this may be due to reversion. Nocareful description has been given of the feralgoat. Dogs which have run wild in variouscountries have hardly anywhere assumed auniform character; but they are probably de-scended from several domestic races, and abo-riginally from several distinct species. Feralcats, both in Europe and La Plata, are regularlystriped; in some cases they have grown to anunusually large size, but do not differ from thedomestic animal in any other character. When

variously-coloured tame rabbits are turned outin Europe, they generally reacquire the colour-ing of the wild animal; there can be no doubtthat this does really occur, but we should re-member that oddly- coloured and conspicuousanimals would suffer much from beasts of preyand from being easily shot; this at least was theopinion of a gentleman who tried to stock hiswoods with a nearly white variety; if thus de-stroyed, they would be supplanted by, insteadof being transformed into, the common rabbit.We have seen that the feral rabbits of Jamaica,and especially of Porto Santo, have assumednew colours and other new characters. The bestknown case of reversion, and that on which thewidely spread belief in its universality appar-ently rests, is that of pigs. These animals haverun wild in the West Indies, South America,and the Falkland Islands, and have everywhereacquired the dark colour, the thick bristles, andgreat tusks of the wild boar; and the younghave reacquired longitudinal stripes. But even

in the case of the pig, Roulin describes the half-wild animals in different parts of South Amer-ica as differing in several respects. In Louisianathe pig (13/12. Dureau de la Malle 'ComptesRendus' tome 41 1855 page 807. From thestatements above given, the author concludesthat the wild pigs of Louisiana are not de-scended from the European Sus scrofa.) has runwild, and is said to differ a little in form, andmuch in colour, from the domestic animal, yetdoes not closely resemble the wild boar ofEurope. With pigeons and fowls (13/13. Capt.W. Allen, in his 'Expedition to the Niger' statesthat fowls have run wild on the island of An-nobon, and have become modified in form andvoice. The account is so meagre and vague thatit did not appear to me worth copying; but Inow find that Dureau de la Malle ('ComptesRendus' tome 41 1855 page 690) advances thisas a good instance of reversion to the primitivestock, and as confirmatory of a still more vaguestatement in classical times by Varro.), it is not

known what variety was first turned out, norwhat character the feral birds have assumed.The guinea-fowl in the West Indies, when feral,seems to vary more than in the domesticatedstate.

With respect to plants run wild, Dr. Hooker(13/14. 'Flora of Australia' 1859 Introductionpage 9.) has strongly insisted on what slightevidence the common belief in their reversionto a primitive state rests. Godron (13/15. 'Del'Espece' tome 2 pages 54, 58, 60.) describeswild turnips, carrots, and celery; but theseplants in their cultivated state hardly differfrom their wild prototypes, except in the succu-lency and enlargement of certain parts,— char-acters which would certainly be lost by plantsgrowing in poor soil and struggling with otherplants. No cultivated plant has run wild on soenormous a scale as the cardoon (Cynara car-dunculus) in La Plata. Every botanist who hasseen it growing there, in vast beds, as high as a

horse's back, has been struck with its peculiarappearance; but whether it differs in any im-portant point from the cultivated Spanish form,which is said not to be prickly like its Americandescendant, or whether it differs from the wildMediterranean species, which is said not to besocial (though this may be due merely to thenature of the conditions), I do not know.]

REVERSION TO CHARACTERS DERIVEDFROM A CROSS, IN THE CASE OF SUB-VARIETIES, RACES, AND SPECIES.

When an individual having some recognisablepeculiarity unites with another of the same sub-variety, not having the peculiarity in question,it often reappears in the descendants after aninterval of several generations. Every one musthave noticed, or heard from old people of chil-dren closely resembling in appearance or men-tal disposition, or in so small and complex acharacter as expression, one of their grandpar-

ents, or some more distant collateral relation.Very many anomalies of structure and diseases(13/16. Mr. Sedgwick gives many instances inthe 'British and Foreign Med.-Chirurg. Review'April and July 1863 pages 448, 188.) of whichinstances have been given in the last chapter,have come into a family from one parent, andhave reappeared in the progeny after passingover two or three generations. The followingcase has been communicated to me on goodauthority, and may, I believe, be fully trusted: apointer-bitch produced seven puppies; fourwere marked with blue and white, which is sounusual a colour with pointers that she wasthought to have played false with one of thegreyhounds, and the whole litter was con-demned; but the gamekeeper was permitted tosave one as a curiosity. Two years afterwards afriend of the owner saw the young dog, anddeclared that he was the image of his oldpointer-bitch Sappho, the only blue and whitepointer of pure descent which he had ever seen.

This led to close inquiry, and it was proved thathe was the great-great-grandson of Sappho; sothat, according to the common expression, hehad only 1/16th of her blood in his veins. I maygive one other instance, on the authority of Mr.R. Walker, a large cattle- breeder in Kincardine-shire. He bought a black bull, the son of a blackcow with white legs, white belly and part of thetail white; and in 1870 a calf the gr.-gr.-gr.-gr.-grandchild of this cow was born coloured inthe same very peculiar manner; all the inter-mediate offspring having been black. In thesecases there can hardly be a doubt that a charac-ter derived from a cross with an individual ofthe same variety reappeared after passing overthree generations in the one case, and five inthe other.

When two distinct races are crossed, it is noto-rious that the tendency in the offspring to re-vert to one or both parent-forms is strong, andendures for many generations. I have myself

seen the clearest evidence of this in crossedpigeons and with various plants. Mr. Sidney(13/17. In his edition of 'Youatt on the Pig' 1860page 27.) states that, in a litter of Essex pigs,two young ones appeared which were the im-age of the Berkshire boar that had been usedtwenty-eight years before in giving size andconstitution to the breed. I observed in thefarmyard at Betley Hall some fowls showing astrong likeness to the Malay breed, and wastold by Mr. Tollet that he had forty years beforecrossed his birds with Malays; and that, thoughhe had at first attempted to get rid of this strain,he had subsequently given up the attempt indespair, as the Malay character would reap-pear.

This strong tendency in crossed breeds to re-vert has given rise to endless discussions inhow many generations after a single cross, ei-ther with a distinct breed or merely with aninferior animal, the breed may be considered as

pure, and free from all danger of reversion. Noone supposes that less than three generationssuffices, and most breeders think that six, se-ven, or eight are necessary, and some go to stillgreater lengths. (13/18. Dr. P. Lucas, 'Hered.Nat.' tome 2 pages 314, 892: see a good practicalarticle on the subject in 'Gardener's Chronicle'1856 page 620. I could add a vast number ofreferences, but they would be superfluous.) Butneither in the case of a breed which has beencontaminated by a single cross, nor when, inthe attempt to form an intermediate breed, half-bred animals have been matched together dur-ing many generations, can any rule be laiddown how soon the tendency to reversion willbe obliterated. It depends on the difference inthe strength or prepotency of transmission inthe two parent-forms, on their actual amount ofdifference, and on the nature of the conditionsof life to which the crossed offspring are ex-posed. But we must be careful not to confoundthese cases of reversion to characters which

were gained by a cross, with those under thefirst class, in which characters originally com-mon to BOTH parents, but lost at some formerperiod, reappear; for such characters may recurafter an almost indefinite number of genera-tions.

The law of reversion is as powerful with hy-brids, when they are sufficiently fertile to breedtogether, or when they are repeatedly crossedwith either pure parent-form, as in the case ofmongrels. It is not necessary to give instances.With plants almost every one who has workedon this subject, from the time of Kolreuter tothe present day, has insisted on this tendency.Gartner has recorded some good instances; butno one has given more striking ones than Nau-din. (13/19. Kolreuter gives curious cases in his'Dritte Fortsetzung' 1766 ss. 53, 59; and in hiswell-known 'Memoirs on Lavatera and Jalapa.'Gartner 'Bastarderzeugung' ss. 437, 441, etc.Naudin in his "Recherches sur l'Hybridite"

'Nouvelles Archives du Museum' tome 1 page25.) The tendency differs in degree or strengthin different groups, and partly depends, as weshall presently see, on whether the parent-plants have been long cultivated. Although thetendency to reversion is extremely general withnearly all mongrels and hybrids, it cannot beconsidered as invariably characteristic of them;it may also be mastered by long-continued se-lection; but these subjects will more properly bediscussed in a future chapter on Crossing. Fromwhat we see of the power and scope of rever-sion, both in pure races, and when varieties orspecies are crossed, we may infer that charac-ters of almost every kind are capable of reap-pearing after having been lost for a great lengthof time. But it does not follow from this that ineach particular case certain characters will re-appear; for instance, this will not occur when arace is crossed with another endowed withprepotency of transmission. Sometimes thepower of reversion wholly fails, without our

being able to assign any cause for the failure:thus it has been stated that in a French familyin which 85 out of above 600 members, duringsix generations, had been subject to night-blindness, "there has not been a single exampleof this affection in the children of parents whowere themselves free from it." (13/20. Quotedby Mr. Sedgwick in 'Med.-Chirurg. Review'April 1861 page 485. Dr. H. Dobell in 'Med.-Chirurg. Transactions' volume 46 gives an ana-logous case in which, in a large family, fingerswith thickened joints were transmitted to sev-eral members during five generations; butwhen the blemish once disappeared it neverreappeared.)

REVERSION THROUGH BUD-PROPAGATION—PARTIAL REVERSION, BYSEGMENTS IN THE SAME FLOWER ORFRUIT, OR IN DIFFERENT PARTS OF THEBODY IN THE SAME INDIVIDUAL ANIMAL.

In the eleventh chapter many cases of reversionby buds, independently of seminal generation,were given—as when a leaf-bud on a varie-gated, a curled, or laciniated variety suddenlyreassumes its proper character; or as when aProvence-rose appears on a moss-rose, or apeach on a nectarine-tree. In some of these ca-ses only half the flower or fruit, or a smallersegment, or mere stripes, reassume their formercharacter; and here we have reversion by seg-ments. Vilmorin (13/21. Verlot 'Des Varietes'1865 page 63.) has also recorded several caseswith plants derived from seed, of flowers re-verting by stripes or blotches to their primitivecolours: he states that in all such cases a whiteor pale-coloured variety must first be formed,and, when this is propagated for a length oftime by seed, striped seedlings occasionallymake their appearance; and these can after-wards by care be multiplied by seed.

The stripes and segments just referred to arenot due, as far as is known, to reversion to cha-racters derived from a cross, but to characterslost by variation. These cases, however, asNaudin (13/22. 'Nouvelles Archives du Mu-seum' tome 1 page 25. Alex. Braun (in his 'Re-juvenescence' Ray Soc. 1853 page 315) appar-ently holds a similar opinion.) insists in his dis-cussion on disjunction of character, are closelyanalogous with those given in the eleventhchapter, in which crossed plants have beenknown to produce half-and- half or stripedflowers and fruit, or distinct kinds of flowerson the same root resembling the two parent-forms. Many piebald animals probably comeunder this same head. Such cases, as we shallsee in the chapter on Crossing, apparently re-sult from certain characters not readily blend-ing together, and, as a consequence of this in-capacity for fusion, the offspring either per-fectly resemble one of their two parents, or re-semble one parent in one part, and the other

parent in another part; or whilst young are in-termediate in character, but with advancing agerevert wholly or by segments to either parent-form, or to both. Thus, young trees of the Cyti-sus adami are intermediate in foliage and flow-ers between the two parent-forms; but whenolder the buds continually revert either par-tially or wholly to both forms. The cases givenin the eleventh chapter on the changes whichoccurred during growth in crossed plants ofTropaeolum, Cereus, Datura, and Lathyrus areall analogous. As, however, these plants arehybrids of the first generation, and as theirbuds after a time come to resemble their par-ents and not their grandparents, these cases donot at first appear to come under the law ofreversion in the ordinary sense of the word;nevertheless, as the change is effected througha succession of bud-generations on the sameplant, they may be thus included.

Analogous facts have been observed in the ani-mal kingdom, and are more remarkable, asthey occur in the same individual in the strict-est sense, and not as with plants through a suc-cession of bud-generations. With animals theact of reversion, if it can be so designated, doesnot pass over a true generation, but merelyover the early stages of growth in the same in-dividual. For instance, I crossed several whitehens with a black cock, and many of the chick-ens were, during the first year, perfectly white,but acquired during the second year blackfeathers; on the other hand, some of the chick-ens which were at first black, became duringthe second year piebald with white. A greatbreeder (13/23. Mr. Teebay in 'The PoultryBook' by Mr. Tegetmeier 1866 page 72.) says,that a Pencilled Brahma hen which has any ofthe blood of the Light Brahma in her, will "oc-casionally produce a pullet well pencilled dur-ing the first year, but she will most likely moultbrown on the shoulders and become quite

unlike her original colours in the second year."The same thing occurs with light Brahmas if ofimpure blood. I have observed exactly similarcases with the crossed offspring from differ-ently coloured pigeons. But here is a more re-markable fact: I crossed a turbit, which has afrill formed by the feathers being reversed onits breast, with a trumpeter; and one of theyoung pigeons thus raised at first showed not atrace of the frill, but, after moulting thrice, asmall yet unmistakably distinct frill appearedon its breast. According to Girou (13/24. Quo-ted by Hofacker 'Ueber die Eigenschaften' etc.s. 98.) calves produced from a red cow by ablack bull, or from a black cow by a red bull,are not rarely born red, and subsequently be-come black. I possess a dog, the daughter of awhite terrier by a fox- coloured bulldog; as apuppy she was quite white, but when about sixmonths old a black spot appeared on her nose,and brown spots on her ears. When a little ol-der she was badly wounded on the back, and

the hair which grew on the cicatrix was of abrown colour, apparently derived from herfather. This is the more remarkable, as withmost animals having coloured hair, that whichgrows on a wounded surface is white.

In the foregoing cases, the characters whichwith advancing age reappeared, were presentin the immediately preceding generations; butcharacters sometimes reappear in the samemanner after a much longer interval of time.Thus the calves of a hornless race of cattlewhich originated in Corrientes, though at firstquite hornless, as they become adult sometimesacquire small, crooked, and loose horns; andthese in succeeding years occasionally becomeattached to the skull. (13/25. Azara 'Essais Hist.Nat. de Paraguay' tome 2 1801 page 372.) Whiteand black Bantams, both of which generallybreed true, sometimes assume as they grow olda saffron or red plumage. For instance, a first-rate black bantam has been described, which

during three seasons was perfectly black, butthen annually became more and more red; andit deserves notice that this tendency to change,whenever it occurs in a bantam, "is almost cer-tain to prove hereditary." (13/26. These factsare given on the high authority of Mr. Hewitt in'The Poultry Book' by Mr. Tegetmeier 1866 pa-ge 248.) The cuckoo or blue-mottled Dorkingcock, when old, is liable to acquire yellow ororange hackles in place of his proper bluish-grey hackles. (13/27. 'The Poultry Book' by Te-getmeier 1866 page 97.) Now as Gallus bankivais coloured red and orange, and as Dorkingfowls and bantams are descended from thisspecies, we can hardly doubt that the changewhich occasionally occurs in the plumage ofthese birds as their age advances, results from atendency in the individual to revert to the pri-mitive type.

CROSSING AS A DIRECT CAUSE OF RE-VERSION.

It has long been notorious that hybrids andmongrels often revert to both or to one of theirparent-forms, after an interval of from two toseven or eight, or, according to some authori-ties, even a greater number of generations. Butthat the act of crossing in itself gives an impulsetowards reversion, as shown by the reappear-ance of long-lost characters, has never, I be-lieve, been hitherto proved. The proof lies incertain peculiarities, which do not characterisethe immediate parents, and therefore cannothave been derived from them, frequentlyappearing in the offspring of two breeds whencrossed, which peculiarities never appear, orappear with extreme rarity, in these samebreeds, as long as they are precluded fromcrossing. As this conclusion seems to me highlycurious and novel, I will give the evidence indetail.[My attention was first called to this subject,and I was led to make numerous experiments,by MM. Boitard and Corbie having stated that,

when they crossed certain breeds of pigeons,birds coloured like the wild C. livia, or thecommon dovecote—namely, slaty-blue, withdouble black wing-bars, sometimes chequeredwith black, white loins, the tail barred withblack, with the outer feathers edged with whi-te,—were almost invariably produced. Thebreeds which I crossed, and the remarkableresults attained, have been fully described inthe sixth chapter. I selected pigeons belongingto true and ancient breeds, which had not atrace of blue or any of the above specifiedmarks; but when crossed, and their mongrelsrecrossed, young birds were often produced,more or less plainly coloured slaty-blue, withsome or all of the proper characteristic marks. Imay recall to the reader's memory one case,namely, that of a pigeon, hardly distinguish-able from the wild Shetland species, the grand-child of a red-spot, white fantail, and two blackbarbs, from any of which, when purely-bred,the production of a pigeon coloured like the

wild C. livia would have been almost a prod-igy.

I was thus led to make the experiments, re-corded in the seventh chapter, on fowls. I se-lected long-established pure breeds, in whichthere was not a trace of red, yet in several of themongrels feathers of this colour appeared; andone magnificent bird, the offspring of a blackSpanish cock and white Silk hen, was colouredalmost exactly like the wild Gallus bankiva. Allwho know anything of the breeding of poultrywill admit that tens of thousands of pure Span-ish and of pure white Silk fowls might havebeen reared without the appearance of a redfeather. The fact, given on the authority of Mr.Tegetmeier, of the frequent appearance, inmongrel fowls, of pencilled or transversely-barred feathers, like those common to manygallinaceous birds, is likewise apparently a caseof reversion to a character formerly possessedby some ancient progenitor of the family. I owe

to the kindness of this excellent observer theopportunity of inspecting some neck-hacklesand tail-feathers from a hybrid between thecommon fowl and a very distinct species, theGallus varius; and these feathers are trans-versely striped in a conspicuous manner withdark metallic blue and grey, a character whichcould not have been derived from either im-mediate parent.

I have been informed by Mr. B.P. Brent, that hecrossed a white Aylesbury drake and a blackso-called Labrador duck, both of which are truebreeds, and he obtained a young drake closelylike the mallard (A. boschas). Of the musk-duck (Cairina moschata, Linn.) there are twosub-breeds, namely, white and slate-coloured;and these I am informed breed true, or nearlytrue. But the Rev. W.D. Fox tells me that, byputting a white drake to a slate-coloured duck,black birds, pied with white, like the wildmusk-duck, were always produced. I hear from

Mr. Blyth that hybrids from the canary andgold-finch almost always have streaked feath-ers on their backs; and this streaking must bederived from the original wild canary.

We have seen in the fourth chapter, that the so-called Himalayan rabbit, with its snow-whitebody, black ears, nose, tail, and feet, breedsperfectly true. This race is known to have beenformed by the union of two varieties of silver-grey rabbits. Now, when a Himalayan doe wascrossed by a sandy-coloured buck, a silver-greyrabbit was produced; and this is evidently acase of reversion to one of the parent varieties.The young of the Himalayan rabbit are bornsnow-white, and the dark marks do not appearuntil some time subsequently; but occasionallyyoung Himalayan rabbits are born of a lightsilver-grey, which colour soon disappears; sothat here we have a trace of reversion, duringan early period of life, to the parent varieties,independently of any recent cross.

In the third chapter it was shown that at anancient period some breeds of cattle in the wil-der parts of Britain were white with dark ears,and that the cattle now kept half wild in certainparks, and those which have run quite wild intwo distant parts of the world, are likewise thuscoloured. Now, an experienced breeder, Mr. J.Beasley, of Northamptonshire (13/28. 'Gar-dener's Chronicle and Agricultural Gazette'1866 page 528.), crossed some carefully selectedWest Highland cows with purely-bred short-horn bulls. The bulls were red, red and white,or dark roan; and the Highland cows were allof a red colour, inclining to a light or yellowshade. But a considerable number of the off-spring—and Mr. Beasley calls attention to thisas a remarkable fact—were white, or whitewith red ears. Bearing in mind that none of theparents were white, and that they were purely-bred animals, it is highly probable that here theoffspring reverted, in consequence of the cross,to the colour of some ancient and half-wild

parent-breed. The following case, perhaps, co-mes under the same head: cows in their naturalstate have their udders but little developed,and do not yield nearly so much milk as ourdomesticated animals. Now there is some rea-son to believe (13/29. Ibid 1860 page 343. I amglad to find that so experienced a breeder ofcattle as Mr. Willoughby Wood, 'Gardener'sChronicle' 1869 page 1216, admits my principleof a cross giving a tendency to reversion.) thatcross-bred animals between two kinds, both ofwhich are good milkers, such as Alderneys andShorthorns, often turn out worthless in thisrespect.

In the chapter on the Horse reasons were as-signed for believing that the primitive stockwas striped and dun-coloured; and details weregiven, showing that in all parts of the worldstripes of a dark colour frequently appear alongthe spine, across the legs, and on the shoulders,where they are occasionally double or treble,

and even sometimes on the face and body ofhorses of all breeds and of all colours. But thestripes appear most frequently on the variouskinds of duns. In foals they are sometimesplainly seen, and subsequently disappear. Thedun-colour and the stripes are strongly trans-mitted when a horse thus characterised is cros-sed with any other; but I was not able to provethat striped duns are generally produced fromthe crossing of two distinct breeds, neither ofwhich are duns, though this does sometimesoccur.

The legs of the ass are often striped, and thismay be considered as a reversion to the wildparent form, the Equus taeniopus of Abyssinia(13/30. Sclater in 'Proc. Zoolog. Soc.' 1862 page163.), which is generally thus striped. In thedomestic animal the stripes on the shoulder areoccasionally double, or forked at the extremity,as in certain zebrine species. There is reason tobelieve that the foal is more frequently striped

on the legs than the adult animal. As with thehorse, I have not acquired any distinct evidencethat the crossing of differently-coloured varie-ties of the ass brings out the stripes.

But now let us turn to the result of crossing thehorse and ass. Although mules are not nearlyso numerous in England as asses, I have seen amuch greater number with striped legs, andwith the stripes far more conspicuous than ineither parent-form. Such mules are generallylight-coloured, and might be called fallow-duns. The shoulder-stripe in one instance wasdeeply forked at the extremity, and in anotherinstance was double, though united in themiddle. Mr. Martin gives a figure of a Spanishmule with strong zebra-like marks on its legs(13/31. 'History of the Horse' page 212.), andremarks that mules are particularly liable to bethus striped on their legs. In South America,according to Roulin (13/32. 'Mem. presentespar divers Savans a l'Acad. Royale' tome 6 1835

page 338.), such stripes are more frequent andconspicuous in the mule than in the ass. In theUnited States, Mr. Gosse (13/33. 'Letters fromAlabama' 1859 page 280.), speaking of theseanimals, says, "that in a great number, perhapsin nine out of every ten, the legs are bandedwith transverse dark stripes."

Many years ago I saw in the Zoological Gar-dens a curious triple hybrid, from a bay mare,by a hybrid from a male ass and female zebra.This animal when old had hardly any stripes;but I was assured by the superintendent, thatwhen young it had shoulder-stripes, and faintstripes on its flanks and legs. I mention thiscase more especially as an instance of the stri-pes being much plainer during youth than inold age.

As the zebra has such a conspicuously stripedbody and legs, it might have been expected thatthe hybrids from this animal and the commonass would have had their legs in some degree

striped; but it appears from the figures given inDr. Gray's 'Knowsley Gleanings' and still moreplainly from that given by Geoffroy and F. Cu-vier (13/34. 'Hist. Nat. des Mammiferes' 1820tome 1), that the legs are much more conspicu-ously striped than the rest of the body; and thisfact is intelligible only on the belief that the assaids in giving, through the power of reversion,this character to its hybrid offspring.

The quagga is banded over the whole front partof its body like a zebra, but has no stripes on itslegs, or mere traces of them. But in the famoushybrid bred by Lord Morton (13/35. 'Phi-losoph. Transact.' 1821 page 20.) from a chest-nut, nearly purely-bred, Arabian mare, by amale quagga, the stripes were "more stronglydefined and darker than those on the legs of"the quagga." The mare was subsequently putto a black Arabian horse, and bore two colts,both of which, as formerly stated, were plainly

striped on the legs, and one of them likewisehad stripes on the neck and body.

The Equus indicus (13/36. Sclater in 'Proc. Zo-olog. Soc.' 1862 page 163: this species is theGhor-Khur of N.W. India, and has often beencalled the Hemionus of Pallas. See also Mr.Blyth's excellent paper in 'Journal of AsiaticSoc. of Bengal' volume 28 1860 page 229.) ischaracterised by a spinal stripe, without shoul-der or leg stripes; but traces of these latter stri-pes may occasionally be seen even in the adult(13/37. Another species of wild ass, the true E.hemionus or Kiang, which ordinarily has noshoulder-stripes, is said occasionally to havethem; and these, as with the horse and ass, aresometimes double: see Mr. Blyth in the paperjust quoted and in 'Indian Sporting Review'1856 page 320: and Col. Hamilton Smith in'Nat. Library, Horses' page 318; and 'Dict.Class. d'Hist. Nat.' tome 3 page 563.) and Colo-nel S. Poole, who has had ample opportunities

for observation, informs me that in the foal,when first born, the head and legs are oftenstriped, but the shoulder-stripe is not so dis-tinct as in the domestic ass; all these stripes,excepting that along the spine, soon disappear.Now a hybrid, raised at Knowsley (13/38. Fig-ured in the 'Gleanings from the Knowsley Me-nageries' by Dr. J.E. Gray.) from a female of thisspecies by a male domestic ass, had all four legstransversely and conspicuously striped, hadthree short stripes on each shoulder and hadeven some zebra-like stripes on its face! Dr.Gray informs me that he has seen a second hy-brid of the same parentage, similarly striped.

From these facts we see that the crossing of theseveral equine species tends in a marked man-ner to cause stripes to appear on various partsof the body, especially on the legs. As we donot know whether the parent-form of the genuswas striped, the appearance of the stripes canonly hypothetically be attributed to reversion.

But most persons, after considering the manyundoubted cases of variously coloured marksreappearing by reversion in my experiments oncrossed pigeons and fowls, will come to thesame conclusion with respect to the horse-genus; and if so, we must admit that the pro-genitor of the group was striped on the legs,shoulders, face, and probably over the wholebody, like a zebra.

Lastly, Professor Jaeger has given (13/39. 'Dar-win'sche Theorie und ihre Stellung zu Moralund Religion' page 85.) a good case with pigs.He crossed the Japanese or masked breed withthe common German breed, and the offspringwere intermediate in character. He then re-crossed one of these mongrels with the pureJapanese, and in the litter thus produced one ofthe young resembled in all its characters a wildpig; it had a long snout and upright ears, andwas striped on the back. It should be borne inmind that the young of the Japanese breed are

not striped, and that they have a short muzzleand ears remarkably dependent.]

A similar tendency to the recovery of long lostcharacters holds good even with the instincts ofcrossed animals. There are some breeds offowls which are called "everlasting layers," be-cause they have lost the instinct of incubation;and so rare is it for them to incubate that I haveseen notices published in works on poultry,when hens of such breeds have taken to sit.(13/40. Cases of both Spanish and Polish henssitting are given in the 'Poultry Chronicle' 1855volume 3 page 477.) Yet the aboriginal specieswas of course a good incubator; and with birdsin a state of nature hardly any instinct is sostrong as this. Now, so many cases have beenrecorded of the crossed offspring from two ra-ces, neither of which are incubators, becomingfirst-rate sitters, that the reappearance of thisinstinct must be attributed to reversion fromcrossing. One author goes so far as to say, "that

a cross between two non-sitting varieties al-most invariably produces a mongrel that be-comes broody, and sits with remarkable steadi-ness." (13/41. 'The Poultry Book' by Mr. Teget-meier 1866 pages 119, 163. The author, whoremarks on the two negatives ('Journ. of Hort.'1862 page 325), states that two broods wereraised from a Spanish cock and Silver-pencilledHamburgh hen, neither of which are incuba-tors, and no less than seven out of eight hens inthese two broods "showed a perfect obstinacyin sitting." The Rev. E.S. Dixon ('OrnamentalPoultry' 1848 page 200) says that chickens rea-red from a cross between Golden and BlackPolish fowls, are "good and steady birds to sit."Mr. B.P. Brent informs me that he raised somegood sitting hens by crossing Pencilled Ham-burgh and Polish breeds. A cross-bred birdfrom a Spanish non-incubating cock and Co-chin incubating hen is mentioned in the 'Poul-try Chronicle' volume 3 page 13, as an "exem-plary mother." On the other hand, an excep-

tional case is given in the 'Cottage Gardener'1860 page 388 of a hen raised from a Spanishcock and black Polish hen which did not incu-bate.) Another author, after giving a strikingexample, remarks that the fact can be explainedonly on the principle that "two negatives makea positive." It cannot, however, be maintainedthat hens produced from a cross between twonon-sitting breeds invariably recover their lostinstinct, any more than that crossed fowls orpigeons invariably recover the red or blueplumage of their prototypes. Thus I raised sev-eral chickens from a Polish hen by a Spanishcock,—breeds which do not incubate,—andnone of the young hens at first showed anytendency to sit; but one of them—the only onewhich was preserved—in the third year satwell on her eggs and reared a brood of chick-ens. So that here we have the reappearancewith advancing age of a primitive instinct, inthe same manner as we have seen that the redplumage of the Gallus bankiva is sometimes

reacquired both by crossed and purely-bredfowls of various kinds as they grow old.

The parents of all our domesticated animalswere of course aboriginally wild in disposition;and when a domesticated species is crossedwith a distinct species, whether this is a domes-ticated or only a tamed animal, the hybrids areoften wild to such a degree, that the fact is in-telligible only on the principle that the cross hascaused a partial return to a primitive disposi-tion. Thus, the Earl of Powis formerly importedsome thoroughly domesticated humped cattlefrom India, and crossed them with Englishbreeds, which belong to a distinct species; andhis agent remarked to me, without any ques-tion having been asked, how oddly wild thecross-bred animals were. The European wildboar and the Chinese domesticated pig are al-most certainly specifically distinct: Sir F. Dar-win crossed a sow of the latter breed with awild Alpine boar which had become extremely

tame, but the young, though having half-domesticated blood in their veins, were "ex-tremely wild in confinement, and would not eatswill like common English pigs." Captain Hut-ton, in India, crossed a tame goat with a wildone from the Himalaya, and he remarked to mehow surprisingly wild the offspring were. Mr.Hewitt, who has had great experience in cross-ing tame cock-pheasants with fowls belongingto five breeds, gives as the character of all "ex-traordinary wildness" (13/42. 'The PoultryBook' by Tegetmeier 1866 pages 165, 167.); but Ihave myself seen one exception to this rule. Mr.S. J. Salter (13/43. 'Natural History Review'1863 April page 277.) who raised a large num-ber of hybrids from a bantam-hen by Gallussonneratii, states that "all were exceedinglywild." Mr. Waterton (13/44. 'Essays on NaturalHistory' page 917.) bred some wild ducks fromeggs hatched under a common duck, and theyoung were allowed to cross freely bothamongst themselves and with the tame ducks;

they were "half wild and half tame; they cameto the windows to be fed, but still they had awariness about them quite remarkable."

On the other hand, mules from the horse andass are certainly not in the least wild, thoughnotorious for obstinacy and vice. Mr. Brent,who has crossed canary-birds with many kindsof finches, has not observed, as he informs me,that the hybrids were in any way remarkablywild: but Mr. Jenner Weir who has had stillgreater experience, is of a directly oppositeopinion. He remarks that the siskin is the tam-est of finches, but its mules are as wild, whenyoung, as newly caught birds, and are oftenlost through their continued efforts to escape.Hybrids are often raised between the commonand musk duck, and I have been assured bythree persons, who have kept these crossedbirds, that they were not wild; but Mr. Garnett(13/45. As stated by Mr. Orton in his 'Physiol-ogy of Breeding' page 12.) observed that his

hybrids were wild, and exhibited "migratorypropensities" of which there is not a vestige inthe common or musk duck. No case is knownof this latter bird having escaped and becomewild in Europe or Asia, except, according toPallas, on the Caspian Sea; and the commondomestic duck only occasionally becomes wildin districts where large lakes and fens abound.Nevertheless, a large number of cases havebeen recorded (13/46. M. E. de Selys-Longchamps refers ('Bulletin Acad. Roy. deBruxelles' tome 12 No. 10) to more than sevenof these hybrids shot in Switzerland and Fran-ce. M. Deby asserts ('Zoologist' volume 5 1845-46 page 1254) that several have been shot invarious parts of Belgium and Northern France.Audubon ('Ornitholog. Biography' volume 3page 168), speaking of these hybrids, says that,in North America, they "now and then wanderoff and become quite wild.") of hybrids fromthese two ducks having been shot in a com-pletely wild state, although so few are reared in

comparison with purely-bred birds of eitherspecies. It is improbable that any of these hy-brids could have acquired their wildness fromthe musk-duck having paired with a truly wildduck; and this is known not to be the case inNorth America; hence we must infer that theyhave reacquired, through reversion, their wild-ness, as well as renewed powers of flight.

These latter facts remind us of the statements,so frequently made by travellers in all parts ofthe world, on the degraded state and savagedisposition of crossed races of man. That manyexcellent and kind-hearted mulattos have ex-isted no one will dispute; and a more mild andgentle set of men could hardly be found thanthe inhabitants of the island of Chiloe, whoconsist of Indians commingled with Spaniardsin various proportions. On the other hand, ma-ny years ago, long before I had thought of thepresent subject, I was struck with the fact that,in South America, men of complicated descent

between Negroes, Indians, and Spaniards, sel-dom had, whatever the cause might be, a goodexpression. (13/47. 'Journal of Researches' 1845page 71.) Livingstone—and a more unimpeach-able authority cannot be quoted,—after speak-ing of a half-caste man on the Zambesi, de-scribed by the Portuguese as a rare monster ofinhumanity, remarks, "It is unaccountable whyhalf-castes, such as he, are so much more cruelthan the Portuguese, but such is undoubtedlythe case." An inhabitant remarked to Living-stone, "God made white men, and God madeblack men, but the Devil made halfcastes."(13/48. 'Expedition to the Zambesi' 1865 pages25, 150.) When two races, both low in the scale,are crossed the progeny seems to be eminentlybad. Thus the noble- hearted Humboldt, whofelt no prejudice against the inferior races,speaks in strong terms of the bad and savagedisposition of Zambos, or half-castes betweenIndians and Negroes; and this conclusion hasbeen arrived at by various observers. (13/49.

Dr. P. Broca on 'Hybridity in the Genus Homo'English translation 1864 page 39.) From thesefacts we may perhaps infer that the degradedstate of so many half-castes is in part due toreversion to a primitive and savage condition,induced by the act of crossing, even if mainlydue to the unfavourable moral conditions un-der which they are generally reared.

SUMMARY ON THE PROXIMATE CAU-SES LEADING TO REVERSION.

When purely-bred animals or plants reassumelong-lost characters,—when the common ass,for instance, is born with striped legs, when apure race of black or white pigeons throws aslaty-blue bird, or when a cultivated heartseasewith large and rounded flowers produces aseedling with small and elongated flowers,—we are quite unable to assign any proximatecause. When animals run wild, the tendency toreversion, which, though it has been greatly

exaggerated, no doubt exists, is sometimes to acertain extent intelligible. Thus, with feral pigs,exposure to the weather will probably favourthe growth of the bristles, as is known to be thecase with the hair of other domesticated ani-mals, and through correlation the tusks willtend to be redeveloped. But the reappearanceof coloured longitudinal stripes on young feralpigs cannot be attributed to the direct action ofexternal conditions. In this case, and in manyothers, we can only say that any change in thehabits of life apparently favour a tendency,inherent or latent in the species, to return to theprimitive state.

It will be shown in a future chapter that theposition of flowers on the summit of the axis,and the position of seeds within the capsule,sometimes determine a tendency towards re-version; and this apparently depends on theamount of sap or nutriment which the flower-buds and seeds receive. The position, also, of

buds, either on branches or on roots, sometimesdetermines, as was formerly shown, the trans-mission of the character proper to the variety,or its reversion to a former state.

We have seen in the last section that when tworaces or species are crossed there is the strong-est tendency to the reappearance in the off-spring of long- lost characters, possessed byneither parent nor immediate progenitor. Whentwo white, or red, or black pigeons, of well-established breeds, are united, the offspring arealmost sure to inherit the same colours; butwhen differently-coloured birds are crossed,the opposed forces of inheritance apparentlycounteract each other, and the tendency whichis inherent in both parents to produce slaty-blue offspring becomes predominant. So it is inseveral other cases. But when, for instance, theass is crossed with E. indicus or with the hor-se—animals which have not striped legs—andthe hybrids have conspicuous stripes on their

legs and even on their faces, all that can be saidis, that an inherent tendency to reversion isevolved through some disturbance in the or-ganisation caused by the act of crossing.

Another form of reversion is far commoner,indeed is almost universal with the offspringfrom a cross, namely, to the characters properto either pure parent-form. As a general rule,crossed offspring in the first generation arenearly intermediate between their parents, butthe grandchildren and succeeding generationscontinually revert, in a greater or lesser degree,to one or both of their progenitors. Several au-thors have maintained that hybrids and mon-grels include all the characters of both parents,not fused together, but merely mingled in dif-ferent proportions in different parts of thebody; or, as Naudin (13/50. 'Nouvelles Ar-chives du Museum' tome 1 page 151.) has ex-pressed it, a hybrid is a living mosaic-work, inwhich the eye cannot distinguish the discor-

dant elements, so completely are they inter-mingled. We can hardly doubt that, in a certainsense, this is true, as when we behold in a hy-brid the elements of both species segregatingthemselves into segments in the same flower orfruit, by a process of self-attraction or self-affinity; this segregation taking place either byseminal or bud-propagation. Naudin furtherbelieves that the segregation of the two specificelements or essences is eminently liable to oc-cur in the male and female reproductive matter;and he thus explains the almost universal ten-dency to reversion in successive hybrid genera-tions. For this would be the natural result of theunion of pollen and ovules, in both of whichthe elements of the same species had been seg-regated by self-affinity. If, on the other hand,pollen which included the elements of one spe-cies happened to unite with ovules includingthe elements of the other species, the interme-diate or hybrid state would still be retained,and there would be no reversion. But it would,

as I suspect, be more correct to say that the ele-ments of both parent-species exist in every hy-brid in a double state, namely, blended to-gether and completely separate. How this ispossible, and what the term specific essence orelement may be supposed to express, I shallattempt to show in the chapter on the hypothe-sis of pangenesis.

But Naudin's view, as propounded by him, isnot applicable to the reappearance of characterslost long ago by variation; and it is hardly ap-plicable to races or species which, after havingbeen crossed at some former period with a dis-tinct form, and having since lost all traces of thecross, nevertheless occasionally yield an indi-vidual which reverts (as in the case of the great-great-grandchild of the pointer Sappho) to thecrossing form. The most simple case of rever-sion, namely, of a hybrid or mongrel to itsgrandparents, is connected by an almost perfectseries with the extreme case of a purely-bred

race recovering characters which had been lostduring many ages; and we are thus led to inferthat all the cases must be related by some com-mon bond.

Gartner believed that only highly sterile hybridplants exhibit any tendency to reversion to theirparent-forms. This erroneous belief may per-haps be accounted for by the nature of the gen-era crossed by him, for he admits that the ten-dency differs in different genera. The statementis also directly contradicted by Naudin's obser-vations, and by the notorious fact that perfectlyfertile mongrels exhibit the tendency in a highdegree,—even in a higher degree, according toGartner himself, than hybrids. (13/51. 'Bas-tarderzeugung' s. 582, 438, etc.)

Gartner further states that reversions rarelyoccur with hybrid plants raised from specieswhich have not been cultivated, whilst, withthose which have been long cultivated, they areof frequent occurrence. This conclusion ex-

plains a curious discrepancy: Max Wichura(13/52. 'Die Bastardbefruchtung… der Weiden'1865 s. 23. For Gartner's remarks on this head,see 'Bastarderzeugung' s. 474, 582.) who wor-ked exclusively on willows which had not beensubjected to culture, never saw an instance ofreversion; and he goes so far as to suspect thatthe careful Gartner had not sufficiently pro-tected his hybrids from the pollen of the par-ent-species: Naudin, on the other hand, whochiefly experimented on cucurbitaceous andother cultivated plants, insists more strenu-ously than any other author on the tendency toreversion in all hybrids. The conclusion that thecondition of the parent-species, as affected byculture, is one of the proximate causes leadingto reversion, agrees well with the converse caseof domesticated animals and cultivated plantsbeing liable to reversion when they becomeferal; for in both cases the organisation or con-stitution must be disturbed, though in a verydifferent way. (13/53. Prof. Weismann in his

very curious essay on the different forms pro-duced by the same species of butterfly at differ-ent seasons ('Saison- Dimorphismus derSchmetterlinge' pages 27, 28), has come to asimilar conclusion, namely, that any causewhich disturbs the organisation, such as theexposure of the cocoons to heat or even tomuch shaking, gives a tendency to reversion.)

Finally, we have seen that characters often re-appear in purely-bred races without our beingable to assign any proximate cause; but whenthey become feral this is either indirectly ordirectly induced by the change in their condi-tions of life. With crossed breeds, the act ofcrossing in itself certainly leads to the recoveryof long-lost characters, as well as of those de-rived from either parent-form. Changed condi-tions, consequent on cultivation, and the rela-tive position of buds, flowers, and seeds on theplant, all apparently aid in giving this sametendency. Reversion may occur either through

seminal or bud generation, generally at birth,but sometimes only with an advance of age.Segments or portions of the individual mayalone be thus affected. That a being should beborn resembling in certain characters an ances-tor removed by two or three, and in some casesby hundreds or even thousands of generations,is assuredly a wonderful fact. In these cases thechild is commonly said to inherit such charac-ters directly from its grandparent, or more re-mote ancestors. But this view is hardly conceiv-able. If, however, we suppose that every char-acter is derived exclusively from the father ormother, but that many characters lie latent ordormant in both parents during a long succes-sion of generations, the foregoing facts are in-telligible. In what manner characters may beconceived to lie latent, will be considered in afuture chapter to which I have lately alluded.

LATENT CHARACTERS.

But I must explain what is meant by characterslying latent. The most obvious illustration isafforded by secondary sexual characters. Inevery female all the secondary male characters,and in every male all the secondary female cha-racters, apparently exist in a latent state, readyto be evolved under certain conditions. It iswell known that a large number of femalebirds, such as fowls, various pheasants, par-tridges, peahens, ducks, etc., when old or dis-eased, or when operated on, assume many orall of the secondary male characters of theirspecies. In the case of the hen-pheasant this hasbeen observed to occur far more frequentlyduring certain years than during others. (13/54.Yarrell 'Phil. Transact.' 1827 page 268; Dr. Ham-ilton in 'Proc. Zoolog. Soc.' 1862 page 23.) Aduck ten years old has been known to assumeboth the perfect winter and summer plumageof the drake. (13/55. 'Archiv. Skand. Beitragezur Naturgesch.' 8 s. 397-413.) Waterton (13/56.In his 'Essays on Nat. Hist.' 1838 Mr. Hewitt

gives analogous cases with hen- pheasants in'Journal of Horticulture' July 12, 1864 page 37.Isidore Geoffroy Saint-Hilaire in his 'Essais deZoolog. Gen.' ('Suites a Buffon' 1842 pages 496-513), has collected such cases in ten differentkinds of birds. It appears that Aristotle waswell aware of the change in mental dispositionin old hens. The case of the female deer acquir-ing horns is given at page 513.) gives a curiouscase of a hen which had ceased laying, and hadassumed the plumage, voice, spurs, and war-like disposition of the cock; when opposed toan enemy she would erect her hackles andshow fight. Thus every character, even to theinstinct and manner of fighting, must have laindormant in this hen as long as her ovaria con-tinued to act. The females of two kinds of deer,when old, have been known to acquire horns;and, as Hunter has remarked, we see some-thing of an analogous nature in the human spe-cies.

On the other hand, with male animals, it is no-torious that the secondary sexual characters aremore or less completely lost when they are sub-jected to castration. Thus, if the operation beperformed on a young cock, he never, as Yarrellstates, crows again; the comb, wattles, andspurs do not grow to their full size, and thehackles assume an intermediate appearancebetween true hackles and the feathers of thehen. Cases are recorded of confinement, whichoften affects the reproductive system, causinganalogous results. But characters properly con-fined to the female are likewise acquired by themale; the capon takes to sitting on eggs, andwill bring up chickens; and what is more curi-ous, the utterly sterile male hybrids from thepheasant and the fowl act in the same manner,"their delight being to watch when the hensleave their nests, and to take on themselves theoffice of a sitter." (13/57. 'Cottage Gardener'1860 page 379.) That admirable observerReaumur (13/58. 'Art de faire Eclore' etc. 1749

tome 2 page 8.) asserts that a cock, by beinglong confined in solitude and darkness, can betaught to take charge of young chickens; hethen utters a peculiar cry, and retains duringhis whole life this newly acquired maternalinstinct. The many well-ascertained cases ofvarious male mammals giving milk shows thattheir rudimentary mammary glands retain thiscapacity in a latent condition.

We thus see that in many, probably in all cases,the secondary characters of each sex lie dor-mant or latent in the opposite sex, ready to beevolved under peculiar circumstances. We canthus understand how, for instance, it is possiblefor a good milking cow to transmit her goodqualities through her male offspring to futuregenerations; for we may confidently believethat these qualities are present, though latent,in the males of each generation. So it is with thegame-cock, who can transmit his superiority incourage and vigour through his female to his

male offspring; and with man it is known(13/59. Sir H. Holland 'Medical Notes and Re-flections' 3rd edition 1855 page 31.) that dis-eases, such as hydrocele, necessarily confinedto the male sex, can be transmitted through thefemale to the grandson. Such cases as theseoffer, as was remarked at the commencement ofthis chapter, the simplest possible examples ofreversion; and they are intelligible on the beliefthat characters common to the grandparent andgrandchild of the same sex are present, thoughlatent, in the intermediate parent of the oppo-site sex.

The subject of latent characters is so important,as we shall see in a future chapter, that I willgive another illustration. Many animals havethe right and left sides of their body unequallydeveloped: this is well known to be the casewith flat-fish, in which the one side differs inthickness and colour and in the shape of thefins, from the other, and during the growth of

the young fish one eye is gradually twistedfrom the lower to the upper surface. (13/60. SeeSteenstrup on the 'Obliquity of Flounders' inAnnals and Mag. of Nat. Hist.' May 1865 page361. I have given an abstract of Malm's explana-tion of this wonderful phenomenon in the 'Ori-gin of Species' 6th Edition page 186.) In mostflat-fishes the left is the blind side, but in someit is the right; though in both cases reversed or"wrong fishes," are occasionally developed; andin Platessa flesus the right or left side is indif-ferently the upper one. With gasteropods orshell-fish, the right and left sides are extremelyunlike; the far greater number of species aredextral, with rare and occasional reversals ofdevelopment; and some few are normally sinis-tral; but certain species of Bulimus, and manyAchatinellae (13/61. Dr. E. von Martens in 'An-nals and Mag. of Nat. Hist.' March 1866 page209.) are as often sinistral as dextral. I will givean analogous case in the great articulate king-dom: the two sides of Verruca (13/62. Darwin

'Balanidae' Ray Soc. 1854 page 499: see also theappended remarks on the apparently capri-cious development of the thoracic limbs on theright and left sides in the higher crustaceans.)are so wonderfully unlike, that without carefuldissection it is extremely difficult to recognisethe corresponding parts on the opposite sidesof the body; yet it is apparently a mere matterof chance whether it be the right or the left sidethat undergoes so singular amount of change.One plant is known to me (13/63. Mormodesignea: Darwin 'Fertilisation of Orchids' 1862page 251.) in which the flower, according as itstands on the one or other side of the spike, isunequally developed. In all the foregoing casesthe two sides are perfectly symmetrical at anearly period of growth. Now, whenever a spe-cies is as liable to be unequally developed onthe one as on the other side, we may infer thatthe capacity for such development is present,though latent, in the undeveloped side. And asa reversal of development occasionally occurs

in animals of many kinds, this latent capacity isprobably very common.

The best yet simplest cases of characters lyingdormant are, perhaps, those previously given,in which chickens and young pigeons, raisedfrom a cross between differently colouredbirds, are at first of one colour, but in a year ortwo acquire feathers of the colour of the otherparent; for in this case the tendency to a changeof plumage is clearly latent in the young bird.So it is with hornless breeds of cattle, some ofwhich acquire small horns as they grow old.Purely bred black and white bantams, and so-me other fowls, occasionally assume, with ad-vancing years, the red feathers of the parent-species. I will here add a somewhat differentcase, as it connects in a striking manner latentcharacters of two classes. Mr. Hewitt (13/64.'Journal of Horticulture' July 1864 page 38. Ihave had the opportunity of examining theseremarkable feathers through the kindness of

Mr. Tegetmeier.) possessed an excellent Se-bright gold-laced bantam hen, which, as shebecame old, grew diseased in her ovaria, andassumed male characters. In this breed the ma-les resemble the females in all respects exceptin their combs, wattles, spurs, and instincts;hence it might have been expected that the dis-eased hen would have assumed only thosemasculine characters which are proper to thebreed, but she acquired, in addition, well-arched tail sickle-feathers quite a foot in length,saddle-feathers on the loins, and hackles on theneck,—ornaments which, as Mr. Hewitt re-marks, "would be held as abominable in thisbreed." The Sebright bantam is known (13/65.'The Poultry Book' by Mr. Tegetmeier 1866 pa-ge 241.) to have originated about the year 1800from a cross between a common bantam and aPolish fowl, recrossed by a hen-tailed bantam,and carefully selected; hence there can hardlybe a doubt that the sickle-feathers and hackleswhich appeared in the old hen were derived

from the Polish fowl or common bantam; andwe thus see that not only certain masculinecharacters proper to the Sebright bantam, butother masculine characters derived from thefirst progenitors of the breed, removed by aperiod of above sixty years, were lying latent inthis henbird, ready to be evolved as soon as herovaria became diseased.

From these several facts it must be admittedthat certain characters, capacities, and instincts,may lie latent in an individual, and even in asuccession of individuals, without our beingable to detect the least sign of their presence.When fowls, pigeons, or cattle of different col-ours are crossed, and their offspring changecolour as they grow old, or when the crossedturbit acquired the characteristic frill after itsthird moult, or when rarely-bred bantams par-tially assume the red plumage of their proto-type, we cannot doubt that these qualities werefrom the first present, though latent, in the in-

dividual animal, like the characters of a moth inthe caterpillar. Now, if these animals had pro-duced offspring before they had acquired withadvancing age their new characters, nothing ismore probable than that they would havetransmitted them to some of their offspring,who in this case would in appearance havereceived such characters from their grand- par-ents or more distant progenitors. We shouldthen have had a case of reversion, that is, of thereappearance in the child of an ancestral char-acter, actually present, though during youthcompletely latent, in the parent; and this wemay safely conclude is what occurs in all rever-sions to progenitors, however remote.

This view of the latency in each generation ofall the characters which appear through rever-sion, is also supported by their actual presencein some cases during early youth alone, or bytheir more frequent appearance and greaterdistinctness at this age than during maturity.

We have seen that this is often the case with thestripes on the legs and faces of the several spe-cies of the horse genus. The Himalayan rabbit,when crossed, sometimes produces offspringwhich revert to the parent silver-grey breed,and we have seen that in purely bred animalspale-grey fur occasionally reappears duringearly youth. Black cats, we may feel assured,would occasionally produce by reversion tab-bies; and on young black kittens, with a pedi-gree (13/66. Carl Vogt 'Lectures on Man' Eng-lish translation 1864 page 411.) known to havebeen long pure, faint traces of stripes may al-most always be seen which afterwards disap-pear. Hornless Suffolk cattle occasionally pro-duce by reversion horned animals; and Youatt(13/67. 'On Cattle' page 174.) asserts that evenin hornless individuals "the rudiment of a hornmay be often felt at an early age."

No doubt it appears at first sight in the highestdegree improbable that in every horse of every

generation there should be a latent capacityand tendency to produce stripes, though thesemay not appear once in a thousand genera-tions; that in every white, black, or other col-oured pigeon, which may have transmitted itsproper colour during centuries, there should bea latent capacity in the plumage to become blueand to be marked with certain characteristicbars; that in every child in a six-fingered familythere should be the capacity for the productionof an additional digit; and so in other cases.Nevertheless, there is no more inherent im-probability in this being the case than in a use-less and rudimentary organ, or even in only atendency to the production of a rudimentaryorgan, being inherited during millions of gen-erations, as is well known to occur with a mul-titude of organic beings. There is no more in-herent improbability in each domestic pig, dur-ing a thousand generations, retaining the ca-pacity and tendency to develop great tusks un-der fitting conditions, than in the young calf

having retained, for an indefinite number ofgenerations rudimentary incisor teeth, whichnever protrude through the gums.

I shall give at the end of the next chapter asummary of the three preceding chapters; butas isolated and striking cases of reversion havehere been chiefly insisted on, I wish to guardthe reader against supposing that reversion isdue to some rare or accidental combination ofcircumstances. When a character, lost duringhundreds of generations, suddenly reappears,no doubt some such combination must occur;but reversions, to the immediately precedinggenerations may be constantly observed, atleast, in the offspring of most unions. This hasbeen universally recognised in the case of hy-brids and mongrels, but it has been recognisedsimply from the difference between the unitedforms rendering the resemblance of the off-spring to their grandparents or more remoteprogenitors of easy detection. Reversion is li-

kewise almost invariably the rule, as Mr.Sedgwick has shown, with certain diseases.Hence we must conclude that a tendency to thispeculiar form of transmission is an integral partof the general law of inheritance.

MONSTROSITIES.

A large number of monstrous growths and oflesser anomalies are admitted by every one tobe due to an arrest of development, that is, tothe persistence of an embryonic condition. Butmany monstrosities cannot be thus explained;for parts of which no trace can be detected inthe embryo, but which occur in other membersof the same class of animals occasionally ap-pear, and these may probably with truth beattributed to reversion. As, however, I havetreated this subject as fully as I could in my'Descent of Man' (ch. 1 2nd edition), I will nothere recur to it.

[When flowers which have normally an irregu-lar structure become regular or peloric, thechange is generally looked at by botanists as areturn to the primitive state. But Dr. MaxwellMasters (13/68. 'Natural Hist. Review' April1863 page 258. See also his Lecture, Royal Insti-tution, March 16, 1860. On same subject seeMoquin-Tandon 'Elements de Teratologie' 1841pages 184, 352. Dr. Peyritsch has collected alarge number of very interesting cases 'Sitzb. d.k. Akad. d. Wissensch.' Wien b. 60 and espe-cially b. 66 1872 page 125.), who has ably dis-cussed this subject, remarks that when, for in-stance, all the sepals of a Tropaeolum becomegreen and of the same shape, instead of beingcoloured with one prolonged into a spur, orwhen all the petals of a Linaria become simpleand regular, such cases may be due merely toan arrest of development; for in these flowersall the organs during their earliest condition aresymmetrical, and, if arrested at this stage ofgrowth, they would not become irregular. If,

moreover, the arrest were to take place at a stillearlier period of development, the result wouldbe a simple tuft of green leaves; and no oneprobably would call this a case of reversion. Dr.Masters designates the cases first alluded to asregular peloria; and others, in which all thecorresponding parts assume a similar form ofirregularity, as when all the petals in a Linariabecome spurred, as irregular peloria. We haveno right to attribute these latter cases to rever-sion, until it can be shown that the parent-form,for instance, of the genus Linaria had had all itspetals spurred; for a chance of this naturemight result from the spreading of an anoma-lous structure, in accordance with the law, to bediscussed in a future chapter, of homologousparts tending to vary in the same manner. Butas both forms of peloria frequently occur on thesame individual plant of the Linaria (13/69.Verlot 'Des Varietes' 1865 page 89; Naudin'Nouvelles Archives du Museum' tome 1 page137.), they probably stand in some close rela-

tion to one another. On the doctrine that peloriais simply the result of an arrest of development,it is difficult to understand how an organ ar-rested at a very early period of growth shouldacquire its full functional perfection;—how apetal, supposed to be thus arrested, should ac-quire its brilliant colours, and serve as an enve-lope to the flower, or a stamen produce efficientpollen; yet this occurs with many peloric flow-ers. That pelorism is not due to mere chancevariability, but either to an arrest of develop-ment or to reversion, we may infer from anobservation made by Ch. Morren (13/70. In hisdiscussion on some curious peloric Calceolariasquoted in 'Journal of Horticulture' February 24,1863 page 152.) namely, that families whichhave irregular flowers often "return by thesemonstrous growths to their regular form;whilst we never see a regular flower realise thestructure of an irregular one."

Some flowers have almost certainly becomemore or less completely peloric through rever-sion, as the following interesting case shows.Corydalis tuberosa properly has one of its twonectaries colourless, destitute of nectar, onlyhalf the size of the other, and therefore, to acertain extent, in a rudimentary state; the pistilis curved towards the perfect nectary, and thehood, formed of the inner petals, slips off thepistil and stamen in one direction alone, so that,when a bee sucks the perfect nectary, the stig-ma and stamens are exposed and rubbedagainst the insect's body. In several closely al-lied genera, as in Dielytra, etc., there are twoperfect nectaries, the pistil is straight, and thehood slips off on either side, according as thebee sucks either nectary. Now, I have examinedseveral flowers of Corydalis tuberosa, in whichboth nectaries were equally developed and con-tained nectar; in this we see only the redevel-opment of a partially aborted organ; but withthis redevelopment the pistil becomes straight,

and the hood slips off in either direction, sothat these flowers have acquired the perfectstructure, so well adapted for insect agency, ofDielytra and its allies. We cannot attribute thesecoadapted modifications to chance, or to corre-lated variability; we must attribute them toreversion to a primordial condition of the spe-cies.

The peloric flowers of Pelargonium have theirfive petals in all respects alike, and there is nonectary so that they resemble the symmetricalflowers of the closely allied genus Geranium;but the alternate stamens are also sometimesdestitute of anthers, the shortened filamentsbeing left as rudiments, and in this respect theyresemble the symmetrical flowers of the closelyallied genus Erodium. Hence we may look atthe peloric flowers of Pelargonium as havingreverted to the state of some primordial form,the progenitor of the three closely related gen-era of Pelargonium, Geranium, and Erodium.

In the peloric form of Antirrhinum majus, ap-propriately called the" Wonder," the tubularand elongated flowers differ wonderfully fromthose of the common snapdragon; the calyxand the mouth of the corolla consist of six equallobes, and include six equal instead of four un-equal stamens. One of the two additional sta-mens is manifestly formed by the developmentof a microscopically minute papilla, which maybe found at the base of the upper lip of theflower of the common snapdragons in the nine-teen plants examined by me. That this papilla isa rudiment of a stamen was well shown by itsvarious degrees of development in crossedplants between the common and the peloricAntirrhinum. Again, a peloric Galeobdolonluteum, growing in my garden, had five equalpetals, all striped like the ordinary lower lip,and included five equal instead of four unequalstamens; but Mr. R. Keeley, who sent me thisplant, informs me that the flowers vary greatly,having from four to six lobes to the corolla, and

from three to six stamens. (13/71. For othercases of six divisions in peloric flowers of theLabiatae and Scrophulariaceae see Moquin-Tandon 'Teratologie' page 192.) Now, as themembers of the two great families to which theAntirrhinum and Galeobdolon belong areproperly pentamerous, with some of the partsconfluent and others suppressed, we ought notto look at the sixth stamen and the sixth lobe tothe corolla in either case as due to reversion,any more than the additional petals in doubleflowers in these same two families. But the caseis different with the fifth stamen in the peloricAntirrhinum, which is produced by the rede-velopment of a rudiment always present, andwhich probably reveals to us the state of theflower, as far as the stamens are concerned, atsome ancient epoch. It is also difficult to believethat the other four stamens and the petals, afteran arrest of development at a very early em-bryonic age, would have come to full perfectionin colour, structure, and function, unless these

organs had at some former period normallypassed through a similar course of growth.Hence it appears to me probable that the pro-genitor of the genus Antirrhinum must at someremote epoch have included five stamens andborne flowers in some degree resembling thosenow produced by the peloric form. The conclu-sion that peloria is not a mere monstrosity, irre-spective of any former state of the species, issupported by the fact that this structure is oftenstrongly inherited, as in the case of the peloricAntirrhinum and Gloxinia and sometimes inthat of the peloric Corydalis solida. (13/72.Godron reprinted from the 'Memoires del'Acad. de Stanislas' 1868.)

Lastly I may add that many instances havebeen recorded of flowers, not generally consid-ered as peloric, in which certain organs are ab-normally augmented in number. As an increaseof parts cannot be looked at as an arrest of de-velopment, nor as due to the redevelopment of

rudiments, for no rudiments are present, and asthese additional parts bring the plant intocloser relationship with its natural allies, theyought probably to be viewed as reversions to aprimordial condition.]

These several facts show us in an interestingmanner how intimately certain abnormal statesare connected together; namely, arrests of de-velopment causing parts to become rudimen-tary or to be wholly suppressed,—the redevel-opment of parts now in a more or less rudimen-tary condition,—the reappearance of organs ofwhich not a vestige can be detected,—and tothese may be added, in the case of animals, thepresence during youth, and subsequent disap-pearance, of certain characters which occasion-ally are retained throughout life. Some natural-ists look at all such abnormal structures as areturn to the ideal state of the group to whichthe affected being belongs; but it is difficult toconceive what is meant to be conveyed by this

expression. Other naturalists maintain, withgreater probability and distinctness of view,that the common bond of connection betweenthe several foregoing cases is an actual, thoughpartial, return to the structure of the ancientprogenitor of the group. If this view be correct,we must believe that a vast number of charac-ters, capable of evolution, lie hidden in everyorganic being. But it would be a mistake tosuppose that the number is equally great in allbeings. We know, for instance, that plants ofmany orders occasionally become peloric; butmany more cases have been observed in theLabiatae and Scrophulariaceae than in any ot-her order; and in one genus of the Scrophulari-aceae, namely Linaria, no less than thirteenspecies have been described in this condition(13/73. Moquin- Tandon 'Teratologie' page186.) On this view of the nature of peloric flow-ers, and bearing in mind certain monstrositiesin the animal kingdom, we must conclude thatthe progenitors of most plants and animals

have left an impression, capable of redevelop-ment, on the germs of their descendants, al-though these have since been profoundly modi-fied.

The fertilised germ of one of the higher ani-mals, subjected as it is to so vast a series ofchanges from the germinal cell to old age,—incessantly agitated by what Quatrefages wellcalls the tourbillon vital,—is perhaps the mostwonderful object in nature. It is probable thathardly a change of any kind affects either par-ent, without some mark being left on the germ.But on the doctrine of reversion, as given in thischapter, the germ becomes a far more marvel-lous object, for, besides the visible changeswhich it undergoes, we must believe that it iscrowded with invisible characters, proper toboth sexes, to both the right and left side of thebody, and to a long line of male and femaleancestors separated by hundreds or even thou-sands of generations from the present time: and

these characters, like those written on paperwith invisible ink, lie ready to be evolved whe-never the organisation is disturbed by certainknown or unknown conditions.

CHAPTER 2.XIV.INHERITANCE continued.—FIXEDNESS OFCHARACTER—PREPOTENCY—SEXUALLIMITATION—CORRESPONDENCE OF AGE.

FIXEDNESS OF CHARACTER APPAR-ENTLY NOT DUE TO ANTIQUITY OF IN-HERITANCE. PREPOTENCY OF TRANSMIS-SION IN INDIVIDUALS OF THE SAME FAM-ILY, IN CROSSED BREEDS AND SPECIES;OFTEN STRONGER IN ONE SEX THAN THEOTHER; SOMETIMES DUE TO THE SAMECHARACTER BEING PRESENT AND VISI-BLE IN ONE BREED AND LATENT IN THEOTHER. INHERITANCE AS LIMITED BY SEX.

NEWLY-ACQUIRED CHARACTERS IN OURDOMESTICATED ANIMALS OFTENTRANSMITTED BY ONE SEX ALONE,SOMETIMES LOST BY ONE SEX ALONE.INHERITANCE AT CORRESPONDING PE-RIODS OF LIFE. THE IMPORTANCE OF THEPRINCIPLE WITH RESPECT TO EMBRYOL-OGY; AS EXHIBITED IN DOMESTICATEDANIMALS: AS EXHIBITED IN THE APPEAR-ANCE AND DISAPPEARANCE OF INHER-ITED DISEASES; SOMETIMES SUPERVEN-ING EARLIER IN THE CHILD THAN IN THEPARENT. SUMMARY OF THE THREE PRE-CEDING CHAPTERS.

In the last two chapters the nature and force ofInheritance, the circumstances which interferewith its power, and the tendency to Reversion,with its many remarkable contingencies, werediscussed. In the present chapter some otherrelated phenomena will be treated of, as fullyas my materials permit.

FIXEDNESS OF CHARACTER.

It is a general belief amongst breeders that thelonger any character has been transmitted by abreed, the more fully it will continue to betransmitted. I do not wish to dispute the truthof the proposition that inheritance gainsstrength simply through long continuance, butI doubt whether it can be proved. In one sensethe proposition is little better than a truism; ifany character has remained constant duringmany generations, it will be likely to continueso, if the conditions of life remain the same. So,again, in improving a breed, if care be taken fora length of time to exclude all inferior indi-viduals, the breed will obviously tend to be-come truer, as it will not have been crossedduring many generations by an inferior animal.We have previously seen, but without beingable to assign any cause, that, when a newcharacter appears, it is occasionally from thefirst constant, or fluctuates much, or whollyfails to be transmitted. So it is with the aggre-

gate of slight differences which characterise anew variety, for some propagate their kindfrom the first much truer than others. Evenwith plants multiplied by bulbs, layers, etc.,which may in one sense be said to form parts ofthe same individual, it is well known that cer-tain varieties retain and transmit through suc-cessive bud-generations their newly-acquiredcharacters more truly than others. In none ofthese, nor in the following cases, does thereappear to be any relation between the forcewith which a character is transmitted and thelength of time during which it has been trans-mitted. Some varieties, such as white and yel-low hyacinths and white sweet-peas, transmittheir colours more faithfully than do the varie-ties which have retained their natural colour. Inthe Irish family, mentioned in the twelfth chap-ter, the peculiar tortoiseshell-like colouring ofthe eyes was transmitted far more faithfullythan any ordinary colour. Ancon andMauchamp sheep and niata cattle, which are all

comparatively modern breeds, exhibit re-markably strong powers of inheritance. Manysimilar cases could be adduced.

As all domesticated animals and cultivatedplants have varied, and yet are descended fromaboriginally wild forms, which no doubt hadretained the same character from an immenselyremote epoch, we see that scarcely any degreeof antiquity ensures a character being transmit-ted perfectly true. In this case, however, it maybe said that changed conditions of life inducecertain modifications, and not that the power ofinheritance fails; but in every case of failure,some cause, either internal or external, mustinterfere. It will generally be found that theorgans or parts which in our domesticated pro-ductions have varied, or which still continue tovary,—that is, which fail to retain their formerstate,—are the same with the parts which differin the natural species of the same genus. As, onthe theory of descent with modification, the

species of the same genus have been modifiedsince they branched off from a common pro-genitor, it follows that the characters by whichthey differ from one another have varied,whilst other parts of the organisation have re-mained unchanged; and it might be argued thatthese same characters now vary under domes-tication, or fail to be inherited, from their lesserantiquity. But variation in a state of natureseems to stand in some close relation withchanged conditions of life, and characterswhich have already varied under such condi-tions would be apt to vary under the still grea-ter changes consequent on domestication, in-dependently of their greater or less antiquity.

Fixedness of character, or the strength of inheri-tance, has often been judged of by the prepon-derance of certain characters in the crossed off-spring between distinct races; but prepotencyof transmission here comes into play, and this,as we shall immediately see, is a very different

consideration from the strength or weakness ofinheritance. (14/1. See 'Youatt on Cattle' pages92, 69, 78, 88, 163; and 'Youatt on Sheep' page325. Also Dr. Lucas 'L'Hered. Nat.' tome 2 page310.) It has often been observed that breeds ofanimals inhabiting wild and mountainouscountries cannot be permanently modified byour improved breeds; and as these latter are ofmodern origin, it has been thought that thegreater antiquity of the wilder breeds has beenthe cause of their resistance to improvement bycrossing; but it is more probably due to theirstructure and constitution being better adaptedto the surrounding conditions. When plants arefirst subjected to culture, it has been found that,during several generations, they transmit theircharacters truly, that is, do not vary, and thishas been attributed to ancient characters beingstrongly inherited: but it may with equal orgreater probability be consequent on changedconditions of life requiring a long time for theircumulative action. Notwithstanding these con-

siderations, it would perhaps be rash to denythat characters become more strongly fixed thelonger they are transmitted; but I believe thatthe proposition resolves itself into this,—thatcharacters of all kinds, whether new or old,tend to be inherited, and that those which havealready withstood all counteracting influencesand been truly transmitted, will, as a generalrule, continue to withstand them, and conse-quently be faithfully inherited.

PREPOTENCY IN THE TRANSMISSIONOF CHARACTER.

When individuals, belonging to the same fam-ily, but distinct enough to be recognised, orwhen two well-marked races, or two species,are crossed, the usual result, as stated in theprevious chapter, is, that the offspring in thefirst generation are intermediate between theirparents, or resemble one parent in one part andthe other parent in another part. But this is by

no means the invariable rule; for in many casesit is found that certain individuals, races, andspecies, are prepotent in transmitting their li-keness. This subject has been ably discussed byProsper Lucas (14/2. 'Hered. Nat.' tome 2 pages112-120.), but is rendered extremely complex bythe prepotency sometimes running equally inboth sexes, and sometimes more strongly inone sex than in the other; it is likewise compli-cated by the presence of secondary sexual char-acters, which render the comparison of crossedbreeds with their parents difficult.

It would appear that in certain families someone ancestor, and after him others in the samefamily, have had great power in transmittingtheir likeness through the male line; for we can-not otherwise understand how the same fea-tures should so often be transmitted after mar-riages with many females, as in the case of theAustrian Emperors; and so it was, according toNiebuhr, with the mental qualities of certain

Roman families. (14/3. Sir H. Holland 'Chap-ters on Mental Physiology' 1852 page 234.) Thefamous bull Favourite is believed (14/4. 'Gar-dener's Chronicle' 1860 page 270.) to have had aprepotent influence on the shorthorn race. Ithas also been observed (14/5. Mr. N.H. Smith'Observations on Breeding' quoted in 'Ency-clop. of Rural Sports' page 278.) with Englishracehorses that certain mares have generallytransmitted their own character, whilst othermares of equally pure blood have allowed thecharacter of the sire to prevail. A famous blackgreyhound, Bedlamite, as I hear from Mr. C.M.Brown "invariably got all his puppies black, nomatter what was the colour of the bitch;" butthen Bedlamite "had a preponderance of blackin his blood, both on the sire and dam side."

[The truth of the principle of prepotency comesout more clearly when distinct races are cros-sed. The improved Shorthorns, notwithstand-ing that the breed is comparatively modern, are

generally acknowledged to possess great powerin impressing their likeness on all other breeds;and it is chiefly in consequence of this powerthat they are so highly valued for exportation.(14/6. Quoted by Bronn 'Geshichte der Natur'b. 2 s. 170. See Sturm 'Ueber Racen' 1825 s. 104-107. For the niata cattle see my 'Journal of Re-searches' 1845 page 146.) Godine has given acurious case of a ram of a goat-like breed ofsheep from the Cape of Good Hope, which pro-duced offspring hardly to be distinguishedfrom himself, when crossed with ewes of twel-ve other breeds. But two of these half-bredewes, when put to a merino ram, producedlambs closely resembling the merino breed.Girou de Buzareingues (14/7. Lucas 'L'HerediteNat.' tome 2 page 112.) found that of two racesof French sheep the ewes of one, when crossedduring successive generations with merinorams, yielded up their character far sooner thanthe ewes of the other race. Sturm and Girouhave given analogous cases with other breeds

of sheep and with cattle, the prepotency run-ning in these cases through the male side; but Iwas assured on good authority in South Amer-ica, that when niata cattle are crossed withcommon cattle, though the niata breed is pre-potent whether males or females are used, yetthat the prepotency is strongest through thefemale line. The Manx cat is tailless and haslong hind legs; Dr. Wilson crossed a male Manxwith common cats, and, out of twenty-threekittens, seventeen were destitute of tails; butwhen the female Manx was crossed by com-mon male cats all the kittens had tails, thoughthey were generally short and imperfect. (14/8.Mr. Orton 'Physiology of Breeding' 1855 page9.)

In making reciprocal crosses between pouterand fantail pigeons, the pouter- race seemed tobe prepotent through both sexes over the fan-tail. But this is probably due to weak power inthe fantail rather than to any unusually strong

power in the pouter, for I have observed thatbarbs also preponderate over fantails. Thisweakness of transmission in the fantail, thoughthe breed is an ancient one, is said (14/9. Boi-tard and Corbie 'Les Pigeons' 1824 page 224.) tobe general; but I have observed one exceptionto the rule, namely, in a cross between a fantailand laugher. The most curious instance knownto me of weak power in both sexes is in thetrumpeter pigeon. This breed has been wellknown for at least 130 years: it breeds perfectlytrue, as I have been assured by those who havelong kept many birds: it is characterised by apeculiar tuft of feathers over the beak, by acrest on the head, by a singular coo quite unlikethat of any other breed, and by much-featheredfeet. I have crossed both sexes with turbits oftwo sub-breeds, with almond tumblers, spots,and runts, and reared many mongrels and re-crossed them; and though the crest on the headand feathered feet were inherited (as is gener-ally the case with most breeds), I have never

seen a vestige of the tuft over the beak or heardthe peculiar coo. Boitard and Corbie (14/10.'Les Pigeons' pages 168, 198.) assert that this isthe invariable result of crossing trumpeterswith other breeds: Neumeister (14/11. 'DasGanze' etc. 1837 s. 39.), however, states that inGermany mongrels have been obtained, thoughvery rarely, which were furnished with the tuftand would trumpet: but a pair of these mon-grels with a tuft, which I imported, nevertrumpeted. Mr. Brent states (14/12. 'The PigeonBook' page 46.) that the crossed offspring of atrumpeter were crossed with trumpeters forthree generations, by which time the mongrelshad 7/8ths of this blood in their veins, yet thetuft over the beak did not appear. At the fourthgeneration the tuft appeared, but the birdsthough now having 15-16ths trumpeter's bloodstill did not trumpet. This case well shows thewide difference between inheritance and pre-potency; for here we have a well- establishedold race which transmits its characters faith-

fully, but which, when crossed with any otherrace, has the feeblest power of transmitting itstwo chief characteristic qualities.

I will give one other instance with fowls andpigeons of weakness and strength in the trans-mission of the same character to their crossedoffspring. The Silk fowl breeds true, and thereis reason to believe is a very ancient race; butwhen I reared a large number of mongrels froma Silk hen by a Spanish cock, not one exhibitedeven a trace of the so-called silkiness. Mr. Hew-itt also asserts that in no instance are the silkyfeathers transmitted by this breed whencrossed with any other variety. But three birdsout of many raised by Mr. Orton from a crossbetween a silk cock and a bantam hen had silkyfeathers. (14/13. 'Physiology of Breeding' page22; Mr. Hewitt in 'The Poultry Book' by Teget-meier 1866 page 224.) So that it is certain thatthis breed very seldom has the power of trans-mitting its peculiar plumage to its crossed pro-

geny. On the other hand, there is a silk sub-variety of the fantail pigeon, which has its feat-hers in nearly the same state as in the Silk fowl:now we have already seen that fantails, whencrossed, possess singularly weak power intransmitting their general qualities; but the silksub-variety when crossed with any other small-sized race invariably transmits its silky feath-ers! (14/14. Boitard and Corbie 'Les Pigeons'1824 page 226.)

The well-known horticulturist, Mr. Paul, in-forms me that he fertilised the Black Prince hol-lyhock with pollen of the White Globe and theLemonade and Black Prince hollyhocks recip-rocally; but not one seedling from these threecrosses inherited the black colour of the BlackPrince. So, again, Mr. Laxton, who has had suchgreat experience in crossing peas, writes to methat "whenever a cross has been effected be-tween a white-blossomed and a purple- blos-somed pea, or between a white-seeded and a

purple-spotted, brown or maple- seeded pea,the offspring seems to lose nearly all the char-acteristics of the white-flowered and white-seeded varieties; and this result followswhether these varieties have been used as thepollen-bearing or seed-producing parents."

The law of prepotency comes into action whenspecies are crossed, as with races and individu-als. Gartner has unequivocally shown (14/15.'Bastarderzeugung' s. 256, 290, etc. Naudin'Nouvelles Archives du Museum' tome 1 page149 gives a striking instance of prepotency inDatura stramonium when crossed with twoother species.) that this is the case with plants.To give one instance: when Nicotiana panicu-lata and vincaeflora are crossed, the characterof N. paniculata is almost completely lost in thehybrid; but if N. quadrivalvis be crossed withN. vincaeflora, this latter species, which wasbefore so prepotent, now in its turn almost dis-appears under the power of N. quadrivalvis. It

is remarkable that the prepotency of one spe-cies over another in transmission is quite inde-pendent, as shown by Gartner, of the greater orless facility with which the one fertilises theother.

With animals, the jackal is prepotent over thedog, as is stated by Flourens, who made manycrosses between these animals; and this waslikewise the case with a hybrid which I oncesaw between a jackal and a terrier. I cannotdoubt, from the observations of Colin and oth-ers, that the ass is prepotent over the horse; theprepotency in this instance running more stron-gly through the male than through the femaleass; so that the mule resembles the ass moreclosely than does the hinny. (14/16. Flourens'Longevite Humaine' page 144 on crossed jack-als. With respect to the difference between themule and the hinny I am aware that this hasgenerally been attributed to the sire and damtransmitting their characters differently; but

Colin, who has given in his 'Traite Phys.Comp.' tome 2 pages 537-539, the fullest de-scription which I have met with of these recip-rocal hybrids, is strongly of opinion that the asspreponderates in both crosses, but in an un-equal degree. This is likewise the conclusion ofFlourens, and of Bechstein in his 'Natur-geschichte Deutschlands' b. 1 s. 294. The tail ofthe hinny is much more like that of the horsethan is the tail of the mule, and this is generallyaccounted for by the males of both speciestransmitting with greater power this part oftheir structure; but a compound hybrid which Isaw in the Zoological Gardens, from a mare bya hybrid ass- zebra, closely resembled itsmother in its tail.) The male pheasant, judgingfrom Mr. Hewitt's descriptions (14/17. Mr.Hewitt who has had such great experience inraising these hybrids says ('Poultry Book' byMr. Tegetmeier 1866 pages 165-167) that in all,the head was destitute of wattles, comb, andear-lappets; and all closely resembled the

pheasant in the shape of the tail and generalcontour of the body. These hybrids were raisedfrom hens of several breeds by a cock-pheasant;but another hybrid, described by Mr. Hewitt,was raised from a hen-pheasant, by a silver-laced Bantam cock, and this possessed a rudi-mental comb and wattles.), and from the hy-brids which I have seen, preponderates overthe domestic fowl; but the latter, as far as col-our is concerned, has considerable power oftransmission, for hybrids raised from five dif-ferently coloured hens differed greatly inplumage. I formerly examined some curioushybrids in the Zoological Gardens, between thePenguin variety of the common duck and theEgyptian goose (Anser aegyptiacus); and al-though I will not assert that the domesticatedvariety preponderated over the natural species,yet it had strongly impressed its unnatural up-right figure on these hybrids.

I am aware that such cases as the foregoinghave been ascribed by various authors, not toone species, race, or individual being prepotentover the other in impressing its character on itscrossed offspring, but to such rules as that thefather influences the external characters and themother the internal or vital organs. But thegreat diversity of the rules given by variousauthors almost proves their falseness. Dr. Pros-per Lucas has fully discussed this point, andhas shown (14/18. 'L'Hered. Nat.' tome 2 book2 chapter 1.) that none of the rules (and I couldadd others to those quoted by him) apply to allanimals. Similar rules have been announced forplants, and have been proved by Gartner(14/19. 'Bastarderzeugung' s. 264-266. Naudin'Nouvelles Archives du Museum' tome 1 page148 has arrived at a similar conclusion.) to beall erroneous. If we confine our view to thedomesticated races of a single species, or per-haps even to the species of the same genus,some such rules may hold good; for instance, it

seems that in reciprocally crossing variousbreeds of fowls the male generally gives colour(14/20. 'Cottage Gardener' 1856 pages 101,137.); but conspicuous exceptions have passedunder my own eyes. It seems that the ram usu-ally gives its peculiar horns and fleece to itscrossed offspring, and the bull the presence orabsence of horns.

In the following chapter on Crossing I shallhave occasion to show that certain charactersare rarely or never blended by crossing, but aretransmitted in an unmodified state from eitherparent-form; I refer to this fact here because it issometimes accompanied on the one side byprepotency, which thus acquires the false ap-pearance of unusual strength. In the samechapter I shall show that the rate at which aspecies or breed absorbs and obliterates an-other by repeated crosses, depends in chief parton prepotency in transmission.]

In conclusion, some of the cases above given,—for instance, that of the trumpeter pigeon,—prove that there is a wide difference betweenmere inheritance and prepotency. This latterpower seems to us, in our ignorance, to act inmost cases quite capriciously. The very samecharacter, even though it be an abnormal ormonstrous one, such as silky feathers, may betransmitted by different species, when crossed,either with prepotent force or singular feeble-ness. It is obvious, that a purely-bred form ofeither sex, in all cases in which prepotency doesnot run more strongly in one sex than the other,will transmit its character with prepotent forceover a mongrelised and already variable form.(14/21. See some remarks on this head withrespect to sheep by Mr. Wilson in 'Gardener'sChronicle' 1863 page 15. Many striking in-stances of this result are given by M. Malingie-Nouel 'Journ. R. Agricult. Soc.' volume 14 1853page 220 with respect to crosses between Eng-lish and French sheep. He found that he ob-

tained the desired influence of the Englishbreeds by crossing intentionally mongrelisedFrench breeds with pure English breeds.) Fromseveral of the above-given cases we may con-clude that mere antiquity of character does notby any means necessarily make it prepotent. Insome cases prepotency apparently depends onthe same character being present and visible inone of the two breeds which are crossed, andlatent or invisible in the other breed; and in thiscase it is natural that the character which ispotentially present in both breeds should beprepotent. Thus, we have reason to believe thatthere is a latent tendency in all horses to bedun-coloured and striped; and when a horse ofthis kind is crossed with one of any other col-our, it is said that the offspring are almost sureto be striped. Sheep have a similar latent ten-dency to become dark-coloured, and we haveseen with what prepotent force a ram with afew black spots, when crossed with whitesheep of various breeds, coloured its offspring.

All pigeons have a latent tendency to becomeslaty-blue, with certain characteristic marks,and it is known that, when a bird thus colouredis crossed with one of any other colour, it ismost difficult afterwards to eradicate the bluetint. A nearly parallel case is offered by thoseblack bantams which, as they grow old, de-velop a latent tendency to acquire red feathers.But there are exceptions to the rule: hornlessbreeds of cattle possess a latent capacity to re-produce horns, yet when crossed with hornedbreeds they do not invariably produce off-spring bearing horns.

We meet with analogous cases with plants.Striped flowers, though they can be propagatedtruly by seed, have a latent tendency to becomeuniformly coloured, but when once crossed bya uniformly coloured variety, they ever after-wards fail to produce striped seedlings. (14/22.Verlot 'Des Varietes' 1865 page 66.) Anothercase is in some respects more curious: plants

bearing peloric flowers have so strong a latenttendency to reproduce their normally irregularflowers, that this often occurs by buds when aplant is transplanted into poorer or richer soil.(14/23. Moquin-Tandon 'Teratologie' page 191.)Now I crossed the peloric snapdragon (Antir-rhinum majus), described in the last chapter,with pollen of the common form; and the latter,reciprocally, with peloric pollen. I thus raisedtwo great beds of seedlings, and not one waspeloric. Naudin (14/24. 'Nouvelles Archives duMuseum' tome 1 page 137.) obtained the sameresult from crossing a peloric Linaria with thecommon form. I carefully examined the flowersof ninety plants of the crossed Antirrhinum inthe two beds, and their structure had not beenin the least affected by the cross, except that ina few instances the minute rudiment of the fifthstamen, which is always present, was morefully or even completely developed. It must notbe supposed that this entire obliteration of thepeloric structure in the crossed plants can be

accounted for by any incapacity of transmis-sion; for I raised a large bed of plants from thepeloric Antirrhinum, artificially fertilised by itsown pollen, and sixteen plants, which alonesurvived the winter, were all as perfectly pelo-ric as the parent-plant. Here we have a goodinstance of the wide difference between theinheritance of a character and the power oftransmitting it to crossed offspring. The crossedplants, which perfectly resembled the commonsnapdragon, were allowed to sow themselves,and out of a hundred and twenty-seven seed-lings, eighty-eight proved to be common snap-dragons, two were in an intermediate conditionbetween the peloric and normal state, and thir-ty-seven were perfectly peloric, having re-verted to the structure of their one grand-parent. This case seems at first sight to offer anexception to the rule just given, namely, that acharacter which is present in one form and la-tent in the other is generally transmitted withprepotent force when the two forms are

crossed. For in all the Scrophulariaceae, andespecially in the genera Antirrhinum andLinaria, there is, as was shown in the last chap-ter, a strong latent tendency to become peloric;but there is also, as we have seen, a stillstronger tendency in all peloric plants to reac-quire their normal irregular structure. So thatwe have two opposed latent tendencies in thesame plants. Now, with the crossed Antirrhin-ums the tendency to produce normal or irregu-lar flowers, like those of the common Snap-dragon, prevailed in the first generation; whilstthe tendency to pelorism, appearing to gainstrength by the intermission of a generation,prevailed to a large extent in the second set ofseedlings. How it is possible for a character togain strength by the intermission of a genera-tion, will be considered in the chapter on pan-genesis.

On the whole, the subject of prepotency is ex-tremely intricate,—from its varying so much in

strength, even in regard to the same character,in different animals,—from its running eitherequally in both sexes, or, as frequently is thecase with animals, but not with plants, muchstronger in one sex than the other,—from theexistence of secondary sexual characters,—fromthe transmission of certain characters beinglimited, as we shall immediately see, by sex,—from certain characters not blending to-gether,—and, perhaps, occasionally from theeffects of a previous fertilisation on the mother.It is therefore not surprising that no one hashitherto succeeded in drawing up general ruleson the subject of prepotency.

INHERITANCE AS LIMITED BY SEX.

New characters often appear in one sex, andare afterwards transmitted to the same sex,either exclusively or in a much greater degreethan to the other. This subject is important, be-cause with animals of many kinds in a state of

nature, both high and low in the scale, secon-dary sexual characters, not directly connectedwith the organs of reproduction, are conspicu-ously present. With our domesticated animals,characters of this kind often differ widely fromthose distinguishing the two sexes of the parentspecies; and the principle of inheritance, as lim-ited by sex, explains how this is possible.

[Dr. P. Lucas has shown (14/25. 'L'Hered. Nat.'tome 2 pages 137-165. See also Mr. Sedgwick'sfour memoirs, immediately to be referred to.)that when a peculiarity, in no manner con-nected with the reproductive organs, appearsin either parent, it is often transmitted exclu-sively to the offspring of the same sex, or to amuch greater number of them than of the op-posite sex. Thus, in the family of Lambert, thehorn-like projections on the skin were transmit-ted from the father to his sons and grandsonsalone; so it has been with other cases of ich-thyosis, with supernumerary digits, with a de-

ficiency of digits and phalanges, and in a lesserdegree with various diseases, especially withcolour-blindness and the haemorrhagic diathe-sis, that is, an extreme liability to profuse anduncontrollable bleeding from trifling wounds.On the other hand, mothers have transmitted,during several generations, to their daughtersalone, supernumerary and deficient digits, col-our-blindness and other peculiarities. So thatthe very same peculiarity may become attachedto either sex, and be long inherited by that sexalone; but the attachment in certain cases ismuch more frequent to one than the other sex.The same peculiarities also may be promiscu-ously transmitted to either sex. Dr. Lucas givesother cases, showing that the male occasionallytransmits his peculiarities to his daughtersalone, and the mother to her sons alone; buteven in this case we see that inheritance is to acertain extent, though inversely, regulated bysex. Dr. Lucas, after weighing the whole evi-dence, comes to the conclusion that every pecu-

liarity tends to be transmitted in a greater orlesser degree to that sex in which it first ap-pears. But a more definite rule, as I have else-where shown (14/26. 'Descent of Man' 2nd edi-tion page 32.) generally holds good, namely,that variations which first appear in either sexat a late period of life, when the reproductivefunctions are active, tend to be developed inthat sex alone; whilst variations which first ap-pear early in life in either sex are commonlytransmitted to both sexes. I am, however, farfrom supposing that this is the sole determin-ing cause.

A few details from the many cases collected byMr. Sedgwick (14/27. On Sexual Limitation inHereditary Diseases 'Brit. and For. Med.-Chirurg. Review' April 1861 page 477; July pa-ge 198; April 1863 page 445; and July page 159.Also in 1867 'On the influence of Age in He-reditary Disease.'), may be here given. Colour-blindness, from some unknown cause, shows

itself much oftener in males than in females; inupwards of two hundred cases collected by Mr.Sedgwick, nine-tenths related to men; but it iseminently liable to be transmitted throughwomen. In the case given by Dr. Earle, mem-bers of eight related families were affected dur-ing five generations: these families consisted ofsixty-one individuals, namely, of thirty-twomales, of whom nine-sixteenths were incapableof distinguishing colour, and of twenty-ninefemales, of whom only one-fifteenth were thusaffected. Although colour-blindness thus gen-erally clings to the male sex, nevertheless, inone instance in which it first appeared in a fe-male, it was transmitted during five genera-tions to thirteen individuals, all of whom werefemales. The haemorrhagic diathesis, often ac-companied by rheumatism, has been known toaffect the males alone during five generations,being transmitted, however, through the fe-males. It is said that deficient phalanges in thefingers have been inherited by the females

alone during ten generations. In another case, aman thus deficient in both hands and feet,transmitted the peculiarity to his two sons andone daughter; but in the third generation,—outof nineteen grandchildren, twelve sons had thefamily defect, whilst the seven daughters werefree. In ordinary cases of sexual limitation, thesons or daughters inherit the peculiarity, what-ever it may be, from their father or mother, andtransmit it to their children of the same sex; butgenerally with the haemorrhagic diathesis, andoften with colour-blindness, and in some othercases, the sons never inherit the peculiarity di-rectly from their fathers, but the daughters alo-ne transmit the latent tendency, so that the sonsof the daughters alone exhibit it. Thus the fa-ther, grandson, and great-great-grandson willexhibit a peculiarity,— the grandmother,daughter, and great-grand-daughter havingtransmitted it in a latent state. Hence we have,as Mr. Sedgwick remarks, a double kind of ata-vism or reversion; each grandson apparently

receiving and developing the peculiarity fromhis grandfather, and each daughter apparentlyreceiving the latent tendency from her grand-mother.

From the various facts recorded by Dr. ProsperLucas, Mr. Sedgwick, and others, there can beno doubt that peculiarities first appearing ineither sex, though not in any way necessarily orinvariably connected with that sex, stronglytend to be inherited by the offspring of the sa-me sex, but are often transmitted in a latentstate through the opposite sex.

Turning now to domesticated animals, we findthat certain characters not proper to the parentspecies are often confined to, and inherited by,one sex alone; but we do not know the historyof the first appearance of such characters. In thechapter on Sheep, we have seen that the malesof certain races differ greatly from the femalesin the shape of their horns, these being absentin the ewes of some breeds; they differ also in

the development of fat in the tail and in theoutline of the forehead. These differences, judg-ing from the character of the allied wild species,cannot be accounted for by supposing that theyhave been derived from distinct parent forms.There is, also, a great difference between thehorns of the two sexes in one Indian breed ofgoats. The bull zebu is said to have a largerhump than the cow. In the Scotch deer-houndthe two sexes differ in size more than in anyother variety of the dog (14/28. W. Scrope 'Artof Deer Stalking' page 354.) and, judging fromanalogy, more than in the aboriginal parent-species. The peculiar colour called tortoise-shellis very rarely seen in a male cat; the males ofthis variety being of a rusty tint.

In various breeds of the fowl the males andfemales often differ greatly; and these differ-ences are far from being the same with thosewhich distinguish the two sexes of the parent-species, the Gallus bankiva; and consequently

have originated under domestication. In certainsub-varieties of the Game race we have the un-usual case of the hens differing from each othermore than the cocks. In an Indian breed of awhite colour shaded with black, the hens in-variably have black skins, and their bones arecovered by a black periosteum, whilst the cocksare never or most rarely thus characterised.Pigeons offer a more interesting case; forthroughout the whole great family the twosexes do not often differ much; and the malesand females of the parent-form, the C. livia, areundistinguishable: yet we have seen that withpouters the male has the characteristic qualityof pouting more strongly developed than thefemale; and in certain sub-varieties the malesalone are spotted or striated with black, or oth-erwise differ in colour. When male and femaleEnglish carrier-pigeons are exhibited in sepa-rate pens, the difference in the development ofthe wattle over the beak and round the eyes isconspicuous. So that here we have instances of

the appearance of secondary sexual charactersin the domesticated races of a species in whichsuch differences are naturally quite absent.]

On the other hand, secondary sexual characterswhich belong to the species in a state of natureare sometimes quite lost, or greatly diminished,under domestication. We see this in the smallsize of the tusks in our improved breeds of thepig, in comparison with those of the wild boar.There are sub- breeds of fowls, in which themales have lost the fine-flowing tail-feathersand hackles; and others in which there is nodifference in colour between the two sexes. Insome cases the barred plumage, which in galli-naceous birds is commonly the attribute of thehen, has been transferred to the cock, as in thecuckoo sub-breeds. In other cases masculinecharacters have been partly transferred to thefemale, as with the splendid plumage of thegolden-spangled Hamburgh hen, the enlargedcomb of the Spanish hen, the pugnacious dis-

position of the Game hen, and as in the well-developed spurs which occasionally appear inthe hens of various breeds. In Polish fowls bothsexes are ornamented with a topknot, that ofthe male being formed of hackle-like feathers,and this is a new male character in the genusGallus. On the whole, as far as I can judge, newcharacters are more apt to appear in the malesof our domesticated animals than in the fe-males (14/29. I have given in my 'Descent ofMan' 2nd edition page 223 sufficient evidencethat male animals are usually more variablethan the females.), and afterwards to be inher-ited exclusively or more strongly by the males.Finally, in accordance with the principle of in-heritance as limited by sex, the preservationand augmentation of secondary sexual charac-ters in natural species offers no especial diffi-culty, as this would follow through that form ofselection which I have called sexual selection.

INHERITANCE AT CORRESPONDINGPERIODS OF LIFE.

This is an important subject. Since the publica-tion of my 'Origin of Species' I have seen noreason to doubt the truth of the explanationthere given of one of the most remarkable factsin biology, namely, the difference between theembryo and the adult animal. The explanationis, that variations do not necessarily or gener-ally occur at a very early period of embryonicgrowth, and that such variations are inheritedat a corresponding age. As a consequence ofthis the embryo, even after the parent-form hasundergone great modification, is left only sligh-tly modified; and the embryos of widely-different animals which are descended from acommon progenitor remain in many importantrespects like one another and probably liketheir common progenitor. We can thus under-stand why embryology throws a flood of lighton the natural system of classification, as thisought to be as far as possible genealogical.

When the embryo leads an independent life,that is, becomes a larva, it has to be adapted tothe surrounding conditions in its structure andinstincts, independently of those of its parents;and the principle of inheritance at correspond-ing periods of life renders this possible.

This principle is, indeed, in one way so obviousthat it escapes attention. We possess a numberof races of animals and plants, which, whencompared with one another and with their par-ent-forms, present conspicuous differences,both in their immature and mature states. Lookat the seeds of the several kinds of peas, beans,maize, which can be propagated truly, and seehow they differ in size, colour, and shape,whilst the full-grown plants differ but little.Cabbages, on the other hand, differ greatly infoliage and manner of growth, but hardly at allin their seeds; and generally it will be foundthat the differences between cultivated plantsat different periods of growth are not necessar-

ily closely connected together, for plants maydiffer much in their seeds and little when full-grown, and conversely may yield seeds hardlydistinguishable, yet differ much when full-grown. In the several breeds of poultry, de-scended from a single species, differences in theeggs and chickens whilst covered with down,in the plumage at the first and subsequentmoults, as well as in the comb and wattles, areall inherited. With man peculiarities in the milkand second teeth (of which I have received thedetails) are inheritable, and longevity is oftentransmitted. So again with our improvedbreeds of cattle and sheep, early maturity, in-cluding the early development of the teeth, andwith certain breeds of fowl the early appear-ance of secondary sexual characters, all comeunder the same head of inheritance at corre-sponding periods.

Numerous analogous facts could be given. Thesilk-moth, perhaps, offers the best instance; for

in the breeds which transmit their characterstruly, the eggs differ in size, colour, and shape:the caterpillars differ, in moulting three or fourtimes, in colour, even in having a dark-coloured mark like an eyebrow, and in the lossof certain instincts;—the cocoons differ in size,shape, and in the colour and quality of the silk;these several differences being followed byslight or barely distinguishable differences inthe mature moth.

But it may be said that, if in the above cases anew peculiarity is inherited, it must be at thecorresponding stage of development; for an eggor seed can resemble only an egg or seed, andthe horn in a full-grown ox can resemble only ahorn. The following cases show inheritance atcorresponding periods more plainly, becausethey refer to peculiarities which might havesupervened, as far as we can see, earlier or laterin life, yet are inherited at the same period atwhich they first appeared.

[In the Lambert family the porcupine-like ex-crescences appeared in the father and sons atthe same age, namely, about nine weeks afterbirth. (14/30. Prichard 'Phys. Hist. of Mankind'1851 volume 1 page 349.) In the extraordinaryhairy family described by Mr. Crawfurd(14/31. 'Embassy to the Court of Ava' volume 1page 320. The third generation is described byCapt. Yule in his 'Narrative of the Mission tothe Court of Ava' 1855 page 94.), children wereproduced during three generations with hairyears; in the father the hair began to grow overhis body at six years old; in his daughter so-mewhat earlier, namely, at one year; and inboth generations the milk teeth appeared late inlife, the permanent teeth being afterwards sin-gularly deficient. Greyness of hair at an unusu-ally early age has been transmitted in somefamilies. These cases border on diseases inher-ited at corresponding periods of life, to which Ishall immediately refer.

It is a well-known peculiarity with almond-tumbler pigeons, that the full beauty and pecu-liar character of the plumage does not appearuntil the bird has moulted two or three times.Neumeister describes and figures a brace ofpigeons in which the whole body is white ex-cept the breast, neck, and head; but in their firstplumage all the white feathers have colourededges. Another breed is more remarkable: itsfirst plumage is black, with rusty-red wing-barsand a crescent-shaped mark on the breast; thesemarks then become white, and remain so dur-ing three or four moults; but after this periodthe white spreads over the body, and the birdloses its beauty. (14/32. 'Das Ganze der Tau-benzucht' 1837 s. 24 tab. 4 figure 2 s. 21 tab. 1figure 4.) Prize canary- birds have their wingsand tail black: "this colour, however, is onlyretained until the first moult, so that they mustbe exhibited ere the change takes place. Oncemoulted, the peculiarity has ceased. Of courseall the birds emanating from this stock have

black wings and tails the first year." (14/33.Kidd 'Treatise on the Canary' page 18.) A curi-ous and somewhat analogous account has beengiven (14/34. Charlesworth 'Mag. of Nat. Hist.'volume 1 1837 page 167.) of a family of wildpied rooks which were first observed in 1798,near Chalfont, and which every year from thatdate up to the period of the published notice,viz., 1837 "have several of their brood particol-oured, black and white. This variegation of theplumage, however, disappears with the firstmoult; but among the next young families thereare always a few pied ones." These changes ofplumage, which are inherited at various corre-sponding periods of life in the pigeon, canary-bird, and rook, are remarkable, because theparent-species passes through no such change.

Inherited diseases afford evidence in some re-spects of less value than the foregoing cases,because diseases are not necessarily connectedwith any change in structure; but in other re-

spects of more value, because the periods havebeen more carefully observed. Certain diseasesare communicated to the child apparently by aprocess like inoculation, and the child is fromthe first affected; such cases may be here passedover. Large classes of diseases usually appearat certain ages, such as St. Vitus's dance inyouth, consumption in early mid-life, gout la-ter, and apoplexy still later; and these are natu-rally inherited at the same period. But even indiseases of this class, instances have been re-corded, as with St. Vitus's dance, showing thatan unusually early or late tendency to the dis-ease is inheritable. (14/35. Dr. Prosper Lucas'Hered. Nat.' tome 2 page 713.) In most casesthe appearance of any inherited disease is lar-gely determined by certain critical periods ineach person's life, as well as by unfavourableconditions. There are many other diseases,which are not attached to any particular period,but which certainly tend to appear in the childat about the same age at which the parent was

first attacked. An array of high authorities, an-cient and modern, could be given in support ofthis proposition. The illustrious Hunter be-lieved in it; and Piorry (14/36. 'L'Hered. dansles Maladies' 1840 page 135. For Hunter seeHarlan 'Med. Researches' page 530.) cautionsthe physician to look closely to the child at theperiod when any grave inheritable disease at-tacked the parent. Dr. Prosper Lucas (14/37.'L'Hered. Nat.' tome 2 page 850.), after collect-ing facts from every source, asserts that affec-tions of all kinds, though not related to anyparticular period of life, tend to reappear in theoffspring at whatever period of life they firstappeared in the progenitor.

As the subject is important, it may be well togive a few instances, simply as illustrations, notas proof; for proof, recourse must be had to theauthorities above quoted. Some of the follow-ing cases have been selected for the sake of sho-wing that, when a slight departure from the

rule occurs, the child is affected somewhat ear-lier in life than the parent. In the family of LeCompte blindness was inherited through threegenerations, and no less than twenty-sevenchildren and grandchildren were all affected atabout the same age; their blindness in generalbegan to advance about the fifteenth or six-teenth year, and ended in total deprivation ofsight at the age of about twenty-two. (14/38.Sedgwick 'Brit. and For. Med.-Chirurg. Review'April 1861 page 485. In some accounts thenumber of children and grandchildren is givenas 37; but this seems to be an error judgingfrom the paper first published in the 'BaltimoreMed. and Phys. Reg.' 1809 of which Mr. Sedg-wick has been so kind as to send me a copy.) Inanother case a father and his four children allbecame blind at twenty-one years old; in an-other, a grandmother grew blind at thirty-five,her daughter at nineteen, and three grandchil-dren at the ages of thirteen and eleven. (14/39.Prosper Lucas 'Hered. Nat.' tome 1 page 400.)

So with deafness, two brothers, their father andpaternal grandfather, all became deaf at the ageof forty. (14/40. Sedgwick ibid July 1861 page202.)

Esquirol gives several striking instances of in-sanity coming on at the same age, as that of agrandfather, father, and son, who all commit-ted suicide near their fiftieth year. Many othercases could be given, as of a whole family whobecame insane at the age of forty. (14/41. Pio-rry page 109; Prosper Lucas tome 2 page 759.)Other cerebral affections sometimes follow thesame rule,—for instance, epilepsy and apo-plexy. A woman died of the latter disease whensixty-three years old; one of her daughters atforty-three, and the other at sixty-seven: thelatter had twelve children, who all died fromtubercular meningitis. (14/42. Prosper Lucastome 2 page 748.) I mention this latter case be-cause it illustrates a frequent occurrence, name-ly, a change in the precise nature of an inher-

ited disease, though still affecting the same or-gan.

Asthma has attacked several members of thesame family when forty years old, and otherfamilies during infancy. The most different dis-eases, such as angina pectoris, stone in thebladder, and various affections of the skin,have appeared in successive generations atnearly the same age. The little finger of a manbegan from some unknown cause to grow in-wards, and the same finger in his two sons be-gan at the same age to bend inwards in a simi-lar manner. Strange and inexplicable neuralgicaffections have caused parents and children tosuffer agonies at about the same period of life.(14/43. Prosper Lucas tome 3 pages 678, 700,702; Sedgwick ibid April 1863 page 449 andJuly 1863 page 162. Dr. J. Steinan 'Essay on He-reditary Disease' 1843 pages 27, 34.)

I will give only two other cases, which are in-teresting as illustrating the disappearance as

well as the appearance of disease at the sameage. Two brothers, their father, their paternaluncles, seven cousins, and their paternal grand-father, were all similarly affected by a skin-disease, called pityriasis versicolor; "the dis-ease, strictly limited to the males of the family(though transmitted through the females), usu-ally appeared at puberty, and disappeared atabout the age of forty or forty-five years." Thesecond case is that of four brothers, who whenabout twelve years old suffered almost everyweek from severe headaches, which were re-lieved only by a recumbent position in a darkroom. Their father, paternal uncles, paternalgrandfather, and granduncles all suffered in thesame way from headaches, which ceased at theage of fifty-four or fifty-five in all those wholived so long. None of the females of the familywere affected. (14/44. These cases are given byMr. Sedgwick on the authority of Dr. H. Stew-art in 'Med.-Chirurg. Review' April 1863 pages449, 477.)]

It is impossible to read the foregoing accounts,and the many others which have been re-corded, of diseases coming on during three oreven more generations in several members ofthe same family at the same age, especially inthe case of rare affections in which the coinci-dence cannot be attributed to chance, and todoubt that there is a strong tendency to inheri-tance in disease at corresponding periods oflife. When the rule fails, the disease is apt tocome on earlier in the child than in the parent;the exceptions in the other direction being verymuch rarer. Dr. Lucas (14/45. 'Hered. Nat.' to-me 2 page 852.) alludes to several cases of in-herited diseases coming on at an earlier period.I have already given one striking instance withblindness during three generations; and Mr.Bowman remarks that this frequently occurswith cataract. With cancer there seems to be apeculiar liability to earlier inheritance: Sir J.Paget, who has particularly attended to thissubject, and tabulated a large number of cases,

informs me that he believes that in nine casesout of ten the later generation suffers from thedisease at an earlier period than the previousgeneration. He adds, "In the instances in whichthe opposite relation holds, and the members oflater generations have cancer at a later age thantheir predecessors, I think it will be found thatthe non- cancerous parents have lived to ex-treme old ages." So that the longevity of a non-affected parent seems to have the power of in-fluencing the fatal period in the offspring; andwe thus apparently get another element ofcomplexity in inheritance.

The facts, showing that with certain diseasesthe period of inheritance occasionally or evenfrequently advances, are important with re-spect to the general descent-theory, for theyrender it probable that the same thing wouldoccur with ordinary modifications of structure.The final result of a long series of such ad-vances would be the gradual obliteration of

characters proper to the embryo and larva,which would thus come to resemble more andmore closely the mature parent-form. But anystructure which was of service to the embryo orlarva would be preserved by the destruction atthis stage of growth of each individual whichmanifested any tendency to lose its proper cha-racter at too early an age.

Finally, from the numerous races of cultivatedplants and domestic animals, in which theseeds or eggs, the young or old, differ from oneanother and from those of the parent-species;—from the cases in which new characters haveappeared at a particular period, and afterwardsbeen inherited at the same period;—and fromwhat we know with respect to disease, we mustbelieve in the truth of the great principle ofinheritance at corresponding periods of life.

SUMMARY OF THE THREE PRECEDINGCHAPTERS.

Strong as is the force of inheritance, it allowsthe incessant appearance of new characters.These, whether beneficial or injurious,—of themost trifling importance, such as a shade ofcolour in a flower, a coloured lock of hair, or amere gesture,—or of the highest importance, aswhen affecting the brain, or an organ so perfectand complex as the eye,—or of so grave a na-ture as to deserve to be called a monstrosity,—or so peculiar as not to occur normally in anymember of the same natural class,—are ofteninherited by man, by the lower animals, andplants. In numberless cases it suffices for theinheritance of a peculiarity that one parent alo-ne should be thus characterised. Inequalities inthe two sides of the body, though opposed tothe law of symmetry, may be transmitted. The-re is ample evidence that the effects of mutila-tions and of accidents, especially or perhapsexclusively when followed by disease, are occa-sionally inherited. There can be no doubt thatthe evil effects of the long-continued exposure

of the parent to injurious conditions are some-times transmitted to the offspring. So it is, aswe shall see in a future chapter, with the effectsof the use and disuse of parts, and of mentalhabits. Periodical habits are likewise transmit-ted, but generally, as it would appear, withlittle force.

Hence we are led to look at inheritance as therule, and non-inheritance as the anomaly. Butthis power often appears to us in our ignoranceto act capriciously, transmitting a characterwith inexplicable strength or feebleness. Thevery same peculiarity, as the weeping habit oftrees, silky feathers, etc., may be inherited ei-ther firmly or not at all by different members ofthe same group, and even by different indi-viduals of the same species, though treated inthe same manner. In this latter case we see thatthe power of transmission is a quality which ismerely individual in its attachment. As withsingle characters, so it is with the several con-

current slight differences which distinguishsub-varieties or races; for of these, some can bepropagated almost as truly as species, whilstothers cannot be relied on. The same rule holdsgood with plants, when propagated by bulbs,offsets, etc., which in one sense still form partsof the same individual, for some varieties retainor inherit through successive bud-generationstheir character far more truly than others.

Some characters not proper to the parent-species have certainly been inherited from anextremely remote epoch, and may therefore beconsidered as firmly fixed. But it is doubtfulwhether length of inheritance in itself givesfixedness of character; though the chances areobviously in favour of any character which haslong been transmitted true or unaltered stillbeing transmitted true as long as the conditionsof life remain the same. We know that manyspecies, after having retained the same charac-ter for countless ages, whilst living under their

natural conditions, when domesticated havevaried in the most diversified manner, that is,have failed to transmit their original form; sothat no character appears to be absolutely fixed.We can sometimes account for the failure ofinheritance by the conditions of life being op-posed to the development of certain characters;and still oftener, as with plants cultivated bygrafts and buds, by the conditions causing newand slight modifications incessantly to appear.In this latter case it is not that inheritance who-lly fails, but that new characters are continuallysuperadded. In some few cases, in which bothparents are similarly characterised, inheritanceseems to gain so much force by the combinedaction of the two parents, that it counteracts itsown power, and a new modification is the re-sult.

In many cases the failure of the parents totransmit their likeness is due to the breed hav-ing been at some former period crossed; and

the child takes after his grandparent or moreremote ancestor of foreign blood. In other cases,in which the breed has not been crossed, butsome ancient character has been lost throughvariation, it occasionally reappears throughreversion, so that the parents apparently fail totransmit their own likeness. In all cases, how-ever, we may safely conclude that the childinherits all its characters from its parents, inwhom certain characters are latent, like the sec-ondary sexual characters of one sex in the ot-her. When, after a long succession of bud- gen-erations, a flower or fruit becomes separatedinto distinct segments, having the colours orother attributes of both parent-forms, we can-not doubt that these characters were latent inthe earlier buds, though they could not then bedetected, or could be detected only in an inti-mately commingled state. So it is with animalsof crossed parentage, which with advancingyears occasionally exhibit characters derivedfrom one of their two parents, of which not a

trace could at first be perceived. Certain mon-strosities, which resemble what naturalists callthe typical form of the group in question, ap-parently come under the same law of reversion.It is assuredly an astonishing fact that the maleand female sexual elements, that buds, andeven full-grown animals, should retain charac-ters, during several generations in the case ofcrossed breeds, and during thousands of gen-erations in the case of pure breeds, written as itwere in invisible ink, yet ready at any time tobe evolved under certain conditions.

What these conditions precisely are, we do notknow. But any cause which disturbs the or-ganisation or constitution seems to be suffi-cient. A cross certainly gives a strong tendencyto the reappearance of long-lost characters,both corporeal and mental. In the case ofplants, this tendency is much stronger withthose species which have been crossed afterlong cultivation and which therefore have had

their constitutions disturbed by this cause aswell as by crossing, than with species whichhave always lived under their natural condi-tions and have then been crossed. A return,also, of domesticated animals and cultivatedplants to a wild state favours reversion; but thetendency under these circumstances has beenmuch exaggerated.

When individuals of the same family whichdiffer somewhat, and when races or species arecrossed, the one is often prepotent over theother in transmitting its character. A race maypossess a strong power of inheritance, and yetwhen crossed, as we have seen with trumpeter-pigeons, yield to the prepotency of every otherrace. Prepotency of transmission may be equalin the two sexes of the same species, but oftenruns more strongly in one sex. It plays an im-portant part in determining the rate at whichone race can be modified or wholly absorbedby repeated crosses with another. We can sel-

dom tell what makes one race or species prepo-tent over another; but it sometimes depends onthe same character being present and visible inone parent, and latent or potentially present inthe other.

Characters may first appear in either sex, butoftener in the male than in the female, and af-terwards be transmitted to the offspring of thesame sex. In this case we may feel confidentthat the peculiarity in question is really presentthough latent in the opposite sex! hence thefather may transmit through his daughter anycharacter to his grandson; and the mother con-versely to her granddaughter. We thus learn,and the fact is an important one, that transmis-sion and development are distinct powers. Oc-casionally these two powers seem to be an-tagonistic, or incapable of combination in thesame individual; for several cases have beenrecorded in which the son has not directly in-herited a character from his father, or directly

transmitted it to his son, but has received it bytransmission through his non-affected mother,and transmitted it through his non-affecteddaughter. Owing to inheritance being limitedby sex, we see how secondary sexual charactersmay have arisen under nature; their preserva-tion and accumulation being dependent ontheir service to either sex.

At whatever period of life a new character firstappears, it generally remains latent in the off-spring until a corresponding age is attained,and then is developed. When this rule fails, thechild generally exhibits the character at an ear-lier period than the parent. On this principle ofinheritance at corresponding periods, we canunderstand how it is that most animals displayfrom the germ to maturity such a marvelloussuccession of characters.

Finally, though much remains obscure withrespect to Inheritance, we may look at the fol-lowing laws as fairly well established.

FIRSTLY, a tendency in every character, newand old, to be transmitted by seminal and budgeneration, though often counteracted by vari-ous known and unknown causes.

SECONDLY, reversion or atavism, which de-pends on transmission and development beingdistinct powers: it acts in various degrees andmanners through both seminal and bud genera-tion.

THIRDLY, prepotency of transmission, whichmay be confined to one sex, or be common toboth sexes.

FOURTHLY, transmission, as limited by sex,generally to the same sex in which the inheritedcharacter first appeared; and this in many, pro-bably most cases, depends on the new characterhaving first appeared at a rather late period oflife.

FIFTHLY, inheritance at corresponding periodsof life, with some tendency to the earlier devel-opment of the inherited character.

In these laws of Inheritance, as displayed underdomestication, we see an ample provision forthe production, through variability and naturalselection, of new specific forms.

CHAPTER 2.XV.

ON CROSSING.

FREE INTERCROSSING OBLITERATESTHE DIFFERENCES BETWEEN ALLIEDBREEDS. WHEN THE NUMBERS OF TWOCOMMINGLING BREEDS ARE UNEQUAL,ONE ABSORBS THE OTHER. THE RATE OFABSORPTION DETERMINED BY PREPO-TENCY OF TRANSMISSION, BY THE CON-DITIONS OF LIFE, AND BY NATURAL SE-LECTION. ALL ORGANIC BEINGS OCCA-

SIONALLY INTERCROSS; APPARENT EX-CEPTIONS. ON CERTAIN CHARACTERSINCAPABLE OF FUSION; CHIEFLY OR EX-CLUSIVELY THOSE WHICH HAVE SUD-DENLY APPEARED IN THE INDIVIDUAL.ON THE MODIFICATION OF OLD RACES,AND THE FORMATION OF NEW RACES BYCROSSING. SOME CROSSED RACES HAVEBRED TRUE FROM THEIR FIRST PRODUC-TION. ON THE CROSSING OF DISTINCTSPECIES IN RELATION TO THE FORMA-TION OF DOMESTIC RACES.

In the two previous chapters, when discussingreversion and prepotency, I was necessarily ledto give many facts on crossing. In the presentchapter I shall consider the part which crossingplays in two opposed directions,—firstly, inobliterating characters, and consequently inpreventing the formation of new races; andsecondly, in the modification of old races, or inthe formation of new and intermediate races,by a combination of characters. I shall also

show that certain characters are incapable offusion.

The effects of free or uncontrolled breedingbetween the members of the same variety or ofclosely allied varieties are important; but are soobvious that they need not be discussed atmuch length. It is free intercrossing which chie-fly gives uniformity, both under nature andunder domestication, to the individuals of thesame species or variety, when they live min-gled together and are not exposed to any causeinducing excessive variability. The preventionof free crossing, and the intentional matching ofindividual animals, are the corner-stones of thebreeder's art. No man in his senses would ex-pect to improve or modify a breed in any par-ticular manner, or keep an old breed true anddistinct, unless he separated his animals. Thekilling of inferior animals in each generationcomes to the same thing as their separation. Insavage and semi-civilised countries, where the

inhabitants have not the means of separatingtheir animals, more than a single breed of thesame species rarely or never exists. In formertimes, even in the United States, there were nodistinct races of sheep, for all had been mingledtogether. (15/1. 'Communications to the Boardof Agriculture' volume 1 page 367.) The cele-brated agriculturist Marshall (15/2. 'Review ofReports, North of England' 1808 page 200.) re-marks that "sheep that are kept within fences,as well as shepherded flocks in open countries,have generally a similarity, if not a uniformity,of character in the individuals of each flock;"for they breed freely together, and are pre-vented from crossing with other kinds; whereasin the unenclosed parts of England the unshep-herded sheep, even of the same flock, are farfrom true or uniform, owing to various breedshaving mingled and crossed. We have seen thatthe half-wild cattle in each of the several Britishparks are nearly uniform in character; but inthe different parks, from not having mingled

and crossed during many generations, theydiffer to a certain small extent.

We cannot doubt that the extraordinary num-ber of varieties and sub-varieties of the pigeon,amounting to at least one hundred and fifty, ispartly due to their remaining, differently fromother domesticated birds, paired for life oncematched. On the other hand, breeds of cats im-ported into this country soon disappear, fortheir nocturnal and rambling habits render ithardly possible to prevent free crossing. Reng-ger (15/3. 'Saugethiere von Paraguay' 1830 s.212.) gives an interesting case with respect tothe cat in Paraguay: in all the distant parts ofthe kingdom it has assumed, apparently fromthe effects of the climate, a peculiar character,but near the capital this change has been pre-vented, owing, as he asserts, to the native ani-mal frequently crossing with cats importedfrom Europe. In all cases like the foregoing, theeffects of an occasional cross will be augmented

by the increased vigour and fertility of the cros-sed offspring, of which fact evidence will here-after be given; for this will lead to the mongrelsincreasing more rapidly than the pure parent-breeds.

When distinct breeds are allowed to cross free-ly, the result will be a heterogeneous body; forinstance, the dogs in Paraguay are far from uni-form, and can no longer be affiliated to theirparent-races. (15/4. Rengger 'Saugethiere' etc. s.154.) The character which a crossed body ofanimals will ultimately assume must dependon several contingencies,—namely, on the rela-tive members of the individuals belonging tothe two or more races which are allowed tomingle; on the prepotency of one race over theother in the transmission of character; and onthe conditions of life to which they are exposed.When two commingled breeds exist at first innearly equal numbers, the whole will sooner orlater become intimately blended, but not so

soon, both breeds being equally favoured in allrespects, as might have been expected. The fol-lowing calculation (15/5. White 'Regular Gra-dation in Man' page 146.) shows that this is thecase: if a colony with an equal number of blackand white men were founded, and we assumethat they marry indiscriminately, are equallyprolific, and that one in thirty annually diesand is born; then "in 65 years the number ofblacks, whites, and mulattoes would be equal.In 91 years the whites would be 1-10th, theblacks 1-10th, and the mulattoes, or people ofintermediate degrees of colour, 8-10ths of thewhole number. In three centuries not 1-100thpart of the whites would exist."

When one of two mingled races exceed the ot-her greatly in number, the latter will soon bewholly, or almost wholly, absorbed and lost.(15/6. Dr. W.F. Edwards in his 'Caracteres Phy-siolog. des Races Humaines' page 24 first calledattention to this subject, and ably discussed it.)

Thus European pigs and dogs have been large-ly introduced in the islands of the PacificOcean, and the native races have been absorbedand lost in the course of about fifty or sixtyyears (15/7. Rev. D. Tyerman and Bennett'Journal of Voyages' 1821-1829 volume 1 page300.); but the imported races no doubt werefavoured. Rats may be considered as semi-domesticated animals. Some snake-rats (Musalexandrinus) escaped in the Zoological Gar-dens of London "and for a long time afterwardsthe keepers frequently caught cross-bred rats,at first half-breds, afterwards with less of thecharacter of the snake-rat, till at length all tracesof it disappeared." (15/8. Mr. S.J. Salter 'JournalLinn. Soc.' volume 6 1862 page 71.) On the ot-her hand, in some parts of London, especiallynear the docks, where fresh rats are frequentlyimported, an endless variety of intermediateforms may be found between the brown, black,and snake rat, which are all three usually ran-ked as distinct species.

How many generations are necessary for onespecies or race to absorb another by repeatedcrosses has often been discussed (15/9. Sturm'Ueber Racen, etc.' 1825 s. 107. Bronn 'Geschich-te der Natur' b. 2 s. 170 gives a table of the pro-portions of blood after successive crosses. Dr. P.Lucas 'L'Heredite Nat.' tome 2 page 308.); andthe requisite number has probably been muchexaggerated. Some writers have maintainedthat a dozen or score, or even more genera-tions, are necessary; but this in itself is improb-able, for in the tenth generation there would beonly 1-1024th part of foreign blood in the off-spring. Gartner found (15/10. 'Bastarder-zeugung' s. 463, 470.), that with plants, one spe-cies could be made to absorb another in fromthree to five generations, and he believes thatthis could always be effected in from six to se-ven generations. In one instance, however, Kol-reuter (15/11. 'Nova Acta Petrop.' 1794 page393: see also previous volume.) speaks of theoffspring of Mirabilis vulgaris, crossed during

eight successive generations by M. longiflora,as resembling this latter species so closely, thatthe most scrupulous observer could detect "vixaliquam notabilem differentiam" or, as he says,he succeeded, "ad plenariam fere transmuta-tionem." But this expression shows that the actof absorption was not even then absolutelycomplete, though these crossed plants con-tained only the 1-256th part of M. vulgaris. Theconclusions of such accurate observers as Gart-ner and Kolreuter are of far higher worth thanthose made without scientific aim by breeders.The most precise account which I have metwith is given by Stonehenge (15/12. 'The Dog'1867 pages 179-184.) and is illustrated by pho-tographs. Mr. Hanley crossed a greyhoundbitch with a bulldog; the offspring in each suc-ceeding generation being recrossed with first-rate greyhounds. As Stonehenge remarks, itmight naturally be supposed that it would takeseveral crosses to get rid of the heavy form ofthe bulldog; but Hysterics, the gr-gr-

granddaughter of a bulldog, showed no tracewhatever of this breed in external form. Sheand all of the same litter, however, were "re-markably deficient in stoutness, though fast aswell as clever." I believe clever refers to skill inturning. Hysterics was put to a son of Bedlam-ite, "but the result of the fifth cross is not as yet,I believe, more satisfactory than that of thefourth." On the other hand, with sheep,Fleischmann (15/13. As quoted in the 'TruePrinciples of Breeding' by C.H. Macknight andDr. H. Madden 1865 page 11.) shows how per-sistent the effects of a single cross may be: hesays "that the original coarse sheep (of Ger-many) have 5500 fibres of wool on a squareinch; grades of the third or fourth Merino crossproduced about 8000, the twentieth cross27,000, the perfect pure Merino blood 40,000 to48,000." So that common German sheep crossedtwenty times successively with Merino did notby any means acquire wool as fine as that of thepure breed. But in all cases, the rate of absorp-

tion will depend largely on the conditions oflife being favourable to any particular charac-ter; and we may suspect that there would be aconstant tendency to degeneration in the woolof Merinos under the climate of Germany, un-less prevented by careful selection; and thusperhaps the foregoing remarkable case may beexplained. The rate of absorption must alsodepend on the amount of distinguishable dif-ference between the two forms which are cros-sed, and especially, as Gartner insists, on pre-potency of transmission in the one form overthe other. We have seen in the last chapter thatone of two French breeds of sheep yielded upits character, when crossed with Merinos, verymuch more slowly than the other; and thecommon German sheep referred to by Fleisch-mann may be in this respect analogous. In allcases there will be more or less liability to re-version during many subsequent generations,and it is this fact which has probably led au-thors to maintain that a score or more of gen-

erations are requisite for one race to absorbanother. In considering the final result of thecommingling of two or more breeds, we mustnot forget that the act of crossing in itself tendsto bring back long-lost characters not proper tothe immediate parent-forms.

With respect to the influence of the conditionsof life on any two breeds which are allowed tocross freely, unless both are indigenous andhave long been accustomed to the country whe-re they live, they will, in all probability, be un-equally affected by the conditions, and this willmodify the result. Even with indigenousbreeds, it will rarely or never occur that bothare equally well adapted to the surroundingcircumstances; more especially when permittedto roam freely, and not carefully tended, as isgenerally the case with breeds allowed to cross.As a consequence of this, natural selection willto a certain extent come into action, and thebest fitted will survive, and this will aid in de-

termining the ultimate character of the com-mingled body.

How long a time it would require before such acrossed body of animals would assume a uni-form character within a limited area, no one cansay; that they would ultimately become uni-form from free intercrossing, and from the sur-vival of the fittest, we may feel assured; but thecharacters thus acquired would rarely or never,as may be inferred from the previous consid-erations, be exactly intermediate between thoseof the two parent-breeds. With respect to thevery slight differences by which the individualsof the same sub-variety, or even of allied varie-ties, are characterised, it is obvious that freecrossing would soon obliterate such small dis-tinctions. The formation of new varieties, inde-pendently of selection, would also thus be pre-vented; except when the same variation con-tinually recurred from the action of some stron-gly predisposing cause. We may therefore con-

clude that free crossing has in all cases playedan important part in giving uniformity of char-acter to all the members of the same domesticrace and of the same natural species, thoughlargely governed by natural selection and bythe direct action of the surrounding conditions.

ON THE POSSIBILITY OF ALL ORGANICBEINGS OCCASIONALLY INTERCROSSING.

But it may be asked, can free crossing occurwith hermaphrodite animals and plants? Allthe higher animals, and the few insects whichhave been domesticated, have separate sexes,and must inevitably unite for each birth. Withrespect to the crossing of hermaphrodites, thesubject is too large for the present volume, butin the 'Origin of Species' I have given a shortabstract of the reasons which induce me to be-lieve that all organic beings occasionally cross,though perhaps in some cases only at long in-tervals of time. (15/14. With respect to plants,

an admirable essay on this subject (Die Gesch-lechter-Vertheilung bei den Pflanzen: 1867) hasbeen published by Dr. Hildebrand who arrivesat the same general conclusions as I have done.Various other treatises have since appeared onthe same subject, more especially by HermannMuller and Delpino.) I will merely recall thefact that many plants, though hermaphrodite instructure, are unisexual in function;—such asthose called by C.K. Sprengel DICHOGA-MOUS, in which the pollen and stigma of thesame flower are matured at different periods;or those called by me RECIPROCALLY DI-MORPHIC, in which the flower's own pollen isnot fitted to fertilise its own stigma; or again,the many kinds in which curious mechanicalcontrivances exist, effectually preventing self-fertilisation. There are, however, many her-maphrodite plants which are not in any wayspecially constructed to favour intercrossing,but which nevertheless commingle almost asfreely as animals with separated sexes. This is

the case with cabbages, radishes, and onions, asI know from having experimented on them:even the peasants of Liguria say that cabbagesmust be prevented "from falling in love" witheach other. In the orange tribe, Gallesio (15/15.'Teoria della Riproduzione Vegetal' 1816 page12.) remarks that the amelioration of the vari-ous kinds is checked by their continual andalmost regular crossing. So it is with numerousother plants.

On the other hand, some cultivated plants rare-ly or never intercross, for instance, the commonpea and sweet-pea (Lathyrus odoratus); yettheir flowers are certainly adapted for crossfertilisation. The varieties of the tomato andaubergine (Solanum) and the pimenta (Pimentavulgaris?) are said (15/16. Verlot 'Des Varietes'1865 page 72.) never to cross, even when grow-ing alongside one another. But it should be ob-served that these are all exotic plants, and wedo not know how they would behave in their

native country when visited by the proper in-sects. With respect to the common pea, I haveascertained that it is rarely crossed in this coun-try owing to premature fertilisation. There ex-ist, however, some plants which under theirnatural conditions appear to be always self-fertilised, such as the Bee Ophrys (Ophrys apif-era) and a few other Orchids; yet these plantsexhibit the plainest adaptations for cross-fertilisation. Again, some few plants are be-lieved to produce only closed flowers, calledcleistogene, which cannot possibly be crossed.This was long thought to be the case with theLeersia oryzoides (15/17. Duval Jouve 'Bull.Soc. Bot. de France' tome 10 1863 page 194.With respect to the perfect flowers setting seedsee Dr. Ascherson in 'Bot. Zeitung' 1864 page350.), but this grass is now known occasionallyto produce perfect flowers, which set seed.

Although some plants, both indigenous andnaturalised, rarely or never produce flowers, or

if they flower never produce seeds, yet no onedoubts that phanerogamic plants are adaptedto produce flowers, and the flowers to produceseed. When they fail, we believe that suchplants under different conditions would per-form their proper function, or that they for-merly did so, and will do so again. On analo-gous grounds, I believe that the flowers in theabove specified anomalous cases which do notnow intercross, either would do so occasionallyunder different conditions, or that they for-merly did so—the means for affecting this be-ing generally still retained—and will again in-tercross at some future period, unless indeedthey become extinct. On this view alone, manypoints in the structure and action of the repro-ductive organs in hermaphrodite plants andanimals are intelligible,—for instance, the factof the male and female organs never being socompletely enclosed as to render access fromwithout impossible. Hence we may concludethat the most important of all the means for

giving uniformity to the individuals of thesame species, namely, the capacity of occasion-ally intercrossing, is present, or has been for-merly present, with all organic beings, except,perhaps, some of the lowest.

[ON CERTAIN CHARACTERS NOTBLENDING.

When two breeds are crossed their charactersusually become intimately fused together; butsome characters refuse to blend, and are trans-mitted in an unmodified state either from bothparents or from one. When grey and white mi-ce are paired, the young are piebald, or purewhite or grey, but not of an intermediate tint;so it is when white and common collared turtle-doves are paired. In breeding Game fowls, agreat authority, Mr. J. Douglas, remarks, "I mayhere state a strange fact: if you cross a blackwith a white game, you get birds of both breedsof the clearest colour." Sir R. Heron crossed

during many years white, black, brown, andfawn-coloured Angora rabbits, and never oncegot these colours mingled in the same animal,but often all four colours in the same litter.(15/18. Extract of a letter from Sir R. Heron1838 given me by Mr. Yarrell. With respect tomice see 'Annal. des Sc. Nat.' tome 1 page 180;and I have heard of other similar cases. Forturtle-doves Boitard and Corbie 'Les Pigeons'etc. page 238. For the Game fowl 'The PoultryBook' 1866 page 128. For crosses of taillessfowls see Bechstein 'Naturges. Deutsch.' b. 3 s.403. Bronn 'Geschichte der Natur' b. 2 s. 170gives analogous facts with horses. On the hair-less condition of crossed South American dogssee Rengger 'Saugethiere von Paraguay' s. 152;but I saw in the Zoological Gardens mongrels,from a similar cross, which were hairless, quitehairy, or hairy in patches, that is, piebald withhair. For crosses of Dorking and other fowls see'Poultry Chronicle' volume 2 page 355. Aboutthe crossed pigs, extract of letter from Sir R.

Heron to Mr. Yarrell. For other cases see P. Lu-cas 'L'Hered. Nat.' tome 1 page 212.) From caseslike these, in which the colours of the two par-ents are transmitted quite separately to the off-spring, we have all sorts of gradations, leadingto complete fusion. I will give an instance: agentleman with a fair complexion, light hairbut dark eyes, married a lady with dark hairand complexion: their three children have verylight hair, but on careful search about a dozenblack hairs were found scattered in the midst ofthe light hair on the heads of all three.

When turnspit dogs and ancon sheep, both ofwhich have dwarfed limbs, are crossed withcommon breeds, the offspring are not interme-diate in structure, but take after either parent.When tailless or hornless animals are crossedwith perfect animals, it frequently, but by nomeans invariably, happens that the offspringare either furnished with these organs in a per-fect state, or are quite destitute of them. Ac-

cording to Rengger, the hairless condition ofthe Paraguay dog is either perfectly or not at alltransmitted to its mongrel offspring; but I haveseen one partial exception in a dog of this par-entage which had part of its skin hairy, andpart naked, the parts being distinctly separatedas in a piebald animal. When Dorking fowlswith five toes are crossed with other breeds, thechickens often have five toes on one foot andfour on the other. Some crossed pigs raised bySir R. Heron between the solid- hoofed andcommon pig had not all four feet in an inter-mediate condition, but two feet were furnishedwith properly divided, and two with unitedhoofs.

Analogous facts have been observed withplants: Major Trevor Clarke crossed the little,glabrous-leaved, annual stock (Matthiola), withpollen of a large, red-flowered, rough-leaved,biennial stock, called cocardeau by the French,and the result was that half the seedlings had

glabrous and the other half rough leaves, butnone had leaves in an intermediate state. Thatthe glabrous seedlings were the product of therough-leaved variety, and not accidentally ofthe mother-plant's own pollen, was shown bytheir tall and strong habit of growth. (15/19.'Internat. Hort. and Bot. Congress of London'1866.) in the succeeding generations raisedfrom the rough-leaved crossed seedlings, someglabrous plants appeared, showing that theglabrous character, though incapable of blend-ing with and modifying the rough leaves, wasall the time latent in this family of plants. Thenumerous plants formerly referred to, which Iraised from reciprocal crosses between the pe-loric and common Antirrhinum, offer a nearlyparallel case; for in the first generation all theplants resembled the common form, and in thenext generation, out of one hundred and thirty-seven plants, two alone were in an intermediatecondition, the others perfectly resembling ei-ther the peloric or common form. Major Trevor

Clarke also fertilised the above-mentioned red-flowered stock with pollen from the purpleQueen stock, and about half the seedlings scar-cely differed in habit, and not at all in the redcolour of the flower, from the mother-plant, theother half bearing blossoms of a rich purple,closely like those of the paternal plant. Gartnercrossed many white and yellow-flowered spe-cies and varieties of Verbascum; and these col-ours were never blended, but the offspring boreeither pure white or pure yellow blossoms; theformer in the larger proportion. (15/20. 'Bas-tarderzeugung' s. 307. Kolreuter 'Dritte Fort-setszung' s. 34, 39 however, obtained interme-diate tints from similar crosses in the genusVerbascum. With respect to the turnips seeHerbert 'Amaryllidaceae' 1837 page 370.) Dr.Herbert raised many seedlings, as he informedme, from Swedish turnips crossed by two othervarieties, and these never produced flowers ofan intermediate tint, but always like one oftheir parents. I fertilised the purple sweet-pea

(Lathyrus odoratus), which has a dark reddish-purple standard-petal and violet-colouredwings and keel, with pollen of the painted ladysweet-pea, which has a pale cherry-colouredstandard, and almost white wings and keel;and from the same pod I twice raised plantsperfectly resembling both sorts; the greaternumber resembling the father. So perfect wasthe resemblance, that I should have thoughtthere had been some mistake, if the plantswhich were at first identical with the paternalvariety, namely, the painted-lady, had not laterin the season produced, as mentioned in a for-mer chapter, flowers blotched and streakedwith dark purple. I raised grandchildren andgreat-grandchildren from these crossed plants,and they continued to resemble the painted-lady, but during later generations becamerather more blotched with purple, yet nonereverted completely to the original mother-plant, the purple sweet-pea. The following caseis slightly different, but still shows the same

principle: Naudin (15/21. 'Nouvelles Archivesdu Museum' tome 1 page 100.) raised numer-ous hybrids between the yellow Linaria vul-garis and the purple L. purpurea, and duringthree successive generations the colours keptdistinct in different parts of the same flower.

From cases such as the foregoing, in which theoffspring of the first generation perfectly re-semble either parent, we come by a small stepto those cases in which differently colouredflowers borne on the same root resemble bothparents, and by another step to those in whichthe same flower or fruit is striped or blotchedwith the two parental colours, or bears a singlestripe of the colour or other characteristic qual-ity of one of the parent-forms. With hybridsand mongrels it frequently or even generallyhappens that one part of the body resemblesmore or less closely one parent and anotherpart the other parent; and here again some re-sistence to fusion, or, what comes to the same

thing, some mutual affinity between the or-ganic atoms of the same nature, apparentlycomes into play, for otherwise all parts of thebody would be equally intermediate in charac-ter. So again, when the offspring of hybrids ormongrels, which are themselves nearly inter-mediate in character, revert either wholly or bysegments to their ancestors, the principle of theaffinity of similar, or the repulsion of dissimilaratoms, must come into action. To this principle,which seems to be extremely general, we shallrecur in the chapter on pangenesis.

It is remarkable, as has been strongly insistedupon by Isidore Geoffroy St. Hilaire in regardto animals, that the transmission of characterswithout fusion occurs very rarely when speciesare crossed; I know of one exception alone, na-mely, with the hybrids naturally produced be-tween the common and hooded crow (Corvuscorone and cornix), which, however, are closelyallied species, differing in nothing except col-

our. Nor have I met with any well- ascertainedcases of transmission of this kind, even whenone form is strongly prepotent over another,when two races are crossed which have beenslowly formed by man's selection, and there-fore resemble to a certain extent natural species.Such cases as puppies in the same litter closelyresembling two distinct breeds, are probablydue to superfoetation,—that is, to the influenceof two fathers. All the characters above enu-merated, which are transmitted in a perfectstate to some of the offspring and not to oth-ers,— such as distinct colours, nakedness ofskin, smoothness of leaves, absence of horns ortail, additional toes, pelorism, dwarfed struc-ture, etc.,—have all been known to appear sud-denly in individual animals and plants. Fromthis fact, and from the several slight, aggre-gated differences which distinguish domesticraces and species from one another, not beingliable to this peculiar form of transmission, wemay conclude that it is in some way connected

with the sudden appearance of the charactersin question.]

ON THE MODIFICATION OF OLD RACESAND THE FORMATION OF NEW RACES BYCROSSING.

We have hitherto chiefly considered the effectsof crossing in giving uniformity of character;we must now look to an opposite result. Therecan be no doubt that crossing, with the aid ofrigorous selection during several generations,has been a potent means in modifying old ra-ces, and in forming new ones. Lord Orfordcrossed his famous stud of greyhounds oncewith the bulldog, in order to give them courageand perseverance. Certain pointers have beencrossed, as I hear from the Rev. W.D. Fox, withthe foxhound, to give them dash and speed.Certain strains of Dorking fowls have had aslight infusion of Game blood; and I haveknown a great fancier who on a single occasion

crossed his turbit-pigeons with barbs, for thesake of gaining greater breadth of beak.

In the foregoing cases breeds have been crossedonce, for the sake of modifying some particularcharacter; but with most of the improved racesof the pig, which now breed true, there havebeen repeated crosses,—for instance, the im-proved Essex owes its excellence to repeatedcrosses with the Neapolitan, together probablywith some infusion of Chinese blood. (15/22.Richardson 'Pigs' 1847 pages 37, 42; S. Sidney'sedition of 'Youatt on the Pig' 1860 page 3.) Sowith our British sheep: almost all the races, ex-cept the Southdown, have been largely crossed;"this, in fact, has been the history of our princi-pal breeds." (15/23. See Mr. W.C. Spooner'sexcellent paper on Cross-Breeding 'Journal Ro-yal Agricult. Soc.' volume 20 part 2: see also anequally good article by Mr. Ch. Howard in'Gardener's Chronicle' 1860 page 320.) To givean example, the "Oxfordshire Downs" now

rank as an established breed. (15/24. 'Gar-dener's Chronicle' 1857 pages 649, 652.) Theywere produced about the year 1830 by crossing"Hampshire and in some instances Southdownewes with Cotswold rams:" now the Hamp-shire ram was itself produced by repeated cros-ses between the native Hampshire sheep andSouthdowns; and the long-woolled Cotswoldwere improved by crosses with the Leicester,which latter again is believed to have been across between several long-woolled sheep. Mr.Spooner, after considering the various caseswhich have been carefully recorded, concludes,"that from a judicious pairing of cross-bredanimals it is practicable to establish a newbreed." On the continent the history of severalcrossed races of cattle and of other animals hasbeen well ascertained. To give one instance: theKing of Wurtemburg, after twenty-five years'careful breeding, that is, after six or seven gen-erations, made a new breed of cattle from across between a Dutch and a Swiss breed, com-

bined with other breeds. (15/25. 'Bulletin de LaSoc. d'Acclimat.' 1862 tome 9 page 463. See alsofor other cases MM. Moll and Gayot 'Du Boeuf'1860 page 32.) The Sebright bantam, whichbreeds as true as any other kind of fowl, wasformed about sixty years ago by a complicatedcross. (15/26. 'Poultry Chronicle' volume 2 1854page 36.) Dark Brahmas, which are believed bysome fanciers to constitute a distinct species,were undoubtedly formed (15/27. 'The PoultryBook' by W.B. Tegetmeier 1866 page 58.) in theUnited States, within a recent period, by a crossbetween Chittagongs and Cochins. With plantsthere is little doubt that the Swede-turnip ori-ginated from a cross; and the history of a vari-ety of wheat, raised from two very distinct va-rieties, and which after six years' culture pre-sented an even sample, has been recorded ongood authority. (15/28. 'Gardener's Chronicle'1852 page 765.)

Until lately, cautious and experienced breeders,though not averse to a single infusion of for-eign blood, were almost universally convincedthat the attempt to establish a new race, inter-mediate between two widely distinct races, washopeless "they clung with superstitious tenacityto the doctrine of purity of blood, believing it tobe the ark in which alone true safety could befound." (15/29. Spooner in 'Journal Royal Agri-cult. Soc.' volume 20 part 2) Nor was this con-viction unreasonable: when two distinct racesare crossed, the offspring of the first generationare generally nearly uniform in character; buteven this sometimes fails to be the case, espe-cially with crossed dogs and fowls, the youngof which from the first are sometimes muchdiversified. As cross-bred animals are generallyof large size and vigorous, they have been rai-sed in great numbers for immediate consump-tion. But for breeding they are found utterlyuseless; for though they may themselves beuniform in character, they yield during many

generations astonishingly diversified offspring.The breeder is driven to despair, and concludesthat he will never form an intermediate race.But from the cases already given, and fromothers which have been recorded, it appearsthat patience alone is necessary; as Mr. Spoonerremarks, "nature opposes no barrier to success-ful admixture; in the course of time, by the aidof selection and careful weeding, it is practica-ble to establish a new breed." After six or sevengenerations the hoped-for result will in mostcases be obtained; but even then an occasionalreversion, or failure to keep true, may be ex-pected. The attempt, however, will assuredlyfail if the conditions of life be decidedly unfa-vourable to the characters of either parent-breed. (15/30. See Colin 'Traite de Phys. Comp.des Animaux Domestiques' tome 2 page 536,where this subject is well treated.)

Although the grandchildren and succeedinggenerations of cross-bred animals are generally

variable in an extreme degree, some curiousexceptions to the rule have been observed bothwith crossed races and species. Thus Boitardand Corbie (15/31. 'Les Pigeons' page 37.) as-sert that from a Pouter and a Runt "a Cavalierwill appear, which we have classed amongstpigeons of pure race, because it transmits all itsqualities to its posterity." The editor of the'Poultry Chronicle' (15/32. Volume 1 1854 page101.) bred some bluish fowls from a black Span-ish cock and a Malay hen; and these remainedtrue to colour "generation after generation." TheHimalayan breed of rabbits was certainly for-med by crossing two sub-varieties of the silver-grey rabbit; although it suddenly assumed itspresent character, which differs much from thatof either parent-breed, yet it has ever since beeneasily and truly propagated. I crossed someLabrador and Penguin ducks, and recrossed themongrels with Penguins; afterwards most ofthe ducks reared during three generations werenearly uniform in character, being brown with

a white crescentic mark on the lower part of thebreast, and with some white spots at the base ofthe beak; so that by the aid of a little selection anew breed might easily have been formed.With regard to crossed varieties of plants, Mr.Beaton (15/33. 'Cottage Gardener' 1856 page110.) remarks that "Melville's extraordinarycross between the Scotch kale and an early cab-bage is as true and genuine as any on record;"but in this case no doubt selection was prac-tised. Gartner (15/34. 'Bastarderzeugung' s.553.) has given five cases of hybrids, in whichthe progeny kept constant; and hybrids be-tween Dianthus armeria and deltoides re-mained true and uniform to the tenth genera-tion. Dr. Herbert likewise showed me a hybridfrom two species of Loasa which from its firstproduction had kept constant during severalgenerations.

We have seen in the first chapter, that the sev-eral kinds of dogs are almost certainly de-

scended from more than one species, and so itis with cattle, pigs and some other domesti-cated animals. Hence the crossing of aborigi-nally distinct species probably came into playat an early period in the formation of our pre-sent races. From Rutimeyer's observations therecan be little doubt that this occurred with cattle;but in most cases one form will probably haveabsorbed and obliterated the other, for it is notlikely that semi-civilised men would have ta-ken the necessary pains to modify by selectiontheir commingled, crossed, and fluctuatingstock. Nevertheless, those animals which werebest adapted to their conditions of life wouldhave survived through natural selection; andby this means crossing will often have indi-rectly aided in the formation of primeval do-mesticated breeds. Within recent times, as faras animals are concerned, the crossing of dis-tinct species has done little or nothing towardsthe formation or modification of our races. It isnot yet known whether the several species of

silk-moth which have been recently crossed inFrance will yield permanent races. With plantswhich can be multiplied by buds and cuttings,hybridisation has done wonders, as with manykinds of Roses, Rhododendrons, Pelargoniums,Calceolarias, and Petunias. Nearly all theseplants can be propagated by seed, most of themfreely; but extremely few or none come true byseed.

Some authors believe that crossing is the chiefcause of variability,—that is, of the appearanceof absolutely new characters. Some have goneso far as to look at it as the sole cause; but thisconclusion is disproved by the facts given inthe chapter on Bud-variation. The belief thatcharacters not present in either parent or intheir ancestors frequently originate from cross-ing is doubtful; that they occasionally do so isprobable; but this subject will be more conven-iently discussed in a future chapter on thecauses of Variability.

A condensed summary of this and of the threefollowing chapters, together with some re-marks on Hybridism, will be given in the nine-teenth chapter.

CHAPTER 2.XVI.

CAUSES WHICH INTERFERE WITH THEFREE CROSSING OF VARIETIES—INFLUENCE OF DOMESTICATION ON FER-TILITY.

DIFFICULTIES IN JUDGING OF THE FER-TILITY OF VARIETIES WHEN CROSSED.VARIOUS CAUSES WHICH KEEP VARIETIESDISTINCT, AS THE PERIOD OF BREEDINGAND SEXUAL PREFERENCE. VARIETIES OFWHEAT SAID TO BE STERILE WHEN CROS-SED. VARIETIES OF MAIZE, VERBASCUM,HOLLYHOCK, GOURDS, MELONS, ANDTOBACCO, RENDERED IN SOME DEGREE

MUTUALLY STERILE. DOMESTICATIONELIMINATES THE TENDENCY TO STERIL-ITY NATURAL TO SPECIES WHENCROSSED. ON THE INCREASED FERTILITYOF UNCROSSED ANIMALS AND PLANTSFROM DOMESTICATION AND CULTIVA-TION.

The domesticated races of both animals andplants, when crossed, are, with extremely fewexceptions, quite prolific,—in some cases evenmore so than the purely-bred parent-races. Theoffspring, also, raised from such crosses arelikewise, as we shall see in the following chap-ter, generally more vigorous and fertile thantheir parents. On the other hand, species whencrossed, and their hybrid offspring, are almostinvariably in some degree sterile; and here the-re seems to exist a broad and insuperable dis-tinction between races and species. The impor-tance of this subject as bearing on the origin ofspecies is obvious; and we shall hereafter recurto it.

It is unfortunate how few precise observationshave been made on the fertility of mongrel ani-mals and plants during several successive gen-erations. Dr. Broca (16/1. 'Journal de Physi-olog.' tome 2 1859 page 385.) has remarked thatno one has observed whether, for instance,mongrel dogs, bred inter se, are indefinitelyfertile; yet, if a shade of infertility be detectedby careful observation in the offspring of natu-ral forms when crossed, it is thought that theirspecific distinction is proved. But so manybreeds of sheep, cattle, pigs, dogs, and poultry,have been crossed and recrossed in variousways, that any sterility, if it had existed, wouldfrom being injurious almost certainly have beenobserved. In investigating the fertility of cros-sed varieties many sources of doubt occur.Whenever the least trace of sterility betweentwo plants, however closely allied, was ob-served by Kolreuter, and more especially byGartner, who counted the exact number of seedin each capsule, the two forms were at once

ranked as distinct species; and if this rule befollowed, assuredly it will never be proved thatvarieties when crossed are in any degree sterile.We have formerly seen that certain breeds ofdogs do not readily pair together; but no obser-vations have been made whether, when paired,they produce the full number of young, andwhether the latter are perfectly fertile inter se;but, supposing that some degree of sterilitywere found to exist, naturalists would simplyinfer that these breeds were descended fromaboriginally distinct species; and it would bescarcely possible to ascertain whether or notthis explanation was the true one.

The Sebright Bantam is much less prolific thanany other breed of fowls, and is descendedfrom a cross between two very distinct breeds,recrossed by a third sub-variety. But it wouldbe extremely rash to infer that the loss of fertil-ity was in any manner connected with its cros-sed origin, for it may with more probability be

attributed either to long-continued close inter-breeding, or to an innate tendency to sterilitycorrelated with the absence of hackles and sic-kle tail-feathers.

Before giving the few recorded cases of forms,which must be ranked as varieties, being insome degree sterile when crossed, I may re-mark that other causes sometimes interferewith varieties freely intercrossing. Thus theymay differ too greatly in size, as with somekinds of dogs and fowls: for instance, the editorof the 'Journal of Horticulture, etc.' (16/2. De-cember 1863 page 484.) says that he can keepBantams with the larger breeds without muchdanger of their crossing, but not with the sma-ller breeds, such as Games, Hamburghs, etc.With plants a difference in the period of flower-ing serves to keep varieties distinct, as with thevarious kinds of maize and wheat: thus ColonelLe Couteur (16/3. On 'The Varieties of Wheat'page 66.) remarks, "the Talavera wheat, from

flowering much earlier than any other kind, issure to continue pure." In different parts of theFalkland Islands the cattle are breaking up intoherds of different colours; and those on thehigher ground, which are generally white, usu-ally breed, as I am informed by Sir J. Sulivan,three months earlier than those on the lowland;and this would manifestly tend to keep theherds from blending.

Certain domestic races seem to prefer breedingwith their own kind; and this is a fact of someimportance, for it is a step towards that instinc-tive feeling which helps to keep closely alliedspecies in a state of nature distinct. We havenow abundant evidence that, if it were not forthis feeling, many more hybrids would be na-turally produced than in this case. We haveseen in the first chapter that the alco dog ofMexico dislikes dogs of other breeds; and thehairless dog of Paraguay mixes less readilywith the European races, than the latter do with

each other. In Germany the female Spitz-dog issaid to receive the fox more readily than willother dogs; a female Australian Dingo in Eng-land attracted the wild male foxes. But thesedifferences in the sexual instinct and attractivepower of the various breeds may be wholly dueto their descent from distinct species. In Para-guay the horses have much freedom, and anexcellent observer (16/4. Rengger 'Saugethierevon Paraguay' s. 336.) believes that the nativehorses of the same colour and size prefer asso-ciating with each other, and that the horseswhich have been imported from Entre Rios andBanda Oriental into Paraguay likewise preferassociating together. In Circassia six sub-racesof the horse have received distinct names; and anative proprietor of rank (16/5. See a memoirby MM. Lherbette and De Quatrefages in 'Bull.Soc. d'Acclimat.' tome 8 July 1861 page 312.)asserts that horses of three of these races, whilstliving a free life, almost always refuse to mingleand cross, and will even attack one another.

It has been observed, in a district stocked withheavy Lincolnshire and light Norfolk sheep,that both kinds; though bred together, whenturned out, "in a short time separate to asheep;" the Lincolnshires drawing off to therich soil, and the Norfolks to their own drylight soil; and as long as there is plenty of grass,"the two breeds keep themselves as distinct asrooks and pigeons." In this case different habitsof life tend to keep the races distinct. On one ofthe Faroe islands, not more than half a mile indiameter, the half-wild native black sheep aresaid not to have readily mixed with the im-ported white sheep. It is a more curious factthat the semi-monstrous ancon sheep of mod-ern origin "have been observed to keep to-gether, separating themselves from the rest ofthe flock, when put into enclosures with othersheep." (16/6. For the Norfolk sheep see Mar-shall 'Rural Economy of Norfolk' volume 2page 136. See Rev. L. Landt 'Description ofFaroe' page 66. For the ancon sheep see 'Phil.

Transact.' 1813 page 90.) With respect to fallow-deer, which live in a semi-domesticated condi-tion, Mr. Bennett (16/7. White 'Nat. Hist. ofSelbourne' edited by Bennett page 39. Withrespect to the origin of the dark-coloured deersee 'Some Account of English Deer Parks' byE.P. Shirley, Esq.) states that the dark and palecoloured herds, which have long been kepttogether in the Forest of Dean, in HighMeadow Woods, and in the New Forest, havenever been known to mingle: the dark-coloureddeer, it may be added, are believed to havebeen first brought by James I. from Norway, onaccount of their greater hardiness. I importedfrom the island of Porto Santo two of the feralrabbits, which differ, as described in the fourthchapter, from common rabbits; both proved tobe males, and, though they lived during someyears in the Zoological Gardens, the superin-tendent, Mr. Bartlett, in vain endeavoured tomake them breed with various tame kinds; butwhether this refusal to breed was due to any

change in the instinct, or simply to their ex-treme wildness, or whether confinement hadrendered them sterile, as often occurs, cannotbe determined.

Whilst matching for the sake of experimentmany of the most distinct breeds of pigeons, itfrequently appeared to me that the birds,though faithful to their marriage vow, retainedsome desire after their own kind. Accordingly Iasked Mr. Wicking, who has kept a larger stockof various breeds together than any man inEngland, whether he thought that they wouldprefer pairing with their own kind, supposingthat there were males and females enough ofeach; and he without hesitation answered thathe was convinced that this was the case. It hasoften been noticed that the dovecote pigeonseems to have an actual aversion towards theseveral fancy breeds (16/8. 'The Dovecote' bythe Rev. E.S. Dixon page 155; Bechstein 'Natur-gesch. Deutschlands' b. 4 1795 page 17.) yet all

have certainly sprung from a common progeni-tor. The Rev. W.D. Fox informs me that hisflocks of white and common Chinese geesekept distinct.

These facts and statements, though some ofthem are incapable of proof, resting only on theopinion of experienced observers, show thatsome domestic races are led by different habitsof life to keep to a certain extent separate, andthat others prefer coupling with their ownkind, in the same manner as species in a state ofnature, though in a much less degree.

[With respect to sterility from the crossing ofdomestic races, I know of no well-ascertainedcase with animals. This fact, seeing the greatdifference in structure between some breeds ofpigeons, fowls, pigs, dogs, etc., is extraordi-nary, in contrast with the sterility of manyclosely allied natural species when crossed; butwe shall hereafter attempt to show that it is notso extraordinary as it at first appears. And it

may be well here to recall to mind that theamount of external difference between twospecies is not a safe guide for predictingwhether or not they will breed together,—someclosely allied species when crossed being ut-terly sterile, and others which are extremelyunlike being moderately fertile. I have said thatno case of sterility in crossed races rests on sat-isfactory evidence; but here is one which at firstseems trustworthy. Mr. Youatt (16/9. 'Cattle'page 202.) and a better authority cannot bequoted, states, that formerly in Lancashirecrosses were frequently made between long-horn and shorthorn cattle; the first cross wasexcellent, but the produce was uncertain; in thethird or fourth generation the cows were badmilkers; "in addition to which, there was muchuncertainty whether the cows would conceive;and full one-third of the cows among some ofthese half-breds failed to be in calf." This at firstseems a good case: but Mr. Wilkinson states(16/10. Mr. J. Wilkinson in 'Remarks addressed

to Sir J. Sebright' 1820 page 38.), that a breedderived from this same cross was actually es-tablished in another part of England; and if ithad failed in fertility, the fact would surelyhave been noticed. Moreover, supposing thatMr. Youatt had proved his case, it might beargued that the sterility was wholly due to thetwo parent-breeds being descended from pri-mordially distinct species.

In the case of plants Gartner states that he fertil-ised thirteen heads (and subsequently nine oth-ers) on a dwarf maize bearing yellow seed(16/11. 'Bastarderzeugung' s. 87, 169. See alsothe Table at the end of volume.) with pollen ofa tall maize having red seed; and one head alo-ne produced good seed, but only five in num-ber. Though these plants are monoecious, andtherefore do not require castration, yet I shouldhave suspected some accident in the manipula-tion, had not Gartner expressly stated that hehad during many years grown these two varie-

ties together, and they did not spontaneouslycross; and this, considering that the plants aremonoecious and abound with pollen, and arewell known generally to cross freely, seemsexplicable only on the belief that these two va-rieties are in some degree mutually infertile.The hybrid plants raised from the above fiveseeds were intermediate in structure, extremelyvariable, and perfectly fertile. (16/12. 'Bas-tarderzeugung' s. 87, 577.) In like manner Prof.Hildebrand (16/13. 'Bot. Zeitung' 1868 page327.) could not succeed in fertilising the femaleflowers of a plant bearing brown grains withpollen from a certain kind bearing yellowgrains; although other flowers on the sameplant, which were fertilised with their own pol-len, yielded good seed. No one, I believe, evensuspects that these varieties of maize are dis-tinct species; but had the hybrids been in theleast sterile, no doubt Gartner would at oncehave so classed them. I may here remark, thatwith undoubted species there is not necessarily

any close relation between the sterility of a firstcross and that of the hybrid offspring. Somespecies can be crossed with facility, but pro-duce utterly sterile hybrids; others can becrossed with extreme difficulty, but the hybridswhen produced are moderately fertile. I am notaware, however, of any instance quite like thisof the maize, namely, of a first cross made withdifficulty, but yielding perfectly fertile hybrids.(16/14. Mr. Shirreff formerly thought ('Gar-dener's Chronicle' 1858 page 771) that the off-spring from a cross between certain varieties ofwheat became sterile in the fourth generation;but he now admits ('Improvement of the Cere-als' 1873) that this was an error.)

The following case is much more remarkable,and evidently perplexed Gartner, whose strongwish it was to draw a broad line of distinctionbetween species and varieties. In the genusVerbascum, he made, during eighteen years, avast number of experiments, and crossed no

less than 1085 flowers and counted their seeds.Many of these experiments consisted in cross-ing white and yellow varieties of both V. lych-nitis and V. blattaria with nine other speciesand their hybrids. That the white and yellowflowered plants of these two species are reallyvarieties, no one has doubted; and Gartner ac-tually raised in the case of both species onevariety from the seed of the other. Now in twoof his works (16/15. 'Kenntniss der Befruch-tung' s. 137; 'Bastarderzeugung' s. 92, 181. Onraising the two varieties from seed see s. 307.)he distinctly asserts that crosses between simi-larly-coloured flowers yield more seed thanbetween dissimilarly-coloured; so that the yel-low-flowered variety of either species (andconversely with the white-flowered variety),when crossed with pollen of its own kind,yields more seed than when crossed with thatof the white variety; and so it is when differ-ently coloured species are crossed. The generalresults may be seen in the Table at the end of

his volume. In one instance he gives (16/16.'Bastarderzeugung' s. 216.) the following de-tails; but I must premise that Gartner, to avoidexaggerating the degree of sterility in his cros-ses, always compares the MAXIMUM numberobtained from a cross with the AVERAGEnumber naturally given by the pure mother-plant. The white variety of V. lychnitis, natu-rally fertilised by its own pollen, gave from anAVERAGE of twelve capsules ninety-six goodseeds in each; whilst twenty flowers fertilisedwith pollen from the yellow variety of this sa-me species, gave as the MAXIMUM only eigh-ty-nine good seeds; so that we have the propor-tion of 1000 to 908, according to Gartner's usualscale. I should have thought it possible that sosmall a difference in fertility might have beenaccounted for by the evil effects of the neces-sary castration; but Gartner shows that thewhite variety of V. lychnitis, when fertilisedfirst by the white variety of V. blattaria, andthen by the yellow variety of this species, yiel-

ded seed in the proportion of 622 to 438; and inboth these cases castration was performed.Now the sterility which results from the cross-ing of the differently coloured varieties of thesame species, is fully as great as that whichoccurs in many cases when distinct species arecrossed. Unfortunately Gartner compared theresults of the first unions alone, and not thesterility of the two sets of hybrids producedfrom the white variety of V. lychnitis whenfertilised by the white and yellow varieties ofV. blattaria, for it is probable that they wouldhave differed in this respect.

Mr. J. Scott has given me the results of a seriesof experiments on Verbascum, made by him inthe Botanic Gardens of Edinburgh. (16/17. Theresults have since been published in 'Journ.Asiatic Soc. of Bengal' 1867 page 145.) He re-peated some of Gartner's experiments on dis-tinct species, but obtained only fluctuating re-sults, some confirmatory, the greater number

contradictory; nevertheless these seem hardlysufficient to overthrow the conclusion arrivedat by Gartner from experiments tried on a lar-ger scale. Mr. Scott also experimented on therelative fertility of unions between similarlyand dissimilarly-coloured varieties of the samespecies. Thus he fertilised six flowers of theyellow variety of V. lychnitis by its own pollen,and obtained six capsules; and calling, for thesake of comparison, the average number ofgood seed in each of their capsules one hun-dred, he found that this same yellow variety,when fertilised by the white variety, yieldedfrom seven capsules an average of ninety-fourseed. On the same principle, the white varietyof V. lychnitis by its own pollen (from six cap-sules), and by the pollen of the yellow variety(eight capsules), yielded seed in the proportionof 100 to 82. The yellow variety of V. thapsusby its own pollen (eight capsules), and by thatof the white variety (only two capsules), yiel-ded seed in the proportion of 100 to 94. Lastly,

the white variety of V. blattaria by its own pol-len (eight capsules), and by that of the yellowvariety (five capsules), yielded seed in the pro-portion of 100 to 79. So that in every case theunions of similarly-coloured varieties of thesame species were more fertile than the unionsof dissimilarly-coloured varieties; when all thecases are grouped together, the difference offertility is as 100 to 86. Some additional trialswere made, and altogether thirty-six similarly-coloured unions yielded thirty-five good cap-sules; whilst thirty-five dissimilarly- colouredunions yielded only twenty-six good capsules.Besides the foregoing experiments, the purpleV. phoeniceum was crossed by a rose-colouredand a white variety of the same species; thesetwo varieties were also crossed together, andthese several unions yielded less seed than V.phoeniceum by its own pollen. Hence it followsfrom Mr. Scott's experiments, that in the genusVerbascum the similarly and dissimilarly-coloured varieties of the same species behave,

when crossed, like closely allied but distinctspecies. (16/18. The following facts, given byKolreuter in his 'Dritte Fortsetzung' ss. 34, 39,appear at first sight strongly to confirm Mr.Scott's and Gartner's statements; and to a cer-tain limited extent they do so. Kolreuter asserts,from innumerable observations, that insectsincessantly carry pollen from one species andvariety of Verbascum to another; and I can con-firm this assertion; yet he found that the whiteand yellow varieties of Verbascum lychnitisoften grew wild mingled together: moreover,he cultivated these two varieties in consider-able numbers during four years in his garden,and they kept true by seed; but when he cros-sed them, they produced flowers of an inter-mediate tint. Hence it might have been thoughtthat both varieties must have a stronger electiveaffinity for the pollen of their own variety thanfor that of the other; this elective affinity, I mayadd of each species for its own pollen (Kol-reuter 'Dritte Forts.' s. 39 and Gartner 'Bas-

tarderz.' passim) being a perfectly well-ascertained power. But the force of the forego-ing facts is much lessened by Gartner's numer-ous experiments, for, differently from Kol-reuter, he never once got ('Bastarderz.' s. 307)an intermediate tint when he crossed the yel-low and white flowered varieties of Verbas-cum. So that the fact of the white and yellowvarieties keeping true to their colour by seeddoes not prove that they were not mutuallyfertilised by the pollen carried by insects fromone to the other.)

This remarkable fact of the sexual affinity ofsimilarly-coloured varieties, as observed byGartner and Mr. Scott, may not be of very rareoccurrence; for the subject has not been at-tended to by others. The following case isworth giving, partly to show how difficult it isto avoid error. Dr. Herbert (16/19. 'Amaryllida-ceae' 1837 page 366. Gartner has made a similarobservation.) has remarked that variously-

coloured double varieties of the Hollyhock (Al-thea rosea) may be raised with certainty byseed from plants growing close together. I havebeen informed that nurserymen who raise seedfor sale do not separate their plants; accord-ingly I procured seed of eighteen named varie-ties; of these, eleven varieties produced sixty-two plants all perfectly true to their kind; andseven produced forty-nine plants, half of whichwere true and half false. Mr. Masters of Canter-bury has given me a more striking case; he sa-ved seed from a great bed of twenty-four na-med varieties planted in closely adjoining rows,and each variety reproduced itself truly withonly sometimes a shade of difference in tint.Now in the hollyhock the pollen, which isabundant, is matured and nearly all shed be-fore the stigma of the same flower is ready toreceive it (16/20. Kolreuter first observed thisfact, 'Mem. de l'Acad. de St. Petersburg' volume3 page 127. See also C.K. Sprengel 'Das Ent-deckte Geheimniss' s. 345.); and as bees covered

with pollen incessantly fly from plant to plant,it would appear that adjoining varieties couldnot escape being crossed. As, however, thisdoes not occur, it appeared to me probable thatthe pollen of each variety was prepotent on itsown stigma over that of all other varieties, but Ihave no evidence on this point. Mr. C. Turnerof Slough, well known for his success in thecultivation of this plant, informs me that it isthe doubleness of the flowers which preventsthe bees gaining access to the pollen and stig-ma; and he finds that it is difficult even to crossthem artificially. Whether this explanation willfully account for varieties in close proximitypropagating themselves so truly by seed, I donot know.

The following cases are worth giving, as theyrelate to monoecious forms, which do not re-quire, and consequently cannot have been in-jured by, castration. Girou de Buzareinguescrossed what he designates three varieties of

gourd (16/21. Namely Barbarines, Pastissons,Giraumous: 'Annal. des Sc. Nat.' tome 30 1833pages 398 and 405.), and asserts that their mu-tual fertilisation is less easy in proportion to thedifference which they present. I am aware howimperfectly the forms in this group were untilrecently known; but Sageret (16/22. 'Memoiresur les Cucurbitaceae' 1826 pages 46, 55.), whoranked them according to their mutual fertility,considers the three forms above alluded to asvarieties, as does a far higher authority, name-ly, M. Naudin. (16/23. 'Annales des Sc. Nat.'4th series tome 6. M. Naudin considers theseforms as undoubtedly varieties of Cucurbitapepo.) Sageret (16/24. 'Mem. Cucurb.' page 8.)has observed that certain melons have a greatertendency, whatever the cause may be, to keeptrue than others; and M. Naudin, who has hadsuch immense experience in this group, in-forms me that he believes that certain varietiesintercross more readily than others of the samespecies; but he has not proved the truth of this

conclusion; the frequent abortion of the pollennear Paris being one great difficulty. Neverthe-less, he has grown close together, during sevenyears, certain forms of Citrullus, which, as theycould be artificially crossed with perfect facilityand produced fertile offspring, are ranked asvarieties; but these forms when not artificiallycrossed kept true. Many other varieties, on theother hand, in the same group cross with suchfacility, as M. Naudin repeatedly insists, thatwithout being grown far apart they cannot bekept in the least true.

Another case, though somewhat different, maybe here given, as it is highly remarkable, and isestablished on excellent evidence. Kolreuterminutely describes five varieties of the commontobacco (16/25. 'Zweite Forts.' s. 53 namely Ni-cotiana major vulgaris; (2) perennis; (3) transyl-vanica; (4) a sub- var. of the last; (5) major lati-fol. fl. alb.) which were reciprocally crossed,and the offspring were intermediate in charac-

ter and as fertile as their parents: from this factKolreuter inferred that they are really varieties;and no one, as far as I can discover, seems tohave doubted that such is the case. He alsocrossed reciprocally these five varieties with N.glutinosa, and they yielded very sterile hybrids;but those raised from the var. perennis, whet-her used as the father or mother plant, were notso sterile as the hybrids from the four othervarieties. (16/26. Kolreuter was so much struckwith this fact that he suspected that a little pol-len of N. glutinosa in one of his experimentsmight have accidentally got mingled with thatof var. perennis, and thus aided its fertilisingpower. But we now know conclusively fromGartner ('Bastarderz.' s. 34, 43) that the pollenof two species never acts CONJOINTLY on athird species; still less will the pollen of a dis-tinct species, mingled with a plant's own pol-len, if the latter be present in sufficient quan-tity, have any effect. The sole effect of minglingtwo kinds of pollen is to produce in the same

capsule seeds which yield plants, some takingafter the one and some after the other parent.)So that the sexual capacity of this one varietyhas certainly been in some degree modified, soas to approach in nature that of N. glutinosa.(16/27. Mr. Scott has made some observationson the absolute sterility of a purple and whiteprimrose (Primula vulgaris) when fertilised bypollen from the common primrose ('Journal ofProc. of Linn. Soc.' volume 8 1864 page 98); butthese observations require confirmation. I rai-sed a number of purple-flowered long- styledseedlings from seed kindly sent me by Mr.Scott, and, though they were all in some degreesterile, they were much more fertile with pollentaken from the common primrose than withtheir own pollen. Mr. Scott has likewise de-scribed a red equal-styled cowslip (P. veris ibidpage 106), which was found by him to be high-ly sterile when crossed with the common cow-slip; but this was not the case with severalequal-styled red seedlings raised by me from

his plant. This variety of the cowslip presentsthe remarkable peculiarity of combining maleorgans in every respect like those of the short-styled form, with female organs resembling infunction and partly in structure those of thelong-styled form; so that we have the singularanomaly of the two forms combined in the sa-me flower. Hence it is not surprising that theseflowers should be spontaneously self-fertile in ahigh degree.)

These facts with respect to plants show that insome few cases certain varieties have had theirsexual powers so far modified, that they crosstogether less readily and yield less seed thanother varieties of the same species. We shallpresently see that the sexual functions of mostanimals and plants are eminently liable to beaffected by the conditions of life to which theyare exposed; and hereafter we shall briefly dis-cuss the conjoint bearing of this fact, and oth-

ers, on the difference in fertility between cros-sed varieties and crossed species.

DOMESTICATION ELIMINATES THETENDENCY TO STERILITY WHICH IS GEN-ERAL WITH SPECIES WHEN CROSSED.

This hypothesis was first propounded by Pallas(16/28. 'Act. Acad. St. Petersburg' 1780 part 2pages 84, 100.), and has been adopted by sev-eral authors. I can find hardly any direct factsin its support; but unfortunately no one hascompared, in the case of either animals orplants, the fertility of anciently domesticatedvarieties, when crossed with a distinct species,with that of the wild parent-species when simi-larly crossed. No one has compared, for in-stance, the fertility of Gallus bankiva and of thedomesticated fowl, when crossed with a dis-tinct species of Gallus or Phasianus; and theexperiment would in all cases be surroundedby many difficulties. Dureau de la Malle, who

has so closely studied classical literature, states(16/29. 'Annales des Sc. Nat.' tome 21 1st seriespage 61.) that in the time of the Romans thecommon mule was produced with more diffi-culty than at the present day; but whether thisstatement may be trusted I know not. A muchmore important, though somewhat different,case is given by M. Groenland (16/30. 'Bull.Bot. Soc. de France' December 27, 1861 tome 8page 612.), namely, that plants, known fromtheir intermediate character and sterility to behybrids between Aegilops and wheat, haveperpetuated themselves under culture since1857, WITH A RAPID BUT VARYING IN-CREASE OF FERTILITY IN EACH GENERA-TION. In the fourth generation the plants, stillretaining their intermediate character, had be-come as fertile as common cultivated wheat.

The indirect evidence in favour of the Pallasiandoctrine appears to me to be extremely strong.In the earlier chapters I have shown that our

various breeds of the dog are descended fromseveral wild species; and this probably is thecase with sheep. There can be no doubt that theZebu or humped Indian ox belongs to a distinctspecies from European cattle: the latter, more-over, are descended from two forms, whichmay be called either species or races. We havegood evidence that our domesticated pigs be-long to at least two specific types, S. scrofa andindicus. Now a widely extended analogy leadsto the belief that if these several allied species,when first reclaimed, had been crossed, theywould have exhibited, both in their first unionsand in their hybrid offspring, some degree ofsterility. Nevertheless, the several domesticatedraces descended from them are now all, as faras can be ascertained, perfectly fertile together.If this reasoning be trustworthy, and it is ap-parently sound, we must admit the Pallasiandoctrine that long- continued domesticationtends to eliminate that sterility which is natural

to species when crossed in their aboriginalstate.

ON INCREASED FERTILITY FROM DO-MESTICATION AND CULTIVATION.

Increased fertility from domestication, withoutany reference to crossing, may be here brieflyconsidered. This subject bears indirectly on twoor three points connected with the modificationof organic beings. As Buffon long ago remarked(16/31. Quoted by Isid. Geoffroy St. Hilaire'Hist. Naturelle Generale' tome 3 page 476. Sin-ce this MS. has been sent to press a full discus-sion on the present subject has appeared in Mr.Herbert Spencer's 'Principles of Biology' vol-ume 2 1867 page 457 et seq.), domestic animalsbreed oftener in the year and produce moreyoung at a birth than wild animals of the samespecies; they, also, sometimes breed at an ear-lier age. The case would hardly have deservedfurther notice, had not some authors lately at-

tempted to show that fertility increases anddecreases in an inverse ratio with the amountof food. This strange doctrine has apparentlyarisen from individual animals when suppliedwith an inordinate quantity of food, and fromplants of many kinds when grown on exces-sively rich soil, as on a dunghill, becoming ster-ile: but to this latter point I shall have occasionpresently to return. With hardly an exception,our domesticated animals, which have beenlong habituated to a regular and copious sup-ply of food, without the labour of searching forit, are more fertile than the corresponding wildanimals. It is notorious how frequently cats anddogs breed, and how many young they pro-duce at a birth. The wild rabbit is said generallyto breed four times yearly, and to produce eachtime at most six young; the tame rabbit breedssix or seven times yearly, producing each timefrom four to eleven young; and Mr. HarrisonWeir tells me of a case of eighteen young hav-ing been produced at a birth, all of which sur-

vived. The ferret, though generally so closelyconfined, is more prolific than its supposedwild prototype. The wild sow is remarkablyprolific; she often breeds twice in the year, andbears from four to eight and sometimes eventwelve young; but the domestic sow regularlybreeds twice a year, and would breed oftener ifpermitted; and a sow that produces less thaneight at a birth "is worth little, and the soonershe is fattened for the butcher the better." Theamount of food affects the fertility of the sameindividual: thus sheep, which on mountainsnever produce more than one lamb at a birth,when brought down to lowland pastures fre-quently bear twins. This difference apparentlyis not due to the cold of the higher land, forsheep and other domestic animals are said to beextremely prolific in Lapland. Hard living, also,retards the period at which animals conceive;for it has been found disadvantageous in thenorthern islands of Scotland to allow cows tobear calves before they are four years old.

(16/32. For cats and dogs etc. see Bellingeri in'Annal. des Sc. Nat.' 2nd series, Zoolog. tome 12page 155. For ferrets Bechstein 'NaturgeschichteDeutschlands' b. 1 1801 s. 786, 795. For rabbitsditto s. 1123, 1131; and Bronn 'Geschichte derNatur.' b. 2 s. 99. For mountain sheep ditto s.102. For the fertility of the wild sow, see Bechs-tein 'Naturgesch. Deutschlands' b. 1 1801 s. 534;for the domestic pig Sidney's edition of 'Youatton the Pig' 1860 page 62. With respect to Lap-land see Acerbi 'Travels to the North Cape'English translation volume 2 page 222. Aboutthe Highland cows see 'Hogg on Sheep' page263.)

[Birds offer still better evidence of increasedfertility from domestication: the hen of the wildGallus bankiva lays from six to ten eggs, anumber which would be thought nothing ofwith the domestic hen. The wild duck lays fromfive to ten eggs; the tame one in the course ofthe year from eighty to one hundred. The wild

grey-lag goose lays from five to eight eggs; thetame from thirteen to eighteen, and she lays asecond time; as Mr. Dixon has remarked, "high-feeding, care, and moderate warmth induce ahabit of prolificacy which becomes in somemeasure hereditary." Whether the semi-domesticated dovecote pigeon is more fertilethan the wild rock-pigeon, C. livia, I know not;but the more thoroughly domesticated breedsare nearly twice as fertile as dovecotes: the lat-ter, however, when caged and highly fed, be-come equally fertile with house pigeons. I hearfrom Judge Caton that the wild turkey in theUnited States does not breed when a year old,as the domesticated turkeys there invariablydo. The peahen alone of domesticated birds israther more fertile, according to some accounts,when wild in its native Indian home, than inEurope when exposed to our much colder cli-mate. (16/33. For the eggs of Gallus bankivasee Blyth in 'Annals and Mag. of Nat. Hist.' 2ndseries volume 1 1848 page 456. For wild and

tame ducks Macgillivray 'British Birds' volume5 page 37; and 'Die Enten' s. 87. For wild geeseL. Lloyd 'Scandinavian Adventures' volume 21854 page 413; and for tame geese 'OrnamentalPoultry' by Rev. E.S. Dixon page 139. On thebreeding of Pigeons Pistor 'Das Ganze der Tau-benzucht' 1831 s. 48; and Boitard and Corbie'Les Pigeons' page 158. With respect to pea-cocks, according to Temminck 'Hist. Nat. Gen.des Pigeons' etc. 1813 tome 2 page 41, the henlays in India even as many as twenty eggs; butaccording to Jerdon and another writer quotedin Tegetmeier 'Poultry Book' 1866 pages 280,282, she there lays only from four to nine or teneggs: in England she is said, in the 'PoultryBook' to lay five or six, but another writer saysfrom eight to twelve eggs.)

With respect to plants, no one would expectwheat to tiller more, and each ear to producemore grain, in poor than in rich soil; or to get inpoor soil a heavy crop of peas or beans. Seeds

vary so much in number that it is difficult toestimate them; but on comparing beds of car-rots in a nursery garden with wild plants, theformer seemed to produce about twice as muchseed. Cultivated cabbages yielded thrice as ma-ny pods by measure as wild cabbages from therocks of South Wales. The excess of berriesproduced by the cultivated asparagus in com-parison with the wild plant is enormous. Nodoubt many highly cultivated plants, such aspears, pineapples, bananas, sugar-cane, etc., arenearly or quite sterile; and I am inclined to at-tribute this sterility to excess of food and toother unnatural conditions; but to this subject Ishall recur.]

In some cases, as with the pig, rabbit, etc., andwith those plants which are valued for theirseed, the direct selection of the more fertile in-dividuals has probably much increased theirfertility; and in all cases this may have occurredindirectly, from the better chance of some of the

numerous offspring from the more fertile indi-viduals having been preserved. But with cats,ferrets, and dogs, and with plants like carrots,cabbages, and asparagus, which are not valuedfor their prolificacy, selection can have playedonly a subordinate part; and their increasedfertility must be attributed to the more favour-able conditions of life under which they havelong existed.

CHAPTER 2.XVII.

ON THE GOOD EFFECTS OF CROSSING,AND ON THE EVIL EFFECTS OF CLOSE IN-TERBREEDING.

DEFINITION OF CLOSE INTERBREED-ING. AUGMENTATION OF MORBID TEN-DENCIES. GENERAL EVIDENCE OF THEGOOD EFFECTS DERIVED FROM CROSS-ING, AND ON THE EVIL EFFECTS FROM

CLOSE INTERBREEDING. CATTLE, CLOSE-LY INTERBRED; HALF-WILD CATTLE LONGKEPT IN THE SAME PARKS. SHEEP. FAL-LOW-DEER. DOGS, RABBITS, PIGS. MAN,ORIGIN OF HIS ABHORRENCE OF INCES-TUOUS MARRIAGES. FOWLS. PIGEONS.HIVE-BEES. PLANTS, GENERAL CONSID-ERATIONS ON THE BENEFITS DERIVEDFROM CROSSING. MELONS, FRUIT-TREES,PEAS, CABBAGES, WHEAT, AND FOREST-TREES. ON THE INCREASED SIZE OF HY-BRID PLANTS, NOT EXCLUSIVELY DUE TOTHEIR STERILITY. ON CERTAIN PLANTSWHICH EITHER NORMALLY OR ABNOR-MALLY ARE SELF-IMPOTENT, BUT AREFERTILE, BOTH ON THE MALE AND FE-MALE SIDE, WHEN CROSSED WITH DIS-TINCT INDIVIDUALS EITHER OF THE SAMEOR ANOTHER SPECIES. CONCLUSION.

The gain in constitutional vigour, derived froman occasional cross between individuals of thesame variety, but belonging to distinct families,

or between distinct varieties, has not been solargely or so frequently discussed, as have theevil effects of too close interbreeding. But theformer point is the more important of the two,inasmuch as the evidence is more decisive. Theevil results from close interbreeding are diffi-cult to detect, for they accumulate slowly, anddiffer much in degree with different species;whilst the good effects which almost invariablyfollow a cross are from the first manifest. Itshould, however, be clearly understood that theadvantage of close interbreeding, as far as theretention of character is concerned, is indisput-able, and often outweighs the evil of a slightloss of constitutional vigour. In relation to thesubject of domestication, the whole question isof some importance, as too close interbreedinginterferes with the improvement of old races. Itis important as indirectly bearing on Hybrid-ism; and possibly on the extinction of species,when any form has become so rare that only afew individuals remain within a confined area.

It bears in an important manner on the influ-ence of free intercrossing, in obliterating indi-vidual differences, and thus giving uniformityof character to the individuals of the same raceor species; for if additional vigour and fertilitybe thus gained, the crossed offspring will mul-tiply and prevail, and the ultimate result will befar greater than otherwise would have oc-curred. Lastly, the question is of high interest,as bearing on mankind. I shall therefore discussthis subject at full length. As the facts whichprove the evil effects of close interbreeding aremore copious, though less decisive, than thoseon the good effects of crossing, I shall, undereach group of beings, begin with the former.

There is no difficulty in defining what is meantby a cross; but this is by no means easy in re-gard to "breeding in and in" or "too close inter-breeding," because, as we shall see, differentspecies of animals are differently affected bythe same degree of interbreeding. The pairing

of a father and daughter, or mother and son, orbrothers and sisters, if carried on during sev-eral generations, is the closest possible form ofinterbreeding. But some good judges, for in-stance Sir J. Sebright, believe that the pairing ofa brother and sister is much closer than that ofparents and children; for when the father ismatched with his daughter he crosses, as issaid, with only half his own blood. The conse-quences of close interbreeding carried on fortoo long a time, are, as is generally believed,loss of size, constitutional vigour, and fertility,sometimes accompanied by a tendency to mal-formation. Manifest evil does not usually fol-low from pairing the nearest relations for two,three, or even four generations; but several cau-ses interfere with our detecting the evil—suchas the deterioration being very gradual, and thedifficulty of distinguishing between such directevil and the inevitable augmentation of anymorbid tendencies which may be latent or ap-parent in the related parents. On the other

hand, the benefit from a cross, even when therehas not been any very close interbreeding, isalmost invariably at once conspicuous. There isgood reason to believe, and this was the opin-ion of that most experienced observer Sir J. Se-bright (17/1. 'The Art of Improving the Breed,etc.' 1809 page 16.), that the evil effects of closeinterbreeding may be checked or quite pre-vented by the related individuals being sepa-rated for a few generations and exposed to dif-ferent conditions of life. This conclusion is nowheld by many breeders; for instance Mr. Carr(17/2. 'The History of the Rise and Progress ofthe Killerby, etc. Herds' page 41.) remarks, it isa well-known "fact that a change of soil andclimate effects perhaps almost as great a changein the constitution as would result from an in-fusion of fresh blood." I hope to show in a fu-ture work that consanguinity by itself countsfor nothing, but acts solely from related organ-isms generally having a similar constitution,

and having been exposed in most cases to simi-lar conditions.

That any evil directly follows from the closestinterbreeding has been denied by many per-sons; but rarely by any practical breeder; andnever, as far as I know, by one who has largelybred animals which propagate their kind quic-kly. Many physiologists attribute the evil exclu-sively to the combination and consequent in-crease of morbid tendencies common to bothparents; and that this is an active source of mis-chief there can be no doubt. It is unfortunatelytoo notorious that men and various domesticanimals endowed with a wretched constitution,and with a strong hereditary disposition to dis-ease, if not actually ill, are fully capable of pro-creating their kind. Close interbreeding, on theother hand, often induces sterility; and this in-dicates something quite distinct from the aug-mentation of morbid tendencies common toboth parents. The evidence immediately to be

given convinces me that it is a great law of na-ture, that all organic beings profit from an occa-sional cross with individuals not closely relatedto them in blood; and that, on the other hand,long-continued close interbreeding is injurious.

Various general considerations have had muchinfluence in leading me to this conclusion; butthe reader will probably rely more on specialfacts and opinions. The authority of experi-enced observers, even when they do not ad-vance the grounds of their belief, is of somelittle value. Now almost all men who have bredmany kinds of animals and have written on thesubject, such as Sir J. Sebright, Andrew Knight,etc. (17/3. For Andrew Knight see A. Walker on'Intermarriage' 1838 page 227. Sir J. Sebright'Treatise' has just been quoted.), have expressedthe strongest conviction on the impossibility oflong-continued close interbreeding. Those whohave compiled works on agriculture, and haveassociated much with breeders, such as the

sagacious Youatt, Low, etc., have strongly de-clared their opinion to the same effect. ProsperLucas, trusting largely to French authorities,has come to a similar conclusion. The distin-guished German agriculturist Hermann vonNathusius, who has written the most able trea-tise on this subject which I have met with, con-curs; and as I shall have to quote from this trea-tise, I may state that Nathusius is not only in-timately acquainted with works on agriculturein all languages, and knows the pedigrees ofour British breeds better than most Englishmen,but has imported many of our improved ani-mals, and is himself an experienced breeder.

Evidence of the evil effects of close interbreed-ing can most readily be acquired in the case ofanimals, such as fowls, pigeons, etc., whichpropagate quickly, and, from being kept in thesame place, are exposed to the same conditions.Now I have inquired of very many breeders ofthese birds, and I have hitherto not met with a

single man who was not thoroughly convincedthat an occasional cross with another strain ofthe same sub-variety was absolutely necessary.Most breeders of highly improved or fancybirds value their own strain, and are most un-willing, at the risk, in their opinion, of deterio-ration, to make a cross. The purchase of a first-rate bird of another strain is expensive, andexchanges are troublesome; yet all breeders, asfar as I can hear, excepting those who keep lar-ge stocks at different places for the sake of cros-sing, are driven after a time to take this step.

Another general consideration which has hadgreat influence on my mind is, that with allhermaphrodite animals and plants, which itmight have been thought would have perpetu-ally fertilised themselves and been thus sub-jected for long ages to the closest interbreeding,there is not a single species, as far as I can dis-cover, in which the structure ensures self-fertilisation. On the contrary, there are in a

multitude of cases, as briefly stated in the fif-teenth chapter, manifest adaptations whichfavour or inevitably lead to an occasional crossbetween one hermaphrodite and another of thesame species; and these adaptive structures areutterly purposeless, as far as we can see, forany other end.

[With CATTLE there can be no doubt that ex-tremely close interbreeding may be long carriedon advantageously with respect to externalcharacters, and with no manifest evil as far asconstitution is concerned. The case of Bake-well's Longhorns, which were closely interbredfor a long period, has often been quoted; yetYouatt says (17/4. 'Cattle' page 199.) the breed"had acquired a delicacy of constitution incon-sistent with common management," and "thepropagation of the species was not always cer-tain." But the Shorthorns offer the most strikingcase of close interbreeding; for instance, thefamous bull Favourite (who was himself the

offspring of a half-brother and sister from Fol-jambe) was matched with his own daughter,granddaughter, and great- granddaughter; sothat the produce of this last union, or the great-great- granddaughter, had 15-16ths, or 93.75per cent of the blood of Favourite in her veins.This cow was matched with the bull Welling-ton, having 62.5 per cent of Favourite blood inhis veins, and produced Clarissa; Clarissa wasmatched with the bull Lancaster, having 68.75of the same blood, and she yielded valuableoffspring. (17/5. I give this on the authority ofNathusius 'Ueber Shorthorn Rindvieh' 1857 s.71, see also 'Gardeners Chronicle' 1860 page270. But Mr. J. Storer, a large breeder of cattle,informs me that the parentage of Clarissa is notwell authenticated. In the first volume of the'Herd Book' she was entered as having six de-scents from Favourite, "which was a palpablemistake," and in all subsequent editions shewas spoken of as having only four descents.Mr. Storer doubts even about the four, as no

names of the dams are given. Moreover, Claris-sa bore "only two bulls and one heifer, and inthe next generation her progeny became ex-tinct." Analogous cases of close interbreedingare given in a pamphlet published by Mr. C.Macknight and Dr. H. Madden 'On the TruePrinciples of Breeding' Melbourne Australia1865.) Nevertheless Collings, who reared theseanimals, and was a strong advocate for closebreeding, once crossed his stock with a Gallo-way, and the cows from this cross realised thehighest prices. Bates's herd was esteemed themost celebrated in the world. For thirteen yearshe bred most closely in and in; but during thenext seventeen years, though he had the mostexalted notion of the value of his own stock, hethrice infused fresh blood into his herd: it issaid that he did this, not to improve the form ofhis animals, but on account of their lessenedfertility. Mr. Bates's own view, as given by acelebrated breeder (17/6. Mr. WilloughbyWood in 'Gardener's Chronicle' 1855 page 411;

and 1860 page 270. See the very clear tables andpedigrees given in Nathusius 'Rindvieh' s. 72-77.), was, that "to breed in-and-in from a badstock was ruin and devastation; yet that thepractice may be safely followed within certainlimits when the parents so related are de-scended from first-rate animals." We thus seethat there has been much close interbreedingwith Shorthorns; but Nathusius, after the mostcareful study of their pedigrees, says that hecan find no instance of a breeder who hasstrictly followed this practice during his wholelife. From this study and his own experience, heconcludes that close interbreeding is necessaryto ennoble the stock; but that in effecting thisthe greatest care is necessary, on account of thetendency to infertility and weakness. It may beadded, that another high authority (17/7. Mr.Wright 'Journal of Royal Agricult. Soc.' volume7 1846 page 204. Mr. J. Downing (a successfulbreeder of Shorthorns in Ireland) informs methat the raisers of the great families of Short-

horns carefully conceal their sterility and wantof constitution. He adds that Mr. Bates, after hehad bred his herd in-and-in for some years,"lost in one season twenty-eight calves solelyfrom want of constitution.") asserts that manymore calves are born cripples from Shorthornsthan from other and less closely interbred racesof cattle.

Although by carefully selecting the best ani-mals (as Nature effectually does by the law ofbattle) close interbreeding may be long carriedon with cattle, yet the good effects of a crossbetween almost any two breeds is at onceshown by the greater size and vigour of theoffspring; as Mr. Spooner writes to me, "cross-ing distinct breeds certainly improves cattle forthe butcher." Such crossed animals are ofcourse of no value to the breeder; but they havebeen raised during many years in several partsof England to be slaughtered (17/8. 'Youatt onCattle' page 202.); and their merit is now so

fully recognised, that at fat-cattle shows a sepa-rate class has been formed for their reception.The best fat ox at the great show at Islington in1862 was a crossed animal.

The half-wild cattle, which have been kept inBritish parks probably for 400 or 500 years, oreven for a longer period, have been advancedby Culley and others as a case of long-continued interbreeding within the limits of thesame herd without any consequent injury. Withrespect to the cattle at Chillingham, the lateLord Tankerville owned that they were badbreeders. (17/9. 'Report British Assoc. Zoolog.Sect.' 1838.) The agent, Mr. Hardy, estimates (ina letter to me, dated May, 1861) that in the herdof about fifty the average number annuallyslaughtered, killed by fighting, and dying, isabout ten, or one in five. As the herd is kept upto nearly the same average number, the annualrate of increase must be likewise about one infive. The bulls, I may add, engage in furious

battles, of which battles the present Lord Tan-kerville has given me a graphic description, sothat there will always be rigorous selection ofthe most vigorous males. I procured in 1855from Mr. D. Gardner, agent to the Duke ofHamilton, the following account of the wildcattle kept in the Duke's park in Lanarkshire,which is about 200 acres in extent. The numberof cattle varies from sixty-five to eighty; and thenumber annually killed (I presume by all cau-ses) is from eight to ten; so that the annual rateof increase can hardly be more than one in six.Now in South America, where the herds arehalf-wild, and therefore offer a nearly fair stan-dard of comparison, according to Azara thenatural increase of the cattle on an estancia isfrom one-third to one-fourth of the total num-ber, or one in between three and four and this,no doubt, applies exclusively to adult animalsfit for consumption. Hence the half-wild Britishcattle which have long interbred within thelimits of the same herd are relatively far less

fertile. Although in an unenclosed country likeParaguay there must be some crossing betweenthe different herds, yet even there the inhabi-tants believe that the occasional introduction ofanimals from distant localities is necessary toprevent "degeneration in size and diminutionof fertility." (17/10. Azara 'Quadrupedes duParaguay' tome 2 pages 354, 368.) The decreasein size from ancient times in the Chillinghamand Hamilton cattle must have been prodi-gious, for Professor Rutimeyer has shown thatthey are almost certainly the descendants of thegigantic Bos primigenius. No doubt this de-crease in size may be largely attributed to lessfavourable conditions of life; yet animals roa-ming over large parks, and fed during severewinters, can hardly be considered as placedunder very unfavourable conditions.

With SHEEP there has often been long-continued interbreeding within the limits of thesame flock; but whether the nearest relations

have been matched so frequently as in the caseof Shorthorn cattle, I do not know. The Messrs.Brown during fifty years have never infusedfresh blood into their excellent flock of Leices-ters. Since 1810 Mr. Barford has acted on thesame principle with the Foscote flock. He as-serts that half a century of experience has con-vinced him that when two nearly related ani-mals are quite sound in constitution, in-and-inbreeding does not induce degeneracy; but headds that he "does not pride himself on bree-ding from the nearest affinities." In France theNaz flock has been bred for sixty years withoutthe introduction of a single strange ram. (17/11.For the case of the Messrs. Brown see 'Garde-ner's Chronicle' 1855 page 26. For the Foscoteflock 'Gardener's Chronicle' 1860 page 416. Forthe Naz flock 'Bull. de la Soc. d'Acclimat.' 1860page 477.) Nevertheless, most great breeders ofsheep have protested against close interbree-ding prolonged for too great a length of time.(17/12. Nathusius 'Rindvieh' s. 65; 'Youatt on

Sheep' page 495.) The most celebrated of recentbreeders, Jonas Webb, kept five separate fami-lies to work on, thus "retaining the requisitedistance of relationship between the sexes"(17/13. 'Gardener's Chronicle' 1861 page 631.);and what is probably of greater importance, theseparate flocks will have been exposed to so-mewhat different conditions.

Although by the aid of careful selection thenear interbreeding of sheep may be long conti-nued without any manifest evil, yet it has oftenbeen the practice with farmers to cross distinctbreeds to obtain animals for the butcher, whichplainly shows that good of some kind is deri-ved from this practice. We have excellent evi-dence on this head from Mr. S. Druce (17/14.'Journal R. Agricult. Soc.' volume 14 1853 page212.), who gives in detail the comparativenumbers of four pure breeds and of a cross-breed which can be supported on the sameground, and he gives their produce in fleece

and carcase. A high authority, Mr. Pusey, sumsup the result in money value during an equallength of time, namely (neglecting shillings),for Cotswolds 248 pounds, for Leicesters 223pounds, for Southdowns 204 pounds, forHampshire Downs 264 pounds, and for thecrossbred 293 pounds. A former celebratedbreeder, Lord Somerville, states that his half-breeds from Ryelands and Spanish sheep werelarger animals than either the pure Ryelands orpure Spanish sheep. Mr. Spooner concludes hisexcellent Essay on Crossing by asserting thatthere is a pecuniary advantage in judiciouscross-breeding, especially when the male islarger than the female. (17/15. Lord Somerville'Facts on Sheep and Husbandry' page 6. Mr.Spooner in 'Journal of Royal Agricult. Soc. ofEngland' volume 20 part 2. See also an excellentpaper on the same subject in 'Gardener's Chro-nicle' 1860 page 321 by Mr. Charles Howard.)

As some of our British parks are ancient, it oc-curred to me that there must have been long-continued close interbreeding with the fallow-deer (Cervus dama) kept in them; but on inqui-ry I find that it is a common practice to infusenew blood by procuring bucks from otherparks. Mr. Shirley (17/16. 'Some Account ofEnglish Deer Parks' by Evelyn P. Shirley 1867.),who has carefully studied the management ofdeer, admits that in some parks there has beenno admixture of foreign blood from a time be-yond the memory of man. But he concludes"that in the end the constant breeding in-and-inis sure to tell to the disadvantage of the wholeherd, though it may take a very long time toprove it; moreover, when we find, as is veryconstantly the case, that the introduction offresh blood has been of the very greatest use todeer, both by improving their size and appea-rance, and particularly by being of service inremoving the taint of 'rickback,' if not of otherdiseases, to which deer are sometimes subject

when the blood has not been changed, therecan, I think, be no doubt but that a judiciouscross with a good stock is of the greatest conse-quence, and is indeed essential, sooner or later,to the prosperity of every well-ordered park."

Mr. Meynell's famous foxhounds have beenadduced, as showing that no ill effects followfrom close interbreeding; and Sir J. Sebrightascertained from him that he frequently bredfrom father and daughter, mother and son, andsometimes even from brothers and sisters. Withgreyhounds also there has been much closeinterbreeding, but the best breeders agree thatit may be carried too far. (17/17. Stonehenge'The Dog' 1867 pages 175-188.) But Sir J. Se-bright declares (17/18. 'The Art of Improvingthe Breed' etc. page 13. With respect to Scotchdeerhounds see Scrope 'Art of Deer Stalking'pages 350-353.), that by breeding in-and-in, bywhich he means matching brothers and sisters,he has actually seen the offspring of strong spa-

niels degenerate into weak and diminutive lap-dogs. The Rev. W.D. Fox has communicated tome the case of a small lot of bloodhounds, longkept in the same family, which had becomevery bad breeders, and nearly all had a bonyenlargement in the tail. A single cross with adistinct strain of bloodhounds restored theirfertility, and drove away the tendency to mal-formation in the tail. I have heard the particu-lars of another case with bloodhounds, inwhich the female had to be held to the male.Considering how rapid is the natural increaseof the dog, it is difficult to understand the largeprice of all highly improved breeds, which al-most implies long-continued close interbree-ding, except on the belief that this process les-sens fertility and increases liability to distemperand other diseases. A high authority, Mr. Scro-pe, attributes the rarity and deterioration in sizeof the Scotch deerhound (the few individualsformerly existing throughout the country beingall related) in large part to close interbreeding.

With all highly-bred animals there is more orless difficulty in getting them to procreate quic-kly, and all suffer much from delicacy of consti-tution. A great judge of rabbits (17/19. 'CottageGardener' 1861 page 327.) says, "the long-eareddoes are often too highly bred or forced in theiryouth to be of much value as breeders, oftenturning out barren or bad mothers." They oftendesert their young, so that it is necessary tohave nurse-rabbits, but I do not pretend to at-tribute all these evil results to close interbree-ding. (17/20. Mr. Huth gives ('The Marriage ofNear Kin' 1875 page 302) from the 'Bulletin del'Acad. R. de Med. de Belgique' (volume 9 1866pages 287, 305), several statements made by aM. Legrain with respect to crossing brother andsister rabbits for five or six successive genera-tions with no consequent evil results. I was somuch surprised at this account, and at M. Le-grain's invariable success in his experiments,that I wrote to a distinguished naturalist in Bel-gium to inquire whether M. Legrain was a

trustworthy observer. In answer, I have heardthat, as doubts were expressed about the aut-henticity of these experiments, a commission ofinquiry was appointed, and that at a succee-ding meeting of the Society ('Bull. de l'Acad. R.de Med. de Belgique' 1867 3rd series tome 1 no.1 to 5), Dr. Crocq reported "qu'il etait materie-llement impossible que M. Legrain ait fait lesexperiences qu'il annonce." To this public accu-sation no satisfactory answer was made.)

With respect to PIGS there is more unanimityamongst breeders on the evil effects of closeinterbreeding than, perhaps, with any otherlarge animal. Mr. Druce, a great and successfulbreeder of the Improved Oxfordshires (a cros-sed race), writes, "without a change of boars ofa different tribe, but of the same breed, consti-tution cannot be preserved." Mr. Fisher Hobbs,the raiser of the celebrated Improved Essexbreed, divided his stock into three separatefamilies, by which means he maintained the

breed for more than twenty years, "by judiciousselection from the THREE DISTINCT FAMI-LIES." (17/21. Sidney's edition of 'Youatt on thePig' 1860 page 30; page 33 quotation from Mr.Druce; page 29 on Lord Western's case.) LordWestern was the first importer of a Neapolitanboar and sow. "From this pair he bred in-and-in, until the breed was in danger of becomingextinct, a sure result (as Mr. Sidney remarks) ofin-and-in breeding." Lord Western then crossedhis Neapolitan pigs with the old Essex, andmade the first great step towards the ImprovedEssex breed. Here is a more interesting case.Mr. J. Wright, well known as a breeder, crossed(17/22. 'Journal of Royal Agricult. Soc. of En-gland' 1846 volume 7 page 205.) the same boarwith the daughter, granddaughter, and great-granddaughter, and so on for seven genera-tions. The result was, that in many instances theoffspring failed to breed; in others they produ-ced few that lived; and of the latter many wereidiotic, without sense, even to suck, and when

attempting to move could not walk straight.Now it deserves especial notice, that the twolast sows produced by this long course of inter-breeding were sent to other boars, and theybore several litters of healthy pigs. The bestsow in external appearance produced duringthe whole seven generations was one in the laststage of descent; but the litter consisted of thisone sow. She would not breed to her sire, yetbred at the first trial to a stranger in blood. Sothat, in Mr. Wright's case, long-continued andextremely close interbreeding did not affect theexternal form or merit of the young; but withmany of them the general constitution andmental powers, and especially the reproductivefunctions, were seriously affected.

Nathusius gives (17/23. 'Ueber Rindvieh' etc. s.78. Col. Le Couteur, who has done so much forthe agriculture of Jersey, writes to me that frompossessing a fine breed of pigs he bred themvery closely, twice pairing brothers and sisters,

but nearly all the young had fits and died sud-denly.) an analogous and even more strikingcase: he imported from England a pregnantsow of the large Yorkshire breed, and bred theproduct closely in-and-in for three generations:the result was unfavourable, as the young wereweak in constitution, with impaired fertility.One of the latest sows, which he esteemed agood animal, produced, when paired with herown uncle (who was known to be productivewith sows of other breeds), a litter of six, and asecond time a litter of only five weak youngpigs. He then paired this sow with a boar of asmall black breed, which he had likewise im-ported from England; this boar, when matchedwith sows of his own breed, produced fromseven to nine young. Now, the sow of the largebreed, which was so unproductive when pairedwith her own uncle, yielded to the small blackboar, in the first litter twenty-one, and in thesecond litter eighteen young pigs; so that in one

year she produced thirty-nine fine young ani-mals!

As in the case of several other animals alreadymentioned, even when no injury is perceptiblefrom moderately close interbreeding, yet, toquote the words of Mr. Coate (who five timeswon the annual gold medal of the SmithfieldClub Show for the best pen of pigs), "Crossesanswer well for profit to the farmer, as you getmore constitution and quicker growth; but forme, who sell a great number of pigs for bree-ding purposes, I find it will not do, as it requi-res many years to get anything like purity ofblood again." (17/24. Sidney on the 'Pig' page36. See also note page 34. Also Richardson onthe 'Pig' 1847 page 26.)]

Almost all the animals as yet mentioned aregregarious, and the males must frequently pairwith their own daughters, for they expel theyoung males as well as all intruders, until for-ced by old age and loss of strength to yield to

some stronger male. It is therefore not impro-bable that gregarious animals may have beenrendered less susceptible than non-social spe-cies to the evil consequences of close interbree-ding, so that they may be enabled to live inherds without injury to their offspring. Unfor-tunately we do not know whether an animallike the cat, which is not gregarious, wouldsuffer from close interbreeding in a greater de-gree than our other domesticated animals. Butthe pig is not, as far as I can discover, strictlygregarious, and we have seen that it appearseminently liable to the evil effects of close in-terbreeding. Mr. Huth, in the case of the pig,attributes (page 285) these effects to theirhaving been "cultivated most for their fat," or tothe selected individuals having had a weakconstitution; but we must remember that it isgreat breeders who have brought forward theabove cases, and who are far more familiarthan ordinary men can be, with the causes

which are likely to interfere with the fertility oftheir animals.

The effects of close interbreeding in the case ofman is a difficult subject, on which I will saybut little. It has been discussed by various aut-hors under many points of view. (17/25. Dr.Dally has published an excellent article (trans-lated in the 'Anthropolog. Review' May 1864page 65), criticising all writers who have main-tained that evil follows from consanguineousmarriages. No doubt on this side of the ques-tion many advocates have injured their causeby inaccuracies: thus it has been stated (Devay'Du Danger des Mariages' etc. 1862 page 141)that the marriages of cousins have been prohi-bited by the legislature of Ohio; but I have beenassured, in answer to inquiries made in theUnited States, that this statement is a mere fa-ble.) Mr. Tylor (17/26. See his interesting workon the 'Early History of Man' 1865 chapter 10.)has shown that with widely different races in

the most distant quarters of the world, marria-ges between relations—even between distantrelations—have been strictly prohibited. Thereare, however, many exceptions to the rule,which are fully given by Mr. Huth (17/27. 'TheMarriage of Near Kin' 1875. The evidence givenby Mr. Huth would, I think, have been evenmore valuable than it is on this and some otherpoints, if he had referred solely to the works ofmen who had long resided in each country re-ferred to, and who showed that they possessedjudgment and caution. See also Mr. W. Adam'On Consanguinity in Marriage' in the 'Fort-nightly Review' 1865 page 710. Also Hofacker'Ueber die Eigenschaften' etc. 1828.) It is a cu-rious problem how these prohibitions aroseduring early and barbarous times. Mr. Tylor isinclined to attribute them to the evil effects ofconsanguineous marriages having been obser-ved; and he ingeniously attempts to explainsome apparent anomalies in the prohibition notextending equally to the relations on the male

and female side. He admits, however, that ot-her causes, such as the extension of friendlyalliances, may have come into play. Mr. W.Adam, on the other hand, concludes that rela-ted marriages are prohibited and viewed withrepugnance, from the confusion which wouldthus arise in the descent of property, and fromother still more recondite reasons. But I cannotaccept these views, seeing that incest is held inabhorrence by savages such as those of Austra-lia and South America (17/28. Sir G. Grey'Journal of Expeditions into Australia' volume 2page 243; and Dobrizhoffer 'On the Abipones ofSouth America.'), who have no property to be-queath, or fine moral feelings to confuse, andwho are not likely to reflect on distant evils totheir progeny. According to Mr. Huth the fee-ling is the indirect result of exogamy, inasmuchas when this practice ceased in any tribe and itbecame endogamous, so that marriages werestrictly confined to the same tribe, it is not unli-kely that a vestige of the former practice would

still be retained, so that closely-related marria-ges would be prohibited. With respect to exo-gamy itself Mr. MacLennan believes that it aro-se from a scarcity of women, owing to femaleinfanticide, aided perhaps by other causes.

It has been clearly shown by Mr. Huth that the-re is no instinctive feeling in man against incestany more than in gregarious animals. We knowalso how readily any prejudice or feeling mayrise to abhorrence, as shown by Hindus in re-gard to objects causing defilement. Althoughthere seems to be no strong inherited feeling inmankind against incest, it seems possible thatmen during primeval times may have beenmore excited by strange females than by thosewith whom they habitually lived; in the samemanner as according to Mr. Cupples (17/29.'Descent of Man' 2nd. edit page 524.), maledeerhounds are inclined towards strange fema-les, while the females prefer dogs with whomthey have associated. If any such feeling for-

merly existed in man, this would have led to apreference for marriages beyond the nearestkin, and might have been strengthened by theoffspring of such marriages surviving in grea-ter numbers, as analogy would lead us to belie-ve would have occurred.

Whether consanguineous marriages, such asare permitted in civilised nations, and whichwould not be considered as close interbreedingin the case of our domesticated animals, causeany injury will never be known with certaintyuntil a census is taken with this object in view.My son, George Darwin, has done what is pos-sible at present by a statistical investigation(17/30. 'Journal of Statistical Soc.' June 1875page 153; and 'Fortnightly Review' June 1875.),and he has come to the conclusion, from hisown researches and those of Dr. Mitchell, thatthe evidence as to any evil thus caused is con-flicting, but on the whole points to the evilbeing very small.

[BIRDS.

In the case of the FOWL a whole array of aut-horities could be given against too close inter-breeding. Sir J. Sebright positively asserts thathe made many trials, and that his fowls, whenthus treated, became long in the legs, small inthe body, and bad breeders. (17/31. 'The Art ofImproving the Breed' page 13.) He producedthe famous Sebright Bantams by complicatedcrosses, and by breeding in-and-in; and sincehis time there has been much close interbree-ding with these animals; and they are now no-toriously bad breeders. I have seen Silver Ban-tams, directly descended from his stock, whichhad become almost as barren as hybrids; fornot a single chicken had been that year hatchedfrom two full nests of eggs. Mr. Hewitt saysthat with these Bantams the sterility of the malestands, with rare exceptions, in the closest rela-tion with their loss of certain secondary malecharacters: he adds, "I have noticed, as a gene-ral rule, that even the slightest deviation from

feminine character in the tail of the male Se-bright—say the elongation by only half an inchof the two principal tail feathers—brings with itimproved probability of increased fertility."(17/32. 'The Poultry Book' by W.B. Tegetmeier1866 page 245.)

Mr. Wright states (17/33. 'Journal Royal Agri-cult. Soc.' 1846 volume 7 page 205; see also Fer-guson on the Fowl pages 83, 317; see also 'ThePoultry Book' by Tegetmeier 1866 page 135with respect to the extent to which cock-fighters found that they could venture to breedin-and-in, viz., occasionally a hen with her ownson; "but they were cautious not to repeat thein-and-in breeding.") that Mr. Clark, "whosefighting-cocks were so notorious, continued tobreed from his own kind till they lost their dis-position to fight, but stood to be cut up withoutmaking any resistance, and were so reduced insize as to be under those weights required forthe best prizes; but on obtaining a cross from

Mr. Leighton, they again resumed their formercourage and weight." It should be borne inmind that game-cocks before they fought werealways weighed, so that nothing was left to theimagination about any reduction or increase ofweight. Mr. Clark does not seem to have bredfrom brothers and sisters, which is the mostinjurious kind of union; and he found, afterrepeated trials, that there was a greater reduc-tion in weight in the young from a father pai-red with his daughter, than from a mother withher son. I may add that Mr. Eyton of Eyton, thewell-known ornithologist, who is a large bree-der of Grey Dorkings, informs me that theycertainly diminish in size, and become less pro-lific, unless a cross with another strain is occa-sionally obtained. So it is with Malays, accor-ding to Mr. Hewitt, as far as size is concerned.(17/34. 'The Poultry Book' by W.B. Tegetmeier1866 page 79.)

An experienced writer (17/35. 'The PoultryChronicle' 1854 volume 1 page 43.) remarksthat the same amateur, as is well known, sel-dom long maintains the superiority of his birds;and this, he adds, undoubtedly is due to all hisstock "being of the same blood;" hence it is in-dispensable that he should occasionally procu-re a bird of another strain. But this is not neces-sary with those who keep a stock of fowls atdifferent stations. Thus, Mr. Ballance, who hasbred Malays for thirty years, and has won moreprizes with these birds than any other fancier inEngland, says that breeding in-and-in does notnecessarily cause deterioration; "but all de-pends upon how this is managed. My plan hasbeen to keep about five or six distinct runs, andto rear about two hundred or three hundredchickens each year, and select the best birdsfrom each run for crossing. I thus secure suffi-cient crossing to prevent deterioration." (17/36.'The Poultry Book' by W.B. Tegetmeier 1866page 79.)

We thus see that there is almost complete una-nimity with poultry-breeders that, when fowlsare kept at the same place, evil quickly followsfrom interbreeding carried on to an extentwhich would be disregarded in the case of mostquadrupeds. Moreover, it is a generally recei-ved opinion that cross- bred chickens are thehardiest and most easily reared. (17/37. 'ThePoultry Chronicle' volume 1 page 89.) Mr. Te-getmeier, who has carefully attended to poultryof all breeds, says (17/38. 'The Poultry Book'1866 page 210.) that Dorking hens, allowed torun with Houdan or Creve-coeur cocks, "pro-duce in the early spring chickens that for size,hardihood, early maturity, and fitness for themarket, surpass those of any pure breed thatwe have ever raised." Mr. Hewitt gives it as ageneral rule with fowls, that crossing the breedincreases their size. He makes this remark afterstating that hybrids from the pheasant and fowlare considerably larger than either progenitor:so again, hybrids from the male golden phea-

sant and female common pheasant "are of farlarger size than either parent-bird." (17/39. Ibid1866 page 167; and 'Poultry Chronicle' volume3 1855 page 15.) To this subject of the increasedsize of hybrids I shall presently return.

With PIGEONS, breeders are unanimous, aspreviously stated, that it is absolutely indispen-sable, notwithstanding the trouble and expensethus caused, occasionally to cross their much-prized birds with individuals of another strain,but belonging, of course, to the same variety. Itdeserves notice that, when size is one of thedesired characters, as with pouters (17/40. 'ATreatise on Fancy Pigeons' by J.M. Eaton page56.) the evil effects of close interbreeding aremuch sooner perceived than when small birds,such as short- faced tumblers, are valued. Theextreme delicacy of the high fancy breeds, suchas these tumblers and improved English ca-rriers, is remarkable; they are liable to manydiseases, and often die in the egg or during the

first moult; and their eggs have generally to behatched under foster-mothers. Although thesehighly-prized birds have invariably been sub-jected to much close interbreeding, yet theirextreme delicacy of constitution cannot perhapsbe thus fully explained. Mr. Yarrell informedme that Sir J. Sebright continued closely inter-breeding some owl-pigeons, until from theirextreme sterility he as nearly as possible lostthe whole family. Mr. Brent (17/41. 'The PigeonBook' page 46.) tried to raise a breed of trumpe-ters, by crossing a common pigeon, and recros-sing the daughter, granddaughter, great-granddaughter, and great-great-granddaughter, with the same male trumpeter,until he obtained a bird with 15/16 of trumpe-ter's blood; but then the experiment failed, for"breeding so close stopped reproduction." Theexperienced Neumeister (17/42. 'Das Ganzeder Taubenzucht' 1837 s. 18.) also asserts thatthe offspring from dovecotes and various otherbreeds are "generally very fertile and hardy

birds:" so again MM. Boitard and Corbie(17/43. 'Les Pigeons' 1824 page 35.), after forty-five years' experience, recommend persons tocross their breeds for amusement; for, if theyfail to make interesting birds, they will succeedunder an economical point of view, "as it isfound that mongrels are more fertile than pi-geons of pure race."

I will refer only to one other animal, namely,the Hive-bee, because a distinguished entomo-logist has advanced this as a case of inevitableclose interbreeding. As the hive is tenanted by asingle female, it might have been thought thather male and female offspring would alwayshave bred together, more especially as bees ofdifferent hives are hostile to each other; a stran-ge worker being almost always attacked whentrying to enter another hive. But Mr. Teget-meier has shown (17/44. 'Proc. Entomolog.Soc.' August 6, 1860 page 126.) that this instinctdoes not apply to drones, which are permitted

to enter any hive; so that there is no a prioriimprobability of a queen receiving a foreigndrone. The fact of the union invariably and ne-cessarily taking place on the wing, during thequeen's nuptial flight, seems to be a specialprovision against continued interbreeding.However this may be, experience has shown,since the introduction of the yellow-bandedLigurian race into Germany and England, thatbees freely cross: Mr. Woodbury, who introdu-ced Ligurian bees into Devonshire, found du-ring a single season that three stocks, at distan-ces of from one to two miles from his hives,were crossed by his drones. In one case the Li-gurian drones must have flown over the city ofExeter, and over several intermediate hives. Onanother occasion several common black queenswere crossed by Ligurian drones at a distanceof from one to three and a half miles. (17/45.'Journal of Horticulture' 1861 pages 39, 77, 158;and 1864 page 206.)

PLANTS.

When a single plant of a new species is intro-duced into any country, if propagated by seed,many individuals will soon be raised, so that ifthe proper insects be present there will be cros-sing. With newly-introduced trees or otherplants not propagated by seed we are not hereconcerned. With old- established plants it is analmost universal practice occasionally to makeexchanges of seed, by which means individualswhich have been exposed to different condi-tions of life,—and this, as we have seen withanimals, diminishes the evil from close inter-breeding,—will occasionally be introduced intoeach district.

With respect to individuals belonging to thesame sub-variety, Gartner, whose accuracy andexperience exceeded that of all other observers,states (17/46. 'Beitrage zur Kenntniss der Be-fruchtung' 1844 s. 366.) that he has many timesobserved good effects from this step, especially

with exotic genera, of which the fertility is so-mewhat impaired, such as Passiflora, Lobelia,Fuchsia. Herbert also says (17/47. 'Amaryllida-ceae' page 371.), "I am inclined to think that Ihave derived advantage from impregnating theflower from which I wished to obtain seed withpollen from another individual of the samevariety, or at least from another flower, ratherthan with its own." Again, Professor Lecoq as-certained that crossed offspring are more vigo-rous and robust than their parents. (17/48. 'Dela Fecondation' 2nd edition 1862 page 79.)

General statements of this kind, however, canseldom be fully trusted: I therefore began along series of experiments, continued for aboutten years, which will I think conclusively showthe good effects of crossing two distinct plantsof the same variety, and the evil effects of long-continued self- fertilisation. A clear light willthus be thrown on such questions, as why flo-wers are almost invariably constructed so as to

permit, or favour, or necessitate the union oftwo individuals. We shall clearly understandwhy monoecious and dioecious,—why dicho-gamous, dimorphic and trimorphic plants exist,and many other such cases. I intend soon topublish an account of these experiments, and Ican here give only a few cases in illustration.The plan which I followed was to grow plantsin the same pot, or in pots of the same size, orclose together in the open ground; carefully toexclude insects; and then to fertilise some of theflowers with pollen from the same flower, andothers on the same plant with pollen from adistinct but adjoining plant. In many of theseexperiments, the crossed plants yielded muchmore seed than the self-fertilised plants; and Ihave never seen the reversed case. The self-fertilised and crossed seeds thus obtained wereallowed to germinate in the same glass vesselon damp sand; and as the seeds germinated,they were planted in pairs on opposite sides ofthe same pot, with a superficial partition bet-

ween them, and were placed so as to be equallyexposed to the light. In other cases the self-fertilised and crossed seeds were simply sownon opposite sides of the same small pot. I have,in short, followed different plans, but in everycase have taken all the precautions which Icould think of, so that the two lots should beequally favoured. The growth of the plants rai-sed from the crossed and self-fertilised seed,were carefully observed from their germinationto maturity, in species belonging to fifty-twogenera; and the difference in their growth, andin withstanding unfavourable conditions, wasin most cases manifest and strongly marked. Itis of importance that the two lots of seedshould be sown or planted on opposite sides ofthe same pot, so that the seedlings may strug-gle against each other; for if sown separately inample and good soil, there is often but littledifference in their growth.

I will briefly describe two of the first cases ob-served by me. Six crossed and six self-fertilisedseeds of Ipomoea purpurea, from plants treatedin the manner above described, were planted assoon as they had germinated, in pairs on oppo-site sides of two pots, and rods of equal thick-ness were given them to twine up. Five of thecrossed plants grew from the first more quicklythan the opposed self-fertilised plants; thesixth, however, was weakly and was for a timebeaten, but at last its sounder constitution pre-vailed and it shot ahead of its antagonist. Assoon as each crossed plant reached the top of itsseven- foot rod its fellow was measured, andthe result was that, when the crossed plantswere seven feet high the self-fertilised had at-tained the average height of only five feet fourand a half inches. The crossed plants flowered alittle before, and more profusely than the self-fertilised plants. On opposite sides of anotherSMALL pot a large number of crossed and self-fertilised seeds were sown, so that they had to

struggle for bare existence; a single rod wasgiven to each lot: here again the crossed plantsshowed from the first their advantage; theynever quite reached the summit of the seven-foot rod, but relatively to the self-fertilisedplants their average height was as seven feet tofive feet two inches. The experiment was repea-ted during several succeeding generations,treated in exactly the same manner, and withnearly the same result. In the second genera-tion, the crossed plants, which were again cros-sed, produced 121 seed-capsules, whilst theself-fertilised, again self-fertilised, producedonly 84 capsules.

Some flowers of the Mimulus luteus were ferti-lised with their own pollen, and others werecrossed with pollen from distinct plants gro-wing in the same pot. The seeds were thicklysown on opposite sides of a pot. The seedlingswere at first equal in height; but when theyoung crossed plants were half an inch, the

self-fertilised plants were only a quarter of aninch high. But this degree of inequality did notlast, for, when the crossed plants were four anda half inches high, the self-fertilised were threeinches, and they retained the same relative dif-ference till their growth was complete. Thecrossed plants looked far more vigorous thanthe uncrossed, and flowered before them; theyproduced also a far greater number of capsules.As in the former case, the experiment was re-peated during several succeeding generations.Had I not watched these plants of Mimulus andIpomoea during their whole growth, I couldnot have believed it possible, that a differenceapparently so slight as that of the pollen beingtaken from the same flower, or from a distinctplant growing in the same pot, could have ma-de so wonderful a difference in the growth andvigour of the plants thus produced. This, undera physiological point of view, is a most remar-kable phenomenon.

With respect to the benefit derived from cros-sing distinct varieties, plenty of evidence hasbeen published. Sageret (17/49. 'Memoire surles Cucurbitacees' pages 36, 28, 30.) repeatedlyspeaks in strong terms of the vigour of melonsraised by crossing different varieties, and addsthat they are more easily fertilised than com-mon melons, and produce numerous goodseed. Here follows the evidence of an Englishgardener (17/50. Loudon's 'Gardener's Mag.'volume 8 1832 page 52.): "I have this summermet with better success in my cultivation ofmelons, in an unprotected state, from the seedsof hybrids (i.e. mongrels) obtained by crossimpregnation, than with old varieties. The offs-pring of three different hybridisations (one mo-re especially, of which the parents were the twomost dissimilar varieties I could select) eachyielded more ample and finer produce thanany one of between twenty and thirty establis-hed varieties."

Andrew Knight (17/51. 'Transact. Hort. Soc.'volume 1 page 25.) believed that his seedlingsfrom crossed varieties of the apple exhibitedincreased vigour and luxuriance; and M. Chev-reul (17/52. 'Annal. des Sc. Nat.' 3rd series, Bot.tome 6 page 189.) alludes to the extreme vigourof some of the crossed fruit- trees raised by Sa-geret.

By crossing reciprocally the tallest and shortestpeas, Knight (17/53. 'Philosophical Transac-tions' 1799 page 200.) says: "I had in this expe-riment a striking instance of the stimulativeeffects of crossing the breeds; for the smallestvariety, whose height rarely exceeded two feet,was increased to six feet: whilst the height ofthe large and luxuriant kind was very little di-minished." Mr. Laxton gave me seed-peas pro-duced from crosses between four distinct kinds;and the plants thus raised were extraordinarilyvigorous, being in each case from one to two or

three feet taller than the parent-forms growingclose alongside them.

Wiegmann (17/54. 'Ueber die Bastarderzeu-gung' 1828 s. 32, 33. For Mr. Chaundy's case seeLoudon's 'Gardener's Mag.' volume 7 1831 page696.) made many crosses between several varie-ties of cabbage; and he speaks with astonish-ment of the vigour and height of the mongrels,which excited the amazement of all the garde-ners who beheld them. Mr. Chaundy raised agreat number of mongrels by planting togethersix distinct varieties of cabbage. These mon-grels displayed an infinite diversity of charac-ter; "But the most remarkable circumstancewas, that, while all the other cabbages and bo-recoles in the nursery were destroyed by a se-vere winter, these hybrids were little injured,and supplied the kitchen when there was noother cabbage to be had."

Mr. Maund exhibited before the Royal Agricul-tural Society (17/55. 'Gardener's Chronicle'

1846 page 601.) specimens of crossed wheat,together with their parent varieties; and theeditor states that they were intermediate incharacter, "united with that greater vigour ofgrowth, which it appears, in the vegetable as inthe animal world, is the result of a first cross."Knight also crossed several varieties of wheat(17/56. 'Philosoph. Transact.' 1799 page 201.),and he says "that in the years 1795 and 1796,when almost the whole crop of corn in the is-land was blighted, the varieties thus obtained,and these only, escaped in this neighbourhood,though sown in several different soils and si-tuations."

Here is a remarkable case: M. Clotzsch (17/57.Quoted in 'Bull. Bot. Soc. France' volume 2 1855page 327.) crossed Pinus sylvestris and nigri-cans, Quercus robur and pedunculata, Alnusglutinosa and incana, Ulmus campestris andeffusa; and the cross-fertilised seeds, as well asseeds of the pure parent-trees, were all sown at

the same time and in the same place. The resultwas, that after an interval of eight years, thehybrids were one-third taller than the puretrees!

The facts above given refer to undoubted varie-ties, excepting the trees crossed by Clotzsch,which are ranked by various botanists as stron-gly-marked races, sub-species, or species. Thattrue hybrids raised from entirely distinct spe-cies, though they lose in fertility, often gain insize and constitutional vigour, is certain. Itwould be superfluous to quote any facts; for allexperimenters, Kolreuter, Gartner, Herbert,Sageret, Lecoq, and Naudin, have been struckwith the wonderful vigour, height, size, tenaci-ty of life, precocity, and hardiness of their hy-brid productions. Gartner (17/58. Gartner 'Bas-tarderzeugung' s. 259, 518, 526 et seq.) sums uphis conviction on this head in the strongestterms. Kolreuter (17/59. 'Fortsetzung' 1763 s.29; 'Dritte Fortsetzung' s. 44, 96; 'Act. Acad. St.

Petersburg' 1782 part 2 page 251; 'Nova Acta'1793 pages 391, 394; 'Nova Acta' 1795 pages316, 323.) gives numerous precise measure-ments of the weight and height of his hybridsin his comparison with measurements of bothparent-forms; and speaks with astonishment oftheir "statura portentosa," their "ambitus vastis-simus ac altitudo valde conspicua." Some ex-ceptions to the rule in the case of very sterilehybrids have, however, been noticed by Gart-ner and Herbert; but the most striking excep-tions are given by Max Wichura (17/60. 'DieBastardbefruchtung' etc. 1865 s. 31, 41, 42.) whofound that hybrid willows were generally ten-der in constitution, dwarf, and short-lived.

Kolreuter explains the vast increase in the sizeof the roots, stems, etc., of his hybrids, as theresult of a sort of compensation due to theirsterility, in the same way as many emasculatedanimals are larger than the perfect males. Thisview seems at first sight extremely probable,

and has been accepted by various authors(17/61. Max Wichura fully accepts this view('Bastardbefruchtung' s. 43), as does the Rev.M.J. Berkeley in 'Journal of Hort. Soc.' January1866 page 70.); but Gartner (17/62. 'Bastarder-zeugung' s. 394, 526, 528.) has well remarkedthat there is much difficulty in fully admittingit; for with many hybrids there is no parallelismbetween the degree of their sterility and theirincreased size and vigour. The most strikinginstances of luxuriant growth have been obser-ved with hybrids which were not sterile in anyextreme degree. In the genus Mirabilis, certainhybrids are unusually fertile, and their extraor-dinary luxuriance of growth, together withtheir enormous roots (17/63. Kolreuter 'NovaActa' 1795 page 316.) have been transmitted totheir progeny. The result in all cases is probablyin part due to the saving of nutriment and vitalforce through the sexual organs acting imper-fectly or not at all, but more especially to thegeneral law of good being derived from a cross.

For it deserves especial attention that mongrelanimals and plants, which are so far from beingsterile that their fertility is often actually aug-mented, have, as previously shown, their size,hardiness, and constitutional vigour generallyincreased. It is not a little remarkable that anaccession of vigour and size should thus ariseunder the opposite contingencies of increasedand diminished fertility.

It is a perfectly well ascertained fact (17/64.Gartner 'Bastarderzeugung' s. 430.) that hybridsinvariably breed with either pure parent, andnot rarely with a distinct species, more readilythan with one another. Herbert is inclined toexplain even this fact by the advantage derivedfrom a cross; but Gartner more justly accountsfor it by the pollen of the hybrid, and probablyits ovules, being in some degree vitiated, whe-reas the pollen and ovules of both pure parentsand of any third species are sound. Nevert-heless, there are some well-ascertained and

remarkable facts, which, as we shall presentlysee, show that a cross by itself undoubtedlytends to increase or re-establish the fertility ofhybrids.

The same law, namely, that the crossed offs-pring both of varieties and species are largerthan the parent-forms, holds good in the moststriking manner with hybrid animals as well aswith mongrels. Mr. Bartlett, who has had suchlarge experience says, "Among all hybrids ofvertebrated animals there is a marked increaseof size." He then enumerates many cases withmammals, including monkeys, and with va-rious families of birds. (17/65. Quoted by Dr.Murie in 'Proc. Zoolog. Soc.' 1870 page 40.)]

ON CERTAIN HERMAPHRODITEPLANTS WHICH, EITHER NORMALLY ORABNORMALLY, REQUIRE TO BE FERTILI-SED BY POLLEN FROM A DISTINCT INDI-VIDUAL OR SPECIES.

The facts now to be given differ from the fore-going, as self-sterility is not here the result oflong-continued close interbreeding. These factsare, however, connected with our present sub-ject, because a cross with a distinct individual isshown to be either necessary or advantageous.Dimorphic and trimorphic plants, though theyare hermaphrodites, must be reciprocally cros-sed, one set of forms by the other, in order to befully fertile, and in some cases to be fertile inany degree. But I should not have noticed theseplants, had it not been for the following casesgiven by Dr. Hildebrand (17/66. 'BotanischeZeitung' January 1864 s. 3.):—

[Primula sinensis is a reciprocally dimorphicspecies: Dr. Hildebrand fertilised twenty-eightflowers of both forms, each by pollen of theother form, and obtained the full number ofcapsules containing on an average 42.7 seed percapsule; here we have complete and normalfertility. He then fertilised forty-two flowers of

both forms with pollen of the same form, buttaken from a distinct plant, and all producedcapsules containing on an average only 19.6seed. Lastly, and here we come to our moreimmediate point, he fertilised forty-eight flo-wers of both forms with pollen of the sameform and taken from the same flower, and nowhe obtained only thirty-two capsules, and thesecontained on an average 18.6 seed, or one lessper capsule than in the former case. So that,with these illegitimate unions, the act of im-pregnation is less assured, and the fertilityslightly less, when the pollen and ovules be-long to the same flower, than when belongingto two distinct individuals of the same form.Dr. Hildebrand has recently made analogousexperiments on the long-styled form of Oxalisrosea, with the same result. (17/67. 'Monatsbe-richt Akad. Wissen.' Berlin 1866 s. 372.)]

It has recently been discovered that certainplants, whilst growing in their native country

under natural conditions, cannot be fertilisedwith pollen from the same plant. They are so-metimes so utterly self-impotent, that, thoughthey can readily be fertilised by the pollen of adistinct species or even distinct genus, yet,wonderful as is the fact, they never produce asingle seed by their own pollen. In some cases,moreover, the plant's own pollen and stigmamutually act on each other in a deleteriousmanner. Most of the facts to be given relate toorchids, but I will commence with a plant be-longing to a widely different family.

[Sixty-three flowers of Corydalis cava, borne ondistinct plants, were fertilised by Dr. Hilde-brand (17/68. International Hort. Congress,London 1866.) with pollen from other plants ofthe same species; and fifty-eight capsules wereobtained, including on an average 4.5 seed ineach. He then fertilised sixteen flowers produ-ced by the same raceme, one with another, butobtained only three capsules, one of which alo-

ne contained any good seeds, namely, two innumber. Lastly, he fertilised twenty-seven flo-wers, each with its own pollen; he left also fifty-seven flowers to be spontaneously fertilised,and this would certainly have ensued if it hadbeen possible, for the anthers not only touchthe stigma, but the pollen-tubes were seen byDr. Hildebrand to penetrate it; neverthelessthese eighty-four flowers did not produce asingle seed-capsule! This whole case is highlyinstructive, as it shows how widely differentthe action of the same pollen is, according as itis placed on the stigma of the same flower, oron that of another flower on the same raceme,or on that of a distinct plant.

With exotic Orchids several analogous caseshave been observed, chiefly by Mr. John Scott.(17/69. 'Proc. Bot. Soc. of Edinburgh' May 1863:these observations are given in abstract, andothers are added, in the 'Journal of Proc. ofLinn. Soc.' volume 8 Bot. 1864 page 162.) Onci-

dium sphacelatum has effective pollen, for Mr.Scott fertilised two distinct species with it; theovules are likewise capable of impregnation,for they were readily fertilised by the pollen ofO. divaricatum; nevertheless, between one andtwo hundred flowers fertilised by their ownpollen did not produce a single capsule, thoughthe stigmas were penetrated by the pollen-tubes. Mr. Robertson Munro, of the Royal Bo-tanic Gardens of Edinburgh, also informs me(1864) that a hundred and twenty flowers ofthis same species were fertilised by him withtheir own pollen, and did not produce a capsu-le, but eight flowers, fertilised by the pollen ofO. divaricatum, produced four fine capsules:again, between two and three hundred flowersof O. divaricatum, fertilised by their own po-llen, did not set a capsule, but twelve flowersfertilised by O. flexuosum produced eight finecapsules: so that here we have three utterlyself-impotent species, with their male and fe-male organs perfect, as shown by their mutual

fertilisation. In these cases fertilisation was ef-fected only by the aid of a distinct species. But,as we shall presently see, distinct plants, raisedfrom seed, of Oncidium flexuosum, and proba-bly of the other species, would have been per-fectly capable of fertilising each other, for this isthe natural process. Again, Mr. Scott found thatthe pollen of a plant of O. microchilum waseffective, for with it he fertilised two distinctspecies; he found its ovules good, for theycould be fertilised by the pollen of one of thesespecies, and by the pollen of a distinct plant ofO. microchilum; but they could not be fertilisedby pollen of the same plant, though the pollen-tubes penetrated the stigma. An analogous casehas been recorded by M. Riviere (17/70. Prof.Lecoq 'De la Fecondation' 2nd edition 1862 pa-ge 76.) with two plants of O. cavendishianum,which were both self-sterile, but reciprocallyfertilised each other. All these cases refer to thegenus Oncidium, but Mr. Scott found thatMaxillaria atro-rubens was "totally insuscepti-

ble of fertilisation with its own pollen," but fer-tilised, and was fertilised by, a widely distinctspecies, viz. M. squalens.

As these orchids had been grown under unna-tural conditions in hot-houses, I concluded thattheir self-sterility was due to this cause. ButFritz Muller informs me that at Desterro, inBrazil, he fertilised above one hundred flowersof the above-mentioned Oncidium flexuosum,which is there endemic, with its own pollen,and with that taken from distinct plants: all theformer were sterile, whilst those fertilised bypollen from any OTHER PLANT of the samespecies were fertile. During the first three daysthere was no difference in the action of the twokinds of pollen: that placed on stigma of thesame plant separated in the usual manner intograins, and emitted tubes which penetrated thecolumn, and the stigmatic chamber shut itself;but only those flowers which had been fertili-sed by pollen taken from a distinct plant pro-

duced seed-capsules. On a subsequent occasionthese experiments were repeated on a largescale with the same result. Fritz Muller foundthat four other endemic species of Oncidiumwere in like manner utterly sterile with theirown pollen, but fertile with that from any otherplant: some of them likewise produced seed-capsules when impregnated with pollen of wi-dely distinct genera, such as Cyrtopodium, andRodriguezia. Oncidium crispum, however, dif-fers from the foregoing species in varying muchin its self- sterility; some plants producing finepods with their own pollen, others failing to doso in two or three instances, Fritz Muller obser-ved that the pods produced by pollen takenfrom a distinct flower on the same plant, werelarger than those produced by the flower's ownpollen. In Epidendrum cinnabarinum, an or-chid belonging to another division of the fami-ly, fine pods were produced by the plant's ownpollen, but they contained by weight onlyabout half as much seed as the capsules which

had been fertilised by pollen from a distinctplant, and in one instance from a distinct spe-cies; moreover, a very large proportion, and insome cases nearly all the seeds produced by theplant's own pollen, were destitute of an embr-yo. Some self-fertilised capsules of a Maxillariawere in a similar state.

Another observation made by Fritz Muller ishighly remarkable, namely, that with variousorchids the plant's own pollen not only fails toimpregnate the flower, but acts on the stigma,and is acted on, in an injurious or poisonousmanner. This is shown by the surface of thestigma in contact with the pollen, and by thepollen itself becoming in from three to fivedays dark brown, and then decaying. The dis-coloration and decay are not caused by parasi-tic cryptograms, which were observed by FritzMuller in only a single instance. These changesare well shown by placing on the same stigma,at the same time, the plant's own pollen and

that from a distinct plant of the same species, orof another species, or even of another and wi-dely remote genus. Thus, on the stigma of On-cidium flexuosum, the plant's own pollen andthat from a distinct plant were placed side byside, and in five days' time the latter was per-fectly fresh, whilst the plant's own pollen wasbrown. On the other hand, when the pollen of adistinct plant of the Oncidium flexuosum andof the Epidendrum zebra (nov. spec.?) wereplaced together on the same stigma, they beha-ved in exactly the same manner, the grains se-parating, emitting tubes, and penetrating thestigma, so that the two pollen-masses, after aninterval of eleven days, could not be distin-guished except by the difference of their caudi-cles, which, of course, undergo no change. FritzMuller has, moreover, made a large number ofcrosses between orchids belonging to distinctspecies and genera, and he finds that in all ca-ses when the flowers are not fertilised theirfootstalks first begin to wither; and the wit-

hering slowly spreads upwards until the ger-mens fall off, after an interval of one or twoweeks, and in one instance of between six andseven weeks; but even in this latter case, and inmost other cases, the pollen and stigma remai-ned in appearance fresh. Occasionally, howe-ver, the pollen becomes brownish, generally onthe external surface, and not in contact with thestigma, as is invariably the case when theplant's own pollen is applied.

Fritz Muller observed the poisonous action ofthe plant's own pollen in the above-mentionedOncidium flexuosum, O. unicorne, pubes (?),and in two other unnamed species. Also in twospecies of Rodriguezia, in two of Notylia, inone of Burlingtonia, and of a fourth genus inthe same group. In all these cases, except thelast, it was proved that the flowers were, asmight have been expected, fertile with pollenfrom a distinct plant of the same species. Nu-merous flowers of one species of Notylia were

fertilised with pollen from the same raceme; intwo days' time they all withered, the germensbegan to shrink, the pollen-masses became darkbrown, and not one pollen-grain emitted a tu-be. So that in this orchid the injurious action ofthe plant's own pollen is more rapid than withOncidium flexuosum. Eight other flowers onthe same raceme were fertilised with pollenfrom a distinct plant of the same species: two ofthese were dissected, and their stigmas werefound to be penetrated by numberless pollen-tubes; and the germens of the other six flowersbecame well developed. On a subsequent occa-sion many other flowers were fertilised withtheir own pollen, and all fell off dead in a fewdays; whilst some flowers on the same racemewhich had been left simply unfertilised ad-hered and long remained fresh. We have seenthat in cross-unions between extremely distinctorchids the pollen long remains undecayed; butNotylia behaved in this respect differently; forwhen its pollen was placed on the stigma of

Oncidium flexuosum, both the stigma and po-llen quickly became dark brown, in the samemanner as if the plant's own pollen had beenapplied.

Fritz Muller suggests that, as in all these casesthe plant's own pollen is not only impotent(thus effectually preventing self-fertilisation),but likewise prevents, as was ascertained in thecase of the Notylia and Oncidium flexuosum,the action of subsequently applied pollen froma distinct individual, it would be an advantageto the plant to have its own pollen renderedmore and more deleterious; for the germenswould thus quickly be killed, and dropping off,there would be no further waste in nourishinga part which ultimately could be of no avail.

The same naturalist found in Brazil three plantsof a Bignonia growing near together. He fertili-sed twenty-nine flowerets on one of them withtheir own pollen, and they did not set a singlecapsule. Thirty flowers were then fertilised

with pollen from a distinct plant, one of thethree, and they yielded only two capsules. Las-tly, five flowers were fertilised with pollenfrom a fourth plant growing at a distance, andall five produced capsules. Fritz Muller thinksthat the three plants which grew near oneanother were probably seedlings from the sameparent, and that from being closely related,they acted very feebly on one another. Thisview is extremely probable, for he has sinceshown in a remarkable paper (17/71. 'JenaischeZeitschrift fur Naturwiss.' b. 7 page 22 1872 andpage 441 1873. A large part of this paper hasbeen translated in the 'American Naturalist'1874 page 223.), that in the case of some Brazi-lian species of Abutilon, which are self-sterile,and between which he raised some complexhybrids, that these, if near relatives, were muchless fertile inter se, than when not closely rela-ted.]

We now come to cases closely analogous withthose just given, but different in so far that onlycertain individuals of the species are self-sterile.This self-impotence does not depend on thepollen or ovules being in an unfit state for ferti-lisation, for both have been found effective inunion with other plants of the same or of a dis-tinct species. The fact of plants having acquiredso peculiar a constitution, that they can be ferti-lised more readily by the pollen of a distinctspecies than by their own, is exactly the reverseof what occurs with all ordinary species. For inthe latter the two sexual elements of the sameindividual plant are of course capable of freelyacting on each other; but are so constituted thatthey are more or less impotent when broughtinto union with the sexual elements of a dis-tinct species, and produce more or less sterilehybrids.

[Gartner experimented on two plants of Lobeliafulgens, brought from separate places, and

found (17/72. 'Bastarderzeugung' s. 64, 357.)that their pollen was good, for he fertilised withit L. cardinalis and syphilitica; their ovules we-re likewise good, for they were fertilised by thepollen of these same two species; but these twoplants of L. fulgens could not be fertilised bytheir own pollen, as can generally be effectedwith perfect ease with this species. Again, thepollen of a plant of Verbascum nigrum grownin a pot was found by Gartner (17/73. Ibid s.357.) capable of fertilising V. lychnitis and V.austriacum; the ovules could be fertilised bythe pollen of V. thapsus; but the flowers couldnot be fertilised by their own pollen. Kolreuter,also (17/74. 'Zweite Fortsetzung' s. 10; 'DritteForts.' s. 40. Mr. Scott likewise fertilised fifty-four flowers of Verbascum phoeniceum, inclu-ding two varieties, with their own pollen, andnot a single capsule was produced. Many of thepollen-grains emitted their tubes, but only afew of them penetrated the stigmas; some slighteffect however was produced, as many of the

ovaries became somewhat developed: 'JournalAsiatic Soc. Bengal' 1867 page 150.), gives thecase of three garden plants of Verbascumphoeniceum, which bore during two years ma-ny flowers; these he fertilised successfully withthe pollen of no less than four distinct species,but they produced not a seed with their ownapparently good pollen; subsequently thesesame plants, and others raised from seed, as-sumed a strangely fluctuating condition, beingtemporarily sterile on the male or female side,or on both sides, and sometimes fertile on bothsides; but two of the plants were perfectly ferti-le throughout the summer.

With Reseda odorata I have found certain indi-viduals quite sterile with their own pollen, andso it is with the indigenous Reseda lutea. Theself-sterile plants of both species were perfectlyfertile when crossed with pollen from any otherindividual of the same species. These observa-tions will hereafter be published in another

work, in which I shall also show that seeds sentto me by Fritz Muller produced by plants ofEschscholtzia californica which were quite self-sterile in Brazil, yielded in this country plantswhich were only slightly self-sterile.

It appears (17/75. Duvernoy quoted by Gartner'Bastarderzeugung' s. 334) that certain flowerson certain plants of Lilium candidum can befertilised more freely by pollen from a distinctindividual than by their own. So, again, withthe varieties of the potato. Tinzmann (17/76.'Gardener's Chronicle' 1846 page 183.), whomade many trials with this plant, says that po-llen from another variety sometimes "exerts apowerful influence, and I have found sorts ofpotatoes which would not bear seed from im-pregnation with the pollen of their own flowerswould bear it when impregnated with otherpollen." It does not, however, appear to havebeen proved that the pollen which failed to acton the flower's own stigma was in itself good.

In the genus Passiflora it has long been knownthat several species do not produce fruit, unlessfertilised by pollen taken from distinct species:thus, Mr. Mowbray (17/77. 'Transact. Hort.Soc.' volume 7 1830 page 95.) found that hecould not get fruit from P. alata and racemosaexcept by reciprocally fertilising them witheach other's pollen; and similar facts have beenobserved in Germany and France. (17/78. Prof.Lecoq 'De la Fecondation' 1845 page 70; Gartner'Bastarderzeugung' s. 64.) I have received twoaccounts of P. quadrangularis never producingfruit from its own pollen, but doing so freelywhen fertilised in one case with the pollen of P.coerulea, and in another case with that of P.edulis. But in three other cases this species frui-ted freely when fertilised with its own pollen;and the writer in one case attributed the favou-rable result to the temperature of the househaving been raised from 5 deg to 10 deg Fahr.above the former temperature, after the flowerswere fertilised. (17/79. 'Gardener's Chronicle'

1868 page 1341.) With respect to P. laurifolia, acultivator of much experience has recently re-marked (17/80. 'Gardener's Chronicle' 1866page 1068.) that the flowers "must be fertilisedwith the pollen of P. coerulea, or of some othercommon kind, as their own pollen will not fer-tilise them." But the fullest details on this sub-ject have been given by Messrs. Scott and Ro-bertson Munro (17/81. 'Journal of Proc. of Linn.Soc.' volume 8 1864 page 1168. Mr. RobertsonMunro in 'Trans. Bot. Soc.' of Edinburgh volu-me 9 page 399.): plants of Passiflora racemosa,coerulea, and alata flowered profusely duringmany years in the Botanic Gardens of Edin-burgh, and, though repeatedly fertilised withtheir own pollen, never produced any seed; yetthis occurred at once with all three specieswhen they were crossed together in variousways. In the case of P. coerulea three plants,two of which grew in the Botanic Gardens, we-re all rendered fertile, merely by impregnatingeach with pollen of one of the others. The same

result was attained in the same manner with P.alata, but with only one plant out of three. Asso many self-sterile species of Passiflora havebeen mentioned, it should be stated that theflowers of the annual P. gracilis are nearly asfertile with their own pollen as with that from adistinct plant; thus sixteen flowers sponta-neously self-fertilised produced fruit, each con-taining on an average 21.3 seed, whilst fruitfrom fourteen crossed flowers contained 24.1seed.

Returning to P. alata, I have received (1866)some interesting details from Mr. RobertsonMunro. Three plants, including one in England,have already been mentioned which were inve-terately self-sterile, and Mr. Munro informs meof several others which, after repeated trialsduring many years, have been found in thesame predicament. At some other places,however, this species fruits readily when ferti-lised with its own pollen. At Taymouth Castle

there is a plant which was formerly grafted byMr. Donaldson on a distinct species, nameunknown, and ever since the operation it hasproduced fruit in abundance by its own pollen;so that this small and unnatural change in thestate of this plant has restored its self-fertility!Some of the seedlings from the Taymouth Cas-tle plant were found to be not only sterile withtheir own pollen, but with each other's pollen,and with the pollen of distinct species. Pollenfrom the Taymouth plant failed to fertilise cer-tain plants of the same species, but was success-ful on one plant in the Edinburgh Botanic Gar-dens. Seedlings were raised from this latterunion, and some of their flowers were fertilisedby Mr. Munro with their own pollen; but theywere found to be as self-impotent as the mot-her-plant had always proved, except when fer-tilised by the grafted Taymouth plant, and ex-cept, as we shall see, when fertilised by herown seedlings. For Mr. Munro fertilised eigh-teen flowers on the self-impotent mother-plant

with pollen from these her own self-impotentseedlings, and obtained, remarkable as the factis, eighteen fine capsules full of excellent seed! Ihave met with no case in regard to plantswhich shows so well as this of P. alata, on whatsmall and mysterious causes complete fertilityor complete sterility depends.]

The facts hitherto given relate to the much-lessened or completely destroyed fertility ofpure species when impregnated with their ownpollen, in comparison with their fertility whenimpregnated by distinct individuals or distinctspecies; but closely analogous facts have beenobserved with hybrids.

[Herbert states (17/82. 'Amaryllidaceae' 1837page 371; 'Journal of Hort. Soc.' volume 2 1847page 19.) that having in flower at the same timenine hybrid Hippeastrums, of complicated ori-gin, descended from several species, he foundthat "almost every flower touched with pollenfrom another cross produced seed abundantly,

and those which were touched with their ownpollen either failed entirely, or formed slowly apod of inferior size, with fewer seeds." In the'Horticultural Journal' he adds that "the admis-sion of the pollen of another cross-bred Hip-peastrum (however complicated the cross) toany one flower of the number, is almost sure tocheck the fructification of the others." In a letterwritten to me in 1839, Dr. Herbert says that hehad already tried these experiments during fiveconsecutive years, and he subsequently repea-ted them, with the same invariable result. Hewas thus led to make an analogous trial on apure species, namely, on the Hippeastrum auli-cum, which he had lately imported from Brazil:this bulb produced four flowers, three of whichwere fertilised by their own pollen, and thefourth by the pollen of a triple cross between H.bulbulosum, reginae, and vittatum; the resultwas, that "the ovaries of the three first flowerssoon ceased to grow, and after a few days pe-rished entirely: whereas the pod impregnated

by the hybrid made vigorous and rapid pro-gress to maturity, and bore good seed, whichvegetated freely." This is, indeed, as Herbertremarks, "a strange truth," but not so strange asit then appeared.

As a confirmation of these statements, I mayadd that Mr. M. Mayes (17/83. Loudon's 'Gar-dener's Magazine' volume 11 1835 page 260.)after much experience in crossing the species ofAmaryllis (Hippeastrum), says, "neither thespecies nor the hybrids will, we are well aware,produce seed so abundantly from their ownpollen as from that of others." So, again, Mr.Bidwell, in New South Wales (17/84. 'Garde-ner's Chronicle' 1850 page 470.) asserts thatAmaryllis belladonna bears many more seedswhen fertilised by the pollen of Brunswigia(Amaryllis of some authors) josephinae or of B.multiflora, than when fertilised by its own po-llen. Mr. Beaton dusted four flowers of a Cyr-tanthus with their own pollen, and four with

the pollen of Vallota (Amaryllis) purpurea; onthe seventh day "those which received theirown pollen slackened their growth, and ultima-tely perished; those which were crossed withthe Vallota held on." (17/85. 'Journal Hort. Soc.'volume 5 page 135. The seedlings thus raisedwere given to the Hort. Soc.; but I find, on in-quiry, that they unfortunately died the follo-wing winter.) These latter cases, however, rela-te to uncrossed species, like those before givenwith respect to Passiflora, Orchids, etc., and arehere referred to only because the plants belongto the same group of Amaryllidaceae.

In the experiments on the hybrid Hippeas-trums, if Herbert had found that the pollen oftwo or three kinds alone had been more effi-cient on certain kinds than their own pollen, itmight have been argued that these, from theirmixed parentage, had a closer mutual affinitythan the others; but this explanation is inadmis-sible, for the trials were made reciprocally

backwards and forwards on nine different hy-brids; and a cross, whichever way taken, al-ways proved highly beneficial. I can add a stri-king and analogous case from experiments ma-de by the Rev. A. Rawson, of Bromley Com-mon, with some complex hybrids of Gladiolus.This skilful horticulturist possessed a numberof French varieties, differing from each otheronly in the colour and size of the flowers, alldescended from Gandavensis, a well-knownold hybrid, said to be descended from G. nata-lensis by the pollen of G. oppositiflorus. (17/86.Mr. D. Beaton in 'Journal of Hort.' 1861 page453. Lecoq however ('De la Fecond.' 1862 page369) states that this hybrid is descended fromG. psittacinus and cardinalis; but this is oppo-sed to Herbert's experience, who found that theformer species could not be crossed.) Mr. Raw-son, after repeated trials, found that none of thevarieties would set seed with their own pollen,although taken from distinct plants of the samevariety (which had, of course, been propagated

by bulbs), but that they all seeded freely withpollen from any other variety. To give twoexamples: Ophir did not produce a capsulewith its own pollen, but when fertilised withthat of Janire, Brenchleyensis, Vulcain and Lin-ne, it produced ten fine capsules; but the pollenof Ophir was good, for when Linne was fertili-sed by it seven capsules were produced. Thislatter variety, on the other hand, was utterlybarren with its own pollen, which we have seenwas perfectly efficient on Ophir. Altogether,Mr. Rawson, in the year 1861 fertilised twenty-six flowers borne by four varieties with pollentaken from other varieties, and every singleflower produced a fine seed-capsule; whereasfifty-two flowers on the same plants, fertilisedat the same time with their own pollen, did notyield a single seed-capsule. Mr. Rawson fertili-sed, in some cases, the alternate flowers, and inother cases all those down one side of the spike,with pollen of other varieties, and the remai-ning flowers with their own pollen. I saw these

plants when the capsules were nearly mature,and their curious arrangement at once broughtfull conviction to the mind that an immenseadvantage had been derived from crossing the-se hybrids.

Lastly, I have heard from Dr. E. Bornet, of An-tibes, who has made numerous experiments incrossing the species of Cistus, but has not yetpublished the results, that, when any of thesehybrids are fertile, they may be said to be, inregard to function, dioecious; "for the flowersare always sterile when the pistil is fertilised bypollen taken from the same flower or from flo-wers on the same plant. But they are often ferti-le if pollen be employed from a distinct indivi-dual of the same hybrid nature, or from a hy-brid made by a reciprocal cross."]

CONCLUSION.

That plants should be self-sterile, although bothsexual elements are in a fit state for reproduc-

tion, appears at first sight opposed to all analo-gy. With respect to the species, all the indivi-duals of which are in this state, although livingunder their natural conditions, we may conclu-de that their self-sterility has been acquired forthe sake of effectually preventing self- fertilisa-tion. The case is closely analogous with that ofdimorphic and trimorphic or heterostyledplants, which can be fully fertilised only byplants belonging to a different form, and not, asin the foregoing cases, indifferently by any ot-her individual of the species. Some of thesehetero- styled plants are completely sterile withpollen taken from the same plant or from thesame form. With respect to species living undertheir natural conditions, of which only certainindividuals are self-sterile (as with Reseda lu-tea), it is probable that these have been rende-red self-sterile to ensure occasional cross-fertilisation, whilst other individuals have re-mained self- fertile to ensure the propagation ofthe species. The case seems to be parallel with

that of plants which produce, as Hermann Mu-ller has discovered, two forms—one bearingmore conspicuous flowers with their structureadapted for cross-fertilisation by insects, andthe other form with less conspicuous flowersadapted for self-fertilisation. The self-sterility,however, of some of the foregoing plants isincidental on the conditions to which they havebeen subjected, as with the Eschscholtzia, theVerbascum phoeniceum (the sterility of whichvaried according to the season), and with thePassiflora alata, which recovered its self-fertility when grafted on a different stock.

It is interesting to observe in the above severalcases the graduated series from plants which,when fertilised by their own pollen, yield thefull number of seeds, but with the seedlings alittle dwarfed in stature—to plants which whenself-fertilised yield few seeds—to those whichyield none, but have their ovaria somewhatdeveloped—and, lastly, to those in which the

plant's own pollen and stigma mutually act onone another like poison. It is also interesting toobserve on how slight a difference in the natureof the pollen or of the ovules complete self-sterility or complete self-fertility must dependin some of the above cases. Every individual ofthe self-sterile species appears to be capable ofproducing the full complement of seed whenfertilised by the pollen of any other individual(though judging from the facts given with res-pect to Abutilon the nearest kin must be excep-ted); but not one individual can be fertilised byits own pollen. As every organism differs insome slight degree from every other individualof the same species, so no doubt it is with theirpollen and ovules; and in the above cases wemust believe that complete self-sterility andcomplete self-fertility depend on such slightdifferences in the ovules and pollen, and nottheir having been differentiated in some specialmanner in relation to one another; for it is im-possible that the sexual elements of many thou-

sand individuals should have been specialisedin relation to every other individual. In some,however, of the above cases, as with certainPassifloras, an amount of differentiation bet-ween the pollen and ovules sufficient for fertili-sation is gained only by employing pollen froma distinct species; but this is probably the resultof such plants having been rendered somewhatsterile from the unnatural conditions to whichthey have been exposed.

Exotic animals confined in menageries are so-metimes in nearly the same state as the above-described self-impotent plants; for, as we shallsee in the following chapter, certain monkeys,the larger carnivora, several finches, geese, andpheasants, cross together, quite as freely as, oreven more freely than the individuals of thesame species breed together. Cases will, also, begiven of sexual incompatibility between cer-tain, male and female domesticated animals,

which, nevertheless, are fertile when matchedwith any other individual of the same kind.

In the early part of this chapter it was shownthat the crossing of individuals belonging todistinct families of the same race, or to differentraces or species, gives increased size and consti-tutional vigour to the offspring, and, except inthe case of crossed species, increased fertility.The evidence rests on the universal testimonyof breeders (for it should be observed that I amnot here speaking of the evil results of closeinterbreeding), and is practically exemplified inthe higher value of cross- bred animals for im-mediate consumption. The good results of cros-sing have also been demonstrated with someanimals and with numerous plants, by actualweight and measurement. Although animals ofpure blood will obviously be deteriorated bycrossing, as far as their characteristic qualitiesare concerned, there seems to be no exceptionto the rule that advantages of the kind just

mentioned are thus gained, even when therehas not been any previous close interbreeding;and the rule applies to such animals as cattleand sheep, which can long resist breeding in-and-in between the nearest blood-relations.

In the case of crossed species, although size,vigour, precocity, and hardiness are, with rareexceptions, gained, fertility, in a greater or lessdegree, is lost; but the gain in the above res-pects can hardly be attributed to the principleof compensation; for there is no close paralle-lism between the increased size and vigour ofhybrid offspring and their sterility. Moreover, ithas been clearly proved that mongrels whichare perfectly fertile gain these same advantagesas well as sterile hybrids.

With the higher animals no special adaptationsfor ensuring occasional crosses between distinctfamilies seem to exist. The eagerness of the ma-les, leading to severe competition betweenthem, is sufficient; for even with gregarious

animals, the old and dominant males will bedispossessed after a time and it would be a me-re chance if a closely related member of thesame family were to be the victorious succes-sor. The structure of many of the lower ani-mals, when they are hermaphrodites, is such asto prevent the ovules being fertilised by themale element of the same individual; so thatthe concourse of two individuals is necessary.In other cases the access of the male element ofa distinct individual is at least possible. Withplants, which are affixed to the ground andcannot wander from place to place like animals,the numerous adaptations for cross-fertilisationare wonderfully perfect, as has been admittedby every one who has studied the subject.

The evil consequences of long-continued closeinterbreeding are not so easily recognised asthe good effects from crossing, for the deterio-ration is gradual. Nevertheless, it is the generalopinion of those who have had most experien-

ce, especially with animals which propagatequickly, that evil does inevitably follow sooneror later, but at different rates with differentanimals. No doubt a false belief may, like a su-perstition, prevail widely; yet it is difficult tosuppose that so many acute observers have allbeen deceived at the expense of much cost andtrouble. A male animal may sometimes be pai-red with his daughter, granddaughter, and soon, even for seven generations, without anymanifest bad result: but the experiment hasnever been tried of matching brothers and sis-ters, which is considered the closest form ofinterbreeding, for an equal number of genera-tions. There is good reason to believe that bykeeping the members of the same family indistinct bodies, especially if exposed to so-mewhat different conditions of life, and by oc-casionally crossing these families, the evil re-sults of interbreeding may be much diminishedor quite eliminated. These results are loss ofconstitutional vigour, size, and fertility; but

there is no necessary deterioration in the gene-ral form of the body, or in other good qualities.We have seen that with pigs first-rate animalshave been produced after long-continued closeinterbreeding, though they had become extre-mely infertile when paired with their near rela-tions. The loss of fertility, when it occurs, seemsnever to be absolute, but only relative to ani-mals of the same blood; so that this sterility isto a certain extent analogous with that of self-impotent plants which cannot be fertilised bytheir own pollen, but are perfectly fertile withpollen of any other individual of the same spe-cies. The fact of infertility of this peculiar natu-re being one of the results of long-continuedinterbreeding, shows that interbreeding doesnot act merely by combining and augmentingvarious morbid tendencies common to bothparents; for animals with such tendencies, if notat the time actually ill, can generally propagatetheir kind. Although offspring descended fromthe nearest blood-relations are not necessarily

deteriorated in structure, yet some authors be-lieve that they are eminently liable to malfor-mations; and this is not improbable, as everyt-hing which lessens the vital powers acts in thismanner. Instances of this kind have been recor-ded in the case of pigs, bloodhounds, and someother animals.

Finally, when we consider the various factsnow given which plainly show that good fo-llows from crossing, and less plainly that evilfollows from close interbreeding, and when webear in mind that with very many organismselaborate provisions have been made for theoccasional union of distinct individuals, theexistence of a great law of nature is almost pro-ved; namely, that the crossing of animals andplants which are not closely related to eachother is highly beneficial or even necessary, andthat interbreeding prolonged during many ge-nerations is injurious.

CHAPTER 2.XVIII.

ON THE ADVANTAGES AND DISADVAN-TAGES OF CHANGED CONDITIONS OF LI-FE: STERILITY FROM VARIOUS CAUSES.

ON THE GOOD DERIVED FROM SLIGHTCHANGES IN THE CONDITIONS OF LIFE.STERILITY FROM CHANGED CONDITIONS,IN ANIMALS, IN THEIR NATIVE COUNTRYAND IN MENAGERIES. MAMMALS, BIRDS,AND INSECTS. LOSS OF SECONDARYSEXUAL CHARACTERS AND OF INSTINCTS.CAUSES OF STERILITY. STERILITY OF DO-MESTICATED ANIMALS FROM CHANGEDCONDITIONS. SEXUAL INCOMPATIBILITYOF INDIVIDUAL ANIMALS. STERILITY OFPLANTS FROM CHANGED CONDITIONS OFLIFE. CONTABESCENCE OF THE ANTHERS.MONSTROSITIES AS A CAUSE OF STERILI-TY. DOUBLE FLOWERS. SEEDLESS FRUIT.STERILITY FROM THE EXCESSIVE DEVE-LOPMENT OF THE ORGANS OF VEGETA-

TION. FROM LONG-CONTINUED PROPA-GATION BY BUDS. INCIPIENT STERILITYTHE PRIMARY CAUSE OF DOUBLE FLO-WERS AND SEEDLESS FRUIT.

ON THE GOOD DERIVED FROM SLIGHTCHANGES IN THE CONDITIONS OF LIFE.

In considering whether any facts were knownwhich might throw light on the conclusion arri-ved at in the last chapter, namely, that benefitsensue from crossing, and that it is a law of na-ture that all organic beings should occasionallycross, it appeared to me probable that the goodderived from slight changes in the conditions oflife, from being an analogous phenomenon,might serve this purpose. No two individuals,and still less no two varieties, are absolutelyalike in constitution and structure; and whenthe germ of one is fertilised by the male ele-ment of another, we may believe that it is actedon in a somewhat similar manner as an indivi-dual when exposed to slightly changed condi-

tions. Now, every one must have observed theremarkable influence on convalescents of achange of residence, and no medical mandoubts the truth of this fact. Small farmers whohold but little land are convinced that their cat-tle derive great benefit from a change of pastu-re. In the case of plants, the evidence is strongthat a great advantage is derived from exchan-ging seeds, tubers, bulbs, and cuttings from onesoil or place to another as different as possible.

[The belief that plants are thus benefited, whet-her or not well founded, has been firmly main-tained from the time of Columella, who wroteshortly after the Christian era, to the presentday; and it now prevails in England, France,and Germany. (18/1. For England see below.For Germany see Metzger 'Getreidearten' 1841s. 63. For France Loiseleur-Deslongchamps('Consid. sur les Cereales' 1843 page 200) givesnumerous references on this subject. For Sout-hern France see Godron 'Florula Juvenalis' 1854

page 28.) A sagacious observer, Bradley, wri-ting in 1724 (18/2. 'A General Treatise of Hus-bandry' volume 3 page 58.), says, "When weonce become Masters of a good Sort of Seed, weshould at least put it into Two or Three Hands,where the Soils and Situations are as differentas possible; and every Year the Parties shouldchange with one another; by which Means, Ifind the Goodness of the Seed will be maintai-ned for several Years. For Want of this Use ma-ny Farmers have failed in their Crops and beengreat Losers." He then gives his own practicalexperience on this head. A modern writer(18/3. 'Gardener's Chronicle and Agricult. Ga-zette' 1858 page 247; and for the second state-ment, Ibid 1850 page 702. On this same subjectsee also Rev. D. Walker 'Prize Essay of High-land Agricult. Soc.' volume 2 page 200. AlsoMarshall 'Minutes of Agriculture' November1775.) asserts, "Nothing can be more clearlyestablished in agriculture than that the conti-nual growth of any one variety in the same dis-

trict makes it liable to deterioration either inquality or quantity." Another writer states thathe sowed close together in the same field twolots of wheat-seed, the product of the same ori-ginal stock, one of which had been grown onthe same land and the other at a distance, andthe difference in favour of the crop from thelatter seed was remarkable. A gentleman inSurrey who has long made it his business toraise wheat to sell for seed, and who has cons-tantly realised in the market higher prices thanothers, assures me that he finds it indispensablecontinually to change his seed; and that for thispurpose he keeps two farms differing much insoil and elevation.

With respect to the tubers of the potato, I findthat at the present day the practice of exchan-ging sets is almost everywhere followed. Thegreat growers of potatoes in Lancashire former-ly used to get tubers from Scotland, but theyfound that "a change from the moss-lands, and

vice versa, was generally sufficient." In formertimes in France the crop of potatoes in the Vos-ges had become reduced in the course of fiftyor sixty years in the proportion from 120-150 to30-40 bushels; and the famous Oberlin attribu-ted the surprising good which he effected inlarge part to changing the sets. (18/4. Oberlin'Memoirs' English translation page 73. For Lan-cashire see Marshall 'Review of Reports' 1808page 295.)

A well-known practical gardener, Mr. Robson(18/5. 'Cottage Gardener' 1856 page 186. ForMr. Robson's subsequent statements see 'Jour-nal of Horticulture' February 18, 1866 page 121.For Mr. Abbey's remarks on grafting etc. IbidJuly 18, 1865 page 44.) positively states that hehas himself witnessed decided advantage fromobtaining bulbs of the onion, tubers of the pota-to, and various seeds, all of the same kind, fromdifferent soils and distant parts of England. Hefurther states that with plants propagated by

cuttings, as with the Pelargonium, and especia-lly the Dahlia, manifest advantage is derivedfrom getting plants of the same variety, whichhave been cultivated in another place; or, "whe-re the extent of the place allows, to take cut-tings from one description of soil to plant onanother, so as to afford the change that seemsso necessary to the well-being of the plants." Hemaintains that after a time an exchange of thisnature is "forced on the grower, whether he beprepared for it or not." Similar remarks havebeen made by another excellent gardener, Mr.Fish, namely, that cuttings of the same varietyof Calceolaria, which he obtained from aneighbour, "showed much greater vigour thansome of his own that were "treated in exactlythe same manner," and he attributed this solelyto his own plants having become "to a certainextent worn out or tired of their quarters." So-mething of this kind apparently occurs in graf-ting and budding fruit-trees; for, according toMr. Abbey, grafts or buds generally take with

greater facility on a distinct variety or evenspecies, or on a stock previously grafted, thanon stocks raised from seeds of the varietywhich is to be grafted; and he believes this can-not be altogether explained by the stocks inquestion being better adapted to the soil andclimate of the place. It should, however, be ad-ded, that varieties grafted or budded on verydistinct kinds, though they may take more rea-dily and grow at first more vigorously thanwhen grafted on closely allied stocks, after-wards often become unhealthy.

I have studied M. Tessier's careful and elabora-te experiments (18/6. 'Mem. de l'Acad. desSciences' 1790 page 209.) made to disprove thecommon belief that good is derived from achange of seed; and he certainly shows that thesame seed may with care be cultivated on thesame farm (it is not stated whether on exactlythe same soil) for ten consecutive years withoutloss. Another excellent observer, Colonel Le

Couteur (18/7. 'On the Varieties of Wheat' page52.) has come to the same conclusion; but thenhe expressly adds, if the same seed be used,"that which is grown on land manured from themixen one year becomes seed for land preparedwith lime, and that again becomes seed for landdressed with ashes, then for land dressed withmixed manure, and so on." But this in effect is asystematic exchange of seed, within the limitsof the same farm.]

On the whole the belief, which has long beenheld by many cultivators, that good followsfrom exchanging seed, tubers, etc., seems to befairly well founded. It seems hardly crediblethat the advantage thus derived can be due tothe seeds, especially if very small ones, obtai-ning in one soil some chemical element defi-cient in the other and in sufficient quantity toinfluence the whole after-growth of the plant.As plants after once germinating are fixed tothe same spot, it might have been anticipated

that they would show the good effects of achange more plainly than do animals whichcontinually wander about; and this apparentlyis the case. Life depending on, or consisting in,an incessant play of the most complex forces, itwould appear that their action is in some waystimulated by slight changes in the circumstan-ces to which each organism is exposed. All for-ces throughout nature, as Mr. Herbert Spencer(18/8. Mr. Spencer has fully and ably discussedthis whole subject in his 'Principles of Biology'1864 volume 2 chapter 10. In the first edition ofmy 'Origin of Species' 1859 page 267, I spoke ofthe good effects from slight changes in the con-ditions of life and from cross-breeding, and ofthe evil effects from great changes in the condi-tions and from crossing widely distinct forms,as a series of facts "connected together by somecommon but unknown bond, which is essentia-lly related to the principle of life.) remarks,tend towards an equilibrium, and for the life ofeach organism it is necessary that this tendency

should be checked. These views and the fore-going facts probably throw light, on the onehand, on the good effects of crossing the breed,for the germ will be thus slightly modified oracted on by new forces; and on the other hand,on the evil effects of close interbreeding pro-longed during many generations, during whichthe germ will be acted on by a male havingalmost identically the same constitution.

STERILITY FROM CHANGED CONDI-TIONS OF LIFE.

I will now attempt to show that animals andplants, when removed from their natural condi-tions, are often rendered in some degree inferti-le or completely barren; and this occurs evenwhen the conditions have not been greatlychanged. This conclusion is not necessarily op-posed to that at which we have just arrived,namely, that lesser changes of other kinds areadvantageous to organic beings. Our present

subject is of some importance, from having anintimate connection with the causes of variabili-ty. Indirectly it perhaps bears on the sterility ofspecies when crossed: for as, on the one hand,slight changes in the conditions of life are fa-vourable to plants and animals, and the cros-sing of varieties adds to the size, vigour, andfertility of their offspring; so, on the other hand,certain other changes in the conditions of lifecause sterility; and as this likewise ensues fromcrossing much-modified forms or species, wehave a parallel and double series of facts, whichapparently stand in close relation to each other.

It is notorious that many animals, though per-fectly tamed, refuse to breed in captivity. Isido-re Geoffroy St.-Hilaire (18/9. 'Essais de Zoolo-gie Generale' 1841 page 256. ) consequently hasdrawn a broad distinction between tamed ani-mals which will not breed under captivity, andtruly domesticated animals which breed free-ly—generally more freely, as shown in the six-

teenth chapter, than in a state of nature. It ispossible and generally easy to tame most ani-mals; but experience has shown that it is diffi-cult to get them to breed regularly, or even atall. I shall discuss this subject in detail; but willgive only those cases which seem most illustra-tive. My materials are derived from noticesscattered through various works, and especia-lly from a Report, kindly drawn up for me bythe officers of the Zoological Society of London,which has especial value, as it records all thecases, during nine years from 1838-46, in whichthe animals were seen to couple but producedno offspring, as well as the cases in which theynever, as far as known, coupled. This MS. Re-port I have corrected by the annual Reportssubsequently published up to the year 1865.(18/10. Since the appearance of the first editionof this work, Mr. Sclater has published ('Proc.Zoolog. Soc.' 1868 page 623) a list of the speciesof mammals which have bred in the gardensfrom 1848 to 1867 inclusive. Of the Artiodactyla

85 species have been kept, and of these 1 spe-cies in 1.9 have bred at least once during the 20years; of 28 Marsupialia, 1 in 2.5 have bred; of74 Carnivora, 1 in 3.0 have bred; of 52 Rodentia,1 in 4.7 have bred; and of Quadrumana 75 spe-cies have been kept, and 1 in 6.2 have bred.)Many facts are given on the breeding of theanimals in that magnificent work, 'Gleaningsfrom the Menageries of Knowsley Hall' by Dr.Gray. I made, also, particular inquiries from theexperienced keeper of the birds in the old Su-rrey Zoological Gardens. I should premise thata slight change in the treatment of animals so-metimes makes a great difference in their ferti-lity; and it is probable that the results observedin different menageries would differ. Indeed,some animals in our Zoological Gardens havebecome more productive since the year 1846. Itis, also, manifest from F. Cuvier's account of theJardin des Plantes (18/11. Du Rut 'Annales duMuseum' 1807 tome 9 page 120.) that the ani-mals formerly bred much less freely there than

with us; for instance, in the Duck tribe, which ishighly prolific, only one species had at thatperiod produced young.

[The most remarkable cases, however, are af-forded by animals kept in their native country,which, though perfectly tamed, quite healthy,and allowed some freedom, are absolutely in-capable of breeding. Rengger (18/12. 'Sauget-hiere von Paraguay' 1830 s. 49, 106, 118, 124,201, 208, 249, 265, 327.), who in Paraguay parti-cularly attended to this subject, specifies sixquadrupeds in this condition; and he mentionstwo or three others which most rarely breed.Mr. Bates, in his admirable work on the Ama-zons, strongly insists on similar cases (18/13.'The Naturalist on the Amazons' 1863 volume 1pages 99, 193; volume 2 page 113.); and he re-marks, that the fact of thoroughly tamed nativemammals and birds not breeding when kept bythe Indians, cannot be wholly accounted for bytheir negligence or indifference, for the turkey

and fowl are kept and bred by various remotetribes. In almost every part of the world—forinstance, in the interior of Africa, and in severalof the Polynesian islands —the natives are ex-tremely fond of taming the indigenous qua-drupeds and birds; but they rarely or neversucceed in getting them to breed.

The most notorious case of an animal not bree-ding in captivity is that of the elephant. Elep-hants are kept in large numbers in their nativeIndian home, live to old age, and are vigorousenough for the severest labour; yet, with a veryfew exceptions, they have never been knowneven to couple, though both males and femaleshave their proper periodical seasons. If, howe-ver, we proceed a little eastward to Ava, wehear from Mr. Crawfurd (18/14. 'Embassy tothe Court of Ava' volume 1 page 534.) that their"breeding in the domestic state, or at least in thehalf-domestic state in which the female elep-hants are generally kept, is of everyday occu-

rrence;" and Mr. Crawfurd informs me that hebelieves that the difference must be attributedsolely to the females being allowed to roam theforest with some degree of freedom. The capti-ve rhinoceros, on the other hand, seems fromBishop Heber's account (18/15. 'Journal' volu-me 1 page 213.) to breed in India far more rea-dily than the elephant. Four wild species of thehorse genus have bred in Europe, though hereexposed to a great change in their naturalhabits of life; but the species have generallybeen crossed one with another. Most of themembers of the pig family breed readily in ourmenageries; even the Red River hog (Potamo-choerus penicillatus), from the swelteringplains of West Africa, has bred twice in theZoological Gardens. Here also the Peccary (Di-cotyles torquatus) has bred several times; butanother species, the D. labiatus, though rende-red so tame as to be half-domesticated, is saidto breed so rarely in its native country of Para-guay, that according to Rengger (18/16. 'Sau-

gethiere' s. 327.) the fact requires confirmation.Mr. Bates remarks that the tapir, though oftenkept tame in Amazonia by the Indians, neverbreeds.

Ruminants generally breed quite freely in En-gland, though brought from widely differentclimates, as may be seen in the Annual Reportsof the Zoological Gardens, and in the Gleaningsfrom Lord Derby's menagerie.

The Carnivora, with the exception of the Plan-tigrade division, breed (though with capriciousexceptions) about half as freely as ruminants.Many species of Felidae have bred in variousmenageries, although imported from diverseclimates and closely confined. Mr. Bartlett, thepresent superintendent of the Zoological Gar-dens (18/17. On the Breeding of the Larger Fe-lidae 'Proc. Zoolog. Soc.' 1861 page 140.) re-marks that the lion appears to breed more fre-quently and to bring forth more young at abirth than any other species of the family. He

adds that the tiger has rarely bred; "but thereare several well-authenticated instances of thefemale tiger breeding with the lion." Strange asthe fact may appear, many animals under con-finement unite with distinct species and produ-ce hybrids quite as freely as, or even more free-ly than, with their own species. On inquiringfrom Dr. Falconer and others, it appears thatthe tiger when confined in India does notbreed, though it has been known to couple. Thechetah (Felis jubata) has never been known byMr. Bartlett to breed in England, but it has bredat Frankfort; nor does it breed in India, where itis kept in large numbers for hunting; but nopains would be taken to make them breed, asonly those animals which have hunted forthemselves in a state of nature are serviceableand worth training. (18/18. Sleeman's 'Ramblesin India' volume 2 page 10.) According toRengger, two species of wild cats in Paraguay,though thoroughly tamed, have never bred.Although so many of the Felidae breed readily

in the Zoological Gardens, yet conception by nomeans always follows union: in the nine-yearReport, various species are specified whichwere observed to couple seventy-three times,and no doubt this must have passed many ti-mes unnoticed; yet from the seventy- threeunions only fifteen births ensued. The Carnivo-ra in the Zoological Gardens were formerly lessfreely exposed to the air and cold than at pre-sent, and this change of treatment, as I was as-sured by the former superintendent, Mr. Miller,greatly increased their fertility. Mr. Bartlett,and there cannot be a more capable judge, says,"it is remarkable that lions breed more freely intravelling collections than in the ZoologicalGardens; probably the constant excitement andirritation produced by moving from place toplace, or change of air, may have considerableinfluence in the matter."

Many members of the Dog family breed readilywhen confined. The Dhole is one of the most

untamable animals in India, yet a pair kept the-re by Dr. Falconer produced young. Foxes, onthe other hand, rarely breed, and I have neverheard of such an occurrence with the Europeanfox: the silver fox of North America (Canis ar-gentatus), however, has bred several times inthe Zoological Gardens. Even the otter has bredthere. Every one knows how readily the semi-domesticated ferret breeds, though shut up inmiserably small cages; but other species of Vi-verra and Paradoxurus absolutely refuse tobreed in the Zoological Gardens. The Genettahas bred both here and in the Jardin des Plan-tes, and produced hybrids. The Herpestes fas-ciatus has likewise bred; but I was formerlyassured that the H. griseus, though many werekept in the Gardens, never bred.

The Plantigrade Carnivora breed under confi-nement much less freely than other Carnivora,although no reason can be assigned for thisfact. In the nine-year Report it is stated that the

bears had been seen in the Zoological Gardensto couple freely, but previously to 1848 hadmost rarely conceived. In the Reports publishedsince this date three species have producedyoung (hybrids in one case), and, wonderful torelate, the white Polar bear has producedyoung. The badger (Meles taxus) has bred seve-ral times in the Gardens; but I have not heardof this occurring elsewhere in England, and theevent must be very rare, for an instance inGermany has been thought worth recording.(18/19. Wiegmann 'Archiv. fur Naturgesch.'1837 s. 162.) In Paraguay the native Nasua,though kept in pairs during many years andperfectly tamed, has never been known, accor-ding to Rengger, to breed or show any sexualpassion; nor, as I hear from Mr. Bates, does thisanimal, or the Cercoleptes, breed in Amazonia.Two other plantigrade genera, Procyon andGulo, though often kept tame in Paraguay, ne-ver breed there. In the Zoological Gardens spe-

cies of Nasua and Procyon have been seen tocouple; but they did not produce young.

As domesticated rabbits, guinea-pigs, and whi-te mice breed so abundantly when closely con-fined under various climates, it might havebeen thought that most other members of theRodent order would have bred in captivity, butthis is not the case. It deserves notice, as sho-wing how the capacity to breed sometimes goesby affinity, that the one native rodent of Para-guay, which there breeds FREELY and hasyielded successive generations, is the Caviaaperea; and this animal is so closely allied tothe guinea-pig, that it has been erroneouslythought to be the parent form. (18/20. Rengger'Saugethiere' etc. s. 276. On the parentage of theguinea-pig, see also Isid. Geoffroy St.- Hilaire'Hist. Nat. Gen.' I sent to Mr. H. Denny of Leedsthe lice which I collected from the wild apereain La Plata, and he informs me that they belongto a genus distinct from those found on the

guinea-pig. This is important evidence that theaperea is not the parent of the guinea-pig; andis worth giving, as some authors erroneouslysuppose that the guinea-pig since being domes-ticated has become sterile when crossed withthe aperea.) In the Zoological Gardens, somerodents have coupled, but have never produ-ced young; some have neither coupled norbred; but a few have bred, as the porcupinemore than once, the Barbary mouse, lemming,chinchilla, and agouti (Dasyprocta aguti) seve-ral times. This latter animal has also producedyoung in Paraguay, though they were borndead and ill-formed; but in Amazonia, accor-ding to Mr. Bates, it never breeds, though oftenkept tame about the houses. Nor does the paca(Coelogenys paca) breed there. The commonhare when confined has, I believe, never bredin Europe; though, according to a recent state-ment, it has crossed with the rabbit. (18/21.Although the existence of the Leporides, asdescribed by Dr. Broca ('Journal de Phys.' tome

2 page 370), has been positively denied, yet Dr.Pigeaux ('Annals and Mag. of Nat. Hist.' volu-me 20 1867 page 75) affirms that the hare andrabbit have produced hybrids.) I have neverheard of the dormouse breeding in confine-ment. But squirrels offer a more curious case:with one exception, no species has bred in theZoological Gardens, yet as many as fourteenindividuals of S. palmarum were kept togetherduring several years. The S. cinera has beenseen to couple, but it did not produce young;nor has this species, when rendered extremelytame in its native country, North America, beenever known to breed. (18/22. 'Quadrupeds ofNorth America' by Audubon and Bachman1846 page 268.) At Lord Derby's menageriesquirrels of many kinds were kept in numbers,but Mr. Thompson, the superintendent, told methat none had ever bred there, or elsewhere asfar as he knew. I have never heard of the En-glish squirrel breeding in confinement. But thespecies which has bred more than once in the

Zoological Gardens is the one which perhapsmight have been least expected, namely, theflying squirrel (Sciuropterus volucella): it has,also, bred several times near Birmingham; butthe female never produced more than twoyoung at a birth, whereas in its native Ameri-can home she bears from three to six young.(18/23. Loudon's 'Mag. of Nat. Hist.' volume 91836 page 571; Audubon and Bachman 'Qua-drupeds of North America' page 221.)

Monkeys, in the nine-year Report from theZoological Gardens, are stated to unite mostfreely, but during this period, though manyindividuals were kept, there were only sevenbirths. I have heard of only one Americanmonkey, the Ouistiti, breeding in Europe.(18/24. Flourens 'De l'Instinct' etc. 1845 page88.) A Macacus, according to Flourens, bred inParis; and more than one species of this genushas produced young in London, especially theMacacus rhesus, which everywhere shows a

special capacity to breed under confinement.Hybrids have been produced both in Paris andLondon from this same genus. The Arabianbaboon, or Cynocephalus hamadryas (18/25.See 'Annual Reports Zoolog. Soc.' 1855, 1858,1863, 1864; 'Times' newspaper August 10, 1847;Flourens 'De l'Instinct' page 85.), and a Cerco-pithecus have bred in the Zoological Gardens,and the latter species at the Duke of Northum-berland's. Several members of the family ofLemurs have produced hybrids in the Zoologi-cal Gardens. It is much more remarkable thatmonkeys very rarely breed when confined intheir native country; thus the Cay (Cebus azara)is frequently and completely tamed in Para-guay, but Rengger (18/26. 'Saugethiere' etc. s.34, 49.) says that it breeds so rarely, that he ne-ver saw more than two females which had pro-duced young. A similar observation has beenmade with respect to the monkeys which arefrequently tamed by the aborigines in Brazil.(18/27. Art. Brazil 'Penny Cyclop.' page 363.) In

Amazonia, these animals are so often kept in atame state, that Mr. Bates in walking throughthe streets of Para counted thirteen species; but,as he asserts, they have never been known tobreed in captivity. (18/28. 'The Naturalist onthe Amazons' volume 1 page 99.)

BIRDS.

Birds offer in some respects better evidencethan quadrupeds, from their breeding morerapidly and being kept in greater numbers.(18/29. A list of the species of birds which havebred in the Zoological Gardens from 1848 to1867 inclusive has been published by Mr. Scla-ter in 'Proc. Zoolog. Soc.' 1869 page 626, sincethe first edition of this work appeared. Of Co-lumbae 51 species have been kept, and of Anse-res 80 species, and in both these families 1 spe-cies in 2.6 have bred at least once in the 20years. Of Gallinae 83 species have been keptand 1 in 27 have bred; of 57 Grallae 1 in 9 have

bred; of 110 Prehensores 1 in 22 have bred; of178 Passeres 1 in 25.4 have bred; of 94 Accipi-tres 1 in 47 have bred; of 25 Picariae and of 35Herodiones not one species in either group hasbred.) We have seen that carnivorous animalsare more fertile under confinement than mostother mammals. The reverse holds good withcarnivorous birds. It is said (18/30. 'Encyclop.of Rural Sports' page 691.) that as many as eigh-teen species have been used in Europe forhawking, and several others in Persia and India(18/31. According to Sir A. Burnes 'Cabool' etc.page 51, eight species are used for hawking inSinde.); they have been kept in their nativecountry in the finest condition, and have beenflown during six, eight, or nine years (18/32.Loudon's 'Mag. of Nat. Hist.' volume 6 1833page 110.); yet there is no record of their havingever produced young. As these birds were for-merly caught whilst young, at great expense,being imported from Iceland, Norway, andSweden, there can be little doubt that, if possi-

ble, they would have been propagated. In theJardin des Plantes, no bird of prey has beenknown to couple. (18/33. F. Cuvier 'Annal. duMuseum' tome 9 page 128.) No hawk, vulture,or owl has ever produced fertile eggs in theZoological Gardens, or in the old Surrey Gar-dens, with the exception, in the former place onone occasion, of a condor and a kite (Milvusniger). Yet several species, namely, the Aquilafusca, Haliaetus leucocephalus, Falco tinnuncu-lus, F. subbuteo, and Buteo vulgaris, have beenseen to couple in the Zoological Gardens. Mr.Morris (18/34. 'The Zoologist' volume 7-8 1849-50 page 2648.) mentions as a unique fact that akestrel (Falco tinnunculus) bred in an aviary.The one kind of owl which has been known tocouple in the Zoological Gardens was the EagleOwl (Bubo maximus); and this species shows aspecial inclination to breed in captivity; for apair at Arundel Castle, kept more nearly in astate of nature "than ever fell to the lot of ananimal deprived of its liberty" (18/35. Knox

'Ornithological Rambles in Sussex' page 91.),actually reared their young. Mr. Gurney hasgiven another instance of this same owl bree-ding in confinement; and he records the case ofa second species of owl, the Strix passerina,breeding in captivity. (18/36. 'The Zoologist'volume 7-8 1849-50 page 2566; volume 9-101851-2 page 3207.)

Of the smaller graminivorous birds, manykinds have been kept tame in their native coun-tries, and have lived long; yet, as the highestauthority on cage- birds (18/37. Bechstein 'Na-turgesch. der Stubenvogel' 1840 s. 20.) remarks,their propagation is "uncommonly difficult."The canary-bird shows that there is no inherentdifficulty in these birds breeding freely in con-finement; and Audubon says (18/38. 'Ornit-hological Biography' volume 5 page 517.) thatthe Fringilla (Spiza) ciris of North Americabreeds as perfectly as the canary. The difficultywith the many finches which have been kept in

confinement is all the more remarkable as morethan a dozen species could be named whichhave yielded hybrids with the canary; but hard-ly any of these, with the exception of the siskin(Fringilla spinus), have reproduced their ownkind. Even the bullfinch (Loxia pyrrhula) hasbred as frequently with the canary, though be-longing to a distinct genus, as with its own spe-cies. (18/39. A case is recorded in 'The Zoolo-gist' volume 1-2 1843-45 page 453. For the siskinbreeding, volume 3-4 1845-46 page 1075. Bechs-tein 'Stubenvogel' s. 139 speaks of bullfinchesmaking nests, but rarely producing young.)With respect to the skylark (Alauda arvensis), Ihave heard of birds living for seven years in anaviary, which never produced young; and agreat London bird-fancier assured me that hehad never known an instance of their breeding;nevertheless one case has been recorded.(18/40. Yarrell 'Hist. British Birds' 1839 volume1 page 412.) In the nine-year Report from theZoological Society, twenty-four insessorial spe-

cies are enumerated which had not bred, and ofthese only four were known to have coupled.

Parrots are singularly long-lived birds; andHumboldt mentions the curious fact of a parrotin South America, which spoke the language ofan extinct Indian tribe, so that this bird preser-ved the sole relic of a lost language. Even inthis country there is reason to believe (18/41.Loudon's 'Mag. of Nat. History' volume 19 1836page 347.) that parrots have lived to the age ofnearly one hundred years; yet they breed sorarely, though many have been kept in Europe,that the event has been thought worth recor-ding in the gravest publications. (18/42. 'Me-moires du Museum d'Hist. Nat.' tome 10 page314: five cases of parrots breeding in France arehere recorded. See also 'Report Brit. Assoc.Zoolog.' 1843.) Nevertheless, when Mr. Buxtonturned out a large number of parrots in Nor-folk, three pairs bred and reared ten youngbirds in the course of two seasons; and this suc-

cess may be attributed to their free life. (18/43.'Annals and Mag. of Nat. Hist.' November 1868page 311.) According to Bechstein (18/44. 'Stu-benvogel' s. 105, 83.) the African Psittacus erit-hacus breeds oftener than any other species inGermany: the P. macoa occasionally lays fertileeggs, but rarely succeeds in hatching them; thisbird, however, has the instinct of incubationsometimes so strongly developed, that it willhatch the eggs of fowls or pigeons. In the Zoo-logical Gardens and in the old Surrey Gardenssome few species have coupled, but, with theexception of three species of parakeets, nonehave bred. It is a much more remarkable factthat in Guiana parrots of two kinds, as I aminformed by Sir R. Schomburgk, are often takenfrom the nests by the Indians and reared inlarge numbers; they are so tame that they flyfreely about the houses, and come when calledto be fed, like pigeons; yet he has never heardof a single instance of their breeding. (18/45.Dr. Hancock remarks ('Charlesworth's Mag. of

Nat. Hist.' volume 2 1838 page 492) "it is singu-lar that, amongst the numerous useful birdsthat are indigenous to Guiana, none are foundto propagate among the Indians; yet the com-mon fowl is reared in abundance throughoutthe country.") In Jamaica, a resident naturalist,Mr. R. Hill (18/46. 'A Week at Pert Royal' 1855page 7.), says, "no birds more readily submit tohuman dependence than the parrot-tribe, butno instance of a parrot breeding in this tame lifehas been known yet." Mr. Hill specifies a num-ber of other native birds kept tame in the WestIndies, which never breed in this state.

The great pigeon family offers a striking con-trast with the parrots: in the nine-year Reportthirteen species are recorded as having bred,and, what is more noticeable, only two wereseen to couple without any result. Since theabove date every annual Report gives manycases of various pigeons breeding. The twomagnificent crowned pigeons (Goura coronata

and victoriae) produced hybrids; nevertheless,of the former species more than a dozen birdswere kept, as I am informed by Mr. Crawfurd,in a park at Penang, under a perfectly well-adapted climate, but never once bred. The Co-lumba migratoria in its native country, NorthAmerica, invariably lays two eggs, but in LordDerby's menagerie never more than one. Thesame fact has been observed with the C. leuco-cephala. (18/47. Audubon 'American Ornit-hology' volume 5 pages 552, 557.)

Gallinaceous birds of many genera likewiseshow an eminent capacity for breeding undercaptivity. This is particularly the case withpheasants, yet our English species seldom laysmore than ten eggs in confinement; whilst fromeighteen to twenty is the usual number in thewild state. (18/48. Mowbray on 'Poultry' 7thedition page 133.) With the Gallinaceae, as withall other orders, there are marked and inexpli-cable exceptions in regard to the fertility of cer-

tain species and genera under confinement.Although many trials have been made with thecommon partridge, it has rarely bred, evenwhen reared in large aviaries; and the hen willnever hatch her own eggs. (18/49. Temminck'Hist. Nat. Gen. des Pigeons' etc. 1813 tome 3pages 288, 382; 'Annals and Mag. of Nat. Hist.'volume 12 1843 page 453. Other species of par-tridge have occasionally bred; as the red-legged(P. rubra), when kept in a large court in France(see Journal de Physique' tome 25 page 294),and in the Zoological Gardens in 1856.) TheAmerican tribe of Guans or Cracidae are tamedwith remarkable ease, but are very shy breedersin this country (18/50. Rev. E.S. Dixon 'TheDovecote' 1851 pages 243-252.); but with carevarious species were formerly made to breedrather freely in Holland. (18/51. Temminck'Hist. Nat. Gen. des Pigeons' etc. tome 2 pages456, 458; tome 3 pages 2, 13, 47.) Birds of thistribe are often kept in a perfectly tamed condi-tion in their native country by the Indians, but

they never breed. (18/52. Bates 'The Naturaliston the Amazons' volume 1 page 193; volume 2page 112.) It might have been expected thatgrouse from their habits of life would not havebred in captivity, more especially as they aresaid soon to languish and die. (18/53. Tem-minck 'Hist. Nat. Gen.' etc. tome 2 page 125. ForTetrao urogallus see L. Lloyd 'Field Sports ofNorth of Europe' volume 1 pages 287, 314; andBull. de la Soc. d'Acclimat.' tome 7 1860 page600. For T. scoticus Thompson 'Nat. Hist. ofIreland' volume 2 1850 page 49. For T. cupido'Boston Journal of Nat. Hist.' volume 3 page199.) But many cases are recorded of theirbreeding: the capercailzie (Tetrao urogallus)has bred in the Zoological Gardens; it breedswithout much difficulty when confined inNorway, and in Russia five successive genera-tions have been reared: Tetrao tetrix has likewi-se bred in Norway; T. scoticus in Ireland; T.umbellus at Lord Derby's; and T. cupido inNorth America.

It is scarcely possible to imagine a greaterchange in habits than that which the membersof the ostrich family must suffer, when coopedup in small enclosures under a temperate cli-mate, after freely roaming over desert and tro-pical plains or entangled forests; yet almost allthe kinds have frequently produced young inthe various European menageries, even themooruk (Casuarius bennetii) from New Ire-land. The African ostrich, though perfectlyhealthy and living long in the South of France,never lays more than from twelve to fifteeneggs, though in its native country it lays fromtwenty-five to thirty. (18/54. Marcel de Serres'Annales des Sc. Nat.' 2nd series Zoolog. tome13 page 175.) Here we have another instance offertility impaired, but not lost, under confine-ment, as with the flying squirrel, the hen-pheasant, and two species of American pi-geons.

Most Waders can be tamed, as the Rev. E.S.Dixon informs me, with remarkable facility; butseveral of them are short-lived under confine-ment, so that their sterility in this state is notsurprising. The cranes breed more readily thanother genera: Grus montigresia has bred seve-ral times in Paris and in the Zoological Gar-dens, as has G. cinerea at the latter place, andG. antigone at Calcutta. Of other members ofthis great order, Tetrapteryx paradisea has bredat Knowsley, a Porphyrio in Sicily, and the Ga-llinula chloropus in the Zoological Gardens. Onthe other hand, several birds belonging to thisorder will not breed in their native country,Jamaica; and the Psophia, though often kept bythe Indians of Guiana about their houses, "isseldom or never known to breed." (18/55. Dr.Hancock in 'Charlesworth's Mag. of Nat. Hist.'volume 2 1838 page 491; R. Hill 'A Week at PortRoyal' page 8; 'Guide to the Zoological Gar-dens' by P.L. Sclater 1859 pages 11, 12; 'The

Knowsley Menagerie' by D. Gray 1846 p1. 14; E.Blyth 'Report Asiatic Soc. of Bengal' May 1855.)

The members of the great Duck family breed asreadily in confinement as do the Columbae andGallinae and this, considering their aquatic andwandering habits, and the nature of their food,could not have been anticipated. Even sometime ago above two dozen species had bred inthe Zoological Gardens; and M. Selys-Longchamps has recorded the production ofhybrids from forty-four different members ofthe family; and to these Professor Newton hasadded a few more cases. (18/56. Prof. Newtonin 'Proc. Zoolog. Soc.' 1860 page 336.) "There isnot," says Mr. Dixon (18/57. 'The Dovecote andAviary' page 428.), "in the wide world, a goosewhich is not in the strict sense of the word do-mesticable;" that is, capable of breeding underconfinement; but this statement is probably toobold. The capacity to breed sometimes varies inindividuals of the same species; thus Audubon

(18/58. 'Ornithological Biography' volume 3page 9.) kept for more than eight years somewild geese (Anser canadensis), but they wouldnot mate; whilst other individuals of the samespecies produced young during the secondyear. I know of but one instance in the wholefamily of a species which absolutely refuses tobreed in captivity, namely, the Dendrocygnaviduata, although, according to Sir R. Schom-burgk (18/59. 'Geograph. Journal' volume 131844 page 32.), it is easily tamed, and is fre-quently kept by the Indians of Guiana. Lastly,with respect to Gulls, though many have beenkept in the Zoological Gardens and in the oldSurrey Gardens, no instance was known beforethe year 1848 of their coupling or breeding; butsince that period the herring gull (Larus argen-tatus) has bred many times in the ZoologicalGardens and at Knowsley.

There is reason to believe that insects are affec-ted by confinement like the higher animals. It is

well known that the Sphingidae rarely breedwhen thus treated. An entomologist (18/60.Loudon's 'Mag. of Nat. Hist.' volume 5 1832page 153.) in Paris kept twenty-five specimensof Saturnia pyri, but did not succeed in gettinga single fertile egg. A number of females of Ort-hosia munda and of Mamestra suasa reared inconfinement were unattractive to the males.(18/61. 'Zoologist' volumes 5-6 1847-48 page1660.) Mr. Newport kept nearly a hundred in-dividuals of two species of Vanessa, but notone paired; this, however, might have been dueto their habit of coupling on the wing. (18/62.'Transact. Entomolog. Soc.' volume 4 1845 page60.) Mr. Atkinson could never succeed in Indiain making the Tarroo silk-moth breed in confi-nement. (18/63. 'Transact. Linn. Soc.' volume 7page 40.) It appears that a number of moths,especially the Sphingidae, when hatched in theautumn out of their proper season, are comple-tely barren; but this latter case is still involvedin some obscurity. (18/64. See an interesting

paper by Mr. Newman in the 'Zoologist' 1857page 5764; and Dr. Wallace in 'Proc. Entomo-log. Soc.' June 4, 1860 page 119.)]

Independently of the fact of many animals un-der confinement not coupling, or, if they cou-ple, not producing young, there is evidence ofanother kind that their sexual functions aredisturbed. For many cases have been recordedof the loss by male birds when confined of theircharacteristic plumage. Thus the common lin-net (Linota cannabina) when caged does notacquire the fine crimson colour on its breast,and one of the buntings (Emberiza passerina)loses the black on its head. A Pyrrhula and anOriolus have been observed to assume thequiet plumage of the hen-bird; and the Falcoalbidus returned to the dress of an earlier age.(18/65. Yarrell 'British Birds' volume 1 page506; Bechstein 'Stubenvogel' s. 185; 'Philosoph.Transact.' 1772 page 271. Bronn 'Geschichte derNatur' b. 2 s. 96 has collected a number of cases.

For the case of the deer see 'Penny Cyclop.' vo-lume 8 page 350.) Mr. Thompson, the superin-tendent of the Knowsley menagerie, informedme that he had often observed analogous facts.The horns of a male deer (Cervus canadensis)during the voyage from America were badlydeveloped; but subsequently in Paris perfecthorns were produced.

When conception takes place under confine-ment, the young are often born dead, or diesoon, or are ill-formed. This frequently occursin the Zoological Gardens, and, according toRengger, with native animals confined in Para-guay. The mother's milk often fails. We mayalso attribute to the disturbance of the sexualfunctions the frequent occurrence of that mons-trous instinct which leads the mother to devourher own offspring,—a mysterious case of per-version, as it at first appears.

Sufficient evidence has now been advanced toprove that animals when first confined are emi-

nently liable to suffer in their reproductive sys-tems. We feel at first naturally inclined to attri-bute the result to loss of health, or at least toloss of vigour; but this view can hardly be ad-mitted when we reflect how healthy, long-lived, and vigorous many animals are undercaptivity, such as parrots, and hawks whenused for hawking, cheetahs when used for hun-ting, and elephants. The reproductive organsthemselves are not diseased; and the diseases,from which animals in menageries usually pe-rish, are not those which in any way affect theirfertility. No domestic animal is more subject todisease than the sheep, yet it is remarkably pro-lific. The failure of animals to breed under con-finement has been sometimes attributed exclu-sively to a failure in their sexual instincts: thismay occasionally come into play, but there isno obvious reason why this instinct should beespecially liable to be affected with perfectlytamed animals, except, indeed, indirectlythrough the reproductive system itself being

disturbed. Moreover, numerous cases havebeen given of various animals which couplefreely under confinement, but never conceive;or, if they conceive and produce young, theseare fewer in number than is natural to the spe-cies. In the vegetable kingdom instinct of cour-se can play no part; and we shall presently seethat plants when removed from their naturalconditions are affected in nearly the same man-ner as animals. Change of climate cannot be thecause of the loss of fertility, for, whilst manyanimals imported into Europe from extremelydifferent climates breed freely, many otherswhen confined in their native land are comple-tely sterile. Change of food cannot be the chiefcause; for ostriches, ducks, and many otheranimals, which must have undergone a greatchange in this respect, breed freely. Carnivo-rous birds when confined are extremely sterile,whilst most carnivorous mammals, exceptplantigrades, are moderately fertile. Nor canthe amount of food be the cause; for a sufficient

supply will certainly be given to valuable ani-mals; and there is no reason to suppose thatmuch more food would be given to them thanto our choice domestic productions which re-tain their full fertility. Lastly, we may inferfrom the case of the elephant, cheetah, varioushawks, and of many animals which are allowedto lead an almost free life in their native land,that want of exercise is not the sole cause.

It would appear that any change in the habitsof life, whatever these habits may be, if greatenough, tends to affect in an inexplicable man-ner the powers of reproduction. The result de-pends more on the constitution of the speciesthan on the nature of the change; for certainwhole groups are affected more than others;but exceptions always occur, for some speciesin the most fertile groups refuse to breed, andsome in the most sterile groups breed freely.Those animals which usually breed freely un-der confinement, rarely breed, as I was assured,

in the Zoological Gardens, within a year or twoafter their first importation. When an animalwhich is generally sterile under confinementhappens to breed, the young apparently do notinherit this power: for had this been the case,various quadrupeds and birds, which are va-luable for exhibition, would have become com-mon. Dr. Broca even affirms (18/66. 'Journal dePhysiologie' tome 2 page 347.) that many ani-mals in the Jardin des Plantes, after having pro-duced young for three or four successive gene-rations, become sterile; but this may be the re-sult of too close interbreeding. It is a remarka-ble circumstance that many mammals andbirds have produced hybrids under confine-ment quite as readily as, or even more readilythan, they have procreated their own kind. Ofthis fact many instances have been given(18/67. For additional evidence on this subjectsee F. Cuvier in 'Annales du Museum' tome 12page 119.); and we are thus reminded of thoseplants which when cultivated refuse to be ferti-

lised by their own pollen, but can easily be fer-tilised by that of a distinct species. Finally, wemust conclude, limited as the conclusion is, thatchanged conditions of life have an especial po-wer of acting injuriously on the reproductivesystem. The whole case is quite peculiar, forthese organs, though not diseased, are thusrendered incapable of performing their properfunctions, or perform them imperfectly.

[STERILITY OF DOMESTICATED ANI-MALS FROM CHANGED CONDITIONS.

With respect to domesticated animals, as theirdomestication mainly depends on the accidentof their breeding freely under captivity, weought not to expect that their reproductive sys-tem would be affected by any moderate degreeof change. Those orders of quadrupeds andbirds, of which the wild species breed mostreadily in our menageries, have afforded us thegreatest number of domesticated productions.

Savages in most parts of the world are fond oftaming animals (18/68. Numerous instancescould be given. Thus Livingstone ('Travels' pa-ge 217) states that the King of the Barotse, aninland tribe which never had any communica-tion with white men, was extremely fond oftaming animals, and every young antelope wasbrought to him. Mr. Galton informs me that theDamaras are likewise fond of keeping pets. TheIndians of South America follow the samehabit. Capt. Wilkes states that the Polynesiansof the Samoan Islands tamed pigeons; and theNew Zealanders, as Mr. Mantell informs me,kept various kinds of birds.); and if any of theseregularly produced young, and were at thesame time useful, they would be at once do-mesticated. If, when their masters migratedinto other countries, they were in additionfound capable of withstanding various clima-tes, they would be still more valuable; and itappears that the animals which breed readily incaptivity can generally withstand different cli-

mates. Some few domesticated animals, such asthe reindeer and camel, offer an exception tothis rule. Many of our domesticated animalscan bear with undiminished fertility the mostunnatural conditions; for instance, rabbits, gui-nea-pigs, and ferrets breed in miserably confi-ned hutches. Few European dogs of any kindwithstand the climate of India without degene-rating, but as long as they survive, they retain,as I hear from Dr. Falconer, their fertility; so itis, according to Dr. Daniell, with English dogstaken to Sierra Leone. The fowl, a native of thehot jungles of India, becomes more fertile thanits parent-stock in every quarter of the world,until we advance as far north as Greenland andNorthern Siberia, where this bird will notbreed. Both fowls and pigeons, which I recei-ved during the autumn direct from Sierra Leo-ne, were at once ready to couple. (18/69. Foranalogous cases with the fowl see Reaumur'L'Art de faire Eclore' etc. 1749 page 243; andCol. Sykes in 'Proc. Zoolog. Soc.' 1832 etc. With

respect to the fowl not breeding in northernregions see Latham 'Hist. of Birds' volume 81823 page 169.) I have, also, seen pigeons bree-ding as freely as the common kinds within ayear after their importation from the upperNile. The guinea- fowl, an aboriginal of the hotand dry deserts of Africa, whilst living underour damp and cool climate, produces a largesupply of eggs.

Nevertheless, our domesticated animals undernew conditions occasionally show signs of les-sened fertility. Roulin asserts that in the hotvalleys of the equatorial Cordillera sheep arenot fully fecund (18/70. "Mem. par divers Sa-vans" 'Acad. des Sciences' tome 6 1835 page347.); and according to Lord Somerville (18/71.'Youatt on Sheep' page 181.) the merino-sheepwhich he imported from Spain were not at firstperfectly fertile, it is said (18/72. J. Mills 'Treati-se on Cattle' 1776 page 72.) that mares broughtup on dry food in the stable, and turned out to

grass, do not at first breed. The peahen, as wehave seen, is said not to lay so many eggs inEngland as in India. It was long before the ca-nary-bird was fully fertile, and even now first-rate breeding birds are not common. (18/73.Bechstein 'Stubenvogel' s. 242.) In the hot anddry province of Delhi, as I hear from Dr. Falco-ner, the eggs of the turkey, though placed un-der a hen, are extremely liable to fail. Accor-ding to Roulin, geese taken to the lofty plateauof Bogota, at first laid seldom, and then only afew eggs; of these scarcely a fourth were hat-ched, and half the young birds died; in the se-cond generation they were more fertile; andwhen Roulin wrote they were becoming as fer-tile as our geese in Europe. With respect to thevalley of Quito, Mr. Orton says (18/74. 'TheAndes and the Amazon' 1870 page 107.) "theonly geese in the valley are a few importedfrom Europe, and these refuse to propagate." Inthe Philippine Archipelago the goose, it is as-serted, will not breed or even lay eggs. (18/75.

Crawford 'Descriptive Dict. of the Indian Is-lands' 1856 page 145.) A more curious case isthat of the fowl, which, according to Roulin,when first introduced would not breed at Cus-co in Bolivia, but subsequently became quitefertile; and the English Game fowl, lately intro-duced, had not as yet arrived at its full fertility,for to raise two or three chickens from a nest ofeggs was thought fortunate. In Europe closeconfinement has a marked effect on the fertilityof the fowl: it has been found in France thatwith fowls allowed considerable freedom onlytwenty per cent of the eggs failed; when allo-wed less freedom forty per cent failed; and inclose confinement sixty out of the hundred we-re not hatched. (18/76. 'Bull. de la Soc. d'Accli-mat.' tome 9 1862 pages 380, 384.) So we seethat unnatural and changed conditions of lifeproduce some effect on the fertility of our mostthoroughly domesticated animals, in the samemanner, though in a far less degree, as withcaptive wild animals.

It is by no means rare to find certain males andfemales which will not breed together, thoughboth are known to be perfectly fertile with ot-her males and females. We have no reason tosuppose that this is caused by these animalshaving been subjected to any change in theirhabits of life; therefore such cases are hardlyrelated to our present subject. The cause appa-rently lies in an innate sexual incompatibility ofthe pair which are matched. Several instanceshave been communicated to me by Mr. W.C.Spooner (well known for his essay on Cross-breeding), by Mr. Eyton of Eyton, by Mr.Wicksted and other breeders, and especially byMr. Waring of Chelsfield, in relation to horses,cattle, pigs, foxhounds, other dogs, and pi-geons. (18/77. For pigeons see Dr. Chapuis 'LePigeon Voyageur Belge' 1865 page 66.) In thesecases, females, which either previously or sub-sequently were proved to be fertile, failed tobreed with certain males, with whom it wasparticularly desired to match them. A change in

the constitution of the female may sometimeshave occurred before she was put to the secondmale; but in other cases this explanation ishardly tenable, for a female, known not to bebarren, has been unsuccessfully paired seven oreight times with the same male likewise knownto be perfectly fertile. With cart-mares, whichsometimes will not breed with stallions of pureblood, but subsequently have bred with cart-stallions, Mr. Spooner is inclined to attributethe failure to the lesser sexual power of the ra-cehorse. But I have heard from the greatestbreeder of racehorses at the present day,through Mr. Waring, that "it frequently occurswith a mare to be put several times during oneor two seasons to a particular stallion of ack-nowledged power, and yet prove barren; themare afterwards breeding at once with someother horse." These facts are worth recording,as they show, like so many previous facts, onwhat slight constitutional differences the fertili-ty of an animal often depends.]

STERILITY OF PLANTS FROM CHANGEDCONDITIONS OF LIFE, AND FROM OTHERCAUSES.

In the vegetable kingdom cases of sterility fre-quently occur, analogous with those previouslygiven in the animal kingdom. But the subject isobscured by several circumstances, presently tobe discussed, namely, the contabescence of theanthers, as Gartner has named a certain affec-tion—monstrosities— doubleness of the flo-wer—much-enlarged fruit—and long-continued or excessive propagation by buds.

[It is notorious that many plants in our gardensand hot-houses, though preserved in the mostperfect health, rarely or never produce seed. Ido not allude to plants which run to leaves,from being kept too damp, or too warm, or toomuch manured; for these do not flower, andthe case may be wholly different. Nor do Iallude to fruit not ripening from want of heat orrotting from too much moisture. But many exo-

tic plants, with their ovules and pollen appea-ring perfectly sound, will not set any seed. Thesterility in many cases, as I know from my ownobservation, is simply due to the absence of theproper insects for carrying the pollen to thestigma. But after excluding the several casesjust specified, there are many plants in whichthe reproductive system has been seriouslyaffected by the altered conditions of life towhich they have been subjected.

It would be tedious to enter on many details.Linnaeus long ago observed (18/78. 'SwedishActs' volume 1 1739 page 3. Pallas makes thesame remark in his 'Travels' English translationvolume 1 page 292.) that Alpine plants, alt-hough naturally loaded with seed, produceeither few or none when cultivated in gardens.But exceptions often occur: the Draba sylves-tris, one of our most thoroughly Alpine plants,multiplies itself by seed in Mr. H.C. Watson'sgarden, near London; and Kerner, who has

particularly attended to the cultivation of Alpi-ne plants, found that various kinds, when cul-tivated, spontaneously sowed themselves.(18/79. A. Kerner 'Die Cultur der Alpenpflan-zen' 1864 s. 139; Watson 'Cybele Britannica'volume 1 page 131; Mr. D. Cameron also haswritten on the culture of Alpine plants in 'Gard.Chronicle' 1848 pages 253, 268, and mentions afew which seed.) Many plants which naturallygrow in peat-earth are entirely sterile in ourgardens. I have noticed the same fact with se-veral liliaceous plants, which nevertheless grewvigorously.

Too much manure renders some kinds utterlysterile, as I have myself observed. The tendencyto sterility from this cause runs in families;thus, according to Gartner (18/80. 'Beitrage zurKenntniss der Befruchtung' 1844 s. 333.), it ishardly possible to give too much manure tomost Gramineae, Cruciferae, and Leguminosae,whilst succulent and bulbous-rooted plants are

easily affected. Extreme poverty of soil is lessapt to induce sterility; but dwarfed plants ofTrifolium minus and repens, growing on alawn often mown and never manured, werefound by me not to produce any seed. Thetemperature of the soil, and the season at whichplants are watered, often have a marked effecton their fertility, as was observed by Kolreuterin the case of Mirabilis. (18/81. 'Nova Acta Pe-trop.' 1793 page 391.) Mr. Scott, in the BotanicGardens of Edinburgh, observed that Onci-dium divaricatum would not set seed whengrown in a basket in which it throve, but wascapable of fertilisation in a pot where it was alittle damper. Pelargonium fulgidum, for manyyears after its introduction, seeded freely; itthen became sterile; now it is fertile (18/82.'Cottage Gardener' 1856 pages 44, 109.) if keptin a dry stove during the winter. Other varietiesof pelargonium are sterile and others fertilewithout our being able to assign any cause.Very slight changes in the position of a plant,

whether planted on a bank or at its base, some-times make all the difference in its producingseed. Temperature apparently has a much morepowerful influence on the fertility of plantsthan on that of animals. Nevertheless it is won-derful what changes some few plants willwithstand with undiminished fertility: thus theZephyranthes candida, a native of the modera-tely warm banks of the Plata, sows itself in thehot dry country near Lima, and in Yorkshireresists the severest frosts, and I have seen seedsgathered from pods which had been coveredwith snow during three weeks. (18/83. Dr. Her-bert 'Amaryllidaceae' page 176.) Berberis walli-chii, from the hot Khasia range in India, isuninjured by our sharpest frosts, and ripens itsfruit under our cool summers. Nevertheless, Ipresume we must attribute to change of climatethe sterility of many foreign plants; thus, thePersian and Chinese lilacs (Syringa persica andchinensis), though perfectly hardy here, neverproduce a seed; the common lilac (S. vulgaris)

seeds with us moderately well, but in parts ofGermany the capsules never contain seed.(18/84. Gartner 'Beitrage zur Kenntniss' etc. s.560, 564.) Some few of the cases, given in thelast chapter, of self-impotent plants, might havebeen here introduced, as their state seems dueto the conditions to which they have been sub-jected.

The liability of plants to be affected in theirfertility by slightly changed conditions is themore remarkable, as the pollen when once inprocess of formation is not easily injured; aplant may be transplanted, or a branch withflower-buds be cut off and placed in water, andthe pollen will be matured. Pollen, also, whenonce mature, may be kept for weeks or evenmonths. (18/85. 'Gardener's Chronicle' 1844page 215; 1850 page 470. Faivre gives a goodresume on this subject in his 'La Variabilite desEspeces' 1868 page 155.) The female organs aremore sensitive, for Gartner (18/86. 'Beitrage zur

Kenntniss' etc. s. 252, 338.) found that dicotyle-donous plants, when carefully removed so thatthey did not in the least flag, could seldom befertilised; this occurred even with potted plantsif the roots had grown out of the hole at thebottom. In some few cases, however, as withDigitalis, transplantation did not prevent ferti-lisation; and according to the testimony ofMawz, Brassica rapa, when pulled up by itsroots and placed in water, ripened its seed.Flower-stems of several monocotyledonousplants when cut off and placed in water likewi-se produce seed. But in these cases I presumethat the flowers had been already fertilised, forHerbert (18/87. 'Journal of Hort. Soc.' volume 21847 page 83.) found with the Crocus that theplants might be removed or mutilated after theact of fertilisation, and would still perfect theirseeds; but that, if transplanted before beingfertilised, the application of pollen was power-less.

Plants which have been long cultivated cangenerally endure with undiminished fertilityvarious and great changes; but not in most ca-ses so great a change of climate as domesticatedanimals. It is remarkable that many plants un-der these circumstances are so much affectedthat the proportion and the nature of theirchemical ingredients are modified, yet theirfertility is unimpaired. Thus, as Dr. Falconerinforms me, there is a great difference in thecharacter of the fibre in hemp, in the quantity ofoil in the seed of the Linum, in the proportionof narcotin to morphine in the poppy, in glutento starch in wheat, when these plants are culti-vated on the plains and on the mountains ofIndia; nevertheless, they all remain fully fertile.

CONTABESCENCE.

Gartner has designated by this term a peculiarcondition of the anthers in certain plants, inwhich they are shrivelled, or become brown

and tough, and contain no good pollen. Whenin this state they exactly resemble the anthers ofthe most sterile hybrids. Gartner (18/88. 'Bei-trage zur Kenntniss' etc. s. 117 et seq.; Kolreuter'Zweite Fortsetzung' s. 10, 121; 'Dritte Fortset-zung' s. 57. Herbert 'Amaryllidaceae' page 355.Wiegmann 'Ueber die Bastarderzeugung' s.27.), in his discussion on this subject, has shownthat plants of many orders are occasionally thusaffected; but the Caryophyllaceae and Liliaceaesuffer most, and to these orders, I think, theEricaceae may be added. Contabescence variesin degree, but on the same plant all the flowersare generally affected to nearly the same extent.The anthers are affected at a very early periodin the flower-bud, and remain in the same state(with one recorded exception) during the life ofthe plant. The affection cannot be cured by anychange of treatment, and is propagated by la-yers, cuttings, etc., and perhaps even by seed.In contabescent plants the female organs areseldom affected, or merely become precocious

in their development. The cause of this affec-tion is doubtful, and is different in differentcases. Until I read Gartner's discussion I attri-buted it, as apparently did Herbert, to the un-natural treatment of the plants; but its perma-nence under changed conditions, and the fema-le organs not being affected, seem incompatiblewith this view. The fact of several endemicplants becoming contabescent in our gardensseems, at first sight, equally incompatible withthis view; but Kolreuter believes that this is theresult of their transplantation. The contabescentplants of Dianthus and Verbascum, found wildby Wiegmann, grew on a dry and sterile bank.The fact that exotic plants are eminently liableto this affection also seems to show that it is insome manner caused by their unnatural treat-ment. In some instances, as with Silene, Gart-ner's view seems the most probable, namely,that it is caused by an inherent tendency in thespecies to become dioecious. I can add anothercause, namely, the illegitimate unions of hete-

rostyled plants, for I have observed seedlings ofthree species of Primula and of Lythrum salica-ria, which had been raised from plants illegiti-mately fertilised by their own-form pollen, withsome or all their anthers in a contabescent state.There is perhaps an additional cause, namely,self-fertilisation; for many plants of Dianthusand Lobelia, which had been raised from self-fertilised seeds, had their anthers in this state;but these instances are not conclusive, as bothgenera are liable from other causes to this affec-tion.

Cases of an opposite nature likewise occur,namely, plants with the female organs struckwith sterility, whilst the male organs remainperfect. Dianthus japonicus, a Passiflora, andNicotiana, have been described by Gartner(18/89. 'Bastarderzengung' s. 356.) as being inthis unusual condition.

MONSTROSITIES AS A CAUSE OF STERI-LITY.

Great deviations of structure, even when thereproductive organs themselves are not se-riously affected, sometimes cause plants to be-come sterile. But in other cases plants may be-come monstrous to an extreme degree and yetretain their full fertility. Gallesio, who certainlyhad great experience (18/90. 'Teoria della Ri-produzione' 1816 page 84; 'Traite du Citrus'1811 page 67.), often attributes sterility to thiscause; but it may be suspected that in some ofhis cases sterility was the cause, and not theresult, of the monstrous growths. The curiousSt. Valery apple, although it bears fruit, rarelyproduces seed. The wonderfully anomalousflowers of Begonia frigida, formerly described,though they appear fit for fructification, aresterile. (18/91. Mr. C.W. Crocker in 'Gardener'sChronicle' 1861 page 1092.) Species of Primulain which the calyx is brightly coloured are said(18/92. Verlot 'Des Varietes' 1865 page 80.) to

be often sterile, though I have known them tobe fertile. On the other hand, Verlot gives seve-ral cases of proliferous flowers which can bepropagated by seed. This was the case with apoppy, which had become monopetalous bythe union of its petals. (18/93. Verlot ibid page88.) Another extraordinary poppy, with thestamens replaced by numerous small supple-mentary capsules, likewise reproduces itself byseed. This has also occurred with a plant ofSaxifraga geum, in which a series of adventi-tious carpels, bearing ovules on their margins,had been developed between the stamens andthe normal carpels (18/94. Prof. Allman, Brit.Assoc., quoted in the 'Phytologist' volume 2page 483. Prof. Harvey, on the authority of Mr.Andrews, who discovered the plant, informedme that this monstrosity could be propagatedby seed. With respect to the poppy see Prof.Goeppert as quoted in 'Journal of Horticulture'July 1, 1863 page 171.) Lastly, with respect topeloric flowers, which depart wonderfully from

the natural structure,—those of Linaria vulgarisseem generally to be more or less sterile, whilstthose before described of Antirrhinum majus,when artificially fertilised with their own po-llen, are perfectly fertile, though sterile whenleft to themselves, for bees are unable to crawlinto the narrow tubular flower. The peloricflowers of Corydalis solida, according to Go-dron (18/95. 'Comptes Rendus' December 19,1864 page 1039.), are sometimes barren andsometimes fertile; whilst those of Gloxinia arewell known to yield plenty of seed. In ourgreenhouse Pelargoniums, the central flower ofthe truss is often peloric, and Mr. Masters in-forms me that he tried in vain during severalyears to get seed from these flowers. I likewisemade many vain attempts, but sometimes suc-ceeded in fertilising them with pollen from anormal flower of another variety; and converse-ly I several times fertilised ordinary flowerswith peloric pollen. Only once I succeeded inraising a plant from a peloric flower fertilised

by pollen from a peloric flower borne by anot-her variety; but the plant, it may be added, pre-sented nothing particular in its structure. Hen-ce we may conclude that no general rule can belaid down; but any great deviation from thenormal structure, even when the reproductiveorgans themselves are not seriously affected,certainly often leads to sexual impotence.

DOUBLE FLOWERS.

When the stamens are converted into petals,the plant becomes on the male side sterile;when both stamens and pistils are thus chan-ged, the plant becomes completely barren.Symmetrical flowers having numerous stamensand petals are the most liable to become dou-ble, as perhaps follows from all multiple organsbeing the most subject to variability. But flo-wers furnished with only a few stamens, andothers which are asymmetrical in structure,sometimes become double, as we see with the

double gorse or Ulex, and Antirrhinum. TheCompositae bear what are called double flo-wers by the abnormal development of the coro-lla of their central florets. Doubleness is some-times connected with prolification (18/96.'Gardener's Chronicle' 1866 page 681.), or thecontinued growth of the axis of the flower.Doubleness is strongly inherited. No one hasproduced, as Lindley remarks (18/97. 'Theoryof Horticulture' page 333.), double flowers bypromoting the perfect health of the plant. Onthe contrary, unnatural conditions of life favourtheir production. There is some reason to belie-ve that seeds kept during many years, andseeds believed to be imperfectly fertilised, yielddouble flowers more freely than fresh and per-fectly fertilised seed. (18/98. Mr. Fairweather'Transact. Hort. Soc.' volume 3 page 406: Bossequoted by Bronn 'Geschichte der Natur' b. 2 s.77. On the effects of the removal of the antherssee Mr. Leitner in Silliman's 'North AmericanJourn. of Science' volume 23 page 47; and Ver-

lot 'Des Varietes' 1865 page 84.) Long-continuedcultivation in rich soil seems to be the commo-nest exciting cause. A double narcissus and adouble Anthemis nobilis, transplanted intovery poor soil, has been observed to becomesingle (18/99. Lindley's 'Theory of Horticulture'page 3?3.); and I have seen a completely doublewhite primrose rendered permanently singleby being divided and transplanted whilst infull flower. It has been observed by Professor E.Morren that doubleness of the flowers and va-riegation of the leaves are antagonistic states;but so many exceptions to the rule have latelybeen recorded (18/100. 'Gardener's Chronicle'1865 page 626; 1866 pages 290, 730; and Verlot'Des Varietes' page 75.), that, though general, itcannot be looked at as invariable. Variegationseems generally to result from a feeble oratrophied condition of the plant, and a largeproportion of the seedlings raised from parents,if both are variegated, usually perish at an earlyage; hence we may perhaps infer that double-

ness, which is the antagonistic state, commonlyarises from a plethoric condition. On the otherhand, extremely poor soil sometimes, thoughrarely, appears to cause doubleness: I formerlydescribed (18/101. 'Gardener's Chronicle' 1843page 628. In this article I suggested the theoryabove given on the doubleness of flowers. Thisview is adopted by Carriere 'Production et Fix.des Varietes' 1865 page 67.) some completelydouble, bud-like, flowers produced in largenumbers by stunted wild plants of Gentianaamarella growing on a poor chalky bank. I havealso noticed a distinct tendency to doublenessin the flowers of a Ranunculus, Horse-chestnut,and Bladder-nut (Ranunculus repens, Aesculuspavia, and Staphylea), growing under veryunfavourable conditions. Professor Lehmann(18/102. Quoted by Gartner 'Bastarderzeugung's. 567.) found several wild plants growing neara hot spring with double flowers. With respectto the cause of doubleness, which arises, as wesee, under widely different circumstances, I

shall presently attempt to show that the mostprobable view is that unnatural conditions firstgive a tendency to sterility, and that then, onthe principle of compensation, as the reproduc-tive organs do not perform their proper func-tions, they either become developed into petals,or additional petals are formed. This view haslately been supported by Mr. Laxton (18/103.'Gardener's Chronicle' 1866 page 901.) who ad-vances the case of some common peas, which,after long-continued heavy rain, flowered asecond time, and produced double flowers.

SEEDLESS FRUIT.

Many of our most valuable fruits, althoughconsisting in a homological sense of widelydifferent organs, are either quite sterile, or pro-duce extremely few seeds. This is notoriouslythe case with our best pears, grapes, and figs,with the pine-apple, banana, bread-fruit, po-megranate, azarole, date-palms, and some

members of the orange-tribe. Poorer varieties ofthese same fruits either habitually or occasiona-lly yield seed. (18/104. Lindley 'Theory of Hor-ticulture' pages 175-179; Godron 'De l'Espece'tome 2 page 106; Pickering 'Races of Man;' Ga-llesio 'Teoria della Riproduzione' l816 pages101-110. Meyen 'Reise um Erde' Th. 2 s. 214states that at Manilla one variety of the bananais full of seeds: and Chamisso (Hooker's 'Bot.Misc.' volume 1 page 310) describes a variety ofthe bread-fruit in the Mariana Islands withsmall fruit, containing seeds which are frequen-tly perfect. Burnes in his 'Travels in Bokhara'remarks on the pomegranate seeding in Ma-zenderan, as a remarkable peculiarity.) Mosthorticulturists look at the great size and anoma-lous development of the fruit as the cause, andsterility as the result; but the opposite view, aswe shall presently see, is more probable.

STERILITY FROM THE EXCESSIVE DEVE-LOPMENT OF THE ORGANS OF GROWTHOR VEGETATION.

Plants which from any cause grow too luxu-riantly, and produce leaves, stems, runners,suckers, tubers, bulbs, etc., in excess, sometimesdo not flower, or if they flower do not yieldseed. To make European vegetables under thehot climate of India yield seed, it is necessary tocheck their growth; and, when one-thirdgrown, they are taken up, and their stems andtap-roots are cut or mutilated. (18/105. Ingle-dew in 'Transact. of Agricult. and Hort. Soc. ofIndia' volume 2.) So it is with hybrids; for ins-tance, Prof. Lecoq (18/106. 'De la Fecondation'1862 page 308.) had three plants of Mirabilis,which, though they grew luxuriantly and flo-wered, were quite sterile; but after beating onewith a stick until a few branches alone wereleft, these at once yielded good seed. The sugar-cane, which grows vigorously and produces alarge supply of succulent stems, never, accor-

ding to various observers, bears seed in theWest Indies, Malaga, India, Cochin China,Mauritius, or the Malay Archipelago. (18/107.Hooker 'Bot. Misc.' volume 1 page 99; Gallesio'Teoria della Riproduzione' page 110. Dr. J. deCordemoy in 'Transact. of the R. Soc. of Mauri-tius' new series volume 6 1873 pages 60-67, gi-ves a large number of cases of plants whichnever seed, including several species indige-nous in Mauritius.) Plants which produce alarge number of tubers are apt to be sterile, asoccurs, to a certain extent, with the commonpotato; and Mr. Fortune informs me that thesweet potato (Convolvulus batatas) in Chinanever, as far as he has seen, yields seed. Dr.Royle remarks (18/108. 'Transact. Linn. Soc.'volume 17 page 563.) that in India the Agavevivipara, when grown in rich soil, invariablyproduces bulbs, but no seeds; whilst a poor soiland dry climate lead to an opposite result. InChina, according to Mr. Fortune, an extraordi-nary number of little bulbs are developed in the

axils of the leaves of the yam, and this plantdoes not bear seed. Whether in these cases, asin those of double flowers and seedless fruit,sexual sterility from changed conditions of lifeis the primary cause which leads to the excessi-ve development of the organs of vegetation, isdoubtful; though some evidence might be ad-vanced in favour of this view. It is perhaps amore probable view that plants which propaga-te themselves largely by one method, namelyby buds, have not sufficient vital power or or-ganised matter for the other method of sexualgeneration.

Several distinguished botanists and good prac-tical judges believe that long- continued propa-gation by cuttings, runners, tubers, bulbs, etc.,independently of any excessive development ofthese parts, is the cause of many plants failingto produce flowers, or producing only barrenflowers,—it is as if they had lost the habit ofsexual generation. (18/109. Godron 'De l'Espe-

ce' tome 2 page 106; Herbert on Crocus in'Journal of Hort. Soc.' volume 1 1846 page 254:Dr. Wight, from what he has seen in India, be-lieves in this view; 'Madras Journal of Lit. andScience' volume 4 1836 page 61.) That manyplants when thus propagated are sterile therecan be no doubt, but as to whether the longcontinuance of this form of propagation is theactual cause of their sterility, I will not venture,from the want of sufficient evidence, to expressan opinion.

That plants may be propagated for long periodsby buds, without the aid of sexual generation,we may safely infer from this being the casewith many plants which must have long survi-ved in a state of nature. As I have had occasionbefore to allude to this subject, I will here givesuch cases as I have collected. Many alpineplants ascend mountains beyond the height atwhich they can produce seed. (18/110. Wah-lenberg specifies eight species in this state on

the Lapland Alps: see Appendix to Linnaeus'Tour in Lapland' translated by Sir J.E. Smithvolume 2 pages 274-280.) Certain species of Poaand Festuca, when growing on mountain-pastures, propagate themselves, as I hear fromMr. Bentham, almost exclusively by bulblets.Kalm gives a more curious instance (18/111.'Travels in North America' English translationvolume 3 page 175.) of several American trees,which grow so plentifully in marshes or inthick woods, that they are certainly well adap-ted for these stations, yet scarcely ever produceseeds; but when accidentally growing on theoutside of the marsh or wood, are loaded withseed. The common ivy is found in NorthernSweden and Russia, but flowers and fruits onlyin the southern provinces. The Acorus calamusextends over a large portion of the globe, but sorarely perfects fruit that this has been seen onlyby a few botanists; according to Caspary, all itspollen-grains are in a worthless condition.(18/112. With respect to the ivy and Acorus see

Dr. Broomfield in the 'Phytologist' volume 3page 376. Also Lindley and Vaucher on theAcorus and see Caspary as below.) The Hyperi-cum calycinum, which propagates itself so free-ly in our shrubberies by rhizomes, and is natu-ralised in Ireland, blossoms profusely, but rare-ly sets any seed, and this only during certainyears; nor did it set any when fertilised in mygarden by pollen from plants growing at a dis-tance. The Lysimachia nummularia, which isfurnished with long runners, so seldom produ-ces seed-capsules, that Prof. Decaisne (18/113.'Annal. des Sc. Nat.' 3rd series Zool. tome 4page 280. Prof. Decaisne refers also to analo-gous cases with mosses and lichens near Paris.),who has especially attended to this plant, hasnever seen it in fruit. The Carex rigida oftenfails to perfect its seed in Scotland, Lapland,Greenland, Germany, and New Hampshire inthe United States. (18/114. Mr. Tuckermann inSilliman's 'American Journal of Science' volume65 page 1.) The periwinkle (Vinca minor),

which spreads largely by runners, is said scar-cely ever to produce fruit in England (18/115.Sir J.E. Smith 'English Flora' volume 1 page339.); but this plant requires insect-aid for itsfertilisation, and the proper insects may be ab-sent or rare. The Jussiaea grandiflora has beco-me naturalised in Southern France, and hasspread by its rhizomes so extensively as to im-pede the navigation of the waters, but neverproduces fertile seed. (18/116. G. Planchon'Flora de Montpellier' 1864 page 20.) The horse-radish (Cochleria armoracia) spreads pertina-ciously and is naturalised in various parts ofEurope; though it bears flowers, these rarelyproduce capsules: Professor Caspary informsme that he has watched this plant since 1851,but has never seen its fruit; 65 per cent of itspollen-grains are bad. The common Ranuncu-lus ficaria rarely bears seed in England, France,or Switzerland; but in 1863 I observed seeds onseveral plants growing near my house. (18/117.On the non-production of seeds in England see

Mr. Crocker in 'Gardener's Weekly Magazine'1852 page 70; Vaucher 'Hist. Phys. Plantesd'Europe' tome 1 page 33; Lecoq 'Geograph.Bot. d'Europe' tome 4 page 466; Dr. D. Clos in'Annal. des Sc. Nat.' 3rd series Bot. tome 171852 page 129: this latter author refers to otheranalogous cases. See more especially on thisplant and on other allied cases Prof. Caspary"Die Nuphar" 'Abhand. Naturw. Gesellsch. zuHalle' b. 11 1870 page 40, 78.) Other cases ana-logous with the foregoing could be given; forinstance, some kinds of mosses and lichenshave never been seen to fructify in France.

Some of these endemic and naturalised plantsare probably rendered sterile from excessivemultiplication by buds, and their consequentincapacity to produce and nourish seed. But thesterility of others more probably depends onthe peculiar conditions under which they live,as in the case of the ivy in the northern part ofEurope, and of the trees in the swamps of the

United States; yet these plants must be in somerespects eminently well adapted for the stationswhich they occupy, for they hold their placesagainst a host of competitors.]

Finally, the high degree of sterility which oftenaccompanies the doubling of flowers, or anexcessive development of fruit, seldom super-venes at once. An incipient tendency is obser-ved, and continued selection completes theresult. The view which seems the most proba-ble, and which connects together all the fore-going facts and brings them within our presentsubject, is, that changed and unnatural condi-tions of life first give a tendency to sterility; andin consequence of this, the organs of reproduc-tion being no longer able fully to perform theirproper functions, a supply of organised matter,not required for the development of the seed,flows either into these organs and renders themfoliaceous, or into the fruit, stems, tubers, etc.,increasing their size and succulency. But it is

probable that there exists, independently of anyincipient sterility, an antagonism between thetwo forms of reproduction, namely, by seedand buds, when either is carried to an extremedegree. That incipient sterility plays an impor-tant part in the doubling of flowers, and in theother cases just specified, I infer chiefly fromthe following facts. When fertility is lost from awholly different cause, namely, from hybri-dism, there is a strong tendency, as Gartner(18/118. 'Bastarderzeugung' s. 565. Kolreuter'Dritte Fortsetzung' s. 73, 87, 119) also showsthat when two species, one single and the otherdouble, are crossed, the hybrids are apt to beextremely double.) affirms, for flowers to be-come double, and this tendency is inherited.Moreover, it is notorious that with hybrids themale organs become sterile before the femaleorgans, and with double flowers the stamensfirst become foliaceous. This latter fact is wellshown by the male flowers of dioecious plants,which, according to Gallesio (18/119. 'Teoria

della Riproduzione Veg.' 1816 page 73.) firstbecome double. Again, Gartner (18/120. 'Bas-tarderzeugung' s. 573.) often insists that theflowers of even utterly sterile hybrids, whichdo not produce any seed, generally yield per-fect capsules or fruit,—a fact which has likewi-se been repeatedly observed by Naudin withthe Cucurbitaceae; so that the production offruit by plants rendered sterile through anycause is intelligible. Kolreuter has also expres-sed his unbounded astonishment at the sizeand development of the tubers in certain hy-brids; and all experimentalists (18/121. Ibid s.527.) have remarked on the strong tendency inhybrids to increase by roots, runners, and suc-kers. Seeing that hybrid plants, which fromtheir nature are more or less sterile, thus tend toproduce double flowers; that they have theparts including the seed, that is the fruit, per-fectly developed, even when containing noseed; that they sometimes yield gigantic roots;that they almost invariably tend to increase

largely by suckers and other such means;—seeing this, and knowing, from the many factsgiven in the earlier parts of this chapter, thatalmost all organic beings when exposed to un-natural conditions tend to become more or lesssterile, it seems much the most probable viewthat with cultivated plants sterility is the exci-ting cause, and double flowers, rich seedlessfruit, and in some cases largely-developed or-gans of vegetation, etc., are the indirect re-sults—these results having been in most caseslargely increased through continued selectionby man.

CHAPTER 2.XIX.

SUMMARY OF THE FOUR LAST CHAP-TERS, WITH REMARKS ON HYBRIDISM.

ON THE EFFECTS OF CROSSING. THEINFLUENCE OF DOMESTICATION ON FER-

TILITY. CLOSE INTERBREEDING. GOODAND EVIL RESULTS FROM CHANGEDCONDITIONS OF LIFE. VARIETIES WHENCROSSED NOT INVARIABLY FERTILE. ONTHE DIFFERENCE IN FERTILITY BETWEENCROSSED SPECIES AND VARIETIES. CON-CLUSIONS WITH RESPECT TO HYBRIDISM.LIGHT THROWN ON HYBRIDISM BY THEILLEGITIMATE PROGENY OF HETEROSTY-LED PLANTS. STERILITY OF CROSSEDSPECIES DUE TO DIFFERENCES CONFINEDTO THE REPRODUCTIVE SYSTEM. NOTACCUMULATED THROUGH NATURAL SE-LECTION. REASONS WHY DOMESTIC VA-RIETIES ARE NOT MUTUALLY STERILE.TOO MUCH STRESS HAS BEEN LAID ONTHE DIFFERENCE IN FERTILITY BETWEENCROSSED SPECIES AND CROSSED VARIE-TIES. CONCLUSION.

It was shown in the fifteenth chapter that whenindividuals of the same variety, or even of adistinct variety, are allowed freely to intercross,

uniformity of character is ultimately acquired.Some few characters, however, are incapable offusion, but these are unimportant, as they areoften of a semi-monstrous nature, and havesuddenly appeared. Hence, to preserve ourdomesticated breeds true, or to improve themby methodical selection, it is obviously necessa-ry that they should be kept separate. Nevert-heless, a whole body of individuals may beslowly modified, through unconscious selec-tion, as we shall see in a future chapter, withoutseparating them into distinct lots. Domesticraces have often been intentionally modified byone or two crosses, made with some allied race,and occasionally even by repeated crosses withvery distinct races; but in almost all such cases,long-continued and careful selection has beenabsolutely necessary, owing to the excessivevariability of the crossed offspring, due to theprinciple of reversion. In a few instances,however, mongrels have retained a uniformcharacter from their first production.

When two varieties are allowed to cross freely,and one is much more numerous than the ot-her, the former will ultimately absorb the latter.Should both varieties exist in nearly equalnumbers, it is probable that a considerable pe-riod would elapse before the acquirement of auniform character; and the character ultimatelyacquired would largely depend on prepotencyof transmission and on the conditions of life;for the nature of these conditions would gene-rally favour one variety more than another, sothat a kind of natural selection would comeinto play. Unless the crossed offspring wereslaughtered by man without the least discrimi-nation, some degree of unmethodical selectionwould likewise come into action. From theseseveral considerations we may infer, that whentwo or more closely allied species first cameinto the possession of the same tribe, their cros-sing will not have influenced, in so great a de-gree as has often been supposed, the characterof the offspring in future times; although in

some cases it probably has had a considerableeffect.

Domestication, as a general rule, increases theprolificness of animals and plants. It eliminatesthe tendency to sterility which is common tospecies when first taken from a state of natureand crossed. On this latter head we have nodirect evidence; but as our races of dogs, cattle,pigs etc., are almost certainly descended fromaboriginally distinct stocks, and as these racesare now fully fertile together, or at least incom-parably more fertile than most species whencrossed, we may with entire confidence acceptthis conclusion.

Abundant evidence has been given that cros-sing adds to the size, vigour, and fertility of theoffspring. This holds good when there has beenno previous close interbreeding. It applies tothe individuals of the same variety but belon-ging to different families, to distinct varieties,sub-species, and even to species. In the latter

case, though size is gained, fertility is lost; butthe increased size, vigour, and hardiness ofmany hybrids cannot be accounted for solelyon the principle of compensation from the inac-tion of the reproductive system. Certain plantswhilst growing under their natural conditions,others when cultivated, and others of hybridorigin, are completely self-impotent, thoughperfectly healthy; and such plants can be stimu-lated to fertility only by being crossed withother individuals of the same or of a distinctspecies.

On the other hand, long-continued close inter-breeding between the nearest relations dimi-nishes the constitutional vigour, size, and ferti-lity of the offspring; and occasionally leads tomalformations, but not necessarily to generaldeterioration of form or structure. This failureof fertility shows that the evil results of inter-breeding are independent of the augmentationof morbid tendencies common to both parents,

though this augmentation no doubt is oftenhighly injurious. Our belief that evil followsfrom close interbreeding rests to a certain ex-tent on the experience of practical breeders,especially of those who have reared many ani-mals of quickly propagating kinds; but it like-wise rests on several carefully recorded expe-riments. With some animals close interbreedingmay be carried on for a long period with impu-nity by the selection of the most vigorous andhealthy individuals; but sooner or later evilfollows. The evil, however, comes on so slowlyand gradually that it easily escapes observa-tion, but can be recognised by the almost ins-tantaneous manner in which size, constitutionalvigour, and fertility are regained when animalsthat have long been interbred are crossed witha distinct family.

These two great classes of facts, namely, thegood derived from crossing, and the evil fromclose interbreeding, with the consideration of

the innumerable adaptations throughout naturefor compelling, or favouring, or at least permit-ting, the occasional union of distinct indivi-duals, taken together, lead to the conclusionthat it is a law of nature that organic beingsshall not fertilise themselves for perpetuity.This law was first plainly hinted at in 1799,with respect to plants, by Andrew Knight(19/1. 'Transactions Phil. Soc.' 1799 page 202.For Kolreuter see 'Mem. de l'Acad. de St.-Petersbourg' tome 3 1809 published 1811 page197. In reading C.K. Sprengel's remarkablework, 'Das entdeckte Geheimniss' etc. 1793, it iscurious to observe how often this wonderfullyacute observer failed to understand the fullmeaning of the structure of the flowers whichhe has so well described, from not alwayshaving before his mind the key to the problem,namely, the good derived from the crossing ofdistinct individual plants.) and, not long after-wards, that sagacious observer Kolreuter, aftershowing how well the Malvaceae are adapted

for crossing, asks, "an id aliquid in recessuhabeat, quod hujuscemodi flores nunquamproprio suo pulvere, sed semper eo aliarum suspeciei impregnentur, merito quaritur? Certenatura nil facit frustra." Although we may de-mur to Kolreuter's saying that nature does not-hing in vain, seeing how many rudimentaryand useless organs there are, yet undoubtedlythe argument from the innumerable contrivan-ces, which favour crossing, is of the greatestweight. The most important result of this law isthat it leads to uniformity of character in theindividuals of the same species. In the case ofcertain hermaphrodites, which probably inter-cross only at long intervals of time, and withunisexual animals inhabiting somewhat separa-ted localities, which can only occasionally comeinto contact and pair, the greater vigour andfertility of the crossed offspring will ultimatelytend to give uniformity of character. But whenwe go beyond the limits of the same species,

free intercrossing is barred by the law of sterili-ty.

In searching for facts which might throw lighton the cause of the good effects from crossing,and of the evil effects from close interbreeding,we have seen that, on the one hand, it is a wide-ly prevalent and ancient belief, that animalsand plants profit from slight changes in theircondition of life; and it would appear that thegerm, in a somewhat analogous manner, is mo-re effectually stimulated by the male element,when taken from a distinct individual, and the-refore slightly modified in nature, than whentaken from a male having the same identicalconstitution. On the other hand, numerousfacts have been given, showing that when ani-mals are first subjected to captivity, even intheir native land, and although allowed muchliberty, their reproductive functions are oftengreatly impaired or quite annulled. Somegroups of animals are more affected than ot-

hers, but with apparently capricious exceptionsin every group. Some animals never or rarelycouple under confinement; some couple freely,but never or rarely conceive. The secondarymale characters, the maternal functions andinstincts, are occasionally affected. With plants,when first subjected to cultivation, analogousfacts have been observed. We probably owe ourdouble flowers, rich seedless fruits, and in so-me cases greatly developed tubers, etc., to inci-pient sterility of the above nature combinedwith a copious supply of nutriment. Animalswhich have long been domesticated, and plantswhich have long been cultivated, can generallywithstand, with unimpaired fertility, greatchanges in their conditions of life; though bothare sometimes slightly affected. With animalsthe somewhat rare capacity of breeding freelyunder confinement, together with their utility,mainly determine the kinds which have beendomesticated.

We can in no case precisely say what is the cau-se of the diminished fertility of an animal whenfirst captured, or of a plant when first cultiva-ted; we can only infer that it is caused by achange of some kind in the natural conditionsof life. The remarkable susceptibility of the re-productive system to such changes,—a suscep-tibility not common to any other organ,—apparently has an important bearing on Varia-bility, as we shall see in a future chapter.

It is impossible not to be struck with the doubleparallelism between the two classes of facts justalluded to. On the one hand, slight changes inthe conditions of life, and crosses betweenslightly modified forms or varieties, are benefi-cial as far as prolificness and constitutional vi-gour are concerned. On the other hand, chan-ges in the conditions greater in degree, or of adifferent nature, and crosses between formswhich have been slowly and greatly modifiedby natural means,—in other words, between

species,—are highly injurious, as far as the re-productive system is concerned, and in somefew instances as far as constitutional vigour isconcerned. Can this parallelism be accidental?Does it not rather indicate some real bond ofconnection? As a fire goes out unless it be sti-rred up, so the vital forces are always tending,according to Mr. Herbert Spencer, to a state ofequilibrium, unless disturbed and renovatedthrough the action of other forces.

In some few cases varieties tend to keep dis-tinct, by breeding at different seasons, by greatdifference in size, or by sexual preference. Butthe crossing of varieties, far from diminishing,generally adds to the fertility of the first unionand of the mongrel offspring. Whether all themore widely distinct domestic varieties areinvariably quite fertile when crossed, we do notpositively know; much time and trouble wouldbe requisite for the necessary experiments, andmany difficulties occur, such as the descent of

the various races from aboriginally distinctspecies, and the doubts whether certain formsought to be ranked as species or varieties. Ne-vertheless, the wide experience of practicalbreeders proves that the great majority of varie-ties, even if some should hereafter prove not tobe indefinitely fertile inter se, are far more ferti-le when crossed, than the vast majority of close-ly allied natural species. A few remarkable ca-ses have, however, been given on the authorityof excellent observers, showing that with plantscertain forms, which undoubtedly must be ran-ked as varieties, yield fewer seeds when cros-sed than is natural to the parent-species. Othervarieties have had their reproductive powers sofar modified that they are either more or lessfertile than their parents, when crossed with adistinct species.

Nevertheless, the fact remains indisputable thatdomesticated varieties, of animals and ofplants, which differ greatly from one another in

structure, but which are certainly descendedfrom the same aboriginal species, such as theraces of the fowl, pigeon, many vegetables, anda host of other productions, are extremely ferti-le when crossed; and this seems to make abroad and impassable barrier between domes-tic varieties and natural species. But, as I willnow attempt to show, the distinction is not sogreat and overwhelmingly important as it atfirst appears.

ON THE DIFFERENCE IN FERTILITYBETWEEN VARIETIES AND SPECIES WHENCROSSED.

This work is not the proper place for fully trea-ting the subject of hybridism, and I have alrea-dy given in my 'Origin of Species' a moderatelyfull abstract. I will here merely enumerate thegeneral conclusions which may be relied on,and which bear on our present point.

FIRSTLY.

The laws governing the production of hybridsare identical, or nearly identical, in the animaland vegetable kingdoms.

SECONDLY.

The sterility of distinct species when first uni-ted, and that of their hybrid offspring, gradua-te, by an almost infinite number of steps, fromzero, when the ovule is never impregnated anda seed-capsule is never formed, up to completefertility. We can only escape the conclusion thatsome species are fully fertile when crossed, bydetermining to designate as varieties all theforms which are quite fertile. This high degreeof fertility is, however, rare. Nevertheless,plants, which have been exposed to unnaturalconditions, sometimes become modified in sopeculiar a manner, that they are much morefertile when crossed with a distinct species thanwhen fertilised by their own pollen. Success in

effecting a first union between two species, andthe fertility of their hybrids, depend in an emi-nent degree on the conditions of life being fa-vourable. The innate sterility of hybrids of thesame parentage and raised from the same seed-capsule often differs much in degree.

THIRDLY.

The degree of sterility of a first cross betweentwo species does not always run strictly para-llel with that of their hybrid offspring. Manycases are known of species which can be cros-sed with ease, but yield hybrids excessivelysterile; and conversely some which can be cros-sed with great difficulty, but produce fairlyfertile hybrids. This is an inexplicable fact, onthe view that species have been specially en-dowed with mutual sterility in order to keepthem distinct.

FOURTHLY.

The degree of sterility often differs greatly intwo species when reciprocally crossed; for thefirst will readily fertilise the second; but thelatter is incapable, after hundreds of trials, offertilising the former. Hybrids produced fromreciprocal crosses between the same two spe-cies likewise sometimes differ in their degree ofsterility. These cases also are utterly inexplica-ble on the view of sterility being a special en-dowment.

FIFTHLY.

The degree of sterility of first crosses and ofhybrids runs, to a certain extent, parallel withthe general or systematic affinity of the formswhich are united. For species belonging to dis-tinct genera can rarely, and those belonging todistinct families can never, be crossed. The pa-rallelism, however, is far from complete; for amultitude of closely allied species will not uni-te, or unite with extreme difficulty, whilst other

species, widely different from one another, canbe crossed with perfect facility. Nor does thedifficulty depend on ordinary constitutionaldifferences, for annual and perennial plants,deciduous and evergreen trees, plants flowe-ring at different seasons, inhabiting differentstations, and naturally living under the mostopposite climates, can often be crossed withease. The difficulty or facility apparently de-pends exclusively on the sexual constitution ofthe species which are crossed; or on their sexualelective affinity, i.e. Wahlverwandtschaft ofGartner. As species rarely or never becomemodified in one character, without being at thesame time modified in many characters, and assystematic affinity includes all visible similari-ties and dissimilarities, any difference in sexualconstitution between two species would natu-rally stand in more or less close relation withtheir systematic position.

SIXTHLY.

The sterility of species when first crossed, andthat of hybrids, may possibly depend to a cer-tain extent on distinct causes. With pure speciesthe reproductive organs are in a perfect condi-tion, whilst with hybrids they are often plainlydeteriorated. A hybrid embryo which partakesof the constitution of its father and mother isexposed to unnatural conditions, as long as it isnourished within the womb, or egg, or seed ofthe mother-form; and as we know that unnatu-ral conditions often induce sterility, the repro-ductive organs of the hybrid might at this earlyage be permanently affected. But this cause hasno bearing on the infertility of first unions. Thediminished number of the offspring from firstunions may often result, as is certainly someti-mes the case, from the premature death of mostof the hybrid embryos. But we shall immediate-ly see that a law of an unknown nature appa-rently exists, which leads to the offspring fromunions, which are infertile, being themselves

more or less infertile; and this at present is allthat can be said.

SEVENTHLY.

Hybrids and mongrels present, with the onegreat exception of fertility, the most strikingaccordance in all other respects; namely, in thelaws of their resemblance to their two parents,in their tendency to reversion, in their variabili-ty, and in being absorbed through repeatedcrosses by either parent- form.

After arriving at these conclusions, I was led toinvestigate a subject which throws considerablelight on hybridism, namely, the fertility of hete-rostyled or dimorphic and trimorphic plants,when illegitimately united. I have had occasionseveral times to allude to these plants, and Imay here give a brief abstract of my observa-tions. Several plants belonging to distinct or-ders present two forms, which exist in aboutequal numbers, and which differ in no respect

except in their reproductive organs; one formhaving a long pistil with short stamens, theother a short pistil with long stamens; bothwith differently sized pollen-grains. With tri-morphic plants there are three forms likewisediffering in the lengths of their pistils and sta-mens, in the size and colour of the pollen-grains, and in some other respects; and as ineach of the three forms there are two sets ofstamens, there are altogether six sets of stamensand three kinds of pistils. These organs are soproportioned in length to one another that, inany two of the forms, half the stamens in eachstand on a level with the stigma of the thirdform. Now I have shown, and the result hasbeen confirmed by other observers, that, in or-der to obtain full fertility with these plants, it isnecessary that the stigma of the one formshould be fertilised by pollen taken from thestamens of corresponding height in the otherform. So that with dimorphic species twounions, which may be called legitimate, are

fully fertile, and two, which may be called ille-gitimate, are more or less infertile. With tri-morphic species six unions are legitimate, orfully fertile, and twelve are illegitimate, or mo-re or less infertile. (19/2. My observations 'Onthe Character and hybrid-like nature of theoffspring from the illegitimate union of Di-morphic and Trimorphic Plants' were publis-hed in the 'Journal of the Linnean Soc.' volume10 page 393. The abstract here given is nearlythe same with that which appeared in the 6thedition of my 'Origin of Species.')

The infertility which may be observed in va-rious dimorphic and trimorphic plants, whenillegitimately fertilised, that is, by pollen takenfrom stamens not corresponding in height withthe pistil, differs much in degree, up to absoluteand utter sterility; just in the same manner asoccurs in crossing distinct species. As the de-gree of sterility in the latter case depends in aneminent degree on the conditions of life being

more or less favourable, so I have found it withillegitimate unions. It is well known that if po-llen of a distinct species be placed on the stigmaof a flower, and its own pollen be afterwards,even after a considerable interval of time, pla-ced on the same stigma, its action is so stronglyprepotent that it generally annihilates the effectof the foreign pollen; so it is with the pollen ofthe several forms of the same species, for legi-timate pollen is strongly prepotent over illegi-timate pollen, when both are placed on the sa-me stigma. I ascertained this by fertilising seve-ral flowers, first illegitimately, and twenty-fourhours afterwards legitimately, with pollen ta-ken from a peculiarly coloured variety, and allthe seedlings were similarly coloured; thisshows that the legitimate pollen, though ap-plied twenty-four hours subsequently, hadwholly destroyed or prevented the action of thepreviously applied illegitimate pollen. Again,as, in making reciprocal crosses between thesame two species, there is occasionally a great

difference in the result, so the same thing oc-curs with trimorphic plants; for instance, themid-styled form of Lythrum salicaria could beillegitimately fertilised with the greatest ease bypollen from the longer stamens of the short-styled form, and yielded many seeds; but theshort-styled form did not yield a single seedwhen fertilised by the longer stamens of themid-styled form.

In all these respects the forms of the same un-doubted species, when illegitimately united,behave in exactly the same manner as do twodistinct species when crossed. This led me care-fully to observe during four years many seed-lings, raised from several illegitimate unions.The chief result is that these illegitimate plants,as they may be called, are not fully fertile. It ispossible to raise from dimorphic species, bothlong-styled and short-styled illegitimate plants,and from trimorphic plants all three illegitimateforms. These can then be properly united in a

legitimate manner. When this is done, there isno apparent reason why they should not yieldas many seeds as did their parents when legi-timately fertilised. But such is not the case; theyare all infertile, but in various degrees; somebeing so utterly and incurably sterile that theydid not yield during four seasons a single seedor even seed-capsule. These illegitimate plants,which are so sterile, although united with eachother in a legitimate manner, may be strictlycompared with hybrids when crossed inter se,and it is well known how sterile these lattergenerally are. When, on the other hand, a hy-brid is crossed with either pure parent- species,the sterility is usually much lessened: and so itis when an illegitimate plant is fertilised by alegitimate plant. In the same manner as thesterility of hybrids does not always run parallelwith the difficulty of making the first crossbetween the two parent-species, so the sterilityof certain illegitimate plants was unusuallygreat, whilst the sterility of the union from

which they were derived was by no meansgreat. With hybrids raised from the same seed-capsule the degree of sterility is innately varia-ble, so it is in a marked manner with illegitima-te plants. Lastly, many hybrids are profuse andpersistent flowerers, whilst other and moresterile hybrids produce few flowers, and areweak, miserable dwarfs; exactly similar casesoccur with the illegitimate offspring of variousdimorphic and trimorphic plants.

Although there is the closest identity in charac-ter and behaviour between illegitimate plantsand hybrids, it is hardly an exaggeration tomaintain that the former are hybrids, but pro-duced within the limits of the same species bythe improper union of certain forms, whilstordinary hybrids are produced from an impro-per union between so-called distinct species.We have already seen that there is the closestsimilarity in all respects between first illegiti-mate unions, and first crosses between distinct

species. This will perhaps be made more fullyapparent by an illustration:—we may supposethat a botanist found two well-marked varieties(and such occur) of the long-styled form of thetrimorphic Lithrum salicaria, and that he de-termined to try by crossing whether they werespecifically distinct. He would find that theyyielded only about one-fifth of the propernumber of seed, and that they behaved in allthe other above-specified respects as if they hadbeen two distinct species. But to make the casesure, he would raise plants from his supposedhybridised seed, and he would find that theseedlings were miserably dwarfed and utterlysterile, and that they behaved in all other res-pects like ordinary hybrids, he might thenmaintain that he had actually proved, in accor-dance with the common view, that his two va-rieties were as good and as distinct species asany in the world; but he would be completelymistaken.

The facts now given on dimorphic and tri-morphic plants are important, because theyshow us, first, that the physiological test of les-sened fertility, both in first crosses and in hy-brids, is no criterion of specific distinction; se-condly, because we may conclude that there issome unknown bond which connects the infer-tility of illegitimate unions with that of theirillegitimate offspring, and we are led to extendthe same view to first crosses and hybrids;thirdly, because we find, and this seems to meof especial importance, that two or three formsof the same species may exist and may differ inno respect whatever, either in structure or inconstitution, relatively to external conditions,and yet be sterile when united in certain ways.For we must remember that it is the union ofthe sexual elements of individuals of the sameform, for instance, of two long-styled forms,which results in sterility; whilst it is the unionof the sexual element proper to two distinctforms which is fertile. Hence the case appears

at first sight exactly the reverse of what occursin the ordinary unions of the individuals of thesame species, and with crosses between distinctspecies. It is, however, doubtful whether this isreally so; but I will not enlarge on this obscuresubject.

We may, however, infer as probable from theconsideration of dimorphic and trimorphicplants, that the sterility of distinct species whencrossed, and of their hybrid progeny, dependsexclusively on the nature of their sexual ele-ments, and not on any difference in their struc-ture or general constitution. We are also led tothis same conclusion by considering reciprocalcrosses, in which the male of one species cannotbe united, or only with great difficulty, with thefemale of a second species, whilst the conversecross can be effected with perfect facility. Thatexcellent observer, Gartner, likewise concludedthat species when crossed are sterile owing to

differences confined to their reproductive sys-tems.

On the principle which makes it necessary forman, whilst he is selecting and improving hisdomestic varieties, to keep them separate, itwould clearly be advantageous to varieties in astate of nature, that is to incipient species, ifthey could be kept from blending, eitherthrough sexual aversion, or by becoming mu-tually sterile. Hence it at one time appeared tome probable, as it has to others, that this sterili-ty might have been acquired through naturalselection. On this view we must suppose that ashade of lessened fertility first spontaneouslyappeared, like any other modification, in cer-tain individuals of a species when crossed withother individuals of the same species; and thatsuccessive slight degrees of infertility, frombeing advantageous, were slowly accumulated.This appears all the more probable, if we admitthat the structural differences between the

forms of dimorphic and trimorphic plants, asthe length and curvature of the pistil, etc., havebeen co-adapted through natural selection; forif this be admitted, we can hardly avoid exten-ding the same conclusion to their mutual infer-tility. Sterility, moreover, has been acquiredthrough natural selection for other and widelydifferent purposes, as with neuter insects inreference to their social economy. In the case ofplants, the flowers on the circumference of thetruss in the guelder rose (Viburnum opulus)and those on the summit of the spike in thefeather-hyacinth (Muscari comosum) have beenrendered conspicuous, and apparently in con-sequence sterile, in order that insects mighteasily discover and visit the perfect flowers. Butwhen we endeavour to apply the principle ofnatural selection to the acquirement by distinctspecies of mutual sterility, we meet with greatdifficulties. In the first place, it may be remar-ked that separate regions are often inhabited bygroups of species or by single species, which

when brought together and crossed are foundto be more or less sterile; now it could clearlyhave been no advantage to such separated spe-cies to have been rendered mutually sterile,and consequently this could not have been ef-fected through natural selection; but it mayperhaps be argued, that, if a species were ren-dered sterile with some one compatriot, sterili-ty with other species would follow as a neces-sary consequence. In the second place, it is asmuch opposed to the theory of natural selec-tion, as to the theory of special creation, that inreciprocal crosses the male element of one formshould have been rendered utterly impotent ona second form, whilst at the same time the maleelement of this second form is enabled freely tofertilise the first form; for this peculiar state ofthe reproductive system could not possiblyhave been advantageous to either species.

In considering the probability of natural selec-tion having come into action in rendering spe-

cies mutually sterile, one of the greatest diffi-culties will be found to lie in the existence ofmany graduated steps from slightly lessenedfertility to absolute sterility. It may be admitted,on the principle above explained, that it wouldprofit an incipient species if it were rendered insome slight degree sterile when crossed with itsparent-form or with some other variety; forthus fewer bastardised and deteriorated offs-pring would be produced to commingle theirblood with the new species in process of forma-tion. But he who will take the trouble to reflecton the steps by which this first degree of sterili-ty could be increased through natural selectionto that higher degree which is common to somany species, and which is universal with spe-cies which have been differentiated to a genericor family rank, will find the subject extraordi-narily complex. After mature reflection it seemsto me that this could not have been effectedthrough natural selection. Take the case of anytwo species which, when crossed, produce few

and sterile offspring; now, what is there whichcould favour the survival of those individualswhich happened to be endowed in a slightlyhigher degree with mutual infertility, andwhich thus approached by one small step to-wards absolute sterility? Yet an advance of thiskind, if the theory of natural selection bebrought to bear, must have incessantly occu-rred with many species, for a multitude aremutually quite barren. With sterile neuter in-sects we have reason to believe that modifica-tions in their structure and fertility have beenslowly accumulated by natural selection, froman advantage having been thus indirectly givento the community to which they belonged overother communities of the same species; but anindividual animal not belonging to a socialcommunity, if rendered slightly sterile whencrossed with some other variety, would notthus itself gain any advantage or indirectly giveany advantage to the other individuals of thesame variety, thus leading to their preservation.

But it would be superfluous to discuss thisquestion in detail; for with plants we have con-clusive evidence that the sterility of crossedspecies must be due to some principle, quiteindependent of natural selection. Both Gartnerand Kolreuter have proved that in general in-cluding numerous species, a series can be for-med from species which when crossed yieldfewer and fewer seeds, to species which neverproduce a single seed, but yet are affected bythe pollen of certain other species, for the ger-men swells. It is here manifestly impossible toselect the more sterile individuals, which havealready ceased to yield seeds; so that this acmeof sterility, when the germen alone is affected,cannot have been gained through selection; andfrom the laws governing the various grades ofsterility being so uniform throughout the ani-mal and vegetable kingdoms, we may infer thatthe cause, whatever it may be, is the same ornearly the same in all cases.

As species have not been rendered mutuallyinfertile through the accumulative action ofnatural selection, and as we may safely conclu-de, from the previous as well as from other andmore general considerations, that they have notbeen endowed through an act of creation withthis quality, we must infer that it has arisenincidentally during their slow formation in con-nection with other and unknown changes intheir organisation. By a quality arising inciden-tally, I refer to such cases as different species ofanimals and plants being differently affected bypoisons to which they are not naturally expo-sed; and this difference in susceptibility is clear-ly incidental on other and unknown differencesin their organisation. So again the capacity indifferent kinds of trees to be grafted on eachother, or on a third species, differs much, and isof no advantage to these trees, but is incidentalon structural or functional differences in theirwoody tissues. We need not feel surprise atsterility incidentally resulting from crosses bet-

ween distinct species,—the modified descen-dants of a common progenitor,—when we bearin mind how easily the reproductive system isaffected by various causes—often by extremelyslight changes in the conditions of life, by tooclose interbreeding, and by other agencies. It iswell to bear in mind such cases as that of thePassiflora alata, which recovered its self-fertility from being grafted on a distinct spe-cies—the cases of plants which normally orabnormally are self-impotent, but can readilybe fertilised by the pollen of a distinct species—and lastly the cases of individual domesticatedanimals which evince towards each othersexual incompatibility.

We now at last come to the immediate pointunder discussion: how is it that, with some fewexceptions in the case of plants, domesticatedvarieties, such as those of the dog, fowl, pigeon,several fruit-trees, and culinary vegetables,which differ from each other in external charac-

ters more than many species, are perfectly ferti-le when crossed, or even fertile in excess, whilstclosely allied species are almost invariably insome degree sterile? We can, to a certain extent,give a satisfactory answer to this question. Pas-sing over the fact that the amount of externaldifference between two species is no sure guideto their degree of mutual sterility, so that simi-lar differences in the case of varieties would beno sure guide, we know that with species thecause lies exclusively in differences in theirsexual constitution. Now the conditions towhich domesticated animals and cultivatedplants have been subjected have had so littletendency towards modifying the reproductivesystem in a manner leading to mutual sterility,that we have very good grounds for admittingthe directly opposite doctrine of Pallas, namely,that such conditions generally eliminate thistendency; so that the domesticated descendantsof species, which in their natural state wouldhave been in some degree sterile when crossed,

become perfectly fertile together. With plants,so far is cultivation from giving a tendency to-wards mutual sterility, that in several well-authenticated cases, already often alluded to,certain species have been affected in a very dif-ferent manner, for they have become self- im-potent, whilst still retaining the capacity of fer-tilising, and being fertilised by, distinct species.If the Pallasian doctrine of the elimination ofsterility through long-continued domesticationbe admitted, and it can hardly be rejected, itbecomes in the highest degree improbable thatsimilar circumstances should commonly bothinduce and eliminate the same tendency;though in certain cases, with species having apeculiar constitution, sterility might occasiona-lly be thus induced. Thus, as I believe, we canunderstand why with domesticated animalsvarieties have not been produced which aremutually sterile; and why with plants only afew such cases have been observed, namely, byGartner, with certain varieties of maize and

verbascum, by other experimentalists with va-rieties of the gourd and melon, and by Kolreu-ter with one kind of tobacco.

With respect to varieties which have originatedin a state of nature, it is almost hopeless to ex-pect to prove by direct evidence that they havebeen rendered mutually sterile; for if even atrace of sterility could be detected, such varie-ties would at once be raised by almost everynaturalist to the rank of distinct species. If, forinstance, Gartner's statement were fully con-firmed, that the blue and red flowered forms ofthe pimpernel (Anagallis arvensis) are sterilewhen crossed, I presume that all the botanistswho now maintain on various grounds thatthese two forms are merely fleeting varieties,would at once admit that they were specificallydistinct.

The real difficulty in our present subject is not,as it appears to me, why domestic varietieshave not become mutually infertile when cros-

sed, but why this has so generally occurredwith natural varieties as soon as they have beenmodified in a sufficient and permanent degreeto take rank as species. We are far from precise-ly knowing the cause; but we can see that thespecies, owing to their struggle for existencewith numerous competitors, must have beenexposed to more uniform conditions of life du-ring long periods of time than domestic varie-ties have been, and this may well make a widedifference in the result. For we know howcommonly wild animals and plants, when ta-ken from their natural conditions and subjectedto captivity, are rendered sterile; and the re-productive functions of organic beings whichhave always lived and been slowly modifiedunder natural conditions would probably inlike manner be eminently sensitive to the in-fluence of an unnatural cross. Domesticatedproductions, on the other hand, which, asshown by the mere fact of their domestication,were not originally highly sensitive to changes

in their conditions of life, and which can nowgenerally resist with undiminished fertilityrepeated changes of conditions, might be ex-pected to produce varieties, which would belittle liable to have their reproductive powersinjuriously affected by the act of crossing withother varieties which had originated in a likemanner.

Certain naturalists have recently laid too greatstress, as it appears to me, on the difference infertility between varieties and species whencrossed. Some allied species of trees cannot begrafted on one another, whilst all varieties canbe so grafted. Some allied animals are affectedin a very different manner by the same poison,but with varieties no such case until recentlywas known; whilst now it has been proved thatimmunity from certain poisons sometimesstands in correlation with the colour of the in-dividuals of the same species. The period ofgestation generally differs much in distinct spe-

cies, but with varieties until lately no such dif-ference had been observed. Here we have va-rious physiological differences, and no doubtothers could be added, between one speciesand another of the same genus, which do notoccur, or occur with extreme rarity, in the caseof varieties; and these differences are apparen-tly wholly or in chief part incidental on otherconstitutional differences, just in the samemanner as the sterility of crossed species is in-cidental on differences confined to the sexualsystem. Why, then, should these latter differen-ces, however serviceable they may indirectly bein keeping the inhabitants of the same countrydistinct, be thought of such paramount impor-tance, in comparison with other incidental andfunctional differences? No sufficient answer tothis question can be given. Hence the fact thatwidely distinct domestic varieties are, with rareexceptions, perfectly fertile when crossed, andproduce fertile offspring, whilst closely alliedspecies are, with rare exceptions, more or less

sterile, is not nearly so formidable an objectionas it appears at first to the theory of the com-mon descent of allied species.

CHAPTER 2.XX.

SELECTION BY MAN.

SELECTION A DIFFICULT ART. MET-HODICAL, UNCONSCIOUS, AND NATURALSELECTION. RESULTS OF METHODICALSELECTION. CARE TAKEN IN SELECTION.SELECTION WITH PLANTS. SELECTIONCARRIED ON BY THE ANCIENTS AND BYSEMI-CIVILISED PEOPLE. UNIMPORTANTCHARACTERS OFTEN ATTENDED TO. UN-CONSCIOUS SELECTION. AS CIRCUMS-TANCES SLOWLY CHANGE, SO HAVE OURDOMESTICATED ANIMALS CHANGEDTHROUGH THE ACTION OF UNCONS-CIOUS SELECTION. INFLUENCE OF DIFFE-

RENT BREEDERS ON THE SAME SUB-VARIETY. PLANTS AS AFFECTED BY UN-CONSCIOUS SELECTION. EFFECTS OF SE-LECTION AS SHOWN BY THE GREATAMOUNT OF DIFFERENCE IN THE PARTSMOST VALUED BY MAN.

The power of Selection, whether exercised byman, or brought into play under naturethrough the struggle for existence and the con-sequent survival of the fittest, absolutely de-pends on the variability of organic beings. Wit-hout variability nothing can be effected; slightindividual differences, however, suffice for thework, and are probably the chief or sole meansin the production of new species. Hence ourdiscussion on the causes and laws of variabilityought in strict order to have preceded the pre-sent subject, as well as inheritance, crossing,etc.; but practically the present arrangementhas been found the most convenient. Man doesnot attempt to cause variability; though heunintentionally effects this by exposing orga-

nisms to new conditions of life, and by crossingbreeds already formed. But variability beinggranted, he works wonders. Unless some de-gree of selection be exercised, the free commin-gling of the individuals of the same varietysoon obliterates, as we have previously seen,the slight differences which arise, and givesuniformity of character to the whole body ofindividuals. In separated districts, long- conti-nued exposure to different conditions of lifemay produce new races without the aid of se-lection; but to this subject of the direct action ofthe conditions of life I shall recur in a futurechapter.

When animals or plants are born with someconspicuous and firmly inherited new charac-ter, selection is reduced to the preservation ofsuch individuals, and to the subsequent pre-vention of crosses; so that nothing more needbe said on the subject. But in the great majorityof cases a new character, or some superiority in

an old character, is at first faintly pronounced,and is not strongly inherited; and then the fulldifficulty of selection is experienced. Indomita-ble patience, the finest powers of discrimina-tion, and sound judgment must be exercisedduring many years. A clearly predeterminedobject must be kept steadily in view. Few menare endowed with all these qualities, especiallywith that of discriminating very slight differen-ces; judgment can be acquired only by longexperience; but if any of these qualities be wan-ting, the labour of a life may be thrown away. Ihave been astonished when celebrated bree-ders, whose skill and judgment have been pro-ved by their success at exhibitions, have shownme their animals, which appeared all alike, andhave assigned their reasons for matching thisand that individual. The importance of thegreat principle of Selection mainly lies in thispower of selecting scarcely appreciable diffe-rences, which nevertheless are found to betransmissible, and which can be accumulated

until the result is made manifest to the eyes ofevery beholder.

The principle of selection may be convenientlydivided into three kinds. METHODICAL SE-LECTION is that which guides a man who sys-tematically endeavours to modify a breed ac-cording to some predetermined standard. UN-CONSCIOUS SELECTION is that which fo-llows from men naturally preserving the mostvalued and destroying the less valued indivi-duals, without any thought of altering thebreed; and undoubtedly this process slowlyworks great changes. Unconscious selectiongraduates into methodical, and only extremecases can be distinctly separated; for he whopreserves a useful or perfect animal will gene-rally breed from it with the hope of gettingoffspring of the same character; but as long ashe has not a predetermined purpose to improvethe breed, he may be said to be selecting un-consciously. (20/1. The term "unconscious se-

lection" has been objected to as a contradiction;but see some excellent observations on thishead by Prof. Huxley ('Nat. Hist. Review' Oc-tober 1864 page 578) who remarks that whenthe wind heaps up sand-dunes it sifts and UN-CONSCIOUSLY SELECTS from the gravel onthe beach grains of sand of equal size.) Lastly,we have NATURAL SELECTION, which im-plies that the individuals which are best fittedfor the complex, and in the course of ageschanging conditions to which they are exposed,generally survive and procreate their kind.With domestic productions, natural selectioncomes to a certain extent into action, indepen-dently of, and even in opposition to, the will ofman.

METHODICAL SELECTION.

What man has effected within recent times inEngland by methodical selection is clearlyshown by our exhibitions of improved quadru-

peds and fancy birds. With respect to cattle,sheep, and pigs, we owe their great improve-ment to a long series of well-known names—Bakewell, Coiling, Ellman, Bates, Jonas Webb,Lords Leicester and Western, Fisher Hobbs,and others. Agricultural writers are unanimouson the power of selection: any number of sta-tements to this effect could be quoted; a fewwill suffice. Youatt, a sagacious and experien-ced observer, writes (20/2. 'On Sheep' 1838page 60.) the principle of selection is "thatwhich enables the agriculturist, not only tomodify the character of his flock, but to changeit altogether." A great breeder of Shorthorns(20/3. Mr. J. Wright on Shorthorn Cattle in'Journal of Royal Agricult. Soc.' volume 7 pages208, 209.) says, "In the anatomy of the shouldermodern breeders have made great improve-ment on the Ketton shorthorns by correctingthe defect in the knuckle or shoulder-joint, andby laying the top of the shoulder more snuglyin the crop, and thereby filling up the hollow

behind it…The eye has its fashion at differentperiods: at one time the eye high and outstan-ding from the head, and at another time thesleepy eye sunk into the head; but these extre-mes have merged into the medium of a full,clear and prominent eye with a placid look."

Again, hear what an excellent judge of pigs(20/4. H.D. Richardson 'On Pigs' 1847 page 44.)says: "The legs should be no longer than just toprevent the animal's belly from trailing on theground. The leg is the least profitable portion ofthe hog, and we therefore require no more of itthan is absolutely necessary for the support ofthe rest." Let any one compare the wild-boarwith any improved breed, and he will see howeffectually the legs have been shortened.

Few persons, except breeders, are aware of thesystematic care taken in selecting animals, andof the necessity of having a clear and almostprophetic vision into futurity. Lord Spencer'sskill and judgment were well known; and he

writes (20/5. 'Journal of Royal Agricult. Soc.'volume 1 page 24.), "It is therefore very desira-ble, before any man commences to breed eithercattle or sheep, that he should make up hismind to the shape and qualities he wishes toobtain, and steadily pursue this object." LordSomerville, in speaking of the marvellous im-provement of the New Leicester sheep, effectedby Bakewell and his successors, says, "It wouldseem as if they had first drawn a perfect form,and then given it life." Youatt (20/6. 'On Sheep'pages 520, 319.) urges the necessity of annuallydrafting each flock, as many animals will cer-tainly degenerate "from the standard of exce-llence which the breeder has established in hisown mind." Even with a bird of such little im-portance as the canary, long ago (1780-1790)rules were established, and a standard of per-fection was fixed according to which the Lon-don fanciers tried to breed the several sub- va-rieties. (20/7. Loudon's 'Mag. of Nat. Hist.' vo-lume 8 1835 page 618.) A great winner of prizes

at the Pigeon-shows (20/8. 'A treatise on theArt of Breeding the Almond Tumbler' 1851 pa-ge 9.), in describing the short-faced AlmondTumbler, says, "There are many first-rate fan-ciers who are particularly partial to what iscalled the goldfinch-beak, which is very beauti-ful; others say, take a full-size round cherrythen take a barleycorn, and judiciously placingand thrusting it into the cherry, form as it wereyour beak; and that is not all, for it will form agood head and beak, provided, as I said before,it is judiciously done; others take an oat; but asI think the goldfinch-beak the handsomest, Iwould advise the inexperienced fancier to getthe head of a goldfinch, and keep it by him forhis observation." Wonderfully different as arethe beaks of the rock pigeon and goldfinch, theend has undoubtedly been nearly gained, as faras external shape and proportions are concer-ned.

Not only should our animals be examined withthe greatest care whilst alive, but, as Andersonremarks (20/9. 'Recreations in Agriculture' vo-lume 2 page 409.) their carcases should be scru-tinised, "so as to breed from the descendants ofsuch only as, in the language of the butcher, cutup well." The "grain of the meat" in cattle, andits being well marbled with fat (20/10. 'Youatton Cattle' pages 191, 227.), and the greater orless accumulation of fat in the abdomen of oursheep, have been attended to with success. Sowith poultry, a writer (20/11. Ferguson 'PrizePoultry' 1854 page 208.), speaking of Cochin-China fowls, which are said to differ much inthe quality of their flesh, says, "the best mode isto purchase two young brother-cocks, kill,dress, and serve up one; if he be indifferent,similarly dispose of the other, and try again; if,however, he be fine and well-flavoured, hisbrother will not be amiss for breeding purposesfor the table."

The great principle of the division of labour hasbeen brought to bear on selection. In certaindistricts (20/12. Wilson in 'Transact. HighlandAgricult. Soc.' quoted in 'Gardener's Chronicle'1844 page 29.) "the breeding of bulls is confinedto a very limited number of persons, who bydevoting their whole attention to this depart-ment, are able from year to year to furnish aclass of bulls which are steadily improving thegeneral breed of the district." The rearing andletting of choice rams has long been, as is wellknown, a chief source of profit to several emi-nent breeders. In parts of Germany this princi-ple is carried with merino sheep to an extremepoint. (20/13. Simmonds quoted in 'Gardener'sChronicle' 1855 page 637. And for the secondquotation see 'Youatt on Sheep' page 171.) So"important is the proper selection of breedinganimals considered, that the best flock-mastersdo not trust to their own judgment or to that oftheir shepherds, but employ persons called'sheep-classifiers' who make it their special bu-

siness to attend to this part of the managementof several flocks, and thus to preserve, or if pos-sible to improve, the best qualities of both pa-rents in the lambs." In Saxony, "when the lambsare weaned, each in his turn is placed upon atable that his wool and form may be minutelyobserved. The finest are selected for breedingand receive a first mark. When they are oneyear old, and prior to shearing them, anotherclose examination of those previously markedtakes place: those in which no defect can befound receive a second mark, and the rest arecondemned. A few months afterwards a thirdand last scrutiny is made; the prime rams andewes receive a third and final mark, but theslightest blemish is sufficient to cause the rejec-tion of the animal." These sheep are bred andvalued almost exclusively for the fineness oftheir wool; and the result corresponds with thelabour bestowed on their selection. Instrumentshave been invented to measure accurately thethickness of the fibres; and "an Austrian fleece

has been produced of which twelve hairs equa-lled in thickness one from a Leicester sheep."

Throughout the world, wherever silk is produ-ced, the greatest care is bestowed on selectingthe cocoons from which the moths for breedingare to be reared. A careful cultivator (20/14.Robinet 'Vers a Soie' 1848 page 271.) likewiseexamines the moths themselves, and destroysthose that are not perfect. But what more im-mediately concerns us is that certain families inFrance devote themselves to raising eggs forsale. (20/15. Quatrefages 'Les Maladies du Vera Soie' 1859 page 101.) In China, near Shanghai,the inhabitants of two small districts have theprivilege of raising eggs for the whole surroun-ding country, and that they may give up theirwhole time to this business, they are interdictedby law from producing silk. (20/16. M. Simonin 'Bull. de la Soc. d'Acclimat.' tome 9 1862 page221.)

The care which successful breeders take in mat-ching their birds is surprising. Sir John Se-bright, whose fame is perpetuated by the "Se-bright Bantam," used to spend "two and threedays in examining, consulting, and disputingwith a friend which were the best of five or sixbirds." (20/17. 'The Poultry Chronicle' volume 11854 page 607.) Mr. Bult, whose pouter-pigeonswon so many prizes, and were exported toNorth America under the charge of a man senton purpose, told me that he always deliberatedfor several days before he matched each pair.Hence we can understand the advice of aneminent fancier, who writes (20/18. J.M. Eaton'A Treatise on Fancy Pigeons' 1852 page 14 and'A Treatise on the Almond Tumbler' 1851 page11.) "I would here particularly guard youagainst having too great a variety of pigeons,otherwise you will know a little of all, but not-hing about one as it ought to be known." Appa-rently it transcends the power of the humanintellect to breed all kinds: "it is possible that

there may be a few fanciers that have a goodgeneral knowledge of fancy pigeons; but thereare many more who labour under the delusionof supposing they know what they do not." Theexcellence of one sub- variety, the AlmondTumbler, lies in the plumage, carriage, head,beak, and eye; but it is too presumptuous in thebeginner to try for all these points. The greatjudge above quoted says, "There are someyoung fanciers who are over-covetous, who gofor all the above five properties at once; theyhave their reward by getting nothing." We thussee that breeding even fancy pigeons is no sim-ple art: we may smile at the solemnity of theseprecepts, but he who laughs will win no prizes.

What methodical selection has effected for ouranimals is sufficiently proved, as already re-marked, by our Exhibitions. So greatly were thesheep belonging to some of the earlier breeders,such as Bakewell and Lord Western, changed,that many persons could not be persuaded that

they had not been crossed. Our pigs, as Mr.Corringham remarks (20/19. 'Journal RoyalAgricultural Soc.' volume 6 page 22.) duringthe last twenty years have undergone, throughrigorous selection together with crossing, acomplete metamorphosis. The first exhibitionfor poultry was held in the Zoological Gardensin 1845; and the improvement effected sincethat time has been great. As Mr. Bailey, thegreat judge, remarked to me, it was formerlyordered that the comb of the Spanish cockshould be upright, and in four or five years allgood birds had upright combs; it was orderedthat the Polish cock should have no comb orwattles, and now a bird thus furnished wouldbe at once disqualified; beards were ordered,and out of fifty-seven pens lately (1860) exhibi-ted at the Crystal Palace, all had beards. So ithas been in many other cases. But in all casesthe judges order only what is occasionally pro-duced and what can be improved and renderedconstant by selection. The steady increase in

weight during the last few years in our fowls,turkeys, ducks, and geese is notorious; "six-pound ducks are now common, whereas fourpounds was formerly the average." As the timerequired to make a change has not often beenrecorded, it may be worth mentioning that ittook Mr. Wicking thirteen years to put a cleanwhite head on an almond tumbler's body, "atriumph," says another fancier, "of which hemay be justly proud." (20/20. 'Poultry Chroni-cle' volume 2 1855 page 596.)

Mr. Tollet, of Betley Hall, selected cows, andespecially bulls, descended from good milkers,for the sole purpose of improving his cattle forthe production of cheese; he steadily tested themilk with the lactometer, and in eight years heincreased, as I was informed by him, the pro-duct in proportion of four to three. Here is acurious case (20/21. Isid. Geoffroy St.-Hilaire'Hist. Nat. Gen.' tome 3 page 254.) of steady butslow progress, with the end not as yet fully

attained: in 1784 a race of silkworms was intro-duced into France, in which one hundred in thethousand failed to produce white cocoons; butnow after careful selection during sixty-fivegenerations, the proportion of yellow cocoonshas been reduced to thirty-five in the thousand.

With plants selection has been followed withthe same good result as with animals. But theprocess is simpler, for plants in the great majo-rity of cases bear both sexes. Nevertheless, withmost kinds it is necessary to take as much careto prevent crosses as with animals or unisexualplants; but with some plants, such as peas, thiscare is not necessary. With all improved plants,excepting of course those which are propagatedby buds, cuttings, etc., it is almost indispensa-ble to examine the seedlings and destroy thosewhich depart from the proper type. This is ca-lled "roguing," and is, in fact, a form of selec-tion, like the rejection of inferior animals. Expe-rienced horticulturists and agriculturists inces-

santly urge every one to preserve the finestplants for the production of seed.

Although plants often present much more cons-picuous variations than animals, yet the closestattention is generally requisite to detect eachslight and favourable change. Mr. Masters rela-tes (20/22. 'Gardener's Chronicle' 1850 page198.) how "many a patient hour was devoted,"whilst he was young, to the detection of diffe-rences in peas intended for seed. Mr. Barnet(20/23. 'Transact. Hort. Soc.' volume 6 page152.) remarks that the old scarlet Americanstrawberry was cultivated for more than a cen-tury without producing a single variety; andanother writer observes how singular it wasthat when gardeners first began to attend tothis fruit it began to vary; the truth no doubtbeing that it had always varied, but that, untilslight variations were selected and propagatedby seed, no conspicuous result was obtained.The finest shades of difference in wheat have

been discriminated and selected with almost asmuch care as, in the case of the higher animals,for instance by Col. Le Couteur and more espe-cially by Major Hallett.

It may be worth while to give a few examplesof methodical selection with plants; but in factthe great improvement of all our anciently cul-tivated plants may be attributed to selectionlong carried on, in part methodically, and inpart unconsciously. I have shown in a formerchapter how the weight of the gooseberry hasbeen increased by systematic selection and cul-ture. The flowers of the Heartsease have beensimilarly increased in size and regularity ofoutline. With the Cineraria, Mr. Glenny (20/24.'Journal of Horticulture' 1862 page 369.) "wasbold enough when the flowers were raggedand starry and ill defined in colour, to fix astandard which was then considered outra-geously high and impossible, and which, evenif reached, it was said, we should be no gainers

by, as it would spoil the beauty of the flowers.He maintained that he was right; and the eventhas proved it to be so." The doubling of flowershas several times been effected by careful selec-tion: the Rev. W. Williamson (20/25 'Transact.Hort. Soc.' volume 4 page 381.), after sowingduring several years seed of Anemone corona-ria, found a plant with one additional petal; hesowed the seed of this, and by perseverance inthe same course obtained several varieties withsix or seven rows of petals. The single Scotchrose was doubled, and yielded eight good va-rieties in nine or ten years. (20/26. 'Transact.Hort. Soc.' volume 4 page 285.) The Canterburybell (Campanula medium) was doubled bycareful selection in four generations. (20/27.Rev. W. Bromehead in 'Gardener's Chronicle'1857 page 550.) In four years Mr. Buckman(20/28. 'Gardener's Chronicle' 1862 page 721.),by culture and careful selection, convertedparsnips, raised from wild seed, into a new andgood variety. By selection during a long course

of years, the early maturity of peas has beenhastened by between ten and twenty-one days.(20/29. Dr. Anderson in 'The Bee' volume 6page 96; Mr. Barnes in 'Gardener's Chronicle'1844 page 476.) A more curious case is offeredby the beet plant, which since its cultivation inFrance, has almost exactly doubled its yield ofsugar. This has been effected by the most care-ful selection; the specific gravity of the rootsbeing regularly tested, and the best roots savedfor the production of seed. (20/30. Godron 'Del'Espece' 1859 tome 2 page 69; 'Gardener'sChronicle' 1854 page 258.)

SELECTION BY ANCIENT AND SEMI-CIVILISED PEOPLE.

In attributing so much importance to the selec-tion of animals and plants, it may be objected,that methodical selection would not have beencarried on during ancient times. A distinguis-hed naturalist considers it as absurd to suppose

that semi-civilised people should have practi-sed selection of any kind. Undoubtedly theprinciple has been systematically acknowled-ged and followed to a far greater extent withinthe last hundred years than at any former pe-riod, and a corresponding result has been gai-ned; but it would be a greater error to suppose,as we shall immediately see, that its importancewas not recognised and acted on during themost ancient times, and by semi- civilised peo-ple. I should premise that many facts now to begiven only show that care was taken in bree-ding; but when this is the case, selection is al-most sure to be practised to a certain extent. Weshall hereafter be enabled better to judge howfar selection, when only occasionally carriedon, by a few of the inhabitants of a country, willslowly produce a great effect.

In a well-known passage in the thirtieth chapterof Genesis, rules are given for influencing, aswas then thought possible, the colour of sheep;

and speckled and dark breeds are spoken of asbeing kept separate. By the time of David thefleece was likened to snow. Youatt (20/31. 'OnSheep' page 18.), who has discussed all the pas-sages in relation to breeding in the Old Testa-ment, concludes that at this early period "someof the best principles of breeding must havebeen steadily and long pursued." It was orde-red, according to Moses, that "Thou shalt not letthy cattle gender with a diverse kind;" but mu-les were purchased (20/32. Volz 'Beitrage zurKulturgeschichte' 1852 s. 47.) so that at this ear-ly period other nations must have crossed thehorse and ass. It is said (20/33. Mitford 'Historyof Greece' volume 1 page 73.) that Erichthonius,some generations before the Trojan war, hadmany brood-mares, "which by his care andjudgment in the choice of stallions produced abreed of horses superior to any in the surroun-ding countries." Homer (Book 5) speaks of Ae-neas' horses as bred from mares which wereput to the steeds of Laomedon. Plato, in his

'Republic' says to Glaucus, "I see that you raiseat your house a great many dogs for the chase.Do you take care about breeding and pairingthem? Among animals of good blood, are therenot always some which are superior to therest?" To which Glaucus answers in the affirma-tive. (20/34. Dr. Dally translated in 'Anthropo-logical Review' May 1864 page 101.) Alexanderthe Great selected the finest Indian cattle tosend to Macedonia to improve the breed.(20/35. Volz 'Beitrage' etc. 1852 s. 80.) Accor-ding to Pliny (20/36 'History of the World'chapter 45.), King Pyrrhus had an especiallyvaluable breed of oxen: and he did not sufferthe bulls and cows to come together till fouryears old, that the breed might not degenerate.Virgil, in his Georgics (lib. 3), gives as strongadvice as any modern agriculturist could do,carefully to select the breeding stock; "to notethe tribe, the lineage, and the sire; whom toreserve for husband of the herd;"—to brand theprogeny;—to select sheep of the purest white,

and to examine if their tongues are swarthy.We have seen that the Romans kept pedigreesof their pigeons, and this would have been asenseless proceeding had not great care beentaken in breeding them. Columella gives detai-led instructions about breeding fowls: "Let thebreeding hens therefore be of a choice colour, arobust body, square-built, full-breasted, withlarge heads, with upright and bright-red co-mbs. Those are believed to be the best bredwhich have five toes." (20/37. 'Gardener'sChronicle' 1848 page 323.) According to Taci-tus, the Celts attended to the races of their do-mestic animals; and Caesar states that they paidhigh prices to merchants for fine imported hor-ses. (20/38. Reynier 'De l'Economie des Celtes'1818 pages 487, 503.) In regard to plants, Virgilspeaks of yearly culling the largest seeds; andCelsus says, "where the corn and crop is butsmall, we must pick out the best ears of corn,and of them lay up our seed separately by it-self." (20/39. Le Couteur on 'Wheat' page 15.)

Coming down the stream of time, we may bebrief. At about the beginning of the ninth cen-tury Charlemagne expressly ordered his offi-cers to take great care of his stallions; and if anyproved bad or old, to forewarn him in goodtime before they were put to the mares. (20/40.Michel 'Des Haras' 1861 page 84.) Even in acountry so little civilised as Ireland during theninth century, it would appear from some an-cient verses (20/41. Sir W. Wilde an 'Essay onUnmanufactured Animal Remains' etc. 1860page 11.), describing a ransom demanded byCormac, that animals from particular places, orhaving a particular character, were valued.Thus it is said,—

Two pigs of the pigs of Mac Lir,A ram and ewe both round and red,I brought with me from Aengus.I brought with me a stallion and a mareFrom the beautiful stud of Manannan,A bull and a white cow from Druim Cain.

Athelstan, in 930, received running-horses as apresent from Germany; and he prohibited theexportation of English horses. King John impor-ted "one hundred chosen stallions from Flan-ders." (20/42. Col. Hamilton Smith 'Nat. Libra-ry' volume 12 Horses, pages 135, 140.) On June16th, 1305, the Prince of Wales wrote to theArchbishop of Canterbury, begging for the loanof any choice stallion, and promising its returnat the end of the season. (20/43. Michel 'DesHaras' page 90.) There are numerous records atancient periods in English history of the impor-tation of choice animals of various kinds, andof foolish laws against their exportation. In thereigns of Henry VII. and VIII. it was orderedthat the magistrates, at Michaelmas, shouldscour the heaths and commons, and destroy allmares beneath a certain size. (20/44. Mr. Baker'History of the Horse' 'Veterinary' volume 13page 423.) Some of our earlier kings passedlaws against the slaughtering rams of any goodbreed before they were seven years old, so that

they might have time to breed. In Spain Cardi-nal Ximenes issued, in 1509, regulations on theSELECTION of good rams for breeding. (20/45.M. l'Abbe Carlier in 'Journal de Physique' vo-lume 24 1784 page 181; this memoir containsmuch information on the ancient selection ofsheep; and is my authority for rams not beingkilled young in England.)

The Emperor Akbar Khan before the year l600is said to have "wonderfully improved" his pi-geons by crossing the breeds; and this necessa-rily implies careful selection. About the sameperiod the Dutch attended with the greatestcare to the breeding of these birds. Belon in1555 says that good managers in France exami-ned the colour of their goslings in order to getgeese of a white colour and better kinds. Mar-kham in 1631 tells the breeder "to elect the lar-gest and goodliest conies," and enters into mi-nute details. Even with respect to seeds ofplants for the flower-garden, Sir J. Hanmer wri-

ting about the year 1660 (20/46. 'Gardener'sChronicle' 1843 page 389.) says, in "choosingseed, the best seed is the most weighty, and ishad from the lustiest and most vigorous stems;"and he then gives rules about leaving only afew flowers on plants for seed; so that evensuch details were attended to in our flower-gardens two hundred years ago. In order toshow that selection has been silently carried onin places where it would not have been expec-ted, I may add that in the middle of the lastcentury, in a remote part of North America, Mr.Cooper improved by careful selection all hisvegetables, "so that they were greatly superiorto those of any other person. When his radis-hes, for instance, are fit for use, he takes ten ortwelve that he most approves, and plants themat least 100 yards from others that blossom atthe same time. In the same manner he treats allhis other plants, varying the circumstances ac-cording to their nature." (20/47. 'Communica-

tions to Board of Agriculture' quoted in Dr.Darwin 'Phytologia' 1800 page 451.)

In the great work on China published in the lastcentury by the Jesuits, and which is chieflycompiled from ancient Chinese encyclopaedias,it is said that with sheep "improving the breedconsists in choosing with particular care thelambs which are destined for propagation, innourishing them well, and in keeping the flocksseparate." The same principles were applied bythe Chinese to various plants and fruit-trees.(20/48. 'Memoire sur les Chinois' 1786 tome 11page 55; tome 5 page 507.) An imperial edictrecommends the choice of seed of remarkablesize; and selection was practised even by impe-rial hands, for it is said that the Ya-mi, or impe-rial rice, was noticed at an ancient period in afield by the Emperor Khang-hi, was saved andcultivated in his garden, and has since becomevaluable from being the only kind which willgrow north of the Great Wall. (20/49. 'Recher-

ches sur l'Agriculture des Chinois' par L.D'Hervey Saint-Denys 1850 page 229. With res-pect to Khang-hi see Huc's 'Chinese Empire'page 311.) Even with flowers, the tree paeony(P. moutan) has been cultivated, according toChinese traditions, for 1400 years; between 200and 300 varieties have been raised, which arecherished like tulips formerly were by theDutch. (20/50. Anderson in 'Linn. Transact.'volume 12 page 253.)

Turning now to semi-civilised people and tosavages: it occurred to me, from what I hadseen of several parts of South America, wherefences do not exist, and where the animals areof little value, that there would be absolutely nocare in breeding or selecting them; and this to alarge extent is true. Roulin (20/51. 'Mem. del'Acad.' (divers savants), tome 6 1835 page333.), however, describes in Columbia a nakedrace of cattle, which are not allowed to increase,on account of their delicate constitution. Ac-

cording to Azara (20/52. 'Des Quadrupedes duParaguay' 1801 tome 2 pages 333, 371.) horsesare often born in Paraguay with curly hair; but,as the natives do not like them, they are destro-yed. On the other hand, Azara states that ahornless bull, born in 1770, was preserved andpropagated its race. I was informed of the exis-tence in Banda Oriental of a breed with rever-sed hair; and the extraordinary niata cattle firstappeared and have since been kept distinct inLa Plata. Hence certain conspicuous variationshave been preserved, and others have beenhabitually destroyed, in these countries, whichare so little favourable for careful selection. Wehave also seen that the inhabitants sometimesintroduce fresh cattle on their estates to preventthe evil effects of close interbreeding. On theother hand, I have heard on reliable authoritythat the Gauchos of the Pampas never take anypains in selecting the best bulls or stallions forbreeding; and this probably accounts for thecattle and horses being remarkably uniform in

character throughout the immense range of theArgentine republic.

Looking to the Old World, in the Sahara Desert"The Touareg is as careful in the selection of hisbreeding Mahari (a fine race of the dromedary)as the Arab is in that of his horse. The pedi-grees are handed down, and many a dromeda-ry can boast a genealogy far longer than thedescendants of the Darley Arabian." (20/53.'The Great Sahara' by the Rev. H.B. Tristram1860 page 238.) According to Pallas the Mongo-lians endeavour to breed the Yaks or horse-tailed buffaloes with white tails, for these aresold to the Chinese mandarins as fly-flappers;and Moorcroft, about seventy years after Pallas,found that white-tailed animals were still selec-ted for breeding. (20/54. Pallas 'Act. Acad. St.Petersburg' 1777 page 249. Moorcroft and Tre-beck 'Travels in the Himalayan Provinces'1841.)

We have seen in the chapter on the Dog thatsavages in different parts of North America andin Guiana cross their dogs with wild Canidae,as did the ancient Gauls, according to Pliny.This was done to give their dogs strength andvigour, in the same way as the keepers in largewarrens now sometimes cross their ferrets (as Ihave been informed by Mr. Yarrell) with thewild polecat, "to give them more devil." Accor-ding to Varro, the wild ass was formerly caughtand crossed with the tame animal to improvethe breed, in the same manner as at the presentday the natives of Java sometimes drive theircattle into the forests to cross with the wildBanteng (Bos sondaicus). (20/55. Quoted fromRaffles in the 'Indian Field' 1859 page 196: forVarro see Pallas ut supra.) In Northern Siberia,among the Ostyaks, the dogs vary in markingsin different districts, but in each place they arespotted black and white in a remarkably uni-form manner (20/56. Erman 'Travels in Siberia'English translation volume 1 page 453.); and

from this fact alone we may infer careful bree-ding, more especially as the dogs of one localityare famed throughout the country for their su-periority. I have heard of certain tribes of Es-quimaux who take pride in their teams of dogsbeing uniformly coloured. In Guiana, as Sir H.Schomburgk informs me (20/57. See also 'Jour-nal of R. Geograph. Soc.' volume 13 part 1 page65.), the dogs of the Turuma Indians are highlyvalued and extensively bartered: the price of agood one is the same as that given for a wife:they are kept in a sort of cage, and the Indians"take great care when the female is in season toprevent her uniting with a dog of an inferiordescription." The Indians told Sir Robert that, ifa dog proved bad or useless, he was not killed,but was left to die from sheer neglect. Hardlyany nation is more barbarous than the Fue-gians, but I hear from Mr. Bridges, the Cate-chist to the Mission, that, "when these savageshave a large, strong, and active bitch, they takecare to put her to a fine dog, and even take care

to feed her well, that her young may be strongand well favoured."

In the interior of Africa, negroes, who have notassociated with white men, show great anxietyto improve their animals; they "always choosethe larger and stronger males for stock;" theMalakolo were much pleased at Livingstone'spromise to send them a bull, and some Bakalo-lo carried a live cock all the way from Loandainto the interior. (20/58. Livingstone 'First Tra-vels' pages 191, 439, 565; see also 'Expedition tothe Zambesi' 1865 page 495, for an analogouscase respecting a good breed of goats.) At Fala-ba Mr. Winwood Reade noticed an unusuallyfine horse, and the negro King informed himthat "the owner was noted for his skill in bree-ding horses." Further south on the same conti-nent, Andersson states that he has known aDamara give two fine oxen for a dog whichstruck his fancy. The Damaras take great de-light in having whole droves of cattle of the

same colour, and they prize their oxen in pro-portion to the size of their horns. "The Nama-quas have a perfect mania for a uniform team;and almost all the people of Southern Africavalue their cattle next to their women, and takea pride in possessing animals that look high-bred. They rarely or never make use of a hand-some animal as a beast of burden." (20/59. An-dersson 'Travels in South Africa' pages 232, 318,319.) The power of discrimination which thesesavages possess is wonderful, and they canrecognise to which tribe any cattle belong. Mr.Andersson further informs me that the nativesfrequently match a particular bull with a parti-cular cow.

The most curious case of selection by semi-civilised people, or indeed by any people,which I have found recorded, is that given byGarcilazo de la Vega, a descendant of the Incas,as having been practised in Peru before thecountry was subjugated by the Spaniards.

(20/60. Dr. Vavasseur in 'Bull. de La Soc. d'Ac-climat.' tome 8 1861 page 136.) The Incas an-nually held great hunts, when all the wild ani-mals were driven from an immense circuit to acentral point. The beasts of prey were first des-troyed as injurious. The wild Guanacos andVicunas were sheared; the old males and fema-les killed, and the others set at liberty. The va-rious kinds of deer were examined; the old ma-les and females were likewise killed, "but theyoung females, with a certain number of males,selected from the most beautiful and strong,"were given their freedom. Here, then, we haveselection by man aiding natural selection. Sothat the Incas followed exactly the reverse sys-tem of that which our Scottish sportsman areaccused of following, namely, of steadily killingthe finest stags, thus causing the whole race todegenerate. (20/61. 'The Natural History of DeeSide' 1855 page 476.) In regard to the domesti-cated llamas and alpacas, they were separatedin the time of the Incas according to colour: and

if by chance one in a flock was born of thewrong colour, it was eventually put into anot-her flock.

In the genus Auchenia there are four forms,—the Guanaco and Vicuna, found wild and un-doubtedly distinct species; the Llama and Al-paca, known only in a domesticated condition.These four animals appear so different, thatmost naturalists, especially those who havestudied these animals in their native country,maintain that they are specifically distinct,notwithstanding that no one pretends to haveseen a wild llama or alpaca. Mr. Ledger, howe-ver, who has closely studied these animals bothin Peru and during their exportation to Austra-lia, and who has made many experiments ontheir propagation, adduces arguments (20/62.'Bull. de la Soc. d'Acclimat.' tome 7 1860 page457.) which seem to me conclusive, that thellama is the domesticated descendant of theguanaco, and the alpaca of the vicuna. And

now that we know that these animals were sys-tematically bred and selected many centuriesago, there is nothing surprising in the greatamount of change which they have undergone.

It appeared to me at one time probable that,though ancient and semi-civilised people mighthave attended to the improvement of their mo-re useful animals in essential points, yet thatthey would have disregarded unimportant cha-racters. But human nature is the same throug-hout the world: fashion everywhere reigns su-preme, and man is apt to value whatever hemay chance to possess. We have seen that inSouth America the niata cattle, which certainlyare not made useful by their shortened facesand upturned nostrils, have been preserved.The Damaras of South Africa value their cattlefor uniformity of colour and enormously longhorns. And I will now show that there is hardlyany peculiarity in our most useful animalswhich, from fashion, superstition, or some ot-

her motive, has not been valued, and conse-quently preserved. With respect to cattle, "anearly record," according to Youatt (20/63. 'Cat-tle' page 48.) "speaks of a hundred white cowswith red ears being demanded as a compensa-tion by the princes of North and South Wales. Ifthe cattle were of a dark or black colour, 150were to be presented." So that colour was at-tended to in Wales before its subjugation byEngland. In Central Africa, an ox that beats theground with its tail is killed; and in South Afri-ca some of the Damaras will not eat the flesh ofa spotted ox. The Kaffirs value an animal with amusical voice; and "at a sale in British Kaffrariathe low "of a heifer excited so much admirationthat a sharp competition sprung up for her pos-session, and she realised a considerable price."(20/64. Livingstone 'Travels' page 576; Anders-son 'Lake Ngami' 1856 page 222. With respectto the sale in Kaffraria see 'Quarterly Review'1860 page 139.) With respect to sheep, the Chi-nese prefer rams without horns; the Tartars

prefer them with spirally wound horns, becau-se the hornless are thought to lose courage.(20/65. 'Memoire sur les Chinois' by the Jesuits1786 tome 11 page 57.) Some of the Damaraswill not eat the flesh of hornless sheep. In re-gard to horses, at the end of the fifteenth centu-ry animals of the colour described as liartpomme were most valued in France. The Arabshave a proverb, "Never buy a horse with fourwhite feet, for he carries his shroud with him"(20/66. F. Michel 'Des Haras' pages 47, 50.); theArabs also, as we have seen, despise dun- co-loured horses. So with dogs, Xenophon andothers at an ancient period were prejudiced infavour of certain colours; and "white or slate-coloured hunting dogs were not esteemed."(20/67. Col. Hamilton Smith 'Dogs' in 'Nat.Lib.' volume 10 page 103.)

Turning to poultry, the old Roman gourmandsthought that the liver of a white goose was themost savoury. In Paraguay black-skinned fowls

are kept because they are thought to be moreproductive, and their flesh the most proper forinvalids. (20/68. Azara 'Quadrupedes du Para-guay' tome 2 page 324.) In Guiana, as I am in-formed by Sir R. Schomburgk, the aborigineswill not eat the flesh or eggs of the fowl, buttwo races are kept distinct merely for orna-ment. In the Philippines, no less than nine sub-varieties of the game-cock are kept and named,so that they must be separately bred.

At the present time in Europe, the smallest pe-culiarities are carefully attended to in our mostuseful animals, either from fashion, or as amark of purity of blood. Many examples couldbe given; two will suffice. "In the Westerncounties of England the prejudice against awhite pig is nearly as strong as against a blackone in Yorkshire." In one of the Berkshire sub-breeds, it is said, "the white should be confinedto four white feet, a white spot between theeyes, and a few white hairs behind each shoul-

der." Mr. Saddler possessed three hundredpigs, every one of which was marked in thismanner." (20/69. Sidney's edition of Youatt1860 pages 24, 25.) Marshall, towards the closeof the last century, in speaking of a change inone of the Yorkshire breeds of cattle, says thehorns have been considerably modified, as "aclean, small, sharp horn has been FASHIONA-BLE for the last twenty years." (20/70. 'RuralEconomy of Yorkshire' volume 2 page 182.) In apart of Germany the cattle of the Race deGfoehl are valued for many good qualities, butthey must have horns of a particular curvatureand tint, so much so that mechanical means areapplied if they take a wrong direction; but theinhabitants "consider it of the highest impor-tance that the nostrils of the bull should beflesh-coloured, and the eyelashes light; this isan indispensable condition. A calf with bluenostrils would not be purchased, or purchasedat a very low price." (20/71. Moll et Gayot 'DuBoeuf' 1860 page 547.) Therefore let no man say

that any point or character is too trifling to bemethodically attended to and selected by bree-ders.

UNCONSCIOUS SELECTION.

By this term I mean, as already more than onceexplained, the preservation by man of the mostvalued, and the destruction of the least valuedindividuals, without any conscious intention onhis part of altering the breed. It is difficult tooffer direct proofs of the results which followfrom this kind of selection; but the indirect evi-dence is abundant. In fact, except that in theone case man acts intentionally, and in the ot-her unintentionally, there is little differencebetween methodical and unconscious selection.In both cases man preserves the animals whichare most useful or pleasing to him, and des-troys or neglects the others. But no doubt a farmore rapid result follows from methodical thanfrom unconscious selection. The "roguing" of

plants by gardeners, and the destruction by lawin Henry VIII.'s reign of all under- sized mares,are instances of a process the reverse of selec-tion in the ordinary sense of the word, but lea-ding to the same general result. The influenceof the destruction of individuals having a parti-cular character is well shown by the necessityof killing every lamb with a trace of black aboutit, in order to keep the flock white; or again, bythe effects on the average height of the men ofFrance of the destructive wars of Napoleon, bywhich many tall men were killed, the shortones being left to be the fathers of families. Thisat least is the conclusion of some of those whohave closely studied the effects of the conscrip-tion; and it is certain that since Napoleon's timethe standard for the army has been loweredtwo or three times.

Unconscious selection blends with methodical,so that it is scarcely possible to separate them.When a fancier long ago first happened to noti-

ce a pigeon with an unusually short beak, orone with the tail-feathers unusually developed,although he bred from these birds with the dis-tinct intention of propagating the variety, yethe could not have intended to make a short-faced tumbler or a fantail, and was far fromknowing that he had made the first step to-wards this end. If he could have seen the finalresult, he would have been struck with asto-nishment, but, from what we know of thehabits of fanciers, probably not with admira-tion. Our English carriers, barbs, and short-faced tumblers have been greatly modified inthe same manner, as we may infer both fromthe historical evidence given in the chapters onthe Pigeon, and from the comparison of birdsbrought from distant countries.

So it has been with dogs; our present fox-hounds differ from the old English hound; ourgreyhounds have become lighter: the Scotchdeer-hound has been modified, and is now ra-

re. Our bulldogs differ from those which wereformerly used for baiting bulls. Our pointersand Newfoundlands do not closely resembleany native dog now found in the countrieswhence they were brought. These changes havebeen effected partly by crosses; but in everycase the result has been governed by the stric-test selection. Nevertheless, there is no reasonto suppose that man intentionally and met-hodically made the breeds exactly what theynow are. As our horses became fleeter, and thecountry more cultivated and smoother, fleeterfox-hounds were desired and produced, butprobably without any one distinctly foreseeingwhat they would become. Our pointers andsetters, the latter almost certainly descendedfrom large spaniels, have been greatly modifiedin accordance with fashion and the desire forincreased speed. Wolves have become extinct,and so has the wolf-dog; deer have becomerarer, bulls are no longer baited, and the co-rresponding breeds of the dog have answered

to the change. But we may feel almost sure thatwhen, for instance, bulls were no longer baited,no man said to himself, I will now breed mydogs of smaller size, and thus create the presentrace. As circumstances changed, men uncons-ciously and slowly modified their course ofselection.

With racehorses selection for swiftness hasbeen followed methodically, and our horsesnow easily surpass their progenitors. The in-creased size and different appearance of theEnglish racehorse led a good observer in Indiato ask," Could any one in this year of 1856, loo-king at our racehorses, conceive that they werethe result of the union of the Arab horse andthe African mare?" (20/72. 'The India SportingReview' volume 2 page 181; 'The Stud Farm' byCecil page 58.) This change has, it is probable,been largely effected through unconscious se-lection, that is, by the general wish to breed asfine horses as possible in each generation, com-

bined with training and high feeding, but wit-hout any intention to give to them their presentappearance. According to Youatt (20/73. 'TheHorse' page 22.), the introduction in OliverCromwell's time of three celebrated Easternstallions speedily affected the English breed;"so that Lord Harleigh, one of the old school,complained that the great horse was fast disap-pearing." This is an excellent proof how carefu-lly selection must have been attended to; forwithout such care, all traces of so small an infu-sion of Eastern blood would soon have beenabsorbed and lost. Notwithstanding that theclimate of England has never been esteemedparticularly favourable to the horse, yet long-continued selection, both methodical and un-conscious, together with that practised by theArabs during a still longer and earlier period,has ended in giving us the best breed of horsesin the world. Macaulay (20/74. 'History of En-gland' volume 1 page 316.) remarks, "Two menwhose authority on such subjects was held in

great esteem, the Duke of Newcastle and SirJohn Fenwick, pronounced that the meanesthack ever imported from Tangier would pro-duce a finer progeny than could be expectedfrom the best sire of our native breed. Theywould not readily have believed that a timewould come when the princes and nobles ofneighbouring lands would be as eager to obtainhorses from England as ever the English hadbeen to obtain horses from Barbary."

The London dray-horse, which differs so muchin appearance from any natural species, andwhich from its size has so astonished manyEastern princes, was probably formed by theheaviest and most powerful animals havingbeen selected during many generations inFlanders and England, but without the leastintention or expectation of creating a horsesuch as we now see. If we go back to an earlyperiod of history, we behold in the antiqueGreek statues, as Schaaffhausen has remarked

(20/75. 'Ueber Bestandigkeit der Arten.'), a hor-se equally unlike a race or dray horse, and dif-fering from any existing breed.

The results of unconscious selection, in an earlystage, are well shown in the difference betweenthe flocks descended from the same stock, butseparately reared by careful breeders. Youattgives an excellent instance of this fact in thesheep belonging to Messrs. Buckley and Bur-gess, which "have been purely bred from theoriginal stock of Mr. Bakewell for upwards offifty years. There is not a suspicion existing inthe mind of any one at all acquainted with thesubject that the owner of either flock has devia-ted in any one instance from the pure blood ofMr. Bakewell's flock; yet the difference betweenthe sheep possessed by these two gentlemen isso great, that they have the appearance of beingquite different varieties." (20/76. 'Youatt onSheep' page 315.) I have seen several analogousand well marked cases with pigeons: for ins-

tance, I had a family of barbs descended fromthose long bred by Sir J. Sebright, and anotherfamily long bred by another fancier, and thetwo families plainly differed from each other.Nathusius—and a more competent witnesscould not be cited—observes that, though theShorthorns are remarkably uniform in appea-rance (except in colour), yet the individual cha-racter and wishes of each breeder become im-pressed on his cattle, so that different herdsdiffer slightly from one another. (20/77. 'UeberShorthorn Rindvieh' 1857 s. 51.) The Herefordcattle assumed their present well-marked cha-racter soon after the year 1769, through carefulselection by Mr. Tomkins (20/78. Low 'Domes-ticated Animals' 1845 page 363.) and the breedhas lately split into two strains—one strainhaving a white face, and differing slightly, it issaid (20/79. 'Quarterly Review' 1849 page 392.),in some other points: but there is no reason tobelieve that this split, the origin of which isunknown, was intentionally made; it may with

much more probability be attributed to diffe-rent breeders having attended to differentpoints. So again, the Berkshire breed of swinein the year 1810 had greatly changed from whatit was in 1780; and since 1810 at least two dis-tinct sub-breeds have arisen bearing the samename. (20/80. H. von Nathusius 'Vorstu-dien…Schweineschadel' 1864 s 140.) Keeping inmind how rapidly all animals increase, and thatsome must be annually slaughtered and somesaved for breeding, then, if the same breederduring a long course of years deliberately set-tles which shall be saved and which shall bekilled, it is almost inevitable that his individualturn of mind will influence the character of hisstock, without his having had any intention tomodify the breed.

Unconscious selection in the strictest sense ofthe word, that is, the saving of the more usefulanimals and the neglect or slaughter of the lessuseful, without any thought of the future, must

have gone on occasionally from the remotestperiod and amongst the most barbarous na-tions. Savages often suffer from famines, andare sometimes expelled by war from their ownhomes. In such cases it can hardly be doubtedthat they would save their most useful animals.When the Fuegians are hard pressed by want,they kill their old women for food rather thantheir dogs; for, as we were assured, "old wo-men no use—dogs catch otters." The samesound sense would surely lead them to preser-ve their more useful dogs when still harderpressed by famine. Mr. Oldfield, who has seenso much of the aborigines of Australia, informsme that "they are all very glad to get a Euro-pean kangaroo dog, and several instances havebeen known of the father killing his own infantthat the mother might suckle the much-prizedpuppy." Different kinds of dogs would be use-ful to the Australian for hunting opossums andkangaroos, and to the Fuegian for catching fishand otters; and the occasional preservation in

the two countries of the most useful animalswould ultimately lead to the formation of twowidely distinct breeds.

With plants, from the earliest dawn of civilisa-tion, the best variety which was known wouldgenerally have been cultivated at each periodand its seeds occasionally sown; so that therewill have been some selection from an extreme-ly remote period, but without any prefixedstandard of excellence or thought of the future.We at the present day profit by a course of se-lection occasionally and unconsciously carriedon during thousands of years. This is proved inan interesting manner by Oswald Heer's re-searches on the lake- inhabitants of Switzer-land, as given in a former chapter; for he showsthat the grain and seed of our present varietiesof wheat, barley, oats, peas, beans, lentils, andpoppy, exceed in size those which were cultiva-ted in Switzerland during the Neolithic andBronze periods. These ancient people, during

the Neolithic period, possessed also a crab con-siderably larger than that now growing wild onthe Jura. (20/81. See also Dr. Christ in Rutime-yer's 'Pfahlbauten' 1861 s. 226.) The pears des-cribed by Pliny were evidently extremely infe-rior in quality to our present pears. We can re-alise the effects of long-continued selection andcultivation in another way, for would any onein his senses expect to raise a first-rate applefrom the seed of a truly wild crab, or a lusciousmelting pear from the wild pear? Alphonse deCandolle informs me that he has lately seen onan ancient mosaic at Rome a representation ofthe melon; and as the Rotnans, who were suchgourmands, are silent on this fruit, he infersthat the melon has been greatly amelioratedsince the classical period.

Coming to later times, Buffon (20/82. The pas-sage is given 'Bull. Soc. d'Acclimat.' 1858 page11.) on comparing the flowers, fruit, and vege-tables which were then cultivated with some

excellent drawings made a hundred and fiftyyears previously, was struck with surprise atthe great improvement which had been effec-ted; and remarks that these ancient flowers andvegetables would now be rejected, not only bya florist but by a village gardener. Since thetime of Buffon the work of improvement hassteadily and rapidly gone on. Every florist whocompares our present flowers with those figu-red in books published not long since, is asto-nished at the change. A well-known amateur(20/83. 'Journal of Horticulture' 1862 page394.), in speaking of the varieties of Pelargo-nium raised by Mr. Garth only twenty-twoyears before, remarks, "What a rage they exci-ted: surely we had attained perfection, it wassaid; and now not one of the flowers of thosedays will be looked at. But none the less is thedebt of gratitude which we owe to those whosaw what was to be done, and did it." Mr. Paul,the well-known horticulturist, in writing of thesame flower (20/84. 'Gardener's Chronicle' 1857

page 85.), says he remembers when youngbeing delighted with the portraits in Sweet'swork; "but what are they in point of beautycompared with the Pelargoniums of this day?Here again nature did not advance by leaps;the improvement was gradual, and if we hadneglected those very gradual advances, wemust have foregone the present grand results."How well this practical horticulturist apprecia-tes and illustrates the gradual and accumulati-ve force of selection! The Dahlia has advancedin beauty in a like manner; the line of impro-vement being guided by fashion, and by thesuccessive modifications which the flowerslowly underwent. (20/85. See Mr. Wildman'saddress to the Floricult. Soc. in 'Gardener'sChronicle' 1843 page 86.) A steady and gradualchange has been noticed in many other flowers:thus an old florist (20/86. 'Journal of Horticul-ture' October 24, 1865 page 239.), after descri-bing the leading varieties of the Pink whichwere grown in 1813 adds, "the pinks of those

days would now be scarcely grown as border-flowers." The improvement of so many flowersand the number of the varieties which havebeen raised is all the more striking when wehear that the earliest known flower-garden inEurope, namely at Padua, dates only from theyear 1545. (20/87. Prescott 'Hist. of Mexico'volume 2 page 61.)

EFFECTS OF SELECTION, AS SHOWN BYTHE PARTS MOST VALUED BY MAN PRE-SENTING THE GREATEST AMOUNT OFDIFFERENCE.

The power of long-continued selection, whet-her methodical or unconscious, or both combi-ned, is well shown in a general way, namely,by the comparison of the differences betweenthe varieties of distinct species, which are va-lued for different parts, such as for the leaves,or stems, or tubers, the seed, or fruit, or flo-wers. Whatever part man values most, that part

will be found to present the greatest amount ofdifference. With trees cultivated for their fruit,Sageret remarks that the fruit is larger than inthe parent-species, whilst with those cultivatedfor the seed, as with nuts, walnuts, almonds,chestnuts, etc., it is the seed itself which is lar-ger; and he accounts for this fact by the fruit inthe one case, and by the seed in the other,having been carefully attended to and selectedduring many ages. Gallesio has made the sameobservation. Godron insists on the diversity ofthe tuber in the potato, of the bulb in the onion,and of the fruit in the melon; and on the closesimilarity of the other parts in these sameplants. (20/88. Sagaret 'Pomologie Physiologi-que' 1830 page 47; Gallesio 'Teoria della Ripro-duzione' 1816 page 88; Godron 'De l'Espece'1859 tome 2 pages 63, 67, 70. In my tenth andeleventh chapters I have given details on thepotato; and I can confirm similar remarks withrespect to the onion. I have also shown how far

Naudin concurs in regard to the varieties of themelon.)

In order to judge how far my own impressionon this subject was correct, I cultivated nume-rous varieties of the same species close to oneanother. The comparison of the amount of dif-ference between widely different organs is ne-cessarily vague; I will therefore give the resultsin only a few cases. We have previously seen inthe ninth chapter how greatly the varieties ofthe cabbage differ in their foliage and stems,which are the selected parts, and how closelythey resemble one another in their flowers, cap-sules, and seeds. In seven varieties of the ra-dish, the roots differed greatly in colour andshape, but no difference whatever could bedetected in their foliage, flowers, or seeds. Nowwhat a contrast is presented, if we compare theflowers of the varieties of these two plants withthose of any species cultivated in our flower-gardens for ornament; or if we compare their

seeds with those of the varieties of maize, peas,beans, etc., which are valued and cultivated fortheir seeds. In the ninth chapter it was shownthat the varieties of the pea differ but little ex-cept in the tallness of the plant, moderately inthe shape of the pod, and greatly in the peaitself, and these are all selected points. The va-rieties, however, of the Pois sans parchemindiffer much more in their pods, and these areeaten and valued. I cultivated twelve varietiesof the common bean; one alone, the Dwarf Fan,differed considerably in general appearance;two differed in the colour of their flowers, onebeing an albino, and the other being whollyinstead of partially purple; several differedconsiderably in the shape and size of the pod,but far more in the bean itself, and this is thevalued and selected part. Toker's bean, for ins-tance, is twice-and-a-half as long and broad asthe horse-bean, and is much thinner and of adifferent shape.

The varieties of the gooseberry, as formerlydescribed, differ much in their fruit, but hardlyperceptibly in their flowers or organs of vegeta-tion. With the plum, the differences likewiseappear to be greater in the fruit than in the flo-wers or leaves. On the other hand, the seed ofthe strawberry, which corresponds with thefruit of the plum, differs hardly at all; whilstevery one knows how greatly the fruit—that is,the enlarged receptacle—differs in several va-rieties. In apples, pears, and peaches the flo-wers and leaves differ considerably, but not, asfar as I can judge, in proportion with the fruit.The Chinese double-flowering peaches, on theother hand, show that varieties of this tree havebeen formed, which differ more in flower thanin fruit. If, as is highly probable, the peach isthe modified descent of the almond, a surpri-sing amount of change has been effected in thesame species, in the fleshy covering of the for-mer and in the kernels of the latter.

When parts stand in close relationship to eachother, such as the seed and the fleshy coveringof the fruit (whatever its homological naturemay be), changes in the one are usually accom-panied by modifications in the other, thoughnot necessarily to the same degree. With theplum-tree, for instance, some varieties produceplums which are nearly alike, but include sto-nes extremely dissimilar in shape; whilst con-versely other varieties produce dissimilar fruitwith barely distinguishable stones; and genera-lly the stones, though they have never beensubjected to selection, differ greatly in the seve-ral varieties of the plum. In other cases organswhich are not manifestly related, through someunknown bond vary together, and are conse-quently liable, without any intention on man'spart, to be simultaneously acted on by selec-tion. Thus the varieties of the stock (Matthiola)have been selected solely for the beauty of theirflowers, but the seeds differ greatly in colourand somewhat in size. Varieties of the lettuce

have been selected solely on account of theirleaves, yet produce seeds which likewise differin colour. Generally, through the law of correla-tion, when a variety differs greatly from its fe-llow-varieties in any one character, it differs toa certain extent in several other characters. Iobserved this fact when I cultivated togethermany varieties of the same species, for I usedfirst to make a list of the varieties which diffe-red most from each other in their foliage andmanner of growth, afterwards of those thatdiffered most in their flowers, then in theirseed-capsules, and lastly in their mature seed;and I found that the same names generally oc-curred in two, three, or four of the successivelists. Nevertheless the greatest amount of diffe-rence between the varieties was always exhibi-ted, as far as I could judge, by that part or or-gan for which the plant was cultivated.

When we bear in mind that each plant was atfirst cultivated because useful to man, and that

its variation was a subsequent, often a longsubsequent, event, we cannot explain the grea-ter amount of diversity in the valuable parts bysupposing that species endowed with an espe-cial tendency to vary in any particular mannerwere originally chosen. We must attribute theresult to the variations in these parts havingbeen successively preserved, and thus conti-nually augmented; whilst other variations, ex-cepting such as inevitably appeared throughcorrelation, were neglected and lost. We maytherefore infer that most plants might be made,through long-continued selection, to yield racesas different from one another in any characteras they now are in those parts for which theyare valued and cultivated.

With animals we see nothing of the same kind;but a sufficient number of species have notbeen domesticated for a fair comparison. Sheepare valued for their wool, and the wool differsmuch more in the several races than the hair in

cattle. Neither sheep, goats, European cattle,nor pigs are valued for their fleetness orstrength; and we do not possess breeds diffe-ring in these respects like the racehorse anddray-horse. But fleetness and strength are va-lued in camels and dogs; and we have with theformer the swift dromedary and heavy camel;with the latter the greyhound and mastiff. Butdogs are valued even in a higher degree fortheir mental qualities and senses; and every oneknows how greatly the races differ in these res-pects. On the other hand, where the dog is keptsolely to serve for food, as in the Polynesianislands and China, it is described as an extre-mely stupid animal. (20/89. Godron 'De l'Espe-ce' tome 2 page 27.) Blumenbach remarks that"many dogs, such as the badger-dog, have abuild so marked and so appropriate for particu-lar purposes, that I should find it very difficultto persuade myself that this astonishing figurewas an accidental consequence of degenera-tion." (20/90. 'The Anthropological Treatises of

Blumenbach' 1856 page 292.) Had Blumenbachreflected on the great principle of selection, hewould not have used the term degeneration,and he would not have been astonished thatdogs and other animals should become exce-llently adapted for the service of man.

On the whole we may conclude that whateverpart or character is most valued— whether theleaves, stems, tubers, bulbs, flowers, fruit, orseed of plants, or the size, strength, fleetness,hairy covering, or intellect of animals— thatcharacter will almost invariably be found topresent the greatest amount of difference bothin kind and degree. And this result may be sa-fely attributed to man having preserved duringa long course of generations the variationswhich were useful to him, and neglected theothers.

I will conclude this chapter by some remarkson an important subject. With animals such asthe giraffe, of which the whole structure is ad-

mirably co- ordinated for certain purposes, ithas been supposed that all the parts must havebeen simultaneously modified; and it has beenargued that, on the principle of natural selec-tion, this is scarcely possible. But in thus ar-guing, it has been tacitly assumed that the va-riations must have been abrupt and great. Nodoubt, if the neck of a ruminant were suddenlyto become greatly elongated, the fore limbs andback would have to be simultaneously strengt-hened and modified; but it cannot be deniedthat an animal might have its neck, or head, ortongue, or fore-limbs elongated a very littlewithout any corresponding modification inother parts of the body; and animals thus sligh-tly modified would, during a dearth, have aslight advantage, and be enabled to browse onhigher twigs, and thus survive. A few mouth-fuls more or less every day would make all thedifference between life and death. By the repe-tition of the same process, and by the occasio-nal intercrossing of the survivors, there would

be some progress, slow and fluctuating thoughit would be, towards the admirably coordina-ted structure of the giraffe. If the short- facedtumbler-pigeon, with its small conical beak,globular head, rounded body, short wings, andsmall feet—characters which appear all in har-mony—had been a natural species, its wholestructure would have been viewed as well fit-ted for its life; but in this case we know thatinexperienced breeders are urged to attend topoint after point, and not to attempt improvingthe whole structure at the same time. Look atthe greyhound, that perfect image of grace,symmetry, and vigour; no natural species canboast of a more admirably co-ordinated struc-ture, with its tapering head, slim body, deepchest, tucked- up abdomen, rat-like tail, andlong muscular limbs, all adapted for extremefleetness, and for running down weak prey.Now, from what we see of the variability ofanimals, and from what we know of the met-hod which different men follow in improving

their stock—some chiefly attending to onepoint, others to another point, others again co-rrecting defects by crosses, and so forth—wemay feel assured that if we could see the longline of ancestors of a first-rate greyhound up toits wild wolf-like progenitor, we should beholdan infinite number of the finest gradations, so-metimes in one character and sometimes inanother, but all leading towards our presentperfect type. By small and doubtful steps suchas these, nature, as we may confidently believe,has progressed, on her grand march of impro-vement and development.

A similar line of reasoning is as applicable toseparate organs as to the whole organisation. Awriter (20/91. Mr. J.J. Murphy in his openingaddress to the Belfast Nat. Hist. Soc. as given inthe 'Belfast Northern Whig' November 19, 1866.Mr. Murphy here follows the line of argumentagainst my views previously and more cau-tiously given by the Rev. C. Pritchard, Pres.

Royal Astronomical Soc., in his sermon Ap-pendix page 33 preached before the British As-sociation at Nottingham 1866.) has recentlymaintained that "it is probably no exaggerationto suppose that in order to improve such anorgan as the eye at all, it must be improved inten different ways at once. And the improbabi-lity of any complex organ being produced andbrought to perfection in any such way is animprobability of the same kind and degree asthat of producing a poem or a mathematicaldemonstration by throwing letters at randomon a table." If the eye were abruptly and greatlymodified, no doubt many parts would have tobe simultaneously altered, in order that theorgan should remain serviceable.

But is this the case with smaller changes? Thereare persons who can see distinctly only in adull light, and this condition depends, I believe,on the abnormal sensitiveness of the retina, andis known to be inherited. Now if a bird, for ins-

tance, receive some great advantage fromseeing well in the twilight, all the individualswith the most sensitive retina would succeedbest and be the most likely to survive; and whyshould not all those which happened to havethe eye itself a little larger, or the pupil capableof greater dilatation, be likewise preserved,whether or not these modifications were strictlysimultaneous? These individuals would subse-quently intercross and blend their respectiveadvantages. By such slight successive changes,the eye of a diurnal bird would be brought intothe condition of that of an owl, which has oftenbeen advanced as an excellent instance of adap-tation. Short- sight, which is often inherited,permits a person to see distinctly a minute ob-ject at so near a distance that it would be indis-tinct to ordinary eyes; and here we have a ca-pacity which might be serviceable under cer-tain conditions, abruptly gained. The Fuegianson board the Beagle could certainly see distantobjects more distinctly than our sailors with all

their long practice; I do not know whether thisdepends upon sensitiveness or on the power ofadjustment in the focus; but this capacity fordistant vision might, it is probable, be slightlyaugmented by successive modifications of eit-her kind. Amphibious animals which are ena-bled to see both in the water and in the air, re-quire and possess, as M. Plateau has shown(20/92. On the Vision of Fishes and Amphibia,translated in 'Annals and Mag. of Nat. Hist.'volume 18 1866 page 469.), eyes constructed onthe following plan: "the cornea is always flat, orat least much flattened in the front of the crys-talline and over a space equal to the diameter ofthat lens, whilst the lateral portions may bemuch curved." The crystalline is very nearly asphere, and the humours have nearly the samedensity as water. Now as a terrestrial animalbecame more and more aquatic in its habits,very slight changes, first in the curvature of thecornea or crystalline, and then in the density ofthe humours, or conversely, might successively

occur, and would be advantageous to the ani-mal whilst under water, without serious detri-ment to its power of vision in the air. It is ofcourse impossible to conjecture by what stepsthe fundamental structure of the eye in the Ver-tebrata was originally acquired, for we knownothing about this organ in the first progenitorsof the class. With respect to the lowest animalsin the scale, the transitional states throughwhich the eye at first probably passed, can bythe aid of analogy be indicated, as I have at-tempted to show in my 'Origin of Species.'(20/93. Sixth edition 1872 page 144.)

CHAPTER 2.XXI.SELECTION, continued.

NATURAL SELECTION AS AFFECTINGDOMESTIC PRODUCTIONS. CHARACTERSWHICH APPEAR OF TRIFLING VALUE OF-TEN OF REAL IMPORTANCE. CIRCUMS-

TANCES FAVOURABLE TO SELECTION BYMAN. FACILITY IN PREVENTING CROSSES,AND THE NATURE OF THE CONDITIONS.CLOSE ATTENTION AND PERSEVERANCEINDISPENSABLE. THE PRODUCTION OF ALARGE NUMBER OF INDIVIDUALS ESPE-CIALLY FAVOURABLE. WHEN NO SELEC-TION IS APPLIED, DISTINCT RACES ARENOT FORMED. HIGHLY-BRED ANIMALSLIABLE TO DEGENERATION. TENDENCY INMAN TO CARRY THE SELECTION OF EACHCHARACTER TO AN EXTREME POINT,LEADING TO DIVERGENCE OF CHARAC-TER, RARELY TO CONVERGENCE. CHA-RACTERS CONTINUING TO VARY IN THESAME DIRECTION IN WHICH THEY HAVEALREADY VARIED. DIVERGENCE OF CHA-RACTER, WITH THE EXTINCTION OF IN-TERMEDIATE VARIETIES, LEADS TO DIS-TINCTNESS IN OUR DOMESTIC RACES.LIMIT TO THE POWER OF SELECTION.LAPSE OF TIME IMPORTANT. MANNER IN

WHICH DOMESTIC RACES HAVE ORIGI-NATED. SUMMARY.

NATURAL SELECTION, OR THE SURVI-VAL OF THE FITTEST, AS AFFECTING DO-MESTIC PRODUCTIONS.

We know little on this head. But as animalskept by savages have to provide throughoutthe year their own food either entirely or to alarge extent, it can hardly be doubted that indifferent countries, varieties differing in consti-tution and in various characters would succeedbest, and so be naturally selected. Hence per-haps it is that the few domesticated animalskept by savages partake, as has been remarkedby more than one writer, of the wild appearan-ce of their masters, and likewise resemble natu-ral species. Even in long-civilised countries, atleast in the wilder parts, natural selection mustact on our domestic races. It is obvious thatvarieties having very different habits, constitu-tion, and structure, would succeed best on

mountains and on rich lowland pastures. Forexample, the improved Leicester sheep wereformerly taken to the Lammermuir Hills; butan intelligent sheep-master reported that "ourcoarse lean pastures were unequal to the task ofsupporting such heavy- bodied sheep; and theygradually dwindled away into less and lessbulk: each generation was inferior to the prece-ding one; and when the spring was severe, sel-dom more than two-thirds of the lambs survi-ved the ravages of the storms." (21/1. Quotedby Youatt on 'Sheep' page 325. See also Youatton 'Cattle' pages 62, 69.) So with the mountaincattle of North Wales and the Hebrides, it hasbeen found that they could not withstand beingcrossed with the larger and more delicate low-land breeds. Two French naturalists, in descri-bing the horses of Circassia, remark that, sub-jected as they are to extreme vicissitudes ofclimate, having to search for scanty pasture,and exposed to constant danger from wolves,the strongest and most vigorous alone survive.

(21/2. MM. Lherbette and De Quatrefages in'Bull. Soc. d'Acclimat.' tome 8 1861 page 311.)

Every one must have been struck with the sur-passing grace, strength, and vigour of the Ga-me-cock, with its bold and confident air, itslong, yet firm neck, compact body, powerfuland closely pressed wings, muscular thighs,strong beak massive at the base, dense andsharp spurs set low on the legs for deliveringthe fatal blow, and its compact, glossy, andmail-like plumage serving as a defence. Nowthe English game-cock has not only been im-proved during many years by man's carefulselection, but in addition, as Mr. Tegetmeierhas remarked (21/3. 'The Poultry Book' 1866page 123. Mr. Tegetmeier, 'The Homing or Ca-rrier Pigeon' 1871 pages 45-58.), by a kind ofnatural selection, for the strongest, most activeand courageous birds have stricken down theirantagonists in the cockpit, generation after ge-neration, and have subsequently served as the

progenitors of their race. The same kind ofdouble selection has come into play with thecarrier pigeon, for during their training the in-ferior birds fail to return home and are lost, sothat even without selection by man only thesuperior birds propagate their race.

In Great Britain, in former times, almost everydistrict had its own breed of cattle and sheep;"they were indigenous to the soil, climate, andpasturage of the locality on which they grazed:they seemed to have been formed for it and byit." (21/4. 'Youatt on Sheep' page 312.) But inthis case we are quite unable to disentangle theeffects of the direct action of the conditions oflife,—of use or habit—of natural selection—andof that kind of selection which we have seen isoccasionally and unconsciously followed byman even during the rudest periods of history.

Let us now look to the action of natural selec-tion on special characters. Although nature isdifficult to resist, yet man often strives against

her power, and sometimes with success. Fromthe facts to be given, it will also be seen thatnatural selection would powerfully affect manyof our domestic productions if left unprotected.This is a point of much interest, for we thuslearn that differences apparently of very slightimportance would certainly determine the sur-vival of a form when forced to struggle for itsown existence. It may have occurred to somenaturalists, as it formerly did to me, that,though selection acting under natural condi-tions would determine the structure of all im-portant organs, yet that it could not affect cha-racters which are esteemed by us of little im-portance; but this is an error to which we areeminently liable, from our ignorance of whatcharacters are of real value to each living crea-ture.

When man attempts to make a breed with someserious defect in structure, or in the mutualrelation of the several parts, he will partly or

completely fail, or encounter much difficulty;he is in fact resisted by a form of natural selec-tion. We have seen that an attempt was oncemade in Yorkshire to breed cattle with enor-mous buttocks, but the cows perished so oftenin bringing forth their calves, that the attempthad to be given up. In rearing short- facedtumblers, Mr. Eaton says (21/5. 'Treatise on theAlmond Tumbler' 1851 page 33.), "I am convin-ced that better head and beak birds have peris-hed in the shell than ever were hatched; thereason being that the amazingly short- facedbird cannot reach and break the shell with itsbeak, and so perishes." Here is a more curiouscase, in which natural selection comes into playonly at long intervals of time: during ordinaryseasons the Niata cattle can graze as well asothers, but occasionally, as from 1827 to 1830the plains of La Plata suffer from long-continued droughts and the pasture is burntup; at such times common cattle and horsesperish by the thousand, but many survive by

browsing on twigs, reeds, etc.; this the Niatacattle cannot so well effect from their upturnedjaws and the shape of their lips; consequently,if not attended to, they perish before the othercattle. In Columbia, according to Roulin, thereis a breed of nearly hairless cattle, called Pelo-nes; these succeed in their native hot district,but are found too tender for the Cordillera; inthis case, however, natural selection determinesonly the range of the variety. It is obvious that ahost of artificial races could never survive in astate of nature;—such as Italian greyhounds,—hairless and almost toothless Turkish dogs,—fantail pigeons, which cannot fly well against astrong wind,—barbs and Polish fowls, withtheir vision impeded by their eye wattles andgreat topknots,—hornless bulls and rams,which consequently cannot cope with othermales, and thus have a poor chance of leavingoffspring,—seedless plants, and many othersuch cases.

Colour is generally esteemed by the systematicnaturalist as unimportant: let us, therefore, seehow far it indirectly affects our domestic pro-ductions, and how far it would affect them ifthey were left exposed to the full force of natu-ral selection. In a future chapter I shall have toshow that constitutional peculiarities of thestrangest kind, entailing liability to the action ofcertain poisons, are correlated with the colourof the skin. I will here give a single case, on thehigh authority of Professor Wyman; he informsme that, being surprised at all the pigs in a partof Virginia being black, he made inquiries, andascertained that these animals feed on the rootsof the Lachnanthes tinctoria, which colourstheir bones pink, and, excepting in the case ofthe black varieties, causes the hoofs to drop off.Hence, as one of the squatters remarked, "weselect the black members of the litter for rai-sing, as they alone have a good chance of li-ving." So that here we have artificial and natu-ral selection working hand in hand. I may add

that in the Tarentino the inhabitants keep blacksheep alone, because the Hypericum crispumabounds there; and this plant does not injureblack sheep, but kills the white ones in about afortnight's time. (21/6. Dr. Heusinger 'Wo-chenschrift fur die Hei1kunde' Berlin 1846 s.279.)

Complexion, and liability to certain diseases,are believed to run together in man and thelower animals. Thus white terriers suffer morethan those of any other colour from the fataldistemper. (21/7. Youatt on the 'Dog' page 232.)In North America plum-trees are liable to adisease which Downing (21/8. 'The Fruit-treesof America' 1845 page 270: for peaches page466.) believes is not caused by insects; the kindsbearing purple fruit are most affected, "and wehave never known the green or yellow fruitedvarieties infected until the other sorts had firstbecome filled with the knots." On the otherhand, peaches in North America suffer much

from a disease called the "yellows," whichseems to be peculiar to that continent, and mo-re than nine-tenths of the victims, "when thedisease first appeared, were the yellow-fleshedpeaches. The white-fleshed kinds are muchmore rarely attacked; in some parts of the coun-try never." In Mauritius, the white sugar-caneshave of late years been so severely attacked bya disease, that many planters have been compe-lled to give up growing this variety (althoughfresh plants were imported from China fortrial), and cultivate only red canes. (21/9. 'Proc.Royal Soc. of Arts and Sciences of Mauritius'1852 page 135.) Now, if these plants had beenforced to struggle with other competing plantsand enemies, there cannot be a doubt that thecolour of the flesh or skin of the fruit, unimpor-tant as these characters are considered, wouldhave rigorously determined their existence.

Liability to the attacks of parasites is also con-nected with colour. White chickens are certain-

ly more subject than dark-coloured chickens tothe "gapes," which is caused by a parasiticworm in the trachea. (21/10. 'Gardener's Chro-nicle' 1856 page 379.) On the other hand, expe-rience has shown that in France the caterpillarswhich produce white cocoons resist the deadlyfungus better than those producing yellow co-coons. (21/11. Quatrefages 'Maladies Actuellesdu Ver a Soie' 1859 pages 12, 214.) Analogousfacts have been observed with plants: a newand beautiful white onion, imported fromFrance, though planted close to other kinds,was alone attacked by a parasitic fungus.(21/12. 'Gardener's Chronicle' 1851 page 595.)White verbenas are especially liable to mildew.(21/13. 'Journal of Horticulture' 1862 page 476.)Near Malaga, during an early period of thevine-disease, the green sorts suffered most;"and red and black grapes, even when interwo-ven with the sick plants, suffered not at all." InFrance whole groups of varieties were compa-ratively free, and others, such as the Chasselas,

did not afford a single fortunate exception; butI do not know whether any correlation betweencolour and liability to disease was here obser-ved. (21/14. 'Gardener's Chronicle' 1852 pages435, 691.) In a former chapter it was shown howcuriously liable one variety of the strawberry isto mildew.

It is certain that insects regulate in many casesthe range and even the existence of the higheranimals, whilst living under their natural con-ditions. Under domestication light-colouredanimals suffer most: in Thuringia (21/15.Bechstein 'Naturgesch. Deutschlands' 1801 b. 1s. 310.) the inhabitants do not like grey, white,or pale cattle, because they are much moretroubled by various kinds of flies than thebrown, red, or black cattle. An Albino negro, ithas been remarked (21/16. Prichard 'Phys. Hist.of Mankind' 1851 volume 1 page 224.), was pe-culiarly sensitive to the bites of insects. In theWest Indies (21/17. G. Lewis 'Journal of Resi-

dence in West Indies' 'Home and Col. Library'page 100.) it is said that "the only horned cattlefit for work are those which have a good deal ofblack in them. The white are terribly tormentedby the insects; and they are weak and sluggishin proportion to the white."

In Devonshire there is a prejudice against whitepigs, because it is believed that the sun blistersthem when turned out (21/18. Sidney's editionof Youatt on the 'Pig' page 24. I have given ana-logous facts in the case of mankind in my 'Des-cent of Man' 2nd edition page 195.); and I knewa man who would not keep white pigs in Kent,for the same reason. The scorching of flowersby the sun seems likewise to depend much oncolour; thus, dark pelargoniums suffer most;and from various accounts it is clear that thecloth-of-gold variety will not withstand a de-gree of exposure to sunshine which other varie-ties enjoy. Another amateur asserts that notonly all dark-coloured verbenas, but likewise

scarlets, suffer from the sun: "the paler kindsstand better, and pale blue is perhaps the bestof all." So again with the heartsease (Viola trico-lor); hot weather suits the blotched sorts, whilstit destroys the beautiful markings of some ot-her kinds. (21/19. 'Journal of Horticulture' 1862pages 476, 498; 1865 page 460. With respect tothe heartsease 'Gardener's Chronicle' 1863 page628.) During one extremely cold season inHolland all red-flowered hyacinths were ob-served to be very inferior in quality. It is belie-ved by many agriculturists that red wheat ishardier in northern climates than white wheat.(21/20. 'Des Jacinthes, de leur Culture' 1768page 53: on wheat 'Gardener's Chronicle' 1846page 653.)

With animals, white varieties from being cons-picuous are the most liable to be attacked bybeasts and birds of prey. In parts of France andGermany where hawks abound, persons areadvised not to keep white pigeons; for, as Par-

mentier says, "it is certain that in a flock thewhite always first fall victims to the kite." InBelgium, where so many societies have beenestablished for the flight of carrier-pigeons,white is the one colour which for the same rea-son is disliked. (20/21. W.B. Tegetmeier 'TheField' February 25, 1865. With respect to blackfowls see a quotation in Thompson 'Nat. Hist.of Ireland' 1849 volume 1 page 22.) Prof. G.Jaeger (21/22. 'In Sachen Darwin's contra Wi-gand' 1874 page 70.) whilst fishing found fourpigeons which had been killed by hawks, andall were white; on another occasion he exami-ned the eyrie of a hawk, and the feathers of thepigeons which had been caught were all of awhite or yellow colour. On the other hand, it issaid that the sea-eagle (Falco ossifragus, Linn.)on the west coast of Ireland picks out the blackfowls, so that "the villagers avoid as much aspossible rearing birds of that colour." M. Dau-din (20/23. 'Bull. de la Soc. d'Acclimat.' tome 71860 page 359.), speaking of white rabbits kept

in warrens in Russia, remarks that their colouris a great disadvantage, as they are thus moreexposed to attack, and can be seen duringbright nights from a distance. A gentleman inKent, who failed to stock his woods with anearly white and hardy kind of rabbit, accoun-ted in the same manner for their early disap-pearance. Any one who will watch a white catprowling after her prey will soon perceive un-der what a disadvantage she lies.

The white Tartarian cherry, "owing either to itscolour being so much like that of the leaves, orto the fruit always appearing from a distanceunripe," is not so readily attacked by birds asother sorts. The yellow-fruited raspberry,which generally comes nearly true by seed, "isvery little molested by birds, who evidently arenot fond of it; so that nets may be dispensedwith in places where nothing else will protectthe red fruit." (21/24. 'Transact. Hort. Soc.' vo-lume 1 2nd series 1835 page 275. For raspbe-

rries see 'Gardener's Chronicle' 1855 page 154and 1863 page 245.) This immunity, though abenefit to the gardener, would be a disadvanta-ge in a state of nature both to the cherry andraspberry, as dissemination depends on birds. Inoticed during several winters that some treesof the yellow-berried holly, which were raisedfrom seed from a tree found wild by my fatherremained covered with fruit, whilst not a scar-let berry could be seen on the adjoining trees ofthe common kind. A friend informs me that amountain-ash (Pyrus aucuparia) growing in hisgarden bears berries which, though not diffe-rently coloured, are always devoured by birdsbefore those on the other trees. This variety ofthe mountain-ash would thus be more freelydisseminated, and the yellow-berried variety ofthe holly less freely, than the common varietiesof these two trees.

Independently of colour, trifling differences aresometimes found to be of importance to plants

under cultivation, and would be of paramountimportance if they had to fight their own battleand to struggle with many competitors. Thethin-shelled peas, called pois sans parchemin,are attacked by birds (21/25. 'Gardener's Chro-nicle' 1843 page 806.) much more commonlythan ordinary peas. On the other hand, the pur-ple-podded pea, which has a hard shell, esca-ped the attacks of tomtits (Parus major) in mygarden far better than any other kind. The thin-shelled walnut likewise suffers greatly from thetomtit. (21/26. Ibid 1850 page 732.) These samebirds have been observed to pass over and thusfavour the filbert, destroying only the otherkinds of nuts which grew in the same orchard.(21/27. Ibid 1860 page 956.)

Certain varieties of the pear have soft bark, andthese suffer severely from wood-boring beetles;whilst other varieties are known to resist theirattacks much better. (21/28. J. De Jonghe in'Gardener's Chronicle' 1860 page 120.) In North

America the smoothness, or absence of downon the fruit, makes a great difference in the at-tacks of the weevil, "which is the uncompromi-sing foe of all smooth stone-fruits;" and the cul-tivator "has the frequent mortification of seeingnearly all, or indeed often the whole crop, fallfrom the trees when half or two-thirds grown."Hence the nectarine suffers more than thepeach. A particular variety of the Morello che-rry, raised in North America, is, without anyassignable cause, more liable to be injured bythis same insect than other cherry-trees. (21/29.Downing 'Fruit-trees of North America' pages266, 501: in regard to the cherry page 198.)From some unknown cause, certain varieties ofthe apple enjoy, as we have seen, the great ad-vantage in various parts of the world of notbeing infested by the coccus. On the otherhand, a particular case has been recorded inwhich aphides confined themselves to the Win-ter Nelis pear and touched no other kind in anextensive orchard. (21/30. 'Gardener's Chroni-

cle' 1849 page 755.) The existence of minuteglands on the leaves of peaches, nectarines, andapricots, would not be esteemed by botanists asa character of the least importance for they arepresent or absent in closely-related sub-varieties, descended from the same parent-tree;yet there is good evidence (21/31. 'Journal ofHorticulture' September 26, 1865 page 254; seeother references given in chapter 10.) that theabsence of glands leads to mildew, which ishighly injurious to these trees.

A difference either in flavour or in the amountof nutriment in certain varieties causes them tobe more eagerly attacked by various enemiesthan other varieties of the same species. Bull-finches (Pyrrhula vulgaris) injure our fruit-treesby devouring the flower-buds, and a pair ofthese birds have been seen "to denude a largeplum-tree in a couple of days of almost everybud;" but certain varieties (21/32. Mr. Selby in'Mag. of Zoology and Botany' Edinburgh vo-

lume 2 1838 page 393.) of the apple and thorn(Crataegus oxyacantha) are more especiallyliable to be attacked. A striking instance of thiswas observed in Mr. Rivers's garden, in whichtwo rows of a particular variety of plum(21/33. The Reine Claude de Bavay 'Journal ofHorticulture' December 27, 1864 page 511.) hadto be carefully protected, as they were usuallystripped of all their buds during the winter,whilst other sorts growing near them escaped.The root (or enlarged stem) of Laing's Swedishturnip is preferred by hares, and therefore suf-fers more than other varieties. Hares and rab-bits eat down common rye before St. John's-day-rye, when both grow together. (21/34. Mr.Pusey in 'Journal of R. Agricult. Soc.' volume 6page 179. For Swedish turnips see 'Gardener'sChronicle' 1847 page 91.) In the south of France,when an orchard of almond-trees is formed, thenuts of the bitter variety are sown, "in orderthat they may not be devoured by field- mice"(21/35. Godron 'De l'Espece' tome 2 page 98.);

so we see the use of the bitter principle in al-monds.

Other slight differences, which would bethought quite unimportant, are no doubt some-times of great service both to plants and ani-mals. The Whitesmith's gooseberry, as formerlystated, produces its leaves later than other va-rieties, and, as the flowers are thus left unpro-tected, the fruit often fails. In one variety of thecherry, according to Mr. Rivers (21/36. 'Garde-ner's Chronicle' 1866 page 732.), the petals aremuch curled backwards, and in consequence ofthis the stigmas were observed to be killed by asevere frost; whilst at the same time, in anothervariety with incurved petals, the stigmas werenot in the least injured. The straw of the Fentonwheat is remarkably unequal in height; and acompetent observer believes that this variety ishighly productive, partly because the ears frombeing distributed at various heights above theground are less crowded together. The same

observer maintains that in the upright varietiesthe divergent awns are serviceable by breakingthe shocks when the ears are dashed togetherby the wind. (21/37. 'Gardener's Chronicle'1862 pages 820, 821.) If several varieties of aplant are grown together, and the seed is indis-criminately harvested, it is clear that the har-dier and more productive kinds will, by a sortof natural selection, gradually prevail over theothers; this takes place, as Colonel Le Couteurbelieves (21/38. 'On the Varieties of Wheat'page 59.), in our wheat-fields, for, as formerlyshown, no variety is quite uniform in character.The same thing, as I am assured by nursery-men, would take place in our flower-gardens, ifthe seed of the different varieties were not se-parately saved. When the eggs of the wild andtame duck are hatched together, the youngwild ducks almost invariably perish, frombeing of smaller size and not getting their fairshare of food. (21/39. Mr. Hewitt and others, in'Journal of Hort.' 1862 page 773.)

Facts in sufficient number have now been givenshowing that natural selection often checks, butoccasionally favours, man's power of selection.These facts teach us, in addition, a valuablelesson, namely, that we ought to be extremelycautious in judging what characters are of im-portance in a state of nature to animals andplants, which have to struggle for existencefrom the hour of their birth to that of theirdeath,—their existence depending on condi-tions, about which we are profoundly ignorant.

CIRCUMSTANCES FAVOURABLE TO SE-LECTION BY MAN.

The possibility of selection rests on variability,and this, as we shall see in the following chap-ters, mainly depends on changed conditions oflife, but is governed by infinitely complex andunknown laws. Domestication, even when longcontinued, occasionally causes but a smallamount of variability, as in the case of the goo-

se and turkey. The slight differences, however,which characterise each individual animal andplant would in most, probably in all, cases suf-fice for the production of distinct races throughcareful and prolonged selection. We see whatselection, though acting on mere individualdifferences, can effect when families of cattle,sheep, pigeons, etc., of the same race, have beenseparately bred during a number of years bydifferent men without any wish on their part tomodify the breed. We see the same fact in thedifference between hounds bred for hunting indifferent districts (21/40. 'Encyclop. of RuralSports' page 405.), and in many other such ca-ses.

In order that selection should produce any re-sult, it is manifest that the crossing of distinctraces must be prevented; hence facility in pai-ring, as with the pigeon, is highly favourablefor the work; and difficulty in pairing, as withcats, prevents the formation of distinct breeds.

On nearly the same principle the cattle of thesmall island of Jersey have been improved intheir milking qualities "with a rapidity thatcould not have been obtained in a widely ex-tended country like France." (21/41. Col. LeCouteur 'Journal Roy. Agricult. Soc.' volume 4page 43.) Although free crossing is a danger onthe one side which every one can see, too closeinterbreeding is a hidden danger on the otherside. Unfavourable conditions of life overrulethe power of selection. Our improved heavybreeds of cattle and sheep could not have beenformed on mountainous pastures; nor coulddray-horses have been raised on a barren andinhospitable land, such as the Falkland Islands,where even the light horses of La Plata rapidlydecrease in size. It seems impossible to preser-ve several English breeds of sheep in France;for as soon as the lambs are weaned their vi-gour decays as the heat of the summer increa-ses (21/42. Malingie-Nouel 'Journal R. Agricult.Soc.' volume 14 1853 pages 215, 217.): it would

be impossible to give great length of wool tosheep within the tropics; yet selection has keptthe Merino breed nearly true under diversifiedand unfavourable conditions. The power ofselection is so great, that breeds of the dog,sheep, and poultry, of the largest and smallestsize, long and short beaked pigeons, and otherbreeds with opposite characters, have had theircharacteristic qualities augmented, though trea-ted in every way alike, being exposed to thesame climate and fed on the same food. Selec-tion, however, is either checked or favoured bythe effects of use or habit. Our wonderfully-improved pigs could never have been formed ifthey had been forced to search for their ownfood; the English racehorse and greyhoundcould not have been improved up to their pre-sent high standard of excellence without cons-tant training.

As conspicuous deviations of structure occurrarely, the improvement of each breed is gene-

rally the result of the selection of slight indivi-dual differences. Hence the closest attention,the sharpest powers of observation, and indo-mitable perseverance, are indispensable. It is,also, highly important that many individuals ofthe breed which is to be improved should beraised; for thus there will be a better chance ofthe appearance of variations in the right direc-tion, and individuals varying in an unfavoura-ble manner may be freely rejected or destroyed.But that a large number of individuals shouldbe raised, it is necessary that the conditions oflife should favour the propagation of the spe-cies. Had the peacock been reared as easily asthe fowl, we should probably ere this have hadmany distinct races. We see the importance of alarge number of plants, from the fact of nurserygardeners almost always beating amateurs inthe exhibition of new varieties. In 1845 it wasestimated (21/43. Gardener's Chronicle' 1845page 273.) that between 4000 and 5000 pelargo-niums were annually raised from seed in En-

gland, yet a decidedly improved variety is rare-ly obtained. At Messrs. Carter's grounds, inEssex, where such flowers as the Lobelia, Ne-mophila, Mignonette, etc., are grown by theacre for seed, "scarcely a season passes withoutsome new kinds being raised, or some impro-vement effected on old kinds." (21/44. 'Journalof Horticulture' 1862 page 157.) At Kew, as Mr.Beaton remarks, where many seedlings ofcommon plants are raised, "you see new formsof Laburnums, Spiraeas, and other shrubs."(21/45. 'Cottage Gardener' 1860 page 368.) Sowith animals: Marshall (21/46. 'A Review ofReports' 1808 page 406.), in speaking of thesheep in one part of Yorkshire, remarks, "asthey belong to poor people, and are mostly insmall lots, they never can be improved." LordRivers, when asked how he succeeded in al-ways having first-rate greyhounds, answered,"I breed many, and hang many." This, as anot-her man remarks, "was the secret of his success;and the same will be found in exhibiting

fowls,— successful competitors breed largely,and keep the best." (21/47. 'Gardener's Chroni-cle' 1853 page 45.)

It follows from this that the capacity of bree-ding at an early age and at short intervals, aswith pigeons, rabbits, etc., facilitates selection;for the result is thus soon made visible, andperseverance in the work encouraged. It canhardly be an accident that the great majority ofthe culinary and agricultural plants which haveyielded numerous races are annuals or bien-nials, which therefore are capable of rapid pro-pagation, and thus of improvement. Sea-kale,asparagus, common and Jerusalem artichokes,potatoes, and onions, must be excepted, as theyare perennials: but onions are propagated likeannuals, and of the other plants just specified,none, with the exception of the potato, haveyielded in this country more than one or twovarieties. In the Mediterranean region, whereartichokes are often raised from seed, there are

several kinds, as I hear from Mr. Bentham. Nodoubt fruit- trees, which cannot be propagatedquickly by seed, have yielded a host of varie-ties, though not permanent races; but these,judging from prehistoric remains, have beenproduced at a comparatively late period.

A species may be highly variable, but distinctraces will not be formed, if from any cause se-lection be not applied. It would be difficult toselect slight variations in fishes from their placeof habitation; and though the carp is extremelyvariable and is much attended to in Germany,only one well- marked race has been formed, asI hear from Lord A. Russell, namely the spie-gel-carpe; and this is carefully secluded fromthe common scaly kind. On the other hand, aclosely allied species, the gold-fish, from beingreared in small vessels, and from having beencarefully attended to by the Chinese, has yiel-ded many races. Neither the bee, which hasbeen semi-domesticated from an extremely

remote period, nor the cochineal insect, whichwas cultivated by the aboriginal Mexicans(21/48. Isidore Geoffroy Saint-Hilaire 'Hist.Nat. Gen.' tome 3 page 49. 'On the CochinealInsect' page 46.), has yielded races; and itwould be impossible to match the queen-beewith any particular drone, and most difficult tomatch cochineal insects. Silk-moths, on the ot-her hand, have been subjected to rigorous selec-tion, and have produced a host of races. Cats,which from their nocturnal habits cannot beselected for breeding, do not, as formerly re-marked, yield distinct races within the samecountry. Dogs are held in abomination in theEast, and their breeding is neglected; conse-quently, as Prof. Moritz Wagner (21/49. 'DieDarwin'sche Theorie und das Migrationsgesetzder Organismen' 1868 page 19.) remarks, onekind alone exists there. The ass in England va-ries much in colour and size; but as it is ananimal of little value and bred by poor people,there has been no selection, and distinct races

have not been formed. We must not attributethe inferiority of our asses to climate, for inIndia they are of even smaller size than in Eu-rope. But when selection is brought to bear onthe ass, all is changed. Near Cordova, as I aminformed (February 1860) by Mr. W.E. Webb,C.E., they are carefully bred, as much as 200pounds having been paid for a stallion ass, andthey have been immensely improved. In Ken-tucky, asses have been imported (for breedingmules) from Spain, Malta, and France; these"seldom averaged more than fourteen handshigh: but the Kentuckians, by great care, haveraised them up to fifteen hands, and sometimeseven to sixteen. The prices paid for thesesplendid animals, for such they really are, willprove how much they are in request. One male,of great celebrity, was sold for upwards of onethousand pounds sterling." These choice assesare sent to cattle-shows, a day being given fortheir exhibition. (21/50. Capt. Marryat quoted

by Blyth in 'Journ. Asiatic Soc. of Bengal' volu-me 28 page 229.)

Analogous facts have been observed withplants: the nutmeg-tree in the Malay archipela-go is highly variable, but there has been no se-lection, and there are no distinct races. (21/51.Mr. Oxley 'Journal of the Indian Archipelago'volume 2 1848 page 645.) The common migno-nette (Reseda odorata), from bearing inconspi-cuous flowers, valued solely for their fragrance,"remains in the same unimproved condition aswhen first introduced." (21/52. Mr. Abbey'Journal of Horticulture' December 1, 1863 page430.) Our common forest-trees are very varia-ble, as may be seen in every extensive nursery-ground; but as they are not valued like fruit-trees, and as they seed late in life, no selectionhas been applied to them; consequently, as Mr.Patrick Matthews remarks (21/53. 'On NavalTimber' 1831 page 107.), they have not yieldeddistinct races, leafing at different periods, gro-

wing to different sizes, and producing timberfit for different purposes. We have gained onlysome fanciful and semi- monstrous varieties,which no doubt appeared suddenly as we nowsee them.

Some botanists have argued that plants cannothave so strong a tendency to vary as is genera-lly supposed, because many species longgrown in botanic gardens, or unintentionallycultivated year after year mingled with ourcorn crops, have not produced distinct races;but this is accounted for by slight variations nothaving been selected and propagated. Let aplant which is now grown in a botanic garden,or any common weed, be cultivated on a largescale, and let a sharp-sighted gardener look outfor each slight variety and sow the seed, andthen, if distinct races are not produced, the ar-gument will be valid.

The importance of selection is likewise shownby considering special characters. For instance,

with most breeds of fowls the form of the comband the colour of the plumage have been atten-ded to, and are eminently characteristic of eachrace; but in Dorkings fashion has never de-manded uniformity of comb or colour; and theutmost diversity in these respects prevails. Ro-se-combs, double-combs, cup-combs, etc., andcolours of all kinds, may be seen in purely bredand closely related Dorking fowls, whilst otherpoints, such as the general form of body, andthe presence of an additional toe, have beenattended to, and are invariably present. It hasalso been ascertained that colour can be fixed inthis breed, as well as in any other. (21/54. Mr.Baily in 'The Poultry Chronicle' volume 2 1854page 150. Also volume 1 page 342; volume 3page 245.)

During the formation or improvement of abreed, its members will always be found tovary much in those characters to which especialattention is directed, and of which each slight

improvement is eagerly sought and selected.Thus, with short-faced tumbler-pigeons, theshortness of the beak, shape of head and plu-mage,—with carriers, the length of the beakand wattle,—with fantails, the tail and carria-ge,—with Spanish fowls, the white face andcomb,—with long-eared rabbits, the length ofear, are all points which are eminently variable.So it is in every case; and the large price paidfor first-rate animals proves the difficulty ofbreeding them up to the highest standard ofexcellence. This subject has been discussed byfanciers (21/55. 'Cottage Gardener' 1855 De-cember page 171; 1856 January pages 248, 323.),and the greater prizes given for highly impro-ved breeds, in comparison with those given forold breeds which are not now undergoing ra-pid improvement, have been fully justified.Nathusius makes (21/56. 'Ueber ShorthornRindvieh' 1857 s. 51.) a similar remark whendiscussing the less uniform character of impro-ved Shorthorn cattle and of the English horse,

in comparison, for example, with the unenno-bled cattle of Hungary, or with the horses of theAsiatic steppes. This want of uniformity in theparts which at the time are undergoing selec-tion chiefly depends on the strength of theprinciple of reversion; but it likewise dependsto a certain extent on the continued variabilityof the parts which have recently varied. Thatthe same parts do continue varying in the samemanner we must admit, for if it were not so,there could be no improvement beyond an ear-ly standard of excellence, and we know thatsuch improvement is not only possible, but is ofgeneral occurrence.

As a consequence of continued variability, andmore especially of reversion, all highly impro-ved races, if neglected or not subjected to inces-sant selection, soon degenerate. Youatt gives acurious instance of this in some cattle formerlykept in Glamorganshire; but in this case thecattle were not fed with sufficient care. Mr. Ba-

ker, in his memoir on the Horse, sums up: "Itmust have been observed in the preceding pa-ges that, whenever there has been neglect, thebreed has proportionally deteriorated." (21/57.'The Veterinary' volume 13 page 720. For theGlamorganshire cattle see Youatt on 'Cattle'page 51.) If a considerable number of improvedcattle, sheep, or other animals of the same race,were allowed to breed freely together, with noselection, but with no change in their conditionof life, there can be no doubt that after a scoreor hundred generations they would be very farfrom excellent of their kind; but, from what wesee of the many common races of dogs, cattle,fowls, pigeons, etc., which without any particu-lar care have long retained nearly the same cha-racter, we have no grounds for believing thatthey would altogether depart from their type.

It is a general belief amongst breeders that cha-racters of all kinds become fixed by long-continued inheritance. But I have attempted to

show in the fourteenth chapter that this beliefapparently resolves itself into the followingproposition, namely, that all characters whate-ver, whether recently acquired or ancient, tendto be transmitted, but that those which havealready long withstood all counteracting in-fluences, will, as a general rule, continue towithstand them, and consequently be faithfullytransmitted.

TENDENCY IN MAN TO CARRY THEPRACTICE OF SELECTION TO AN EXTRE-ME POINT.

It is an important principle that in the processof selection man almost invariably wishes to goto an extreme point. Thus, there is no limit tohis desire to breed certain kinds of horses anddogs as fleet as possible, and others as strong aspossible; certain kinds of sheep for extremefineness, and others for extreme length of wool;and he wishes to produce fruit, grain, tubers,

and other useful parts of plants, as large andexcellent as possible. With animals bred foramusement, the same principle is even morepowerful; for fashion, as we see in our dress,always runs to extremes. This view has beenexpressly admitted by fanciers. Instances weregiven in the chapters on the pigeon, but here isanother: Mr. Eaton, after describing a compara-tively new variety, namely, the Archangel, re-marks, "What fanciers intend doing with thisbird I am at a loss to know, whether they in-tend to breed it down to the tumbler's head andbeak, or carry it out to the carrier's head andbeak; leaving it as they found it, is not progres-sing." Ferguson, speaking of fowls, says, "theirpeculiarities, whatever they may be, must ne-cessarily be fully developed: a little peculiarityforms nought but ugliness, seeing it violates theexisting laws of symmetry." So Mr. Brent, indiscussing the merits of the sub-varieties of theBelgian canary-bird, remarks, "Fanciers alwaysgo to extremes; they do not admire indefinite

properties." (21/58. J.M. Eaton 'A Treatise onFancy Pigeons' page 82; Ferguson on 'Rare andPrize Poultry' page 162; Mr. Brent in 'CottageGardener' October 1860 page 13.)

This principle, which necessarily leads to di-vergence of character, explains the present stateof various domestic races. We can thus see howit is that racehorses and dray-horses, greyh-ounds and mastiffs, which are opposed to eachother in every character,—how varieties so dis-tinct as Cochin-china fowls and bantams, orcarrier-pigeons with very long beaks, and tum-blers with excessively short beaks, have beenderived from the same stock. As each breed isslowly improved, the inferior varieties are firstneglected and finally lost. In a few cases, by theaid of old records, or from intermediate varie-ties still existing in countries where other fas-hions have prevailed, we are enabled partiallyto trace the graduated changes through whichcertain breeds have passed. Selection, whether

methodical or unconscious, always tendingtowards an extreme point, together with theneglect and slow extinction of the intermediateand less-valued forms, is the key which unlocksthe mystery of how man has produced suchwonderful results.

In a few instances selection, guided by utilityfor a single purpose, has led to convergence ofcharacter. All the improved and different racesof the pig, as Nathusius has well shown (21/59.'Die Racen des Schweines' 1860 s. 48.), closelyapproach each other in character, in their shor-tened legs and muzzles, their almost hairless,large, rounded bodies, and small tusks. We seesome degree of convergence in the similar ou-tline of the body in well-bred cattle belongingto distinct races. (21/60. See some good re-marks on this head by M. de Quatrefages 'Unitede l'Espece Humaine' 1861 page 119.) I know ofno other such cases.

Continued divergence of character depends on,and is indeed a clear proof, as previously re-marked, of the same parts continuing to vary inthe same direction. The tendency to mere gene-ral variability or plasticity of organisation cancertainly be inherited, even from one parent, ashas been shown by Gartner and Kolreuter, inthe production of varying hybrids from twospecies, of which one alone was variable. It is initself probable that, when an organ has variedin any manner, it will again vary in the samemanner, if the conditions which first caused thebeing to vary remain, as far as can be judged,the same. This is either tacitly or expressly ad-mitted by all horticulturists: if a gardener ob-serves one or two additional petals in a flower,he feels confident that in a few generations hewill be able to raise a double flower, crowdedwith petals. Some of the seedlings from theweeping Moccas oak were so prostrate thatthey only crawled along the ground. A seedlingfrom the fastigiate or upright Irish yew is des-

cribed as differing greatly from the parent-form"by the exaggeration of the fastigiate habit of itsbranches." (21/61. Verlot 'Des Varietes' 1865page 94.) Mr. Shirreff, who has been highlysuccessful in raising new kinds of wheat, re-marks, "A good variety may safely be regardedas the forerunner of a better one." (21/62. Mr.Patrick Shirreff 'Gardener's Chronicle' 1858page 771.) A great rose-grower, Mr. Rivers, hasmade the same remark with respect to roses.Sageret (21/63. 'Pomologie Physiolog.' 1830page 106.), who had large experience, in spea-king of the future progress of fruit-trees, obser-ves that the most important principle is "thatthe more plants have departed from their origi-nal type, the more they tend to depart from it."There is apparently much truth in this remark;for we can in no other way understand the sur-prising amount of difference between varietiesin the parts or qualities which are valued,whilst other parts retain nearly their originalcharacter.

The foregoing discussion naturally leads to thequestion, what is the limit to the possibleamount of variation in any part or quality, and,consequently, is there any limit to what selec-tion can effect? Will a racehorse ever be rearedfleeter than Eclipse? Can our prize-cattle andsheep be still further improved? Will a goose-berry ever weigh more than that produced by"London" in 1852? Will the beet-root in Franceyield a greater percentage of sugar? Will futurevarieties of wheat and other grain produce hea-vier crops than our present varieties? Thesequestions cannot be positively answered; but itis certain that we ought to be cautious in ans-wering them by a negative. In some lines ofvariation the limit has probably been reached.Youatt believes that the reduction of bone insome of our sheep has already been carried sofar that it entails great delicacy of constitution.(21/64. Youatt on 'Sheep' page 521.) But seeingthe great improvement within recent times inour cattle and sheep, and especially in our pigs;

seeing the wonderful increase in weight in ourpoultry of all kinds during the last few years;he would be a bold man who would assert thatperfection has been reached. It has often beensaid that Eclipse never was, and never will be,beaten in speed by any other horse; but on ma-king inquiries I find that the best judges believethat our present racehorses are fleeter. (21/65.See also Stonehenge 'British Rural Sports' edi-tion of 1871 page 384.) The attempt to raise anew variety of wheat more productive than themany old kinds, might have been thought untillately quite hopeless; but this has been effectedby Major Hallett, by careful selection. With res-pect to almost all our animals and plants, thosewho are best qualified to judge do not believethat the extreme point of perfection has yetbeen reached even in the characters which havealready been carried to a high standard. Forinstance, the short-faced tumbler-pigeon hasbeen greatly modified; nevertheless, accordingto Mr. Eaton (21/66. 'A Treatise on the Almond

Tumbler' page 1.) "the field is still as open forfresh competitors as it was one hundred yearsago." Over and over again it has been said thatperfection had been attained with our flowers,but a higher standard has soon been reached.Hardly any fruit has been more improved thanthe strawberry, yet a great authority remarks(21/67. M. J. de Jonghe in 'Gardener's Chroni-cle' 1858 page 173.), "it must not be concealedthat we are far from the extreme limits at whichwe may arrive."

No doubt there is a limit beyond which theorganisation cannot be modified compatiblywith health or life. The extreme degree of fleet-ness, for instance, of which a terrestrial animalis capable, may have been acquired by our pre-sent racehorses; but as Mr. Wallace has wellremarked (21/68. 'Contributions to the Theoryof Natural Selection' 2nd edition 1871 page292.), the question that interests us, "is notwhether indefinite and unlimited change in any

or all directions is possible, but whether suchdifferences as do occur in nature could havebeen produced by the accumulation of varietiesby selection." And in the case of our domesticproductions, there can be no doubt that manyparts of the organisation, to which man hasattended, have been thus modified to a greaterdegree than the corresponding parts in the na-tural species of the same genera or even fami-lies. We see this in the form and size of ourlight and heavy dogs or horses,—in the beakand many other characters of our pigeons,—inthe size and quality of many fruits,—in compa-rison with the species belonging to the samenatural groups.

Time is an important element in the formationof our domestic races, as it permits innumera-ble individuals to be born, and these when ex-posed to diversified conditions are renderedvariable. Methodical selection has been occa-sionally practised from an ancient period to the

present day, even by semi-civilised people, andduring former times will have produced someeffect. Unconscious selection will have beenstill more effective; for during a lengthenedperiod the more valuable individual animalswill occasionally have been saved, and the lessvaluable neglected. In the course of time, diffe-rent varieties, especially in the less civilisedcountries, will also have been more or less mo-dified through natural selection. It is generallybelieved, though on this head we have little orno evidence, that new characters in time beco-me fixed; and after having long remained fixedit seems possible that under new conditionsthey might again be rendered variable.

How great the lapse of time has been since manfirst domesticated animals and cultivatedplants, we begin dimly to see. When the lake-dwellings of Switzerland were inhabited du-ring the Neolithic period, several animals werealready domesticated and various plants culti-

vated. The science of language tells us that theart of ploughing and sowing the land was fo-llowed, and the chief animals had been alreadydomesticated, at an epoch so immensely remo-te, that the Sanskrit, Greek, Latin, Gothic, Cel-tic, and Sclavonic languages had not as yet di-verged from their common parent-tongue.(21/69. Max Muller 'Science of Language' 1861page 223.)

It is scarcely possible to overrate the effects ofselection occasionally carried on in variousways and places during thousands of genera-tions. All that we know, and, in a still strongerdegree, all that we do not know (21/70. 'Youatton Cattle' pages 116, 128.), of the history of thegreat majority of our breeds, even of our moremodern breeds, agrees with the view that theirproduction, through the action of unconsciousand methodical selection, has been almost in-sensibly slow. When a man attends rather moreclosely than is usual to the breeding of his ani-

mals, he is almost sure to improve them to aslight extent. They are in consequence valuedin his immediate neighbourhood, and are bredby others; and their characteristic features,whatever these may be, will then slowly butsteadily be increased, sometimes by methodicaland almost always by unconscious selection. Atlast a strain, deserving to be called a sub-variety, becomes a little more widely known,receives a local name, and spreads. The sprea-ding will have been extremely slow during an-cient and less civilised times, but now is rapid.By the time that the new breed had assumed asomewhat distinct character, its history, hardlynoticed at the time, will have been completelyforgotten; for, as Low remarks (21/71. 'Domes-ticated Animals' page 188.), "we know howquickly the memory of such events is effaced."

As soon as a new breed is thus formed, it isliable through the same process to break upinto new strains and sub-varieties. For different

varieties are suited for, and are valued under,different circumstances. Fashion changes, but,should a fashion last for even a moderatelength of time, so strong is the principle of in-heritance, that some effect will probably beimpressed on the breed. Thus varieties go onincreasing in number, and history shows ushow wonderfully they have increased since theearliest records. (21/72. Volz 'Beitrage zur Kul-turgeschichte' 1852 s. 99 et passim.) As eachnew variety is produced, the earlier, interme-diate, and less valuable forms will be neglected,and perish. When a breed, from not being va-lued, is kept in small numbers, its extinctionalmost inevitably follows sooner or later, eitherfrom accidental causes of destruction or fromclose interbreeding; and this is an event which,in the case of well-marked breeds, excites atten-tion. The birth or production of a new domesticrace is so slow a process that it escapes notice;its death or destruction is comparatively sud-

den, is often recorded, and when too late some-times regretted.

Several authors have drawn a wide distinctionbetween artificial and natural races. The latterare more uniform in character, possessing in ahigh degree the appearance of natural species,and are of ancient origin. They are generallyfound in less civilised countries, and have pro-bably been largely modified by natural selec-tion, and only to a small extent by man's un-conscious and methodical selection. They have,also, during a long period, been directly actedon by the physical conditions of the countrieswhich they inhabit. The so-called artificial ra-ces, on the other hand, are not so uniform incharacter; some have a semi-monstrous charac-ter, such as "the wry-legged terriers so useful inrabbit-shooting" (21/73. Blaine 'Encyclop. ofRural Sports' page 213.), turnspit dogs, anconsheep, niata oxen, Polish fowls, fantail-pigeons,etc.; their characteristic features have generally

been acquired suddenly, though subsequentlyincreased by careful selections in many cases.Other races, which certainly must be calledartificial, for they have been largely modifiedby methodical selection and by crossing, as theEnglish racehorse, terrier-dogs, the Englishgame-cock, Antwerp carrier-pigeons, etc., ne-vertheless cannot be said to have an unnaturalappearance; and no distinct line, as it seems tome, can be drawn between natural and artificialraces.

It is not surprising that domestic races shouldgenerally present a different aspect from natu-ral species. Man selects and propagates modifi-cations solely for his own use or fancy, and notfor the creature's own good. His attention isstruck by strongly marked modifications,which have appeared suddenly, due to somegreat disturbing cause in the organisation. Heattends almost exclusively to external charac-ters; and when he succeeds in modifying inter-

nal organs,—when for instance he reduces thebones and offal, or loads the viscera with fat, orgives early maturity, etc.-the chances are strongthat he will at the same time weaken the consti-tution. On the other hand, when an animal hasto struggle throughout its life with many com-petitors and enemies, under circumstances in-conceivably complex and liable to change, mo-difications of the most varied nature in the in-ternal organs as well as in external characters,in the functions and mutual relations of parts,will be rigorously tested, preserved, or rejected.Natural selection often checks man's compara-tively feeble and capricious attempts at impro-vement; and if it were not so, the result of hiswork, and of nature's work, would be even stillmore different. Nevertheless, we must not ove-rrate the amount of difference between naturalspecies and domestic races; the most experien-ced naturalists have often disputed whether thelatter are descended from one or from severalaboriginal stocks, and this clearly shows that

there is no palpable difference between speciesand races.

Domestic races propagate their kind far moretruly, and endure for munch longer periods,than most naturalists are willing to admit.Breeders feel no doubt on this head: ask a manwho has long reared Shorthorn or Herefordcattle, Leicester or Southdown sheep, Spanishor Game poultry, tumbler or carrier-pigeons,whether these races may not have been derivedfrom common progenitors, and he will proba-bly laugh you to scorn. The breeder admits thathe may hope to produce sheep with finer orlonger wool and with better carcases, or hand-somer fowls, or carrier-pigeons with beaks justperceptibly longer to the practised eye, andthus be successful at an exhibition. Thus far hewill go, but no farther. He does not reflect onwhat follows from adding up during a longcourse of time many slight, successive modifi-cations; nor does he reflect on the former exis-

tence of numerous varieties, connecting thelinks in each divergent line of descent. He con-cludes, as was shown in the earlier chapters,that all the chief breeds to which he has longattended are aboriginal productions. The sys-tematic naturalist, on the other hand, who ge-nerally knows nothing of the art of breeding,who does not pretend to know how and whenthe several domestic races were formed, whocannot have seen the intermediate gradations,for they do not now exist, nevertheless feels nodoubt that these races are sprung from a singlesource. But ask him whether the closely alliednatural species which he has studied may nothave descended from a common progenitor,and he in his turn will perhaps reject the notionwith scorn. Thus the naturalist and breedermay mutually learn a useful lesson from eachother.

SUMMARY ON SELECTION BY MAN.

There can be no doubt that methodical selectionhas effected and will effect wonderful results. Itwas occasionally practised in ancient times, andis still practised by semi-civilised people. Cha-racters of the highest importance, and others oftrifling value, have been attended to, and modi-fied. I need not here repeat what has been sooften said on the part which unconscious selec-tion has played: we see its power in the diffe-rence between flocks which have been separa-tely bred, and in the slow changes, as circums-tances have slowly changed, which many ani-mals have undergone in the same country, orwhen transported into a foreign land. We seethe combined effects of methodical and un-conscious selection, in the great amount of dif-ference in those parts or qualities which arevalued by man in comparison with the partswhich are not valued, and consequently havenot been attended to. Natural selection oftendetermines man's power of selection. We some-times err in imagining that characters, which

are considered as unimportant by the systema-tic naturalist, could not be affected by thestruggle for existence, and could not be actedon by natural selection; but striking cases havebeen given, showing how great an error this is.

The possibility of selection coming into actionrests on variability; and this is mainly caused,as we shall hereafter see, by changes in theconditions of life. Selection is sometimes rende-red difficult, or even impossible, by the condi-tions being opposed to the desired character orquality. It is sometimes checked by the lessenedfertility and weakened constitution which fo-llow from long-continued close interbreeding.That methodical selection may be successful,the closest attention and discernment, combi-ned with unwearied patience, are absolutelynecessary; and these same qualities, though notindispensable, are highly serviceable in the caseof unconscious selection. It is almost necessarythat a large number of individuals should be

reared; for thus there will be a fair chance ofvariations of the desired nature arising, and ofevery individual with the slightest blemish orin any degree inferior being freely rejected.Hence length of time is an important element ofsuccess. Thus, also, reproduction at an earlyage and at short intervals favours the work.Facility in pairing animals, or their inhabiting aconfined area, is advantageous as a check tofree crossing. Whenever and wherever selectionis not practised, distinct races are not formedwithin the same country. When any one part ofthe body or one quality is not attended to, itremains either unchanged or varies in a fluc-tuating manner, whilst at the same time otherparts and other qualities may become perma-nently and greatly modified. But from the ten-dency to reversion and to continued variability,those parts or organs which are now under-going rapid improvement through selection,are likewise found to vary much. Consequentlyhighly-bred animals when neglected soon de-

generate; but we have no reason to believe thatthe effects of long-continued selection would, ifthe conditions of life remained the same, besoon and completely lost.

Man always tends to go to an extreme point inthe selection, whether methodical or uncons-cious, of all useful and pleasing qualities. Thisis an important principle, as it leads to conti-nued divergence, and in some rare cases to con-vergence of character. The possibility of conti-nued divergence rests on the tendency in eachpart or organ to go on varying in the samemanner in which it has already varied; and thatthis occurs, is proved by the steady and gradualimprovement of many animals and plants du-ring lengthened periods. The principle of di-vergence of character, combined with the ne-glect and final extinction of all previous, less-valued, and intermediate varieties, explains theamount of difference and the distinctness ofour several races. Although we may have rea-

ched the utmost limit to which certain charac-ters can be modified, yet we are far fromhaving reached, as we have good reason to be-lieve, the limit in the majority of cases. Finally,from the difference between selection as carriedon by man and by nature, we can understandhow it is that domestic races often, though byno means always, differ in general aspect fromclosely allied natural species.

Throughout this chapter and elsewhere I havespoken of selection as the paramount power,yet its action absolutely depends on what we inour ignorance call spontaneous or accidentalvariability. Let an architect be compelled tobuild an edifice with uncut stones, fallen from aprecipice. The shape of each fragment may becalled accidental; yet the shape of each has beendetermined by the force of gravity, the natureof the rock, and the slope of the precipice,—events and circumstances, all of which dependon natural laws; but there is no relation bet-

ween these laws and the purpose for whicheach fragment is used by the builder. In thesame manner the variations of each creature aredetermined by fixed and immutable laws; butthese bear no relation to the living structurewhich is slowly built up through the power ofselection, whether this be natural or artificialselection.

If our architect succeeded in rearing a nobleedifice, using the rough wedge- shaped frag-ments for the arches, the longer stones for thelintels, and so forth, we should admire his skilleven in a higher degree than if he had usedstones shaped for the purpose. So it is with se-lection, whether applied by man or by nature;for although variability is indispensably neces-sary, yet, when we look at some highly com-plex and excellently adapted organism, variabi-lity sinks to a quite subordinate position in im-portance in comparison with selection, in thesame manner as the shape of each fragment

used by our supposed architect is unimportantin comparison with his skill.

CHAPTER 2.XXII.

CAUSES OF VARIABILITY.

VARIABILITY DOES NOT NECESSARILYACCOMPANY REPRODUCTION. CAUSESASSIGNED BY VARIOUS AUTHORS. INDI-VIDUAL DIFFERENCES. VARIABILITY OFEVERY KIND DUE TO CHANGED CONDI-TIONS OF LIFE. ON THE NATURE OF SUCHCHANGES. CLIMATE, FOOD, EXCESS OFNUTRIMENT. SLIGHT CHANGES SUFFI-CIENT. EFFECTS OF GRAFTING ON THEVARIABILITY OF SEEDLING-TREES. DO-MESTIC PRODUCTIONS BECOME HABI-TUATED TO CHANGED CONDITIONS. ONTHE ACCUMULATIVE ACTION OF CHAN-GED CONDITIONS. CLOSE INTERBREE-

DING AND THE IMAGINATION OF THEMOTHER SUPPOSED TO CAUSE VARIABI-LITY. CROSSING AS A CAUSE OF THE AP-PEARANCE OF NEW CHARACTERS. VA-RIABILITY FROM THE COMMINGLING OFCHARACTERS AND FROM REVERSION. ONTHE MANNER AND PERIOD OF ACTION OFTHE CAUSES WHICH EITHER DIRECTLY,OR INDIRECTLY THROUGH THE REPRO-DUCTIVE SYSTEM, INDUCE VARIABILITY.

We will now consider, as far as we can, the cau-ses of the almost universal variability of ourdomesticated productions. The subject is anobscure one; but it may be useful to probe ourignorance. Some authors, for instance Dr. Pros-per Lucas, look at variability as a necessarycontingent on reproduction, and as much anaboriginal law as growth or inheritance. Othershave of late encouraged, perhaps unintentiona-lly, this view by speaking of inheritance andvariability as equal and antagonistic principles.Pallas maintained, and he has had some follo-

wers, that variability depends exclusively onthe crossing of primordially distinct forms. Ot-her authors attribute variability to an excess offood, and with animals to an excess relativelyto the amount of exercise taken, or again to theeffects of a more genial climate. That these cau-ses are all effective is highly probable. But wemust, I think, take a broader view, and conclu-de that organic beings, when subjected duringseveral generations to any change whatever intheir conditions, tend to vary; the kind of varia-tion which ensues depending in most cases in afar higher degree on the nature or constitutionof the being, than on the nature of the changedconditions.

Those authors who believe that it is a law ofnature that each individual should differ insome slight degree from every other, maymaintain, apparently with truth, that this is thefact, not only with all domesticated animalsand cultivated plants, but likewise with all or-

ganic beings in a state of nature. The Laplanderby long practice knows and gives a name toeach reindeer, though, as Linnaeus remarks, "todistinguish one from another among such mul-titudes was beyond my comprehension, forthey were like ants on an anthill." In Germanyshepherds have won wagers by recognisingeach sheep in a flock of a hundred, which theyhad never seen until the previous fortnight.This power of discrimination, however, is asnothing compared to that which some floristshave acquired. Verlot mentions a gardener whocould distinguish 150 kinds of camellia, whennot in flower; and it has been positively asser-ted that the famous old Dutch florist Voorhelm,who kept above 1200 varieties of the hyacinth,was hardly ever deceived in knowing each va-riety by the bulb alone. Hence we must conclu-de that the bulbs of the hyacinth and the bran-ches and leaves of the camellia, though appea-ring to an unpractised eye absolutely undistin-guishable, yet really differ. (22/1. 'Des Jacint-

hes' etc. Amsterdam 1768 page 43; Verlot 'DesVarietes' etc. page 86. On the reindeer see Lin-naeus 'Tour in Lapland' translated by Sir J.E.Smith volume 1 page 314. The statement in re-gard to German shepherds is given on the aut-hority of Dr. Weinland.)

As Linnaeus has compared the reindeer innumber to ants, I may add that each ant knowsits fellow of the same community. Several timesI carried ants of the same species (Formica rufa)from one ant-hill to another, inhabited appa-rently by tens of thousands of ants; but thestrangers were instantly detected and killed. Ithen put some ants taken from a very large nestinto a bottle strongly perfumed with assafoeti-da, and after an interval of twenty- four hoursreturned them to their home; they were at firstthreatened by their fellows, but were soon re-cognised and allowed to pass. Hence each antcertainly recognised, independently of odour,its fellow; and if all the ants of the same com-

munity have not some countersign or watch-word, they must present to each other's sensessome distinguishable character.

The dissimilarity of brothers or sisters of thesame family, and of seedlings from the samecapsule, may be in part accounted for by theunequal blending of the characters of the twoparents, and by the more or less complete reco-very through reversion of ancestral characterson either side; but we thus only push the diffi-culty further back in time, for what made theparents or their progenitors different? Hencethe belief (22/2. Muller 'Physiology' Englishtranslation, volume 2 page 1662. With respectto the similarity of twins in constitution, Dr.William Ogle has given me the following extra-ct from Professor Trousseau's Lectures 'Clini-que Medicale' tome 1 page 523, in which a cu-rious case is recorded:—"J'ai donne mes soins adeux freres jumeaux, tous deux si extraordinai-rement ressemblants qu'il m'etait impossible de

les reconnaitre, a moin de les voir l'un a cote del'autre. Cette ressemblance physique s'etendaitplus loin: ils avaient, permettez-moi l'expres-sion, une similitude pathologique plus remar-quable encore. Ainsi l'un d'eux que je voyaisaux neothermes a Paris malade d'une ophthal-mie rhumatismale me disait, 'En ce momentmon frere doit avoir une ophthalmie comme lamienne;' et comme je m'etais recrie, il me mon-trait quelques jours apres une lettre qu'il venaitde recevoir de ce frere alors a Vienne, et qui luiecrivait en effet—'J'ai mon ophthalmie, tu doisavoir la tienne.' Quelque singulier que cecipuisse paraitre, le fait n'en est pas moins exact:on ne me l'a pas raconte, je l'ai vu, et j'en ai vud'autres analogues dans ma pratique. Ces deuxjumeaux etaient aussi tous deux asthmatiques,et asthmatiques a un effroyable degre. Origi-naires de Marseille, ils n'ont jamais pu demeu-rer dans cette ville, ou leurs interets les appe-laient souvent, sans etre pris de leurs acces;jamais ils n'en eprouvaient a Paris. Bien mieux,

il leur suffisait de gagner Toulon pour etre gue-ris de leurs attaques de Marseille. Voyageantsans cesse et dans tous pays pour leurs affaires,ils avaient remarque que certaines localites leuretaient funestes, que dans d'autres ils etaientexempts de tout phenomene d'oppression.")that an innate tendency to vary exists, inde-pendently of external differences, seems at firstsight probable. But even the seeds nurtured inthe same capsule are not subjected to absolute-ly uniform conditions, as they draw their nou-rishment from different points; and we shallsee in a future chapter that this difference so-metimes suffices to affect the character of thefuture plant. The greater dissimilarity of thesuccessive children of the same family in com-parison with twins, which often resemble eachother in external appearance, mental disposi-tion, and constitution, in so extraordinary amanner, apparently proves that the state of theparents at the exact period of conception, or thenature of the subsequent embryonic develop-

ment, has a direct and powerful influence onthe character of the offspring. Nevertheless,when we reflect on the individual differencesbetween organic beings in a state of nature, asshown by every wild animal knowing its mate;and when we reflect on the infinite diversity ofthe many varieties of our domesticated produc-tions, we may well be inclined to exclaim,though falsely as I believe, that Variability mustbe looked at as an ultimate fact, necessarilycontingent on reproduction.

Those authors who adopt this latter viewwould probably deny that each separate varia-tion has its own proper exciting cause. Alt-hough we can seldom trace the precise relationbetween cause and effect, yet the considera-tions presently to be given lead to the conclu-sion that each modification must have its owndistinct cause, and is not the result of what weblindly call accident. The following strikingcase has been communicated to me by Dr. Wi-

lliam Ogle. Two girls, born as twins, and in allrespects extremely alike, had their little fingerson both hands crooked; and in both childrenthe second bicuspid tooth of the second denti-tion on the right side in the upper jaw was mis-placed; for, instead of standing in a line withthe others, it grew from the roof of the mouthbehind the first bicuspid. Neither the parentsnor any other members of the family wereknown to have exhibited any similar peculiari-ty; but a son of one of these girls had the sametooth similarly misplaced. Now, as both thegirls were affected in exactly the same manner,the idea of accident is at once excluded: and weare compelled to admit that there must haveexisted some precise and sufficient causewhich, if it had occurred a hundred times,would have given crooked fingers and mispla-ced bicuspid teeth to a hundred children. It isof course possible that this case may have beendue to reversion to some long-forgotten proge-nitor, and this would much weaken the value

of the argument. I have been led to think of theprobability of reversion, from having been toldby Mr. Galton of another case of twin girls bornwith their little fingers slightly crooked, whichthey inherited from their maternal grandmot-her.

We will now consider the general arguments,which appear to me to have great weight, infavour of the view that variations of all kindsand degrees are directly or indirectly caused bythe conditions of life to which each being, andmore especially its ancestors, have been expo-sed.

No one doubts that domesticated productionsare more variable than organic beings whichhave never been removed from their naturalconditions. Monstrosities graduate so insensi-bly into mere variations that it is impossible toseparate them; and all those who have studiedmonstrosities believe that they are far commo-ner with domesticated than with wild animals

and plants (22/3. Isid. Geoffroy St.-Hilaire'Hist. des Anomalies' tome 3 page 352; Moquin-Tandon 'Teratologie Vegetale' 1841 page 115.);and in the case of plants, monstrosities wouldbe equally noticeable in the natural as in thecultivated state. Under nature, the individualsof the same species are exposed to nearly uni-form conditions, for they are rigorously kept totheir proper places by a host of competing ani-mals and plants; they have, also, long beenhabituated to their conditions of life; but it can-not be said that they are subject to quite uni-form conditions, and they are liable to a certainamount of variation. The circumstances underwhich our domestic productions are reared arewidely different: they are protected from com-petition; they have not only been removed fromtheir natural conditions and often from theirnative land, but they are frequently carriedfrom district to district, where they are treateddifferently, so that they rarely remain duringany considerable length of time exposed to clo-

sely similar conditions. In conformity with this,all our domesticated productions, with the ra-rest exceptions, vary far more than natural spe-cies. The hive-bee, which feeds itself and fo-llows in most respects its natural habits of life,is the least variable of all domesticated animals,and probably the goose is the next least varia-ble; but even the goose varies more than almostany wild bird, so that it cannot be affiliatedwith perfect certainty to any natural species.Hardly a single plant can be named, which haslong been cultivated and propagated by seed,that is not highly variable; common rye (Secalecereale) has afforded fewer and less markedvarieties than almost any other cultivated plant(22/4. Metzger 'Die Getreidarten' 1841 s. 39.);but it may be doubted whether the variations ofthis, the least valuable of all our cereals, havebeen closely observed.

Bud-variation, which was fully discussed in aformer chapter, shows us that variability may

be quite independent of seminal reproduction,and likewise of reversion to long-lost ancestralcharacters. No one will maintain that the sud-den appearance of a moss-rose on a Provence-rose is a return to a former state, for mossinessof the calyx has been observed in no naturalspecies; the same argument is applicable tovariegated and laciniated leaves; nor can theappearance of nectarines on peach-trees be ac-counted for on the principle of reversion. Butbud-variations more immediately concern us,as they occur far more frequently on plantswhich have been highly cultivated during alength of time, than on other and less highlycultivated plants; and very few well- markedinstances have been observed with plants gro-wing under strictly natural conditions. I havegiven one instance of an ash-tree growing in agentleman's pleasure-grounds; and occasiona-lly there may be seen, on beech and other trees,twigs leafing at a different period from the ot-her branches. But our forest trees in England

can hardly be considered as living under stric-tly natural conditions; the seedlings are raisedand protected in nursery-grounds, and mustoften be transplanted into places where wildtrees of the kind would not naturally grow. Itwould be esteemed a prodigy if a dog-rosegrowing in a hedge produced by bud-variationa moss-rose, or a wild bullace or wild cherry-tree yielded a branch bearing fruit of a differentshape and colour from the ordinary fruit. Theprodigy would be enhanced if these varyingbranches were found capable of propagation,not only by grafts, but sometimes by seed; yetanalogous cases have occurred with many ofour highly cultivated trees and herbs.

These several considerations alone render itprobable that variability of every kind is direc-tly or indirectly caused by changed conditionsof life. Or, to put the case under another pointof view, if it were possible to expose all the in-dividuals of a species during many generations

to absolutely uniform conditions of life, therewould be no variability.

ON THE NATURE OF THE CHANGES INTHE CONDITIONS OF LIFE WHICH INDUCEVARIABILITY.

From a remote period to the present day, underclimates and circumstances as different as it ispossible to conceive, organic beings of all kinds,when domesticated or cultivated, have varied.We see this with the many domestic races ofquadrupeds and birds belonging to differentorders, with goldfish and silkworms, withplants of many kinds, raised in various quartersof the world. In the deserts of northern Africathe date-palm has yielded thirty-eight varieties;in the fertile plains of India it is notorious howmany varieties of rice and of a host of otherplants exist; in a single Polynesian island, twen-ty-four varieties of the bread-fruit, the samenumber of the banana, and twenty-two varie-

ties of the arum, are cultivated by the natives;the mulberry- tree in India and Europe hasyielded many varieties serving as food for thesilkworm; and in China sixty-three varieties ofthe bamboo are used for various domestic pur-poses. (22/5. On the date-palm see Vogel 'An-nals and Mag. of Nat. Hist.' 1854 page 460. OnIndian varieties Dr. F. Hamilton 'Transact.Linn. Soc.' volume 14 page 296. On the varietiescultivated in Tahiti see Dr. Bennett in Loudon's'Mag. of N. Hist.' volume 5 1832 page 484. AlsoEllis 'Polynesian Researches' volume 1 pages370, 375. On twenty varieties of the Pandanusand other trees in the Marianne Island see'Hooker's Miscellany' volume 1 page 308. Onthe bamboo in China see Huc 'Chinese Empire'volume 2 page 307.) These facts, and innume-rable others which could be added, indicatethat a change of almost any kind in the condi-tions of life suffices to cause variability—different changes acting on different organisms.

Andrew Knight (22/6. 'Treatise on the Cultureof the Apple' etc. page 3.) attributed the varia-tion of both animals and plants to a more abun-dant supply of nourishment, or to a more fa-vourable climate, than that natural to the spe-cies. A more genial climate, however, is farfrom necessary; the kidney- bean, which is of-ten injured by our spring frosts, and peaches,which require the protection of a wall, havevaried much in England, as has the orange-treein northern Italy, where it is barely able to exist.(22/7. Gallesio 'Teoria della Riproduzione Veg.'page 125.) Nor can we overlook the fact,though not immediately connected with ourpresent subject, that the plants and shells of theArctic regions are eminently variable. (22/8.See Dr. Hooker's Memoir on Arctic Plants in'Linn. Transact.' volume 23 part 2. Mr. Wood-ward, and a higher authority cannot be quoted,speaks of the Arctic mollusca in his 'Rudimen-tary Treatise' 1856 page 355 as remarkably sub-ject to variation.) Moreover, it does not appear

that a change of climate, whether more or lessgenial, is one of the most potent causes of va-riability; for in regard to plants Alph. De Can-dolle, in his 'Geographie Botanique' repeatedlyshows that the native country of a plant, wherein most cases it has been longest cultivated, isthat where it has yielded the greatest numberof varieties.

It is doubtful whether a change in the nature ofthe food is a potent cause of variability. Scarce-ly any domesticated animal has varied morethan the pigeon or the fowl, but their food, es-pecially that of highly-bred pigeons, is genera-lly the same. Nor can our cattle and sheep havebeen subjected to any great change in this res-pect. But in all these cases the food probably ismuch less varied in kind than that which wasconsumed by the species in its natural state.(22/9. Bechstein in his 'Naturgeschichte derStubenvogel' 1840 s. 238, has some good re-marks on this subject. He states that his canary-

birds varied in colour, though kept on uniformfood.)

Of all the causes which induce variability, ex-cess of food, whether or not changed in nature,is probably the most powerful. This view washeld with regard to plants by Andrew Knight,and is now held by Schleiden, more especiallyin reference to the inorganic elements of thefood. (22/10. 'The Plant' by Schleiden transla-ted by Henfrey 1848 page 169. See also Alex.Braun in 'Bot. Memoirs' Ray Soc. 1853 page313.) In order to give a plant more food it suffi-ces in most cases to grow it separately, and thusprevent other plants robbing its roots. It is sur-prising, as I have often seen, how vigorouslyour common wild species flourish when plan-ted by themselves, though not in highly manu-red land; separate growth is, in fact, the firststep in cultivation. We see the converse of thebelief that excess of food induces variability inthe following statement by a great raiser of

seeds of all kinds (22/11. Messrs. Hardy andSon of Maldon in 'Gardener's Chronicle' 1856page 458. Carriere 'Production et Fixation desVarietes' 1865 page 31.): "It is a rule invariablywith us, when we desire to keep a true stock ofany one kind of seed, to grow it on poor landwithout dung; but when we grow for quantity,we act contrary, and sometimes have dearly torepent of it." According also to Carriere, whohas had great experience with flower-gardenseeds, "On remarque en general les plantes devigeur moyenne sont celles qui conservent lemieux leurs caracteres."

In the case of animals the want of a properamount of exercise, as Bechstein remarked, hasperhaps played, independently of the directeffects of the disuse of any particular organ, animportant part in causing variability. We cansee in a vague manner that, when the organisedand nutrient fluids of the body are not usedduring growth, or by the wear and tear of the

tissues, they will be in excess; and as growth,nutrition, and reproduction are intimatelyallied processes, this superfluity might disturbthe due and proper action of the reproductiveorgans, and consequently affect the character ofthe future offspring. But it may be argued thatneither an excess of food nor a superfluity inthe organised fluids of the body necessarilyinduces variability. The goose and the turkeyhave been well fed for many generations, yethave varied very little. Our fruit-trees and culi-nary plants, which are so variable, have beencultivated from an ancient period, and, thoughthey probably still receive more nutriment thanin their natural state, yet they must have recei-ved during many generations nearly the sameamount; and it might be thought that theywould have become habituated to the excess.Nevertheless, on the whole, Knight's view, thatexcess of food is one of the most potent causesof variability, appears, as far as I can judge,probable.

Whether or not our various cultivated plantshave received nutriment in excess, all havebeen exposed to changes of various kinds.Fruit-trees are grafted on different stocks, andgrown in various soils. The seeds of culinaryand agricultural plants are carried from placeto place; and during the last century the rota-tion of our crops and the manures used havebeen greatly changed.

Slight changes of treatment often suffice to in-duce variability. The simple fact of almost allour cultivated plants and domesticated animalshaving varied in all places and at all times,leads to this conclusion. Seeds taken fromcommon English forest-trees, grown undertheir native climate, not highly manured orotherwise artificially treated, yield seedlingswhich vary much, as may be seen in every ex-tensive seed-bed. I have shown in a formerchapter what a number of well-marked andsingular varieties the thorn (Crataegus oxy-

cantha) has produced: yet this tree has beensubjected to hardly any cultivation. In Staf-fordshire I carefully examined a large numberof two British plants, namely Geraniumphaeum and pyrenaicum, which have neverbeen highly cultivated. These plants had spreadspontaneously by seed from a common gardeninto an open plantation; and the seedlings va-ried in almost every single character, both intheir flower and foliage, to a degree which Ihave never seen exceeded; yet they could nothave been exposed to any great change in theirconditions.

With respect to animals, Azara has remarkedwith much surprise (22/12. 'Quadrupedes duParaguay' 1801 tome 2 page 319.) that, whilstthe feral horses on the Pampas are always ofone of three colours, and the cattle always of auniform colour, yet these animals, when bredon the unenclosed estancias, though kept in astate which can hardly be called domesticated,

and apparently exposed to almost identicallythe same conditions as when they are feral,nevertheless display a great diversity of colour.So again in India several species of fresh-waterfish are only so far treated artificially, that theyare reared in great tanks; but this small changeis sufficient to induce much variability. (22/13.M'Clelland on Indian Cyprinidae 'Asiatic Re-searches' volume 19 part 2 1839 pages 266, 268,313.)

Some facts on the effects of grafting, in regardto the variability of trees, deserve attention.Cabanis asserts that when certain pears aregrafted on the quince, their seeds yield a grea-ter number of varieties than do the seeds of thesame variety of pear when grafted on the wildpear. (22/14. Quoted by Sageret 'Pom. Phys.'1830 page 43. This statement, however, is notbelieved by Decaisne.) But as the pear andquince are distinct species, though so closelyrelated that the one can be readily grafted and

succeeds admirably on the other, the fact ofvariability being thus caused is not surprising;as we are here enabled to see the cause, namely,the very different nature of the stock and graft.Several North American varieties of the plumand peach are well known to reproduce them-selves truly by seed; but Downing asserts(22/15. 'The Fruits of America' 1845 page 5.),"that when a graft is taken from one of thesetrees and placed upon another stock, this graf-ted tree is found to lose its singular property ofproducing the same variety by seed, and beco-mes like all other worked trees;"—that is, itsseedlings become highly variable. Another caseis worth giving: the Lalande variety of the wal-nut-tree leafs between April 20th and May 15th,and its seedlings invariably inherit the samehabit; whilst several other varieties of the wal-nut leaf in June. Now, if seedlings are raisedfrom the May-leafing Lalande variety, graftedon another May-leafing variety, though bothstock and graft have the same early habit of

leafing, yet the seedlings leaf at various times,even as late as the 5th of June. (22/16. M. Car-dan in 'Comptes Rendus' December 1848 quo-ted in 'Gardener's Chronicle' 1849 page 101.)Such facts as these are well fitted to show onwhat obscure and slight causes variability de-pends.

[I may here just allude to the appearance ofnew and valuable varieties of fruit-trees and ofwheat in woods and waste places, which at firstsight seems a most anomalous circumstance. InFrance a considerable number of the best pearshave been discovered in woods; and this hasoccurred so frequently, that Poiteau asserts that"improved varieties of our cultivated fruits ra-rely originate with nurserymen." (22/17. M.Alexis Jordan mentions four excellent pearsfound in woods in France, and alludes to others('Mem. Acad. de Lyon' tome 2 1852 page 159).Poiteau's remark is quoted in 'Gardener's Mag.'volume 4 1828 page 385. See 'Gardener's Chro-

nicle' 1862 page 335, for another case of a newvariety of the pear found in a hedge in France.Also for another case, see Loudon's 'Encyclop.of Gardening' page 901. Mr. Rivers has givenme similar information.) In England, on theother hand, no instance of a good pear havingbeen found wild has been recorded; and Mr.Rivers informs me that he knows of only oneinstance with apples, namely, the Bess Poole,which was discovered in a wood in Notting-hamshire. This difference between the twocountries may be in part accounted for by themore favourable climate of France, but chieflyfrom the great number of seedlings whichspring up there in the woods. I infer that this isthe case from a remark made by a French gar-dener (22/18. Duval 'Hist. du Poirier' 1849 page2.), who regards it as a national calamity thatsuch a number of pear-trees are periodically cutdown for firewood, before they have bornefruit. The new varieties which thus spring up inthe woods, though they cannot have received

any excess of nutriment, will have been expo-sed to abruptly changed conditions, but whet-her this is the cause of their production is verydoubtful. These varieties, however, are proba-bly all descended (22/19. I infer that this is thefact from Van Mons' statement ('Arbres Frui-tiers' 1835 tome 1 page 446) that he finds in thewoods seedlings resembling all the chief culti-vated races of both the pear and apple. VanMons, however, looked at these wild varietiesas aboriginal species.) from old cultivated kindsgrowing in adjoining orchards— a circumstan-ce which will account for their variability; andout of a vast number of varying trees there willalways be a good chance of the appearance of avaluable kind. In North America, where fruit-trees frequently spring up in waste places, theWashington pear was found in a hedge, andthe Emperor peach in a wood. (22/20. Downing'Fruit-trees of North America' page 422; Foleyin 'Transact. Hort. Soc.' volume 6 page 412.)

With respect to wheat, some writers have spo-ken (22/21. 'Gardener's Chronicle' 1847 page244.) as if it were an ordinary event for newvarieties to be found in waste places; the Fen-ton wheat was certainly discovered growing ona pile of basaltic detritus in a quarry, but insuch a situation the plant would probably re-ceive a sufficient amount of nutriment. TheChidham wheat was raised from an ear foundON a hedge; and Hunter's wheat was discove-red BY the roadside in Scotland, but it is notsaid that this latter variety grew where it wasfound. (22/22. 'Gardener's Chronicle' 1841 page383; 1850 page 700; 1854 page 650.)]

Whether our domestic productions would everbecome so completely habituated to the condi-tions under which they now live, as to ceasevarying, we have no sufficient means for jud-ging. But, in fact, our domestic productions arenever exposed for a great length of time to uni-form conditions, and it is certain that our most

anciently cultivated plants, as well as animals,still go on varying, for all have recently under-gone marked improvement. In some few cases,however, plants have become habituated tonew conditions. Thus, Metzger, who cultivatedin Germany during many years numerous va-rieties of wheat, brought from different coun-tries (22/23. 'Die Getreidearten' 1843 s. 66, 116,117.), states that some kinds were at first ex-tremely variable, but gradually, in one instanceafter an interval of twenty-five years, becameconstant; and it does not appear that this resul-ted from the selection of the more constantforms.

ON THE ACCUMULATIVE ACTION OFCHANGED CONDITIONS OF LIFE.

We have good grounds for believing that theinfluence of changed conditions accumulates,so that no effect is produced on a species untilit has been exposed during several generations

to continued cultivation or domestication. Uni-versal experience shows us that when new flo-wers are first introduced into our gardens theydo not vary; but ultimately all, with the rarestexceptions, vary to a greater or less extent. In afew cases the requisite number of generations,as well as the successive steps in the progress ofvariation, have been recorded, as in the oftenquoted instance of the Dahlia. (22/24. Sabine in'Hort. Transact.' volume 3 page 225; Bronn'Geschichte der Natur' b. 2 s. 119.) After severalyears' culture the Zinnia has only lately (1860)begun to vary in any great degree. "In the firstseven or eight years of high cultivation, theSwan River daisy (Brachycome iberidifolia)kept to its original colour; it then varied intolilac and purple and other minor shades."(22/25. 'Journal of Horticulture' 1861 page 112;on Zinnia 'Gardener's Chronicle' 1860 page852.) Analogous facts have been recorded withthe Scotch rose. In discussing the variability ofplants several experienced horticulturists have

spoken to the same general effect. Mr. Salter(22/26. 'The Chrysanthemum, its History, etc.'1865 page 3.) remarks, "Every one knows thatthe chief difficulty is in breaking through theoriginal form and colour of the species, andevery one will be on the look-out for any natu-ral sport, either from seed or branch; that beingonce obtained, however trifling the change maybe, the result depends upon himself." M. deJonghe, who has had so much success in raisingnew varieties of pears and strawberries (22/27.'Gardener's Chronicle' 1855 page 54; 'Journal ofHorticulture' May 9, 1865 page 363.), remarkswith respect to the former, "There is anotherprinciple, namely, that the more a type has en-tered into a state of variation, the greater is itstendency to continue doing so; and the more ithas varied from the original type, the more it isdisposed to vary still farther." We have, indeed,already discussed this latter point when trea-ting of the power which man possesses,through selection, of continually augmenting in

the same direction each modification; for thispower depends on continued variability of thesame general kind. The most celebrated horti-culturist in France, namely, Vilmorin (22/28.Quoted by Verlot 'Des Varietes' etc. 1865 page28.), even maintains that, when any particularvariation is desired, the first step is to get theplant to vary in any manner whatever, and togo on selecting the most variable individuals,even though they vary in the wrong direction;for the fixed character of the species being oncebroken, the desired variation will sooner orlater appear.

As nearly all our animals were domesticated atan extremely remote epoch, we cannot, of cour-se, say whether they varied quickly or slowlywhen first subjected to new conditions. But Dr.Bachman (22/29. 'Examination of the Characte-ristics of Genera and Species' Charleston 1855page 14.) states that he has seen turkeys raisedfrom the eggs of the wild species lose their me-

tallic tints and become spotted with white inthe third generation. Mr. Yarrell many yearsago informed me that the wild ducks bred onthe ponds in St. James's Park, which had neverbeen crossed, as it is believed, with domesticducks, lost their true plumage after a few gene-rations. An excellent observer (22/30. Mr.Hewitt 'Journal of Hort.' 1863 page 39.), whohas often reared ducks from the eggs of thewild bird, and who took precautions that thereshould be no crossing with domestic breeds,has given, as previously stated, full details onthe changes which they gradually undergo. Hefound that he could not breed these wild duckstrue for more than five or six generations, "asthey then proved so much less beautiful. Thewhite collar round the neck of the mallard be-came much broader and more irregular, andwhite feathers appeared in the ducklings'wings." They increased also in size of body;their legs became less fine, and they lost theirelegant carriage. Fresh eggs were then procu-

red from wild birds; but again the same resultfollowed. In these cases of the duck and turkeywe see that animals, like plants, do not departfrom their primitive type until they have beensubjected during several generations to domes-tication. On the other hand, Mr. Yarrell infor-med me that the Australian dingos, bred in theZoological Gardens, almost invariably produ-ced in the first generation puppies marked withwhite and other colours; but, these introduceddingos had probably been procured from thenatives, who keep them in a semi-domesticatedstate. It is certainly a remarkable fact that chan-ged conditions should at first produce, as far aswe can see, absolutely no effect; but that theyshould subsequently cause the character of thespecies to change. In the chapter on pangenesisI shall attempt to throw a little light on this fact.

Returning now to the causes which are suppo-sed to induce variability. Some authors (22/31.Devay 'Mariages Consanguins' pages 97, 125. In

conversation I have found two or three natura-lists of the same opinion.) believe that closeinterbreeding gives this tendency, and leads tothe production of monstrosities. In the seven-teenth chapter some few facts were advanced,showing that monstrosities are, as it appears,occasionally thus induced; and there can be nodoubt that close interbreeding causes lessenedfertility and a weakened constitution; hence itmay lead to variability: but I have not sufficientevidence on this head. On the other hand, closeinterbreeding, if not carried to an injurious ex-treme, far from causing variability, tends to fixthe character of each breed.

It was formerly a common belief, still held bysome persons, that the imagination of the mot-her affects the child in the womb. (22/32. Mu-ller has conclusively argued against this belief,'Elements of Phys.' English translation volume2 1842 page 1405.) This view is evidently notapplicable to the lower animals, which lay

unimpregnated eggs, or to plants. Dr. WilliamHunter, in the last century, told my father thatduring many years every woman in a largeLondon Lying-in Hospital was asked before herconfinement whether anything had speciallyaffected her mind, and the answer was writtendown; and it so happened that in no one ins-tance could a coincidence be detected betweenthe woman's answer and any abnormal structu-re; but when she knew the nature of the struc-ture, she frequently suggested some fresh cau-se. The belief in the power of the mother's ima-gination may perhaps have arisen from thechildren of a second marriage resembling theprevious father, as certainly sometimes occurs,in accordance with the facts given in the ele-venth chapter.

CROSSING AS A CAUSE OF VARIABILI-TY.

In an early part of this chapter it was stated thatPallas (22/33. 'Act. Acad. St. Petersburg' 1780part 2 page 84 etc.) and a few other naturalistsmaintain that variability is wholly due to cros-sing. If this means that new characters neverspontaneously appear in our domestic races,but that they are all directly derived from cer-tain aboriginal species, the doctrine is little lessthan absurd; for it implies that animals likeItalian greyhounds, pug-dogs, bull-dogs, pou-ter and fantail pigeons, etc., were able to existin a state of nature. But the doctrine may meansomething widely different, namely, that thecrossing of distinct species is the sole cause ofthe first appearance of new characters, and thatwithout this aid man could not have formed hisvarious breeds. As, however, new charactershave appeared in certain cases by bud- varia-tion, we may conclude with certainty that cros-sing is not necessary for variability. It is, mo-reover, certain that the breeds of various ani-mals, such as of the rabbit, pigeon, duck, etc.,

and the varieties of several plants, are the mo-dified descendants of a single wild species. Ne-vertheless, it is probable that the crossing oftwo forms, when one or both have long beendomesticated or cultivated, adds to the variabi-lity of the offspring, independently of thecommingling of the characters derived from thetwo parent-forms; and this implies that newcharacters actually arise. But we must not for-get the facts advanced in the thirteenth chapter,which clearly prove that the act of crossing of-ten leads to the reappearance or reversion oflong- lost characters; and in most cases it wouldbe impossible to distinguish between the reap-pearance of ancient characters and the first ap-pearance of absolutely new characters. Practica-lly, whether new or old, they would be new tothe breed in which they reappeared.

[Gartner declares (22/34. 'Bastarderzeugung' s.249, 255, 295.), and his experience is of the hig-hest value on such a point, that, when he cros-

sed native plants which had not been cultiva-ted, he never once saw in the offspring any newcharacter; but that from the odd manner inwhich the characters derived from the parentswere combined, they sometimes appeared as ifnew. When, on the other hand, he crossed cul-tivated plants, he admits that new charactersoccasionally appeared, but he is strongly incli-ned to attribute their appearance to ordinaryvariability, not in any way to the cross. An op-posite conclusion, however, appears to me themore probable. According to Kolreuter, hybridsin the genus Mirabilis vary almost infinitely,and he describes new and singular charactersin the form of the seeds, in the colour of theanthers, in the cotyledons being of immensesize, in new and highly peculiar odours, in theflowers expanding early in the season, and intheir closing at night. With respect to one lot ofthese hybrids, he remarks that they presentedcharacters exactly the reverse of what mighthave been expected from their parentage.

(22/35. 'Nova Acta, St. Petersburg' 1794 page378; 1795 pages 307, 313, 316; 1787 page 407.)

Prof. Lecoq (22/36. 'De la Fecondation' 1862page 311.) speaks strongly to the same effect inregard to this same genus, and asserts that ma-ny of the hybrids from Mirabilis jalapa andmultiflora might easily be mistaken for distinctspecies, and adds that they differed in a greaterdegree than the other species of the genus, fromM. jalapa. Herbert, also, has described (22/37.'Amaryllidaceae' 1837 page 362.) certain hybridRhododendrons as being "as UNLIKE ALLOTHERS in foliage, as if they had been a sepa-rate species." The common experience of flori-culturists proves that the crossing and recros-sing of distinct but allied plants, such as thespecies of Petunia, Calceolaria, Fuchsia, Verbe-na, etc., induces excessive variability; hence theappearance of quite new characters is probable.M. Carriere (22/38. Abstracted in 'Gardener'sChronicle' 1860 page 1081.) has lately discussed

this subject: he states that Erythrina cristagallihad been multiplied by seed for many years,but had not yielded any varieties: it was thencrossed with the allied E. herbacea, and "theresistance was now overcome, and varietieswere produced with flowers of extremely diffe-rent size, form, and colour."

From the general and apparently well-foundedbelief that the crossing of distinct species, besi-des commingling their characters, adds greatlyto their variability, it has probably arisen thatsome botanists have gone so far as to maintain(22/39. This was the opinion of the elder DeCandolle, as quoted in 'Dic. Class. d'Hist. Nat.'tome 8 page 405. Puvis in his work 'De la De-generation' 1837 page 37, has discussed thissame point.) that, when a genus includes only asingle species, this when cultivated never va-ries. The proposition made so broadly cannotbe admitted; but it is probably true that thevariability of monotypic genera when cultiva-

ted is generally less than that of genera inclu-ding numerous species, and this quite inde-pendently of the effects of crossing. I haveshown in my 'Origin of Species' that the speciesbelonging to small genera generally yield a lessnumber of varieties in a state of nature thanthose belonging to large genera. Hence the spe-cies of small genera would, it is probable, pro-duce fewer varieties under cultivation than thealready variable species of larger genera.

Although we have not at present sufficient evi-dence that the crossing of species, which havenever been cultivated, leads to the appearanceof new characters, this apparently does occurwith species which have been already renderedin some degree variable through cultivation.Hence crossing, like any other change in theconditions of life, seems to be an element, pro-bably a potent one, in causing variability. Butwe seldom have the means of distinguishing, aspreviously remarked, between the appearance

of really new characters and the reappearanceof long-lost characters, evoked through the actof crossing. I will give an instance of the diffi-culty in distinguishing such cases. The speciesof Datura may be divided into two sections,those having white flowers with green stems,and those having purple flowers with brownstems: now Naudin (22/40. 'Comptes Rendus'Novembre 21, 1864 page 838.) crossed Daturalaevis and ferox, both of which belong to thewhite section, and raised from them 205 hy-brids. Of these hybrids, every one had brownstems and bore purple flowers; so that theyresembled the species of the other section of thegenus, and not their own two parents. Naudinwas so much astonished at this fact, that he wasled carefully to observe both parent- species,and he discovered that the pure seedlings of D.ferox, immediately after germination, had darkpurple stems, extending from the young rootsup to the cotyledons, and that this tint remai-ned ever afterwards as a ring round the base of

the stem of the plant when old. Now I haveshown in the thirteenth chapter that the reten-tion or exaggeration of an early character is sointimately related to reversion, that it evidentlycomes under the same principle. Hence proba-bly we ought to look at the purple flowers andbrown stems of these hybrids, not as new cha-racters due to variability, but as a return to theformer state of some ancient progenitor.

Independently of the appearance of new cha-racters from crossing, a few words may be ad-ded to what has been said in former chapterson the unequal combination and transmissionof the characters proper to the two parent-forms. When two species or races are crossed,the offspring of the first generation are genera-lly uniform, but those subsequently produceddisplay an almost infinite diversity of character.He who wishes, says Kolreuter (22/41. 'NovaActa, St. Petersburg' 1794 page 391.), to obtainan endless number of varieties from hybrids

should cross and recross them. There is alsomuch variability when hybrids or mongrels arereduced or absorbed by repeated crosses witheither pure parent-form: and a still higher de-gree of variability when three distinct species,and most of all when four species, are blendedtogether by successive crosses. Beyond thispoint Gartner (22/42. 'Bastarderzeugung' s. 507,516, 572.), on whose authority the foregoingstatements are made, never succeeded in effec-ting a union; but Max Wichura (22/43. 'DieBastardbefruchtung' etc. 1865 s. 24.) united sixdistinct species of willows into a single hybrid.The sex of the parent species affects in an inex-plicable manner the degree of variability ofhybrids; for Gartner (22/44. 'Bastarderzeugung's. 452, 507.) repeatedly found that when a hy-brid was used as a father and either one of thepure parent-species, or a third species, wasused as the mother, the offspring were morevariable than when the same hybrid was usedas the mother, and either pure parent or the

same third species as the father: thus seedlingsfrom Dianthus barbatus crossed by the hybridD. chinensi-barbatus were more variable thanthose raised from this latter hybrid fertilised bythe pure D. barbatus. Max Wichura (22/45. 'DieBastardbefruchtung' s. 56.) insists strongly onan analogous result with his hybrid willows.Again Gartner (22/46. 'Bastarderzeugung' s.423.) asserts that the degree of variability some-times differs in hybrids raised from reciprocalcrosses between the same two species; and herethe sole difference is, that the one species is firstused as the father and then as the mother. Onthe whole we see that, independently of theappearance of new characters, the variability ofsuccessive crossed generations is extremelycomplex, partly from the offspring partakingunequally of the characters of the two parent-forms, and more especially from their unequaltendency to revert to such characters or to thoseof more ancient progenitors.]

ON THE MANNER AND ON THE PERIODOF ACTION OF THE CAUSES WHICH IN-DUCE VARIABILITY.

This is an extremely obscure subject, and weneed here only consider, whether inheritedvariations are due to certain parts being actedon after they have been formed, or through thereproductive system being affected before theirformation; and in the former case at what pe-riod of growth or development the effect isproduced. We shall see in the two followingchapters that various agencies, such as anabundant supply of food, exposure to a diffe-rent climate, increased use or disuse of parts,etc., prolonged during several generations, cer-tainly modify either the whole organisation orcertain organs; and it is clear at least in the caseof bud-variation that the action cannot havebeen through the reproductive system.

[With respect to the part which the reproducti-ve system takes in causing variability, we have

seen in the eighteenth chapter that even slightchanges in the conditions of life have a remar-kable power in causing a greater or less degreeof sterility. Hence it seems not improbable thatbeings generated through a system so easilyaffected should themselves be affected, orshould fail to inherit, or inherit in excess, cha-racters proper to their parents. We know thatcertain groups of organic beings, but with ex-ceptions in each group, have their reproductivesystems much more easily affected by changedconditions than other groups; for instance, car-nivorous birds, more readily than carnivorousmammals, and parrots more readily than pi-geons; and this fact harmonises with the appa-rently capricious manner and degree in whichvarious groups of animals and plants vary un-der domestication.

Kolreuter (22/47. 'Dritte Fortsetzung' etc. 1766s. 85.) was struck with the parallelism betweenthe excessive variability of hybrids when cros-

sed and recrossed in various ways,—these hy-brids having their reproductive powers more orless affected,—and the variability of ancientlycultivated plants. Max Wichura (22/48. 'DieBastardbefruchtung' etc. 1865 s. 92: see also theRev. M.J. Berkeley on the same subject in 'Jour-nal of Royal Hort. Soc.' 1866 page 80.) has goneone step farther, and shows that with many ofour highly cultivated plants, such as the hya-cinth, tulip, auricula, snapdragon, potato, cab-bage, etc., which there is no reason to believehave been hybridised, the anthers contain ma-ny irregular pollen-grains in the same state asin hybrids. He finds also in certain wild forms,the same coincidence between the state of thepollen and a high degree of variability, as inmany species of Rubus; but in R. caesius andidaeus, which are not highly variable species,the pollen is sound. It is also notorious thatmany cultivated plants, such as the banana,pineapple, bread-fruit, and others previouslymentioned, have their reproductive organs so

seriously affected as to be generally quite steri-le; and when they do yield seed, the seedlings,judging from the large number of cultivatedraces which exist, must be variable in an extre-me degree. These facts indicate that there issome relation between the state of the repro-ductive organs and a tendency to variability;but we must not conclude that the relation isstrict. Although many of our highly cultivatedplants may have their pollen in a deterioratedcondition, yet, as we have previously seen, theyyield more seeds, and our anciently domestica-ted animals are more prolific, than the corres-ponding species in a state of nature. The pea-cock is almost the only bird which is believedto be less fertile under domestication than in itsnative state, and it has varied in a remarkablysmall degree. From these considerations itwould seem that changes in the conditions oflife lead either to sterility or to variability, or toboth; and not that sterility induces variability.On the whole it is probable that any cause af-

fecting the organs of reproduction would like-wise affect their product,—that is, the offspringthus generated.

The period of life at which the causes that in-duce variability act, is likewise an obscure sub-ject, which has been discussed by various aut-hors. (22/49. Dr. P. Lucas has given a history ofopinion on this subject 'Hered. Nat.' 1847 tome1 page 175.) In some of the cases, to be given inthe following chapter, of modifications fromthe direct action of changed conditions, whichare inherited, there can be no doubt that thecauses have acted on the mature or nearly ma-ture animal. On the other hand, monstrosities,which cannot be distinctly separated from les-ser variations, are often caused by the embryobeing injured whilst in the mother's womb or inthe egg. Thus I. Geoffroy Saint-Hilaire (22/50.'Hist. des Anomalies' tome 3 page 499.) assertsthat poor women who work hard during theirpregnancy, and the mothers of illegitimate chil-

dren troubled in their minds and forced to con-ceal their state, are far more liable to give birthto monsters than women in easy circumstances.The eggs of the fowl when placed upright orotherwise treated unnaturally frequently pro-duce monstrous chickens. It would, however,appear that complex monstrosities are inducedmore frequently during a rather late than du-ring a very early period of embryonic life; butthis may partly result from some one part,which has been injured during an early period,affecting by its abnormal growth other partssubsequently developed; and this would be lesslikely to occur with parts injured at a later pe-riod. (22/51. Ibid tome 3 pages 392, 502. Theseveral memoirs by M. Dareste hereafter refe-rred to are of special value on this whole sub-ject.) When any part or organ becomes mons-trous through abortion, a rudiment is generallyleft, and this likewise indicates that its deve-lopment had already commenced.

Insects sometimes have their antennae or legsin a monstrous condition, the larvae of whichdo not possess either antennae or legs; and inthese cases, as Quatrefages (22/52. See his inte-resting work 'Metamorphoses de l'Homme' etc.1862 page 129.) believes, we are enabled to seethe precise period at which the normal progressof development was troubled. But the nature ofthe food given to a caterpillar sometimes affectsthe colours of the moth, without the caterpillaritself being affected; therefore it seems possiblethat other characters in the mature insect mightbe indirectly modified through the larvae. The-re is no reason to suppose that organs whichhave been rendered monstrous have alwaysbeen acted on during their development; thecause may have acted on the organisation at amuch earlier stage. It is even probable that eit-her the male or female sexual elements, or both,before their union, may be affected in such amanner as to lead to modifications in organsdeveloped at a late period of life; in nearly the

same manner as a child may inherit from hisfather a disease which does not appear untilold age.

In accordance with the facts above given, whichprove that in many cases a close relation existsbetween variability and the sterility followingfrom changed conditions, we may concludethat the exciting cause often acts at the earliestpossible period, namely, on the sexual ele-ments, before impregnation has taken place.That an affection of the female sexual elementmay induce variability we may likewise infer asprobable from the occurrence of bud- varia-tions; for a bud seems to be the analogue of anovule. But the male element is apparently muchoftener affected by changed conditions, at leastin a visible manner, than the female element orovule and we know from Gartner's and Wichu-ra's statements that a hybrid used as the fatherand crossed with a pure species gives a greaterdegree of variability to the offspring, than does

the same hybrid when used as the mother. Las-tly, it is certain that variability may be transmit-ted through either sexual element, whether ornot originally excited in them, for Kolreuterand Gartner (22/53. 'Dritte Fortsetzung' etc. s.123; 'Bastarderzeugung' s. 249.) found thatwhen two species were crossed, if either onewas variable, the offspring were rendered va-riable.]

SUMMARY.

From the facts given in this chapter, we mayconclude that the variability of organic beingsunder domestication, although so general, isnot an inevitable contingent on life, but resultsfrom the conditions to which the parents havebeen exposed. Changes of any kind in the con-ditions of life, even extremely slight changes,often suffice to cause variability. Excess of nu-triment is perhaps the most efficient single exci-ting cause. Animals and plants continue to be

variable for an immense period after their firstdomestication; but the conditions to which theyare exposed never long remain quite constant.In the course of time they can be habituated tocertain changes, so as to become less variable;and it is possible that when first domesticatedthey may have been even more variable than atpresent. There is good evidence that the powerof changed conditions accumulates; so that two,three, or more generations must be exposed tonew conditions before any effect is visible. Thecrossing of distinct forms, which have alreadybecome variable, increases in the offspring thetendency to further variability, by the unequalcommingling of the characters of the two pa-rents, by the reappearance of long-lost charac-ters, and by the appearance of absolutely newcharacters. Some variations are induced by thedirect action of the surrounding conditions onthe whole organisation, or on certain parts alo-ne; other variations appear to be induced indi-rectly through the reproductive system being

affected, as we know is often the case with va-rious beings, which when removed from theirnatural conditions become sterile. The causeswhich induce variability act on the mature or-ganism, on the embryo, and, probably, on thesexual elements before impregnation has beeneffected.

CHAPTER 2.XXIII.

DIRECT AND DEFINITE ACTION OF THEEXTERNAL CONDITIONS OF LIFE.

SLIGHT MODIFICATIONS IN PLANTS FROMTHE DEFINITE ACTION OF CHANGEDCONDITIONS, IN SIZE, COLOUR, CHEMI-CAL PROPERTIES, AND IN THE STATE OFTHE TISSUES. LOCAL DISEASES. CONSPI-CUOUS MODIFICATIONS FROM CHANGEDCLIMATE OR FOOD, ETC. PLUMAGE OFBIRDS AFFECTED BY PECULIAR NUTRI-MENT, AND BY THE INOCULATION OF

POISON. LAND-SHELLS. MODIFICATIONSOF ORGANIC BEINGS IN A STATE OF NA-TURE THROUGH THE DEFINITE ACTIONOF EXTERNAL CONDITIONS. COMPARI-SON OF AMERICAN AND EUROPEANTREES. GALLS. EFFECTS OF PARASITICFUNGI. CONSIDERATIONS OPPOSED TOTHE BELIEF IN THE POTENT INFLUENCEOF CHANGED EXTERNAL CONDITIONS.PARALLEL SERIES OF VARIETIES. AMOUNTOF VARIATION DOES NOT CORRESPONDWITH THE DEGREE OF CHANGE IN THECONDITIONS. BUD-VARIATION. MONS-TROSITIES PRODUCED BY UNNATURALTREATMENT. SUMMARY.

If we ask ourselves why this or that characterhas been modified under domestication, weare, in most cases, lost in utter darkness. Manynaturalists, especially of the French school, at-tribute every modification to the "monde am-biant," that is, to changed climate, with all its

diversities of heat and cold, dampness anddryness, light and electricity, to the nature ofthe soil, and to varied kinds and amount offood. By the term definite action, as used in thischapter, I mean an action of such a nature that,when many individuals of the same variety areexposed during several generations to any par-ticular change in their conditions of life, all, ornearly all the individuals, are modified in thesame manner. The effects of habit, or of theincreased use and disuse of various organs,might have been included under this head; butit will be convenient to discuss this subject in aseparate chapter. By the term indefinite action Imean an action which causes one individual tovary in one way and another individual inanother way, as we often see with plants andanimals after they have been subjected for somegenerations to changed conditions of life. Butwe know far too little of the causes and laws ofvariation to make a sound classification. Theaction of changed conditions, whether leading

to definite or indefinite results, is a totally dis-tinct consideration from the effects of selection;for selection depends on the preservation byman of certain individuals, or on their survivalunder various and complex natural circums-tances, and has no relation whatever to theprimary cause of each particular variation.

I will first give in detail all the facts which Ihave been able to collect, rendering it probablethat climate, food, etc., have acted so definitelyand powerfully on the organisation of our do-mesticated productions, that new sub- varietiesor races have been thus formed without the aidof selection by man or nature. I will then givethe facts and considerations opposed to thisconclusion, and finally we will weigh, as fairlyas we can, the evidence on both sides.

When we reflect that distinct races of almost allour domesticated animals exist in each king-dom of Europe, and formerly even in each dis-trict of England, we are at first strongly inclined

to attribute their origin to the definite action ofthe physical conditions of each country; andthis has been the conclusion of many authors.But we should bear in mind that man annuallyhas to choose which animals shall be preservedfor breeding, and which shall be slaughtered.We have also seen that both methodical andunconscious selection were formerly practised,and are now occasionally practised by the mostbarbarous races, to a much greater extent thanmight have been anticipated. Hence it is diffi-cult to judge how far differences in the condi-tions between, for instance, the several districtsin England, have sufficed to modify the breedswhich have been reared in each. It may be ar-gued that, as numerous wild animals andplants have ranged during many ages throug-hout Great Britain, and still retain the samecharacter, the difference in conditions betweenthe several districts could not have modified ina marked manner the various native races ofcattle, sheep, pigs, and horses. The same diffi-

culty of distinguishing between the effects ofnatural selection and the definite action of ex-ternal conditions is encountered in a still higherdegree when we compare closely allied speciesinhabiting two countries, such as North Ameri-ca and Europe, which do not differ greatly inclimate, nature of soil, etc., for in this case natu-ral selection will inevitably and rigorously haveacted during a long succession of ages.

Prof. Weismann has suggested (23/1. 'Ueberden Einfluss der Isolirung auf die Artbildung'1872.) that when a variable species enters a newand isolated country, although the variationsmay be of the same general nature as before,yet it is improbable that they should occur inthe same proportional numbers. After a longeror shorter period, the species will tend to be-come nearly uniform in character from the in-cessant crossing of the varying individuals; butowing to the proportion of the individuals var-ying in different ways not being the same in the

two cases, the final result will be the produc-tion of two forms somewhat different from oneanother. In cases of this kind it would falselyappear as if the conditions had induced certaindefinite modifications, whereas they had onlyexcited indefinite variability, but with the va-riations in slightly different proportional num-bers. This view may throw some light on thefact that the domestic animals which formerlyinhabited the several districts in Great Britain,and the half wild cattle lately kept in severalBritish parks, differed slightly from one anot-her; for these animals were prevented fromwandering over the whole country and inter-crossing, but would have crossed freely withineach district or park.

[From the difficulty of judging how far changedconditions have caused definite modificationsof structure, it will be advisable to give as largea body of facts as possible, showing that extre-mely slight differences within the same coun-

try, or during different seasons, certainly pro-duce an appreciable effect, at least on varietieswhich are already in an unstable condition.Ornamental flowers are good for this purpose,as they are highly variable, and are carefullyobserved. All floriculturists are unanimous thatcertain varieties are affected by very slight dif-ferences in the nature of the artificial compostin which they are grown, and by the naturalsoil of the district, as well as by the season.Thus, a skilful judge, in writing on Carnationsand Picotees (23/2. 'Gardener's Chronicle' 1853page 183.) asks "where can Admiral Curzon beseen possessing the colour, size, and strengthwhich it has in Derbyshire? Where can Flora'sGarland be found equal to those at Slough?Where do high-coloured flowers revel betterthan at Woolwich and Birmingham? Yet in notwo of these districts do the same varieties at-tain an equal degree of excellence, althougheach may be receiving the attention of the mostskilful cultivators." The same writer then re-

commends every cultivator to keep five diffe-rent kinds of soil and manure, "and to endea-vour to suit the respective appetites of theplants you are dealing with, for without suchattention all hope of general success will bevain." So it is with the Dahlia (23/3. Mr. Wild-man 'Floricultural Soc.' February 7, 1843 repor-ted in 'Gardener's Chronicle' 1843 page 86.): theLady Cooper rarely succeeds near London, butdoes admirably in other districts; the reverseholds good with other varieties; and again, the-re are others which succeed equally well invarious situations. A skilful gardener (23/4.Mr. Robson in 'Journal of Horticulture' Februa-ry 13, 1866 page 122.) states that he procuredcuttings of an old and well-known variety (pul-chella) of Verbena, which from having beenpropagated in a different situation presented aslightly different shade of colour; the two varie-ties were afterwards multiplied by cuttings,being carefully kept distinct; but in the second

year they could hardly be distinguished, and inthe third year no one could distinguish them.

The nature of the season has an especial in-fluence on certain varieties of the Dahlia: in1841 two varieties were pre-eminently good,and the next year these same two were pre-eminently bad. A famous amateur (23/5. 'Jour-nal of Horticulture' 1861 page 24.) asserts thatin 1861 many varieties of the Rose came so un-true in character, "that it was hardly possible torecognise them, and the thought was not sel-dom entertained that the grower had lost histally." The same amateur (23/6. Ibid 1862 page83.) states that in 1862 two- thirds of his Auri-culas produced central trusses of flowers, andsuch trusses are liable not to keep true; and headds that in some seasons certain varieties ofthis plant all prove good, and the next seasonall prove bad; whilst exactly the reverse hap-pens with other varieties. In 1845 the editor ofthe 'Gardener's Chronicle' (23/7. 'Gardener's

Chronicle' 1845 page 660.) remarked how sin-gular it was that this year many Calceolariastended to assume a tubular form. With Heart-sease (23/8. Ibid 1863 page 628.) the blotchedsorts do not acquire their proper character untilhot weather sets in; whilst other varieties losetheir beautiful marks as soon as this occurs.

Analogous facts have been observed with lea-ves: Mr. Beaton asserts (23/9. 'Journal of Hort.'1861 pages 64, 309.) that he raised at Shrubland,during six years, twenty thousand seedlingsfrom the Punch Pelargonium, and not one hadvariegated leaves; but at Surbiton, in Surrey,one-third, or even a greater proportion, of theseedlings from this same variety were more orless variegated. The soil of another district inSurrey has a strong tendency to cause variega-tion, as appears from information given me bySir F. Pollock. Verlot (23/10. 'Des Varietes' etc.page 76.) states that the variegated strawberryretains its character as long as grown in a

dryish soil, but soon loses it when planted infresh and humid soil. Mr. Salter, who is wellknown for his success in cultivating variegatedplants, informs me that rows of strawberrieswere planted in his garden in 1859, in the usualway; and at various distances in one row, seve-ral plants simultaneously became variegated;and what made the case more extraordinary, allwere variegated in precisely the same manner.These plants were removed, but during thethree succeeding years other plants in the samerow became variegated, and in no instance we-re the plants in any adjoining row affected.

The chemical qualities, odours, and tissues ofplants are often modified by a change whichseems to us slight. The Hemlock is said not toyield conicine in Scotland. The root of the Aco-nitum napellus becomes innocuous in frigidclimates. The medicinal properties of the Digi-talis are easily affected by culture. As the Pista-cia lentiscus grows abundantly in the South of

France, the climate must suit it, but it yields nomastic. The Laurus sassafras in Europe losesthe odour proper to it in North America.(23/11. Engel 'Sur les Prop. Medicales des Plan-tes' 1860 pages 10, 25. On changes in the odoursof plants see Dalibert's Experiments quoted byBeckman 'Inventions' volume 2 page 344; andNees in Ferussac 'Bull. des Sc. Nat.' 1824 tome 1page 60. With respect to the rhubarb etc. seealso 'Gardener's Chronicle' 1849 page 355; 1862page 1123.) Many similar facts could be given,and they are remarkable because it might havebeen thought that definite chemical compoundswould have been little liable to change either inquality or quantity.

The wood of the American Locust-tree (Robi-nia) when grown in England is nearly worth-less, as is that of the Oak-tree when grown atthe Cape of Good Hope. (23/12. Hooker 'FloraIndica' page 32.) Hemp and flax, as I hear fromDr. Falconer, flourish and yield plenty of seed

on the plains of India, but their fibres are brittleand useless. Hemp, on the other hand, fails toproduce in England that resinous matter whichis so largely used in India as an intoxicatingdrug.

The fruit of the Melon is greatly influenced byslight differences in culture and climate. Henceit is generally a better plan, according to Nau-din, to improve an old kind than to introduce anew one into any locality. The seed of the Per-sian Melon produces near Paris fruit inferior tothe poorest market kinds, but at Bordeauxyields delicious fruit. (23/13. Naudin 'Annalesdes Sc. Nat.' 4th series, Bot. tome 11 1859 page81. 'Gardener's Chronicle' 1859 page 464.) Seedis annually brought from Thibet to Kashmir(23/14. Moorcroft 'Travels' etc. volume 2 page143.) and produces fruit weighing from four toten pounds, but plants raised next year fromseed saved in Kashmir give fruit weighing onlyfrom two to three pounds. It is well known that

American varieties of the Apple produce intheir native land magnificent and brightly- co-loured fruit, but these in England are of poorquality and a dull colour. In Hungary there aremany varieties of the kidney-bean, remarkablefor the beauty of their seeds, but the Rev. M.J.Berkeley (23/15. 'Gardener's Chronicle' 1861page 1113.) found that their beauty could hard-ly ever be preserved in England, and in somecases the colour was greatly changed. We haveseen in the ninth chapter, with respect towheat, what a remarkable effect transportalfrom the north to the south of France, and con-versely, produced on the weight of the grain.]

When man can perceive no change in plants oranimals which have been exposed to a newclimate or to different treatment, insects cansometimes perceive a marked change. A cactushas been imported into India from Canton,Manilla Mauritius, and from the hot-houses ofKew, and there is likewise a so-called native

kind which was formerly introduced fromSouth America; all these plants belong to thesame species and are alike in appearance, butthe cochineal insect flourishes only on the nati-ve kind, on which it thrives prodigiously.(23/16. Royle 'Productive Resources of India'page 59.) Humboldt remarks (23/17. 'PersonalNarrative' English translation volume 5 page101. This statement has been confirmed byKarsten 'Beitrag zur Kenntniss der Rhyncho-prion' Moscow 1864 s. 39 and by others.) thatwhite men "born in the torrid zone walk bare-foot with impunity in the same apartment whe-re a European, recently landed, is exposed tothe attacks of the Pulex penetrans." This insect,the too well-known chigoe, must therefore beable to perceive what the most delicate chemi-cal analysis fails to discover, namely, a diffe-rence between the blood or tissues of a Euro-pean and those of a white man born in the tro-pics. But the discernment of the chigoe is not sosurprising as it at first appears; for according to

Liebig (23/18. 'Organic Chemistry' Englishtranslation 1st edition page 369.) the blood ofmen with different complexions, though in-habiting the same country, emits a differentodour.

[Diseases peculiar to certain localities, heights,or climates, may be here briefly noticed, asshowing the influence of external circumstan-ces on the human body. Diseases confined tocertain races of man do not concern us, for theconstitution of the race may play the more im-portant part, and this may have been determi-ned by unknown causes. The Plica Polonicastands, in this respect, in a nearly intermediateposition; for it rarely affects Germans, who in-habit the neighbourhood of the Vistula, whereso many Poles are grievously affected; neitherdoes it affect Russians, who are said to belongto the same original stock as the Poles. (23/19.Prichard 'Phys. Hist. of Mankind' 1851 volume1 page 155.) The elevation of a district often

governs the appearance of diseases; in Mexicothe yellow fever does not extend above 924metres; and in Peru, people are affected withthe verugas only between 600 and 1600 metresabove the sea; many other such cases could begiven. A peculiar cutaneous complaint, calledthe Bouton d'Alep, affects in Aleppo and someneighbouring districts almost every native in-fant, and some few strangers; and it seems fair-ly well established that this singular complaintdepends on drinking certain waters. In thehealthy little island of St. Helena the scarlet-fever is dreaded like the Plague; analogousfacts have been observed in Chili and Mexico.(23/20. Darwin 'Journal of Researches' 1845page 434.) Even in the different departments ofFrance it is found that the various infirmitieswhich render the conscript unfit for serving inthe army, prevail with remarkable inequality,revealing, as Boudin observes, that many ofthem are endemic, which otherwise would ne-ver have been suspected. (23/21. These state-

ments on disease are taken from Dr. Boudin'Geographie et Statistique Medicale' 1857 tome1 pages 44 and 52; tome 2 page 315.) Any onewho will study the distribution of disease willbe struck with surprise at what slight differen-ces in the surrounding circumstances governthe nature and severity of the complaints bywhich man is at least temporarily affected.

The modifications as yet referred to are extre-mely slight, and in most cases have been cau-sed, as far as we can judge, by equally slightdifferences in the conditions. But such condi-tions acting during a series of generationswould perhaps produce a marked effect.

With plants, a considerable change of climatesometimes produces a conspicuous result. Ihave given in the ninth chapter the most re-markable case known to me, namely, that ofvarieties of maize, which were greatly modifiedin the course of only two or three generationswhen taken from a tropical country to a cooler

one, or conversely. Dr. Falconer informs methat he has seen the English Ribston-pippinapple, a Himalayan oak, Prunus and Pyrus, allassume in the hotter parts of India a fastigiateor pyramidal habit; and this fact is the moreinteresting, as a Chinese tropical species of Py-rus naturally grows thus. Although in thesecases the changed manner of growth seems tohave been directly caused by the great heat, weknow that many fastigiate trees have originatedin their temperate homes. In the Botanic Gar-dens of Ceylon the apple-tree (23/22. 'Ceylon'by Sir J.E. Tennent volume 1 1859 page 89.)"sends out numerous runners under ground,which continually rise into small stems, andform a growth around the parent-tree.) Thevarieties of the cabbage which produce headsin Europe fail to do so in certain tropical coun-tries (23/23. Godron 'De l'Espece' tome 2 page52.) The Rhododendron ciliatum produced atKew flowers so much larger and paler-colouredthan those which it bears on its native Himala-

yan mountain, that Dr. Hooker (23/24. 'Journalof Horticultural Soc.' volume 7 1852 page 117.)would hardly have recognised the species bythe flowers alone. Many similar facts with res-pect to the colour and size of flowers could begiven.

The experiments of Vilmorin and Buckman oncarrots and parsnips prove that abundant nu-triment produces a definite and inheritable ef-fect on the roots, with scarcely any change inother parts of the plant. Alum directly influen-ces the colour of the flowers of the Hydrangea.(23/25. 'Journal of Hort. Soc.' volume 1 page160.) Dryness seems generally to favour thehairiness or villosity of plants. Gartner foundthat hybrid Verbascums became extremelywoolly when grown in pots. Mr. Masters, onthe other hand, states that the Opuntia leucotri-cha "is well clothed with beautiful white hairswhen grown in a damp heat, but in a dry heatexhibits none of this peculiarity." (23/26. See

Lecoq on the Villosity of Plants 'Geograph. Bot.'tome 3 pages 287, 291; Gartner 'Bastarderz.' s.261; Mr. Masters on the Opuntia in 'Gardener'sChronicle' 1846 page 444.) Slight variations ofmany kinds, not worth specifying in detail, areretained only as long as plants are grown incertain soils, of which Sageret (23/27. 'Pom.Phys.' page 136.) gives some instances from hisown experience. Odart, who insists strongly onthe permanence of the varieties of the grape,admits (23/28. 'Ampelographie' 1849 page 19.)that some varieties, when grown under a diffe-rent climate or treated differently, vary in aslight degree, as in the tint of the fruit and inthe period of ripening. Some authors have de-nied that grafting causes even the slightest dif-ference in the scion; but there is sufficient evi-dence that the fruit is sometimes slightly affec-ted in size and flavour, the leaves in duration,and the flowers in appearance. (23/29. Gartner'Bastarderz.' s. 606, has collected nearly all re-corded facts. Andrew Knight in 'Transact. Hort.

Soc.' volume 2 page 160, goes so far as to main-tain that few varieties are absolutely permanentin character when propagated by buds orgrafts.)

There can be no doubt, from the facts given inthe first chapter, that European dogs deteriora-te in India, not only in their instincts but instructure; but the changes which they undergoare of such a nature, that they may be partlydue to reversion to a primitive form, as in thecase of feral animals. In parts of India the tur-key becomes reduced in size, "with the pendu-lous appendage over the beak enormously de-veloped." (23/30. Mr. Blyth 'Annals and Mag ofNat. Hist.' volume 20 1847 page 391.) We haveseen how soon the wild duck, when domestica-ted, loses its true character, from the effects ofabundant or changed food, or from taking littleexercise. From the direct action of a humid cli-mate and poor pasture the horse rapidly de-creases in size in the Falkland Islands. From

information which I have received, this seemslikewise to be the case to a certain extent withsheep in Australia.

Climate definitely influences the hairy coveringof animals; in the West Indies a great change isproduced in the fleece of sheep, in about threegenerations. Dr. Falconer states (23/31. 'Natu-ral History Review' 1862 page 113.) that theThibet mastiff and goat, when brought downfrom the Himalaya to Kashmir, lose their finewool. At Angora not only goats, but shepherd-dogs and cats, have fine fleecy hair, and Mr.Ainsworth (23/32. 'Journal of Roy. Geograp-hical Soc.' volume 9 1839 page 275.) attributesthe thickness of the fleece to the severe winters,and its silky lustre to the hot summers. Burnesstates positively (23/33. 'Travels in Bokhara'volume 3 page 151.) that the Karakool sheeplose their peculiar black curled fleeces whenremoved into any other country. Even withinthe limits of England, I have been assured that

the wool of two breeds of sheep was slightlychanged by the flocks being pastured in diffe-rent localities. (23/34. See also on the influenceof marshy pastures on the wool Godron 'L'Es-pece' tome 2 page 22.) It has been asserted ongood authority (23/35. Isidore Geoffroy Saint-Hilaire 'Hist. Nat. Gen.' tome 3 page 438.) thathorses kept during several years in the deepcoal-mines of Belgium become covered withvelvety hair, almost like that on the mole. Thesecases probably stand in close relation to thenatural change of coat in winter and summer.Naked varieties of several domestic animalshave occasionally appeared; but there is noreason to believe that this is in any way relatedto the nature of the climate to which they havebeen exposed. (23/36. Azara has made somegood remarks on this subject 'Quadrupedes duParaguay' tome 2 page 337. See an account of afamily of naked mice produced in England'Proc. Zoolog. Soc.' 1856 page 38.)

It appears at first sight probable that the increa-sed size, the tendency to fatten, the early matu-rity and altered forms of our improved cattle,sheep, and pigs, have directly resulted fromtheir abundant supply of food. This is the opi-nion of many competent judges, and probablyis to a great extent true. But as far as form isconcerned, we must not overlook the more po-tent influence of lessened use on the limbs andlungs. We see, moreover, as far as size is con-cerned, that selection is apparently a more po-werful agent than a large supply of food, for wecan thus only account for the existence, as re-marked to me by Mr. Blyth, of the largest andsmallest breeds of sheep in the same country, ofCochin-China fowls and Bantams, of smallTumbler and large Runt pigeons, all kept to-gether and supplied with abundant nourish-ment. Nevertheless there can be little doubtthat our domesticated animals have been modi-fied, independently of the increased or lesseneduse of parts, by the conditions to which they

have been subjected, without the aid of selec-tion. For instance, Prof. Rutimeyer (23/37. 'DieFauna der Pfahlbauten' 1861 s. 15.) shows thatthe bones of domesticated quadrupeds can bedistinguished from those of wild animals bythe state of their surface and general appearan-ce. It is scarcely possible to read Nathusius'sexcellent 'Vorstudien' (23/38. 'Schweineschadel'1864 s. 99.) and doubt that, with the highly im-proved races of the pig, abundant food hasproduced a conspicuous effect on the generalform of the body, on the breadth of the headand face, and even on the teeth. Nathusius restsmuch on the case of a purely bred Berkshirepig, which when two months old became di-seased in its digestive organs, and was preser-ved for observation until nineteen months old;at this age it had lost several characteristic fea-tures of the breed, and had acquired a long,narrow head, of large size relatively to its smallbody, and elongated legs. But in this case andin some others we ought not to assume that,

because certain characters are lost, perhapsthrough reversion, under one course of treat-ment, therefore that they were at first directlyproduced by an opposite treatment.

In the case of the rabbit, which has become feralon the island of Porto Santo, we are at firststrongly tempted to attribute the whole chan-ge—the greatly reduced size, the altered tints ofthe fur, and the loss of certain characteristicmarks—to the definite action of the new condi-tions to which it has been exposed. But in allsuch cases we have to consider in addition thetendency to reversion to progenitors more orless remote, and the natural selection of thefinest shades of difference.

The nature of the food sometimes either defini-tely induces certain peculiarities, or stands insome close relation with them. Pallas long agoasserted that the fat-tailed sheep of Siberia de-generate and lose their enormous tails whenremoved from certain saline pastures; and re-

cently Erman (23/39. 'Travels in Siberia' En-glish translation volume 1 page 228.) states thatthis occurs with the Kirgisian sheep whenbrought to Orenburgh.

It is well known that hemp-seed causes bullfin-ches and certain other birds to become black.Mr. Wallace has communicated to me somemuch more remarkable facts of the same natu-re. The natives of the Amazonian region feedthe common green parrot (Chrysotis festiva,Linn.) with the fat of large Siluroid fishes, andthe birds thus treated become beautifully va-riegated with red and yellow feathers. In theMalayan archipelago, the natives of Gilolo alterin an analogous manner the colours of anotherparrot, namely, the Lorius garrulus, Linn., andthus produce the Lori rajah or King-Lory. Theseparrots in the Malay Islands and South Ameri-ca, when fed by the natives on natural vegeta-ble food, such as rice and plaintains, retain theirproper colours. Mr. Wallace has, also, recorded

(23/40. A.R. Wallace 'Travels on the Amazonand Rio Negro' page 294.) a still more singularfact. "The Indians (of S. America) have a cu-rious art by which they change the colours ofthe feathers of many birds. They pluck out tho-se from the part they wish to paint, and inocu-late the fresh wound with the milky secretionfrom the skin of a small toad. The feathers growof a brilliant yellow colour, and on being pluc-ked out, it is said, grow again of the same co-lour without any fresh operation."

Bechstein (23/41. 'Naturgeschichte der Stuben-vogel' 1840 s. 262, 308.) does not entertain anydoubt that seclusion from light affects, at leasttemporarily, the colours of cage-birds.

It is well known that the shells of land-molluscaare affected by the abundance of lime in diffe-rent districts. Isidore Geoffroy Saint-Hilaire(23/42. 'Hist. Nat Gen.' tome 3 page 402.) givesthe case of Helix lactea, which has recentlybeen carried from Spain to the South of France

and to the Rio Plata, and in both countries nowpresents a distinct appearance, but whether thishas resulted from food or climate is not known.With respect to the common oyster, Mr. F. Buc-kland informs me that he can generally distin-guish the shells from different districts; youngoysters brought from Wales and laid down inbeds where "natives" are indigenous, in theshort space of two months begin to assume the"native" character. M. Costa (23/43. 'Bull. de LaSoc. Imp. d'Acclimat.' tome 8 page 351.) hasrecorded a much more remarkable case of thesame nature, namely, that young shells takenfrom the shores of England and placed in theMediterranean, at once altered their manner ofgrowth and formed prominent diverging rays,like those on the shells of the proper Medite-rranean oyster. The same individual shell, sho-wing both forms of growth, was exhibited befo-re a society in Paris. Lastly, it is well knownthat caterpillars fed on different food someti-mes either themselves acquire a different co-

lour or produce moths differing in colour.(23/44. See an account of Mr. Gregson's expe-riments on the Abraxus grossulariata 'Proc.Entomolog. Soc.' January 6, 1862: these experi-ments have been confirmed by Mr. Greening in'Proc. of the Northern Entomolog. Soc.' July 28,1862. For the effects of food on caterpillars see acurious account by M. Michely in 'Bull. De LaSoc. Imp. d'Acclimat.' tome 8 page 563. For ana-logous facts from Dahlbom on Hymenopterasee Westwood 'Modern Class. of Insects' volu-me 2 page 98. See also Dr. L. Moller 'DieAbhangigkeit der Insecten' 1867 s. 70.)

It would be travelling beyond my proper limitshere to discuss how far organic beings in a stateof nature are definitely modified by changedconditions. In my 'Origin of Species' I have gi-ven a brief abstract of the facts bearing on thispoint, and have shown the influence of light onthe colours of birds, and of residence near thesea on the lurid tints of insects, and on the suc-

culency of plants. Mr. Herbert Spencer (23/45.'The Principles of Biology' volume 2 1866. Thepresent chapters were written before I had readMr. Herbert Spencer's work, so that I have notbeen able to make so much use of it as I shouldotherwise probably have done.) has recentlydiscussed with much ability this whole subjecton general grounds. He argues, for instance,that with all animals the external and internaltissues are differently acted on by the surroun-ding conditions, and they invariably differ inintimate structure. So again the upper and lo-wer surfaces of true leaves, as well as of stemsand petioles, when these assume the functionand occupy the position of leaves, are differen-tly circumstanced with respect to light, etc., andapparently in consequence differ in structure.But, as Mr. Herbert Spencer admits, it is mostdifficult in all such cases to distinguish betweenthe effects of the definite action of physicalconditions and the accumulation through natu-ral selection of inherited variations which are

serviceable to the organism, and which havearisen independently of the definite action ofthese conditions.]

Although we are not here concerned with thedefinite action of the conditions of life on orga-nisms in a state of nature, I may state that muchevidence has been gained during the last fewyears on this subject. In the United States, forinstance, it has been clearly proved, more espe-cially by Mr. J.A. Allen, that, with birds, manyspecies differ in tint, size of body and of beak,and in length of tail, in proceeding from theNorth to the South; and it appears that thesedifferences must be attributed to the direct ac-tion of temperature. (23/46. Professor Weis-mann comes to the same conclusion with res-pect to certain European butterflies in his va-luable essay 'Ueber den Saison- Dimorphismus'1875. I might also refer to the recent works ofseveral other authors on the present subject; forinstance to Kerner's 'Gute und schlechte Arten'

1866.) With respect to plants I will give a so-mewhat analogous case: Mr. Meehan (23/47.'Proc. Acad. Nat. Soc. of Philadelphia' January28, 1862.), has compared twenty-nine kinds ofAmerican trees with their nearest Europeanallies, all grown in close proximity and underas nearly as possible the same conditions. In theAmerican species he finds, with the rarest ex-ceptions, that the leaves fall earlier in the sea-son, and assume before their fall a brighter tint;that they are less deeply toothed or serrated;that the buds are smaller; that the trees are mo-re diffuse in growth and have fewer branchlets;and, lastly, that the seeds are smaller—all incomparison with the corresponding Europeanspecies. Now considering that these correspon-ding trees belong to several distinct orders, andthat they are adapted to widely different sta-tions, it can hardly be supposed that their diffe-rences are of any special service to them in theNew and Old worlds; and if so such differencescannot have been gained through natural selec-

tion, and must be attributed to the long conti-nued action of a different climate.

GALLS.

Another class of facts, not relating to cultivatedplants, deserves attention. I allude to the pro-duction of galls. Every one knows the curious,bright-red, hairy productions on the wild rose-tree, and the various different galls producedby the oak. Some of the latter resemble fruit,with one face as rosy as the rosiest apple. Thesebright colours can be of no service either to thegall-forming insect or to the tree, and probablyare the direct result of the action of the light, inthe same manner as the apples of Nova Scotiaor Canada are brighter coloured than Englishapples. According to Osten Sacken's latest revi-sion, no less than fifty-eight kinds of galls areproduced on the several species of oak, by Cy-nips with its sub-genera; and Mr. B.D. Walsh(23/48. See Mr. B.D. Walsh's excellent papers in

'Proc. Entomolog. Soc. Philadelphia' December1866 page 284. With respect to the willow seeibid 1864 page 546.) states that he can add ma-ny others to the list. One American species ofwillow, the Salix humilis, bears ten distinctkinds of galls. The leaves which spring fromthe galls of various English willows differ com-pletely in shape from the natural leaves. Theyoung shoots of junipers and firs, when punc-tured by certain insects, yield monstrousgrowths resembling flowers and fir-cones; andthe flowers of some plants become from thesame cause wholly changed in appearance.Galls are produced in every quarter of theworld; of several sent to me by Mr. Thwaitesfrom Ceylon, some were as symmetrical as acomposite flower when in bud, others smoothand spherical like a berry; some protected bylong spines, others clothed with yellow woolformed of long cellular hairs, others with regu-larly tufted hairs. In some galls the internalstructure is simple, but in others it is highly

complex; thus M. Lacaze-Duthiers (23/49. Seehis admirable 'Histoire des Galles' in 'Annal.des Sc. Nat. Bot.' 3rd series tome 19 1853 page273.) has figured in the common ink-gall no lessthan seven concentric layers, composed of dis-tinct tissue, namely, the epidermic, sub-epidermic, spongy, intermediate, and the hardprotective layer formed of curiously thickenedwoody cells, and, lastly, the central mass,abounding with starch-granules on which thelarvae feed.

Galls are produced by insects of various orders,but the greater number by species of Cynips. Itis impossible to read M. Lacaze-Duthiers' dis-cussion and doubt that the poisonous secretionof the insect causes the growth of the gall; andevery one knows how virulent is the poisonsecreted by wasps and bees, which belong tothe same group with Cynips. Galls grow withextraordinary rapidity, and it is said that theyattain their full size in a few days (23/50. Kirby

and Spence 'Entomology' 1818 volume 1 page450; Lacaze- Duthiers ibid page 284.); it is cer-tain that they are almost completely developedbefore the larvae are hatched. Considering thatmany gall-insects are extremely small, the dropof secreted poison must be excessively minute;it probably acts on one or two cells alone,which, being abnormally stimulated, rapidlyincrease by a process of self-division. Galls, asMr. Walsh (23/51. 'Proc. Entomolog. Soc. Phi-ladelphia' 1864 page 558.) remarks, afford good,constant, and definite characters, each kindkeeping as true to form as does any indepen-dent organic being. This fact becomes still moreremarkable when we hear that, for instance,seven out of the ten different kinds of gallsproduced on Salix humilis are formed by gall-gnats (Cecidomyidae) which "though essentia-lly distinct species, yet resemble one another soclosely that in almost all cases it is difficult, andin most cases impossible, to distinguish thefull-grown insects one from the other." (23/52.

Mr. B.D. Walsh ibid page 633 and December1866 page 275.) For in accordance with a wide-spread analogy we may safely infer that thepoison secreted by insects so closely alliedwould not differ much in nature; yet this slightdifference is sufficient to induce widely diffe-rent results. In some few cases the same speciesof gall-gnat produces on distinct species of wi-llows galls which cannot be distinguished; theCynips fecundatrix, also, has been known toproduce on the Turkish oak, to which it is notproperly attached, exactly the same kind of gallas on the European oak. (23/53. Mr. B.D. Walshibid 1864 pages 545, 411, 495; and December1866 page 278. See also Lacaze-Duthiers.) Theselatter facts apparently prove that the nature ofthe poison is a more powerful agent in deter-mining the form of the gall than the specificcharacter of the tree which is acted on.

As the poisonous secretion of insects belongingto various orders has the special power of affec-

ting the growth of various plants; as a slightdifference in the nature of the poison suffices toproduce widely different results; and lastly, aswe know that the chemical compounds secre-ted by plants are eminently liable to be modi-fied by changed conditions of life, we may be-lieve it possible that various parts of a plantmight be modified through the agency of itsown altered secretions. Compare, for instance,the mossy and viscid calyx of a moss-rose,which suddenly appears through bud-variationon a Provence-rose, with the gall of red mossgrowing from the inoculated leaf of a wild rose,with each filament symmetrically branched likea microscopical spruce-fir, bearing a glandulartip and secreting odoriferous gummy matter.(23/54. Lacaze-Duthiers ibid pages 325, 328.)Or compare, on the one hand, the fruit of thepeach, with its hairy skin, fleshy covering, hardshell and kernel, and on the other hand one ofthe more complex galls with its epidermic,spongy, and woody layers, surrounding tissue

loaded with starch granules. These normal andabnormal structures manifestly present a cer-tain degree of resemblance. Or, again, reflect onthe cases above given of parrots which havehad their plumage brightly decorated throughsome change in their blood, caused by havingbeen fed on certain fishes, or locally inoculatedwith the poison of a toad. I am far from wis-hing to maintain that the moss-rose or the hardshell of the peach-stone or the bright colours ofbirds are actually due to any chemical changein the sap or blood; but these cases of galls andof parrots are excellently adapted to show ushow powerfully and singularly external agen-cies may affect structure. With such facts beforeus, we need feel no surprise at the appearanceof any modification in any organic being.

[I may, also, here allude to the remarkable ef-fects which parasitic fungi sometimes produceon plants. Reissek (23/55. 'Linnaea' volume 171843; quoted by Dr. M.T. Masters, Royal Insti-

tution, March 16, 1860.) has described a The-sium, affected by an Oecidium, which wasgreatly modified, and assumed some of thecharacteristic features of certain allied species,or even genera. Suppose, says Reissek, "thecondition originally caused by the fungus tobecome constant in the course of time, the plantwould, if found growing wild, be considered asa distinct species or even as belonging to a newgenus." I quote this remark to show how pro-foundly, yet in how natural a manner, thisplant must have been modified by the parasiticfungus. Mr. Meehan (23/56. 'Proc. Acad. Nat.Sc., Philadelphia' June 16, 1874 and July 23,1875.) also states that three species of Euphor-bia and Portulaca olereacea, which naturallygrow prostrate, become erect when they areattacked by the Oecidium. Euphorbia maculatain this case also becomes nodose, with thebranchlets comparatively smooth and the lea-ves modified in shape, approaching in these

respects to a distinct species, namely, the E.hypericifolia.]

FACTS AND CONSIDERATIONS OPPO-SED TO THE BELIEF THAT THE CONDI-TIONS OF LIFE ACT IN A POTENT MANNERIN CAUSING DEFINITE MODIFICATIONSOF STRUCTURE.

I have alluded to the slight differences in spe-cies naturally living in distinct countries underdifferent conditions; and such differences wefeel at first inclined to attribute, probably oftenwith justice, to the definite action of the su-rrounding conditions. But it must be borne inmind that there exist many animals and plantswhich range widely and have been exposed togreat diversities of climate, yet remain uniformin character. Some authors, as previously re-marked, account for the varieties of our culina-ry and agricultural plants by the definite actionof the conditions to which they have been ex-

posed in the different parts of Great Britain; butthere are about 200 plants (23/57. Hewett C.Watson 'Cybele Britannica' volume 1 1847 page11.) which are found in every single Englishcounty; and these plants must have been expo-sed for an immense period to considerable dif-ferences of climate and soil, yet do not differ.So, again,, some animals and plants range overa large portion of the world, yet retain the samecharacter.

[Notwithstanding the facts previously given onthe occurrence of highly peculiar local diseasesand on the strange modifications of structure inplants caused by the inoculated poison of in-sects, and other analogous cases; still there area multitude of variations—such as the modifiedskull of the niata ox and bulldog, the longhorns of Caffre cattle, the conjoined toes of thesolid-hoofed swine, the immense crest and pro-tuberant skull of Polish fowls, the crop of thepouter-pigeon, and a host of other such cases—

which we can hardly attribute to the definiteaction, in the sense before specified, of the ex-ternal conditions of life. No doubt in every casethere must have been some exciting cause; butas we see innumerable individuals exposed tonearly the same conditions, and one alone isaffected, we may conclude that the constitutionof the individual is of far higher importancethan the conditions to which it has been expo-sed. It seems, indeed, to be a general rule thatconspicuous variations occur rarely, and in oneindividual alone out of millions, though allmay have been exposed, as far as we can judge,to nearly the same conditions. As the moststrongly marked variations graduate insensiblyinto the most trifling, we are led by the sametrain of thought to attribute each slight varia-tion much more to innate differences of consti-tution, however caused, than to the definiteaction of the surrounding conditions.

We are led to the same conclusion by conside-ring the cases, formerly alluded to, of fowls andpigeons, which have varied and will no doubtgo on varying in directly opposite ways,though kept during many generations undernearly the same conditions. Some, for instance,are born with their beaks, wings, tails, legs, etc.,a little longer, and others with these same partsa little shorter. By the long-continued selectionof such slight individual differences which oc-cur in birds kept in the same aviary, widelydifferent races could certainly be formed; andlong-continued selection, important as is theresult, does nothing but preserve the variationswhich arise, as it appears to us, spontaneously.

In these cases we see that domesticated animalsvary in an indefinite number of particulars,though treated as uniformly as is possible. Onthe other hand, there are instances of animalsand plants, which, though they have been ex-posed to very different conditions, both under

nature and domestication, have varied in near-ly the same manner. Mr. Layard informs methat he has observed amongst the Caffres ofSouth Africa a dog singularly like an arctic Es-quimaux dog. Pigeons in India present nearlythe same wide diversities of colour as in Euro-pe; and I have seen chequered and simply ba-rred pigeons, and pigeons with blue and whiteloins, from Sierra Leone, Madeira, England,and India. New varieties of flowers are conti-nually raised in different parts of Great Britain,but many of these are found by the judges atour exhibitions to be almost identical with oldvarieties. A vast number of new fruit-trees andculinary vegetables have been produced inNorth America: these differ from Europeanvarieties in the same general manner as theseveral varieties raised in Europe differ fromone another; and no one has ever pretendedthat the climate of America has given to themany American varieties any general characterby which they can be recognised. Nevertheless,

from the facts previously advanced on the aut-hority of Mr. Meehan with respect to Americanand European forest-trees it would be rash toaffirm that varieties raised in the two countrieswould not in the course of ages assume a dis-tinctive character. Dr. M. Masters has recordeda striking fact (23/58. 'Gardener's Chronicle'1857 page 629.) bearing on this subject: he rai-sed numerous plants of Hybiscus syriacus fromseed collected in South Carolina and the HolyLand, where the parent-plants must have beenexposed to considerably different conditions;yet the seedlings from both localities broke intotwo similar strains, one with obtuse leaves andpurple or crimson flowers, and the other withelongated leaves and more or less pink flowers.

We may, also, infer the prepotent influence ofthe constitution of the organism over the defini-te action of the conditions of life, from the seve-ral cases given in the earlier chapters of parallelseries of varieties,—an important subject,

hereafter to be more fully discussed. Sub-varieties of the several kinds of wheat, gourds,peaches, and other plants, and to a limited ex-tent sub-varieties of the fowl, pigeon, and dog,have been shown either to resemble or to differfrom one another in a closely corresponding orparallel manner. In other cases, a variety of onespecies resembles a distinct species; or the va-rieties of two distinct species resemble oneanother. Although these parallel resemblancesno doubt often result from reversion to the for-mer characters of a common progenitor; yet inother cases, when new characters first appear,the resemblance must be attributed to the in-heritance of a similar constitution, and conse-quently to a tendency to vary in the same man-ner. We see something of a similar kind in thesame monstrosity appearing and reappearingmany times in the same species of animal, and,as Dr. Maxwell Masters has remarked to me, inthe same species of plant.]

We may at least conclude, that the amount ofmodification which animals and plants haveundergone under domestication does not co-rrespond with the degree to which they havebeen subjected to changed circumstances. Aswe know the parentage of domesticated birdsfar better than of most quadrupeds, we willglance through the list. The pigeon has variedin Europe more than almost any other bird; yetit is a native species, and has not been exposedto any extraordinary change of conditions. Thefowl has varied equally, or almost equally, withthe pigeon, and is a native of the hot jungles ofIndia. Neither the peacock, a native of the samecountry, nor the guinea-fowl, an inhabitant ofthe dry deserts of Africa, has varied at all, oronly in colour. The turkey, from Mexico, hasvaried but little. The duck, on the other hand, anative of Europe, has yielded some well-marked races; and as this is an aquatic bird, itmust have been subjected to a far more seriouschange in its habits than the pigeon or even the

fowl, which nevertheless have varied in a muchhigher degree. The goose, a native of Europeand aquatic like the duck, has varied less thanany other domesticated bird, except the pea-cock.

Bud-variation is, also, important under ourpresent point of view, in some few cases, aswhen all the eyes on the same tuber of the pota-to, or all the fruit on the same plum-tree, or allthe flowers on the same plant, have suddenlyvaried in the same manner, it might be arguedthat the variation had been definitely caused bysome change in the conditions to which theplants had been exposed; yet, in other cases,such an admission is extremely difficult. Asnew characters sometimes appear by bud-variation, which do not occur in the parent-species or in any allied species, we may reject,at least in these cases, the idea that they are dueto reversion. Now it is well worth while to re-flect maturely on some striking case of bud-

variation, for instance that of the peach. Thistree has been cultivated by the million in va-rious parts of the world, has been treated diffe-rently, grown on its own roots and grafted onvarious stocks, planted as a standard, trainedagainst a wall, or under glass; yet each bud ofeach sub-variety keeps true to its kind. But oc-casionally, at long intervals of time, a tree inEngland, or under the widely different climateof Virginia, produces a single bud, and thisyields a branch which ever afterwards bearsnectarines. Nectarines differ, as every oneknows, from peaches in their smoothness, size,and flavour; and the difference is so great thatsome botanists have maintained that they arespecifically distinct. So permanent are the cha-racters thus suddenly acquired, that a nectarineproduced by bud-variation has propagateditself by seed. To guard against the suppositionthat there is some fundamental distinction bet-ween bud and seminal variation, it is well tobear in mind that nectarines have likewise been

produced from the stone of the peach; and,reversely, peaches from the stone of the necta-rine. Now is it possible to conceive externalconditions more closely alike than those towhich the buds on the same tree are exposed?Yet one bud alone, out of the many thousandsborne by the same tree, has suddenly, withoutany apparent cause, produced a nectarine. Butthe case is even stronger than this, for the sameflower-bud has yielded a fruit, one-half or one-quarter a nectarine, and the other half or three-quarters a peach. Again, seven or eight varie-ties of the peach have yielded by bud-variationnectarines: the nectarines thus produced, nodoubt, differ a little from one another; but stillthey are nectarines. Of course there must besome cause, internal or external, to excite thepeach-bud to change its nature; but I cannotimagine a class of facts better adapted to forceon our minds the conviction that what we callthe external conditions of life are in many casesquite insignificant in relation to any particular

variation, in comparison with the organisationor constitution of the being which varies.

It is known from the labours of Geoffroy Saint-Hilaire, and recently from those of Dareste andothers, that eggs of the fowl, if shaken, placedupright, perforated, covered in part with var-nish, etc., produce monstrous chickens. Nowthese monstrosities may be said to be directlycaused by such unnatural conditions, but themodifications thus induced are not of a definitenature. An excellent observer, M. Camille Da-reste (23/59. 'Memoire sur la Production Artifi-cielle des Monstruosites' 1862 pages 8-12; 'Re-cherches sur les Conditions, etc., chez les Mons-tres' 1863 page 6. An abstract is given of Geof-froy's Experiments by his son, in his 'Vie, Tra-vaux' etc. 1847 page 290.), remarks "that thevarious species of monstrosities are not deter-mined by specific causes; the external agencieswhich modify the development of the embryoact solely in causing a perturbation—a perver-

sion in the normal course of development." Hecompares the result to what we see in illness: asudden chill, for instance, affects one indivi-dual alone out of many, causing either a cold,or sore-throat, rheumatism, or inflammation ofthe lungs or pleura. Contagious matter acts inan analogous manner. (23/60. Paget 'Lectureson Surgical Pathology' 1853 volume 1 page483.) We may take a still more specific instance:seven pigeons were struck by rattle-snakes(23/61. 'Researches upon the Venom of the Rat-tle-snake' January 1861 by Dr. Mitchell page67.): some suffered from convulsions; some hadtheir blood coagulated, in others it was perfec-tly fluid; some showed ecchymosed spots onthe heart, others on the intestines, etc.; othersagain showed no visible lesion in any organ. Itis well known that excess in drinking causesdifferent diseases in different men; but in thetropics the effects of intemperance differ fromthose caused in a cold climate (23/62. Mr.Sedgwick 'British and Foreign Medico-Chirurg.

Review' July 1863 page 175.); and in this casewe see the definite influence of opposite condi-tions. The foregoing facts apparently give us asgood an idea as we are likely for a long time toobtain, how in many cases external conditionsact directly, though not definitely, in causingmodifications of structure.

SUMMARY.

There can be no doubt, from the facts given inthis chapter, that extremely slight changes inthe conditions of life sometimes, probably of-ten, act in a definite manner on our domestica-ted productions; and, as the action of changedconditions in causing indefinite variability isaccumulative, so it may be with their definiteaction. Hence considerable and definite modifi-cations of structure probably follow from alte-red conditions acting during a long series ofgenerations. In some few instances a markedeffect has been produced quickly on all, or

nearly all, the individuals which have beenexposed to a marked change of climate, food, orother circumstance. This has occurred withEuropean men in the United States, with Euro-pean dogs in India, with horses in the FalklandIslands, apparently with various animals atAngora, with foreign oysters in the Mediterra-nean, and with maize transported from oneclimate to another. We have seen that the che-mical compounds of some plants and the stateof their tissues are readily affected by changedconditions. A relation apparently exists bet-ween certain characters and certain conditions,so that if the latter be changed the character islost—as with the colours of flowers, the state ofsome culinary plants, the fruit of the melon, thetail of fat-tailed sheep, and the peculiar fleecesof other sheep.

The production of galls, and the change of plu-mage in parrots when fed on peculiar food orwhen inoculated by the poison of a toad, prove

to us what great and mysterious changes instructure and colour, may be the definite resultof chemical changes in the nutrient fluids ortissues.

We now almost certainly know that organicbeings in a state of nature may be modified invarious definite ways by the conditions towhich they have been long exposed, as in thecase of the birds and other animals in the nort-hern and southern United States, and of Ameri-can trees in comparison with their representati-ves in Europe. But in many cases it is most dif-ficult to distinguish between the definite resultof changed conditions, and the accumulationthrough natural selection of indefinite varia-tions which have proved serviceable. If it profi-ted a plant to inhabit a humid instead of an aridstation, a fitting change in its constitution mightpossibly result from the direct action of the en-vironment, though we have no grounds forbelieving that variations of the right kind

would occur more frequently with plants in-habiting a station a little more humid thanusual, than with other plants. Whether the sta-tion was unusually dry or humid, variationsadapting the plant in a slight degree for directlyopposite habits of life would occasionally arise,as we have good reason to believe from whatwe actually see in other cases.

The organisation or constitution of the beingwhich is acted on, is generally a much moreimportant element than the nature of the chan-ged conditions, in determining the nature of thevariation. We have evidence of this in the ap-pearance of nearly similar modifications underdifferent conditions, and of different modifica-tions under apparently nearly the same condi-tions. We have still better evidence of this inclosely parallel varieties being frequently pro-duced from distinct races, or even distinct spe-cies; and in the frequent recurrence of the samemonstrosity in the same species. We have also

seen that the degree to which domesticatedbirds have varied, does not stand in any closerelation with the amount of change to whichthey have been subjected.

To recur once again to bud-variations. Whenwe reflect on the millions of buds which manytrees have produced, before some one bud hasvaried, we are lost in wonder as to what theprecise cause of each variation can be. Let usrecall the case given by Andrew Knight of theforty-year-old tree of the yellow magnum bo-num plum, an old variety which has been pro-pagated by grafts on various stocks for a verylong period throughout Europe and NorthAmerica, and on which a single bud suddenlyproduced the red magnum bonum. We shouldalso bear in mind that distinct varieties, andeven distinct species,—as in the case of pea-ches, nectarines, and apricots,—of certain rosesand camellias,— although separated by a vastnumber of generations from any progenitor in

common, and although cultivated under diver-sified conditions, have yielded by bud-variation closely analogous varieties. When wereflect on these facts we become deeply impres-sed with the conviction that in such cases thenature of the variation depends but little on theconditions to which the plant has been expo-sed, and not in any especial manner on its indi-vidual character, but much more on the inheri-ted nature or constitution of the whole group ofallied beings to which the plant in questionbelongs. We are thus driven to conclude that inmost cases the conditions of life play a subor-dinate part in causing any particular modifica-tion; like that which a spark plays, when a massof combustibles bursts into flame—the natureof the flame depending on the combustiblematter, and not on the spark. (23/63. ProfessorWeismann argues strongly in favour of thisview in his 'Saison-Dimorphismus der Schmet-terlinge' 1875 pages 40-43.)

No doubt each slight variation must have itsefficient cause; but it is as hopeless an attemptto discover the cause of each, as to say why achill or a poison affects one man differentlyfrom another. Even with modifications resul-ting from the definite action of the conditions oflife, when all or nearly all the individuals,which have been similarly exposed, are similar-ly affected, we can rarely see the precise rela-tion between cause and effect. In the next chap-ter it will be shown that the increased use ordisuse of various organs produces an inheritedeffect. It will further be seen that certain varia-tions are bound together by correlation as wellas by other laws. Beyond this we cannot at pre-sent explain either the causes or nature of thevariability of organic beings.

CHAPTER 2.XXIV.

LAWS OF VARIATION—USE AND DIS-USE, ETC.

NISUS FORMATIVUS, OR THE CO-ORDINATING POWER OF THE ORGANISA-TION. ON THE EFFECTS OF THE INCREA-SED USE AND DISUSE OF ORGANS. CHAN-GED HABITS OF LIFE. ACCLIMATISATIONWITH ANIMALS AND PLANTS. VARIOUSMETHODS BY WHICH THIS CAN BE EFFEC-TED. ARRESTS OF DEVELOPMENT. RUDI-MENTARY ORGANS.

In this and the two following chapters I shalldiscuss, as well as the difficulty of the subjectpermits, the several laws which govern Variabi-lity. These may be grouped under the effects ofuse and disuse, including changed habits andacclimatisation—arrest of development—correlated variation—the cohesion of homolo-gous parts-the variability of multiple parts—

compensation of growth—the position of budswith respect to the axis of the plant—and lastly,analogous variation. These several subjects sograduate into one another that their distinctionis often arbitrary.

It may be convenient first briefly to discuss thatcoordinating and reparative power which iscommon, in a higher or lower degree, to allorganic beings, and which was formerly desig-nated by physiologists as nisus formativus.

[Blumenbach and others (24/1. 'An Essay onGeneration' English translation page 18; Paget'Lectures on Surgical Pathology' 1853 volume 1page 209.) have insisted that the principlewhich permits a Hydra, when cut into frag-ments, to develop itself into two or more per-fect animals, is the same with that which causesa wound in the higher animals to heal by a cica-trice. Such cases as that of the Hydra are evi-dently analogous to the spontaneous divisionor fissiparous generation of the lowest animals,

and likewise to the budding of plants. Betweenthese extreme cases and that of a mere cicatricewe have every gradation. Spallanzani (24/2.'An Essay on Animal Reproduction' Englishtranslation 1769 page 79.) by cutting off the legsand tail of a Salamander, got in the course ofthree months six crops of these members; sothat 687 perfect bones were reproduced by oneanimal during one season. At whatever pointthe limb was cut off, the deficient part, and nomore, was exactly reproduced. When a disea-sed bone has been removed, a new one some-times "gradually assumes the regular form, andall the attachments of muscles, ligaments, etc.,become as complete as before." (24/3. Carpen-ter 'Principles of Comp. Physiology' 1854 page479.)

This power of regrowth does not, however,always act perfectly; the reproduced tail of alizard differs in the form of the scales from thenormal tail: with certain Orthopterous insects

the large hind legs are reproduced of smallersize (24/4. Charlesworth 'Mag. of Nat. Hist.'volume 1 1837 page 145.): the white cicatricewhich in the higher animals unites the edges ofa deep wound is not formed of perfect skin, forelastic tissue is not produced till long after-wards. (24/5. Paget 'Lectures on Surgical Pat-hology' volume 1 page 239.) "The activity of thenisus formativus," says Blumenbach, "is in aninverse ratio to the age of the organised body."Its power is also greater with animals, the lo-wer they stand in the scale of organisation; andanimals low in the scale correspond with theembryos of higher animals belonging to thesame class. Newport's observations (24/6. Quo-ted by Carpenter 'Comp. Phys.' page 479.) af-ford a good illustration of this fact, for he foundthat "myriapods, whose highest developmentscarcely carries them beyond the larva of per-fect insects, can regenerate limbs and antennaeup to the time of their last moult;" and so canthe larvae of true insects, but, except in one

order, not in the mature insect. Salamanderscorrespond in development with the tadpolesor larvae of the tailless Batrachians, and bothpossess to a large extent the power of regrowth;but not so the mature tailless Batrachians.

Absorption often plays an important part in therepair of injuries. When a bone is broken anddoes not unite, the ends are absorbed and roun-ded, so that a false joint is formed; or if the endsunite, but overlap, the projecting parts are re-moved. (24/7. Prof. Marey's discussion on thepower of co- adaptation in all parts of the orga-nisation is excellent. 'La Machine Animale' 1873chapter 9. See also Paget 'Lectures' etc. page257.) A dislocated bone will form for itself anew socket. Displaced tendons and varicoseveins excavate new channels in the bonesagainst which they press. But absorption comesinto action, as Virchow remarks, during thenormal growth of bones; parts which are solidduring youth become hollowed out for the me-

dullary tissue as the bone increases in size. Intrying to understand the many well- adaptedcases of regrowth when aided by absorption,we should remember that almost all parts ofthe organisation, even whilst retaining the sa-me form, undergo constant renewal; so that apart which is not renewed would be liable toabsorption.

Some cases, usually classed under the so-callednisus formativus, at first appear to come undera distinct head; for not only are old structuresreproduced, but new structures are formed.Thus, after inflammation "false membranes,"furnished with blood-vessels, lymphatics, andnerves, are developed; or a foetus escapes fromthe Fallopian tubes, and falls into the abdomen,"nature pours out a quantity of plastic lymph,which forms itself into organised membrane,richly supplied with blood-vessels," and thefoetus is nourished for a time. In certain casesof hydrocephalus the open and dangerous spa-

ces in the skull are filled up with new bones,which interlock by perfect serrated sutures.(24/8. These cases are given by Blumenbach inhis 'Essay on Generation' pages 52, 54.) Butmost physiologists, especially on the Continent,have now given up the belief in plastic lymphor blastema, and Virchow (24/9. 'Cellular Pat-hology' translation by Dr. Chance 1860 pages27, 441.) maintains that every structure, new orold, is formed by the proliferation of pre-existing cells. On this view false membranes,like cancerous or other tumours, are merelyabnormal developments of normal growths;and we can thus understand how it is that theyresemble adjoining structures; for instance, thata "false membrane in the serous cavities acqui-res a covering of epithelium exactly like thatwhich covers the original serous membrane;adhesions of the iris may become black appa-rently from the production of pigment-cells likethose of the uvea." (24/10. Paget 'Lectures onPathology' volume 1 1853 page 357.)

No doubt the power of reparation, though notalways perfect, is an admirable provision, rea-dy for various emergencies, even for such asoccur only at long intervals of time. (24/11.Paget ibid page 150.) Yet this power is not morewonderful than the growth and developmentof every single creature, more especially of tho-se which are propagated by fissiparous genera-tion. This subject has been here noticed, becau-se we may infer that, when any part or organ iseither greatly increased in size or wholly sup-pressed through variation and continued selec-tion, the co-ordinating power of the organisa-tion will continually tend to bring again all theparts into harmony with one another.]

ON THE EFFECTS OF THE INCREASEDUSE AND DISUSE OF ORGANS.

It is notorious, and we shall immediately addu-ce proofs, that increased use or action strengt-hens muscles, glands, sense-organs, etc.; and

that disuse, on the other hand, weakens them.It has been experimentally proved by Ranke(24/12. 'Die Blutvertheilung, etc. der Organe'1871 as quoted by Jaeger 'In Sachen Darwin's'1874 page 48. See also H. Spencer 'The Princi-ples of Biology' volume 2 1866 chapters 3-5.)that the flow of blood is greatly increased to-wards any part which is performing work, andsinks again when the part is at rest. Consequen-tly, if the work is frequent, the vessels increasein size and the part is better nourished. Paget(24/13. 'Lectures on Pathology' 1853 volume 1page 71.) also accounts for the long, thick, dark-coloured hairs which occasionally grow, evenin young children, near old-standing inflamedsurfaces or fractured bones by an increasedflow of blood to the part. When Hunter inser-ted the spur of a cock into the comb, which iswell supplied with blood-vessels, it grew in onecase spirally to a length of six inches, and inanother case forward, like a horn, so that thebird could not touch the ground with its beak.

According to the interesting observations of M.Sedillot (24/14. 'Comptes Rendus' September26, 1864 page 539.), when a portion of one ofthe bones of the leg of an animal is removed,the associated bone enlarges till it attains a bulkequal to that of the two bones, of which it hasto perform the functions. This is best exhibitedin dogs in which the tibia has been removed;the companion bone, which is naturally almostfiliform and not one-fifth the size of the other,soon acquires a size equal to or greater thanthat of the tibia. Now, it is at first difficult tobelieve that increased weight acting on astraight bone could, by alternately increasingand diminishing the pressure, cause the bloodto flow more freely in the vessels which per-meate the periosteum and thus supply morenutriment to the bone. Nevertheless the obser-vations adduced by Mr. Spencer (24/15. H.Spencer 'The Principles of Biology' volume 2page 243.), on the strengthening of the bowed

bones of rickety children, along their concavesides, leads to the belief that this is possible.

The rocking of the stem of a tree increases in amarked manner the growth of the woody tissuein the parts which are strained. Prof. Sachs be-lieves, from reasons which he assigns, that thisis due to the pressure of the bark being relaxedin such parts, and not as Knight and H. Spencermaintain, to an increased flow of sap caused bythe movement of the trunk. (24/16. Ibid volu-me 2 page 269. Sachs 'Text-book of Botany' 1875page 734.) But hard woody tissue may be deve-loped without the aid of any movement, as wesee with ivy closely attached to an old wall. Inall such cases, it is very difficult to distinguishbetween the effects of long-continued selectionand those which follow from the increased ac-tion of the part, or directly from some othercause. Mr. H. Spencer (24/17. Ibid volume 2page 273.) acknowledges this difficulty, andgives as an instance the thorns on trees and the

shells of nuts. Here we have extremely hardwoody tissue without the possibility of anymovement, and without, as far as we can see,any other directly exciting cause; and as thehardness of these parts is of manifest service tothe plant, we may look at the result as probablydue to the selection of so-called spontaneousvariations. Every one knows that hard workthickens the epidermis on the hands; and whenwe hear that with infants, long before birth, theepidermis is thicker on the palms and soles ofthe feet than on any other part of the body, aswas observed with admiration by Albinus(24/18. Paget 'Lectures on Pathology' volume 2page 209.), we are naturally inclined to attribu-te this to the inherited effects of long-continueduse or pressure. We are tempted to extend thesame view even to the hoofs of quadrupeds;but who will pretend to determine how far na-tural selection may have aided in the formationof structures of such obvious importance to theanimal?

[That use strengthens the muscles may be seenin the limbs of artisans who follow differenttrades; and when a muscle is strengthened, thetendons, and the crests of bone to which theyare attached, become enlarged; and this mustlikewise be the case with the blood-vessels andnerves. On the other hand, when a limb is notused, as by Eastern fanatics, or when the nervesupplying it with nervous power is effectuallydestroyed, the muscles wither. So again, whenthe eye is destroyed the optic nerve becomesatrophied, sometimes even in the course of afew months. (24/19. Muller 'Phys.' Englishtranslation pages 54, 791. Prof. Reed has given('Physiological and Anat. Researches' page 10) acurious account of the atrophy of the limbs ofrabbits after the destruction of the nerve.) TheProteus is furnished with branchiae as well aswith lungs: and Schreibers (24/20. Quoted byLecoq in 'Geograph. Bot.' tome 1 1854 page182.) found that when the animal was compe-lled to live in deep water, the branchiae were

developed to thrice their ordinary size, and thelungs were partially atrophied. When, on theother hand, the animal was compelled to live inshallow water, the lungs became larger andmore vascular, whilst the branchiae disappea-red in a more or less complete degree. Suchmodifications as these are, however, of compa-ratively little value for us, as we do not actuallyknow that they tend to be inherited.

In many cases there is reason to believe that thelessened use of various organs has affected thecorresponding parts in the offspring. But thereis no good evidence that this ever follows in thecourse of a single generation. It appears, as inthe case of general or indefinite variability, thatseveral generations must be subjected to chan-ged habits for any appreciable result. Our do-mestic fowls, ducks, and geese have almostlost, not only in the individual but in the race,their power of flight; for we do not see a youngfowl, when frightened, take flight like a young

pheasant. Hence I was led carefully to comparethe limb-bones of fowls, ducks, pigeons, andrabbits, with the same bones in the wild parent-species. As the measurements and weights we-re fully given in the earlier chapters I need hereonly recapitulate the results. With domesticpigeons, the length of the sternum, the promi-nence of its crest, the length of the scapulae andfurculum, the length of the wings as measuredfrom tip to tip of the radii, are all reduced rela-tively to the same parts in the wild pigeon. Thewing and tail feathers, however, are increasedin length, but this may have as little connectionwith the use of the wings or tail, as the lengt-hened hair on a dog with the amount of exerci-se which it has habitually taken. The feet ofpigeons, except in the long-beaked races, arereduced in size. With fowls the crest of the ster-num is less prominent, and is often distorted ormonstrous; the wing-bones have become ligh-ter relatively to the leg-bones, and are apparen-tly a little shorter in comparison with those of

the parent-form, the Gallus bankiva. Withducks, the crest of the sternum is affected in thesame manner as in the foregoing cases: the fur-culum, coracoids, and scapulae are all reducedin weight relatively to the whole skeleton: thebones of the wings are shorter and lighter, andthe bones of the legs longer and heavier, relati-vely to each other, and relatively to the wholeskeleton, in comparison with the same bones inthe wild-duck. The decreased weight and sizeof the bones, in the foregoing cases, is probablythe indirect result of the reaction of the weake-ned muscles on the bones. I failed to comparethe feathers of the wings of the tame and wildduck; but Gloger (24/21. 'Das Abandern derVogel' 1833 s. 74.) asserts that in the wild duckthe tips of the wing-feathers reach almost to theend of the tail, whilst in the domestic duck theyoften hardly reach to its base. He remarks alsoon the greater thickness of the legs, and saysthat the swimming membrane between the toes

is reduced; but I was not able to detect this lat-ter difference.

With the domesticated rabbit the body, toget-her with the whole skeleton, is generally largerand heavier than in the wild animal, and theleg-bones are heavier in due proportion; butwhatever standard of comparison be taken,neither the leg-bones nor the scapulae haveincreased in length proportionally with theincreased dimensions of the rest of the skeleton.The skull has become in a marked manner na-rrower, and, from the measurements of its ca-pacity formerly given, we may conclude, thatthis narrowness results from the decreased sizeof the brain, consequent on the mentally inacti-ve life led by these closely-confined animals.

We have seen in the eighth chapter that silk-moths, which have been kept during many cen-turies closely confined, emerge from their co-coons with their wings distorted, incapable offlight, often greatly reduced in size, or even,

according to Quatrefages, quite rudimentary.This condition of the wings may be largelyowing to the same kind of monstrosity whichoften affects wild Lepidoptera when artificiallyreared from the cocoon; or it may be in part dueto an inherent tendency, which is common tothe females of many Bombycidae, to have theirwings in a more or less rudimentary state; butpart of the effect may be attributed to long-continued disuse.]

From the foregoing facts there can be no doubtthat with our anciently domesticated animals,certain bones have increased or decreased insize and weight owing to increased or decrea-sed use; but they have not been modified, asshown in the earlier chapters, in shape or struc-ture. With animals living a free life and occa-sionally exposed to severe competition the re-duction would tend to be greater, as it wouldbe an advantage to them to have the develop-ment of every superfluous part saved. With

highly-fed domesticated animals, on the otherhand, there seems to be no economy of growth,nor any tendency to the elimination of super-fluous details. But to this subject I shall recur.

Turning now to more general observations,Nathusius has shown that with the improvedraces of the pig, the shortened legs and snout,the form of the articular condyles of the occi-put, and the position of the jaws with the uppercanine teeth projecting in a most anomalousmanner in front of the lower canines, may beattributed to these parts not having been fullyexercised. For the highly-cultivated races donot travel in search of food, nor root up theground with their ringed muzzles. (24/22. Nat-husius 'Die Racen des Schweines' 1860 s. 53, 57;'Vorstudien…Schweineschadel' 1864 s. 103, 130,133. Prof. Lucae supports and extends the con-clusions of Von Nathusius: 'Der Schadel desMaskenschweines' 1870.) These modificationsof structure, which are all strictly inherited,

characterise several improved breeds, so thatthey cannot have been derived from any singledomestic stock. With respect to cattle, ProfessorTanner has remarked that the lungs and liver inthe improved breeds "are found to be conside-rably reduced in size when compared with tho-se possessed by animals having perfect liberty"(24/23. 'Journal of Agriculture of HighlandSoc.' July 1860 page 321.); and the reduction ofthese organs affects the general shape of thebody. The cause of the reduced lungs in highly-bred animals which take little exercise is ob-vious; and perhaps the liver may be affected bythe nutritious and artificial food on which theylargely subsist. Again, Dr. Wilckens asserts(24/24. 'Landwirth. Wochenblatt' No. 10.) thatvarious parts of the body certainly differ inAlpine and lowland breeds of several domesti-cated animals, owing to their different habits oflife; for instance, the neck and fore-legs inlength, and the hoofs in shape.

[It is well known that, when an artery is tied,the anastomosing branches, from being forcedto transmit more blood, increase in diameter;and this increase cannot be accounted for bymere extension, as their coats gain in strength.With respect to glands, Sir J. Paget observesthat "when one kidney is destroyed the otheroften becomes much larger, and does doublework." (24/25. 'Lectures on Surgical Pathology'1853 volume 1 page 27.) If we compare the sizeof the udders and their power of secretion incows which have been long domesticated, andin certain breeds of the goat in which the ud-ders nearly touch the ground, with these or-gans in wild or half-domesticated animals, thedifference is great. A good cow with us dailyyields more than five gallons, or forty pints ofmilk, whilst a first-rate animal, kept, for instan-ce, by the Damaras of South Africa (24/26. An-dersson 'Travels in South Africa' page 318. Foranalogous cases in South America see Aug. St.-Hilaire 'Voyage dans la Province de Goyaz'

tome 1 page 71.), "rarely gives more than twoor three pints of milk daily, and, should her calfbe taken from her, she absolutely refuses togive any." We may attribute the excellence ofour cows and of certain goats, partly to the con-tinued selection of the best milking animals,and partly to the inherited effects of the increa-sed action, through man's art, of the secretingglands.

It is notorious that short-sight is inherited; andwe have seen in the twelfth chapter from thestatistical researches of M. Giraud-Teulon, thatthe habit of viewing near objects gives a ten-dency to short-sight. Veterinarians are unani-mous that horses are affected with spavins,splints, ringbones, etc., from being shod andfrom travelling on hard roads, and they arealmost equally unanimous that a tendency tothese malformations is transmitted. Formerlyhorses were not shod in North Carolina, and ithas been asserted that they did not then suffer

from these diseases of the legs and feet. (24/27.Brickell 'Nat. Hist. of North Carolina' 1739 page53.)]

Our domesticated quadrupeds are all descen-ded, as far as is known, from species havingerect ears; yet few kinds can be named, ofwhich at least one race has not drooping ears.Cats in China, horses in parts of Russia, sheepin Italy and elsewhere, the guinea-pig formerlyin Germany, goats and cattle in India, rabbits,pigs, and dogs in all long-civilised countrieshave dependent ears. With wild animals, whichconstantly use their ears like funnels to catchevery passing sound, and especially to ascer-tain the direction whence it comes, there is not,as Mr. Blyth has remarked, any species withdrooping ears except the elephant. Hence theincapacity to erect the ears is certainly in somemanner the result of domestication; and thisincapacity has been attributed by various aut-hors (24/28. Livingstone quoted by Youatt on

'Sheep' page 142. Hodgson in 'Journal of AsiaticSoc. of Bengal' volume 16 1847 page 1006 etc.etc. On the other hand Dr. Wilckens arguesstrongly against the belief that the drooping ofthe ears is the result of disuse: 'Jahrbuch derdeutschen Viehzucht' 1866.) to disuse, for ani-mals protected by man are not compelled habi-tually to use their ears. Col. Hamilton Smith(24/29. 'Naturalist's Library' Dogs volume 21840 page 104.) states that in ancient effigies ofthe dog, "with the exception of one Egyptianinstance, no sculpture of the earlier Grecian eraproduces representations of hounds with com-pletely drooping ears; those with them halfpendulous are missing in the most ancient; andthis character increases, by degrees, in theworks of the Roman period." Godron also hasremarked "that the pigs of the ancient Egyp-tians had not their ears enlarged and pendent."(24/30. 'De l'Espece' tome 1 1859 page 367.) Butit is remarkable that the drooping of the ear isnot accompanied by any decrease in size; on

the contrary, animals so different as fancy rab-bits, certain Indian breeds of the goat, our pet-ted spaniels, blood- hounds, and other dogs,have enormously elongated ears, so that itwould appear as if their weight had causedthem to droop, aided perhaps by disuse. Withrabbits, the drooping of the much elongatedears has affected even the structure of the skull.

The tail of no wild animal, as remarked to meby Mr. Blyth, is curled; whereas pigs and someraces of dogs have their tails much curled. Thisdeformity, therefore, appears to be the result ofdomestication, but whether in any way connec-ted with the lessened use of the tail is doubtful.

The epidermis on our hands is easily thickened,as every one knows, by hard work. In a districtof Ceylon the sheep have "horny callosities thatdefend their knees, and which arise from theirhabit of kneeling down to crop the short herba-ge, and this distinguishes the Jaffna flocks fromthose of other portions of the island;" but it is

not stated whether this peculiarity is inherited.(24/31. 'Ceylon' by Sir J.E. Tennent 1859 volu-me 2 page 531.)

The mucous membrane which lines the sto-mach is continuous with the external skin of thebody; therefore it is not surprising that its tex-ture should be affected by the nature of thefood consumed, but other and more interestingchanges likewise follow. Hunter long ago ob-served that the muscular coat of the stomach ofa gull (Larus tridactylus) which had been fedfor a year chiefly on grain was thickened; and,according to Dr. Edmondston, a similar changeperiodically occurs in the Shetland Islands inthe stomach of the Larus argentatus, which inthe spring frequents the cornfields and feeds onthe seed. The same careful observer has noticeda great change in the stomach of a raven whichhad been long fed on vegetable food. In thecase of an owl (Strix grallaria), similarly trea-ted, Menetries states that the form of the sto-

mach was changed, the inner coat became leat-hery, and the liver increased in size. Whetherthese modifications in the digestive organswould in the course of generations become in-herited is not known. (24/32. For the foregoingstatements see Hunter 'Essays and Observa-tions' 1861 volume 2 page 329; Dr. Edmonds-ton, as quoted in Macgillivray 'British Birds'volume 5 page 550: Menetries as quoted inBronn 'Geschichte der Natur' b. 2 s. 110.)

The increased or diminished length of the intes-tines, which apparently results from changeddiet, is a more remarkable case, because it ischaracteristic of certain animals in their domes-ticated condition, and therefore must be inheri-ted. The complex absorbent system, the blood-vessels, nerves, and muscles, are necessarily allmodified together with the intestines. Accor-ding to Daubenton, the intestines of the domes-tic cat are one-third longer than those of thewild cat of Europe; and although this species is

not the parent- stock of the domestic animal,yet, as Isidore Geoffroy has remarked, the seve-ral species of cats are so closely allied that thecomparison is probably a fair one. The increa-sed length appears to be due to the domesticcat being less strictly carnivorous in its dietthan any wild feline species; for instance, I haveseen a French kitten eating vegetables as readi-ly as meat. According to Cuvier, the intestinesof the domesticated pig exceed greatly in pro-portionate length those of the wild boar. In thetame and wild rabbit the change is of an oppo-site nature, and probably results from the nutri-tious food given to the tame rabbit. (24/33.These statements on the intestines are takenfrom Isidore Geoffroy Saint-Hilaire 'Hist. Nat.Gen.' tome 3 pages 427, 441.)

CHANGED AND INHERITED HABITS OFLIFE.

This subject, as far as the mental powers ofanimals are concerned, so blends into instinct,that I will here only remind the reader of suchcases as the tameness of our domesticated ani-mals—the pointing or retrieving of dogs— theirnot attacking the smaller animals kept byman—and so forth. How much of these chan-ges ought to be attributed to mere habit, andhow much to the selection of individuals whichhave varied in the desired manner, irrespecti-vely of the special circumstances under whichthey have been kept, can seldom be told.

We have already seen that animals may behabituated to a changed diet; but some additio-nal instances may be given. In the PolynesianIslands and in China the dog is fed exclusivelyon vegetable matter, and the taste for this kindof food is to a certain extent inherited. (24/34.Gilbert White 'Nat. Hist. Selborne' 1825 volume2 page 121.) Our sporting dogs will not touchthe bones of game birds, whilst most other dogs

devour them with greediness. In some parts ofthe world sheep have been largely fed on fish.The domestic hog is fond of barley, the wildboar is said to disdain it; and the disdain is par-tially inherited, for some young wild pigs bredin captivity showed an aversion for this grain,whilst others of the same brood relished it.(24/35. Burdach 'Traite de Phys.' tome 2 page267 as quoted by Dr. P. Lucas 'L'Hered. Nat.'tome 1 page 388.) One of my relations bred so-me young pigs from a Chinese sow by a wildAlpine boar; they lived free in the park, andwere so tame that they came to the house to befed; but they would not touch swill, which wasdevoured by the other pigs. An animal whenonce accustomed to an unnatural diet, whichcan generally be effected only during youth,dislikes its proper food, as Spallanzani found tobe the case with a pigeon which had been longfed on meat. Individuals of the same speciestake to new food with different degrees of rea-diness; one horse, it is stated, soon learned to

eat meat, whilst another would have perishedfrom hunger rather than have partaken of it.(24/36. This and several other cases are givenby Colin 'Physiologie Comp. des AnimauxDom.' 1854 tome 1 page 426.) The caterpillars ofthe Bombyx hesperus feed in a state of natureon the leaves of the Cafe diable, but, afterhaving been reared on the Ailanthus, theywould not touch the Cafe diable, and actuallydied of hunger. (24/37. M. Michely de Cayennein 'Bull. Soc. d'Acclimat.' tome 8 1861 page 563.)

It has been found possible to accustom marinefish to live in fresh water; but as such changesin fish and other marine animals have beenchiefly observed in a state of nature, they donot properly belong to our present subject. Theperiod of gestation and of maturity, as shownin the earlier chapters,—the season and thefrequency of the act of breeding,—have all beengreatly modified under domestication. With theEgyptian goose the rate of change with respect

to the season has been recorded. (24/38. Qua-trefages 'Unite de l'Espece Humaine' 1861 page79.) The wild drake pairs with one female, thedomestic drake is polygamous. Certain breedsof fowls have lost the habit of incubation. Thepaces of the horse, and the manner of flight ofcertain breeds of the pigeon, have been modi-fied and are inherited. Cattle, horses, and pigshave learnt to browse under water in the St.John's River, East Florida, where the Vallisneriahas been largely naturalised. The cows wereobserved by Prof. Wyman to keep their headsimmersed for "a period varying from fifteen tothirty-five seconds." (24/39. 'The AmericanNaturalist' April 1874 page 237.) The voice dif-fers much in certain kinds of fowls and pi-geons. Some varieties are clamorous and otherssilent, as the Call and common duck, or theSpitz and pointer dog. Every one knows howthe breeds of the dog differ from one another intheir manner of hunting, and in their ardourafter different kinds of game or vermin.

With plants the period of vegetation is easilychanged and is inherited, as in the case of sum-mer and winter wheat, barley, and vetches; butto this subject we shall immediately return un-der acclimatisation. Annual plants sometimesbecome perennial under a new climate, as Ihear from Dr. Hooker is the case with the stockand mignonette in Tasmania. On the otherhand, perennials sometimes become annuals, aswith the Ricinus in England, and as, accordingto Captain Mangles, with many varieties of theheartsease. Von Berg (24/40. 'Flora' 1835 b. 2page 504.) raised from seed of Verbascumphoeniceum, which is usually a biennial, bothannual and perennial varieties. Some deci-duous bushes become evergreen in hot coun-tries. (24/41. Alph. de Candolle 'Geograph.Bot.' tome 2 page 1078.) Rice requires muchwater, but there is one variety in India whichcan be grown without irrigation. (24/42. Royle'Illustrations of the Botany of the Himalaya'page 19.) Certain varieties of the oat and of our

other cereals are best fitted for certain soils.(24/43. 'Gardener's Chronicle' 1850 pages 204,219.) Endless similar facts could be given in theanimal and vegetable kingdoms. They are noti-ced here because they illustrate analogous dif-ferences in closely allied natural species, andbecause such changed habits of life, whetherdue to habit, or to the direct action of externalconditions, or to so-called spontaneous variabi-lity, would be apt to lead to modifications ofstructure.

ACCLIMATISATION.

From the previous remarks we are naturally ledto the much disputed subject of acclimatisation.There are two distinct questions: Do varietiesdescended from the same species differ in theirpower of living under different climates? Andsecondly, if they so differ, how have they be-come thus adapted? We have seen that Euro-pean dogs do not succeed well in India, and it

is asserted (24/44. Rev. R. Everest 'Journal As.Soc. of Bengal' volume 3 page 19.), that no onehas there succeeded in keeping the Newfound-land long alive; but then it may be argued, andprobably with truth, that these northern breedsare specifically distinct from the native dogswhich flourish in India. The same remark maybe made with respect to different breeds ofsheep, of which, according to Youatt (24/45.Youatt on 'Sheep' 1838 page 491.), not onebrought "from a torrid climate lasts out the se-cond year," in the Zoological Gardens. Butsheep are capable of some degree of acclimati-sation, for Merino sheep bred at the Cape ofGood Hope have been found far better adaptedfor India than those imported from England.(24/46. Royle 'Prod. Resources of India' page153.) It is almost certain that all the breeds ofthe fowl are descended from one species; butthe Spanish breed, which there is good reasonto believe originated near the Mediterranean(24/47. Tegetmeier 'Poultry Book' 1866 page

102.), though so fine and vigorous in England,suffers more from frost than any other breed.The Arrindy silk moth introduced from Bengal,and the Ailanthus moth from the temperateprovince of Shan Tung, in China, belong to thesame species, as we may infer from their identi-ty in the caterpillar, cocoon, and mature states(24/48. Dr. R. Paterson in a paper communica-ted to Bot. Soc. of Canada quoted in the 'Rea-der' 1863 November 13.); yet they differ muchin constitution: the Indian form "will flourishonly in warm latitudes," the other is quite har-dy and withstands cold and rain.

[Plants are more strictly adapted to climatethan are animals. The latter when domesticatedwithstand such great diversities of climate, thatwe find nearly the same species in tropical andtemperate countries; whilst the cultivatedplants are widely dissimilar. Hence a largerfield is open for inquiry in regard to the accli-matisation of plants than of animals. It is no

exaggeration to say that with almost everyplant which has long been cultivated, varietiesexist which are endowed with constitutionsfitted for very different climates; I will selectonly a few of the more striking cases, as itwould be tedious to give all. In North Americanumerous fruit-trees have been raised, and inhorticultural publications,—for instance, in thatby Downing,—lists are given of the varietieswhich are best able to withstand the severeclimate of the northern States and Canada. Ma-ny American varieties of the pear, plum, andpeach are excellent in their own country, butuntil recently, hardly one was known that suc-ceeded in England; and with apples (24/49. Seeremarks by Editor in 'Gardener's Chronicle'1848 page 5.), not one succeeds. Though theAmerican varieties can withstand a severerwinter than ours, the summer here is not hotenough. Fruit-trees have also originated in Eu-rope with different constitutions, but they arenot much noticed, because nurserymen here do

not supply wide areas. The Forelle pear flowersearly, and when the flowers have just set, andthis is the critical period, they have been obser-ved, both in France and England, to withstandwith complete impunity a frost of 18 deg andeven 14 deg Fahr., which killed the flowers,whether fully expanded or in bud, of all otherkinds of pears. (24/50. 'Gardener's Chronicle'1860 page 938. Remarks by Editor and quota-tion from Decaisne.) This power in the flowerof resisting cold and afterwards producing fruitdoes not invariably depend, as we know ongood authority (24/51. J. de Jonghe of Brusselsin 'Gardener's Chronicle' 1857 page 612.), ongeneral constitutional vigour. In proceedingnorthward, the number of varieties which arefound capable of resisting the climate rapidlydecreases, as may be seen in the list of the va-rieties of the cherry, apple, and pear, which canbe cultivated in the neighbourhood of Stoc-kholm. (24/52. Ch. Martius 'Voyage Bot. CotesSept. de la Norvege' page 26.) Near Moscow,

Prince Troubetzkoy planted for experiment inthe open ground several varieties of the pear,but one alone, the Poire sans Pepins, withstoodthe cold of winter. (24/53. 'Journal de l'Acad.Hort. de Gand' quoted in 'Gardener's Chronicle'1859 page 7.) We thus see that our fruit-trees,like distinct species of the same genus, certainlydiffer from each other in their constitutionaladaptation to different climates.

With the varieties of many plants, the adapta-tion to climate is often very close. Thus it hasbeen proved by repeated trials "that few if anyof the English varieties of wheat are adaptedfor cultivation in Scotland" (24/54. 'Gardener'sChronicle' 1851 page 396.); but the failure inthis case is at first only in the quantity, thoughultimately in the quality, of the grain produced.The Rev. M.J. Berkeley sowed wheat-seed fromIndia, and got "the most meagre ears," on landwhich would certainly have yielded a goodcrop from English wheat. (24/55. Ibid 1862 pa-

ge 235.) In these cases varieties have been ca-rried from a warmer to a cooler climate; in thereverse case, as "when wheat was importeddirectly from France into the West Indian Is-lands, it produced either wholly barren spikesor furnished with only two or three miserableseeds, while West Indian seed by its side yiel-ded an enormous harvest." (24/56. On the aut-hority of Labat quoted in 'Gardener's Chronicle'1862 page 235.) Here is another case of closeadaptation to a slightly cooler climate; a kind ofwheat which in England may be used indiffe-rently either as a winter or summer variety,when sown under the warmer climate of Grig-nan, in France, behaved exactly as if it had beena true winter wheat. (24/57. MM. Edwards andColin 'Annal. des Sc. Nat.' 2nd series Bot. tome5 page 22.)

Botanists believe that all the varieties of maizebelong to the same species; and we have seenthat in North America, in proceeding north-

ward, the varieties cultivated in each zone pro-duce their flowers and ripen their seed withinshorter and shorter periods. So that the tall,slowly maturing southern varieties do not suc-ceed in New England, and the New Englishvarieties do not succeed in Canada. I have notmet with any statement that the southern varie-ties are actually injured or killed by a degree ofcold which the northern varieties can withstandwith impunity, though this is probable; but theproduction of early flowering and early see-ding varieties deserves to be considered as oneform of acclimatisation. Hence it has beenfound possible, according to Kalm, to cultivatemaize further and further northwards in Ame-rica. In Europe, also, as we learn from the evi-dence given by Alph. De Candolle, the cultureof maize has extended since the end of the lastcentury thirty leagues north of its formerboundary. (24/58. 'Geograph. Bot.' page 337.)On the authority of Linnaeus (24/59. 'SwedishActs' English translation 1739-40 volume 1.

Kalm in his 'Travels' volume 2 page 166 givesan analogous case with cotton-plants raised inNew Jersey from Carolina seed.), I may quotean analogous case, namely, that in Sweden to-bacco raised from home-grown seed ripens itsseed a month sooner and is less liable to misca-rry than plants raised from foreign seed.

With the Vine, differently from the maize, theline of practical culture has retreated a littlesouthward since the middle ages (24/60. DeCandolle 'Geograph. Bot.' page 339.); but thisseems due to commerce being now easier, sothat it is better to import wine from the souththan to make it in northern districts. Nevert-heless the fact of the vine not having spreadnorthward shows that acclimatisation has madeno progress during several centuries. There is,however, a marked difference in the constitu-tion of the several varieties,— some being har-dy, whilst others, like the muscat of Alexandria,require a very high temperature to come to

perfection. According to Labat (24/61. 'Garde-ner's Chronicle' 1862 page 235.), vines takenfrom France to the West Indies succeed withextreme difficulty, whilst those imported fromMadeira or the Canary Islands thrive admira-bly.

Gallesio gives a curious account of the naturali-sation of the Orange in Italy. During many cen-turies the sweet orange was propagated exclu-sively by grafts, and so often suffered fromfrosts, that it required protection. After the se-vere frost of 1709, and more especially afterthat of 1763, so many trees were destroyed, thatseedlings from the sweet orange were raised,and, to the surprise of the inhabitants, theirfruit was found to be sweet. The trees thus rai-sed were larger, more productive, and hardierthan the old kinds; and seedlings are now con-tinually raised. Hence Gallesio concludes thatmuch more was effected for the naturalisationof the orange in Italy by the accidental produc-

tion of new kinds during a period of about six-ty years, than had been effected by grafting oldvarieties during many ages. (24/62. Gallesio'Teoria della Riproduzione Veg.' 1816 page 125;and 'Traite du Citrus' 1811 page 359.) I may addthat Risso (24/63. 'Essai sur l'Hist. des Oran-gers' 1813 page 20 etc.) describes some Portu-guese varieties of the orange as extremely sen-sitive to cold, and as much tenderer than cer-tain other varieties.

The peach was known to Theophrastus, 322B.C. (24/64. Alph. de Candolle 'Geograph. Bot.'page 882.) According to the authorities quotedby Dr. F. Rolle (24/65. 'Ch. Darwin's Lehre vonder Entstehung' etc. 1862 s. 87.), it was tenderwhen first introduced into Greece, and even inthe island of Rhodes only occasionally borefruit. If this be correct, the peach, in spreadingduring the last two thousand years over themiddle parts of Europe, must have becomemuch hardier. At the present day different va-

rieties differ much in hardiness: some Frenchvarieties will not succeed in England; and nearParis, the Pavie de Bonneuil does not ripen itsfruit till very late in the season, even whengrown on a wall; "it is, therefore, only fit for avery hot southern climate." (24/66. Decaisnequoted in 'Gardener's Chronicle' 1865 page271.)

I will briefly give a few other cases. A variety ofMagnolia grandiflora, raised by M. Roy, withs-tands a temperature several degrees lower thanthat which any other variety can resist. Withcamellias there is much difference in hardiness.One particular variety of the Noisette rosewithstood the severe frost of 1860 "untouchedand hale amidst a universal destruction of ot-her Noisettes." In New York the "Irish yew isquite hardy, but the common yew is liable to becut down." I may add that there are varieties ofthe sweet potato (Convolvulus batatas) whichare suited for warmer, as well as for colder,

climates. (24/67. For the magnolia see Loudon's'Gardener's Mag.' volume 13 1837 page 21. Forcamellias and roses see 'Gardener's Chronicle'1860 page 384. For the yew 'Journal of Hort.'March 3, 1863 p 174. For sweet potatoes see Col.von Siebold in 'Gardener's Chronicle' 1855 page822.)]

The plants as yet mentioned have been foundcapable of resisting an unusual degree of coldor heat, when fully grown. The following casesrefer to plants whilst young. In a large bed ofyoung Araucarias of the same age, growingclose together and equally exposed, it was ob-served (24/68. The Editor 'Gardener's Chroni-cle' 1861 page 239.), after the unusually severewinter of 1860-61, that, "in the midst of thedying, numerous individuals remained onwhich the frost had absolutely made no kind ofimpression." Dr. Lindley, after alluding to thisand other similar cases, remarks, "Among thelessons which the late formidable winter has

taught us, is that, even in their power of resis-ting cold, individuals of the same species ofplants are remarkably different." Near Salisbu-ry, there was a sharp frost on the night of May24, 1836, and all the French beans (Phaseolusvulgaris) in a bed were killed except about onein thirty, which completely escaped. (24/69.Loudon's 'Gardener's Mag.' volume 12 1836page 378.) On the same day of the month, butin the year 1864, there was a severe frost inKent, and two rows of scarlet- runners (P. mul-tiflorus) in my garden, containing 390 plants ofthe same age and equally exposed, were allblackened and killed except about a dozenplants. In an adjoining row of "Fulmer's dwarfbean" (P. vulgaris), one single plant escaped. Astill more severe frost occurred four days af-terwards, and of the dozen plants which hadpreviously escaped only three survived; thesewere not taller or more vigorous than the otheryoung plants, but they escaped completely,with not even the tips of their leaves browned.

It was impossible to behold these three plants,with their blackened, withered, and deadbrethren all around, and not see at a glance thatthey differed widely in constitutional power ofresisting frost.

This work is not the proper place to show thatwild plants of the same species, naturally gro-wing at different altitudes or under differentlatitudes, become to a certain extent acclimati-sed, as is proved by the different behaviour oftheir seedlings when raised in another country.In my 'Origin of Species' I have alluded to somecases, and I could add many others. One ins-tance must suffice: Mr. Grigor, of Forres (24/70.'Gardener's Chronicle' 1865 page 699. Mr. G.Maw gives ('Gardener's Chronicle' 1870 page895) a number of striking cases; he broughthome from southern Spain and northern Africaseveral plants, which he cultivated in Englandalongside specimens from northern districts;and he found a great difference not only in

their hardiness during the winter, but in thebehaviour of some of them during the sum-mer.), states that seedlings of the Scotch fir (Pi-nus sylvestris), raised from seed from the Con-tinent and from the forests of Scotland, differmuch. "The difference is perceptible in one-year-old, and more so in two-year- old seed-lings; but the effects of the winter on the secondyear's growth almost uniformly make thosefrom the Continent quite brown, and so dama-ged, that by the month of March they are quiteunsaleable, while the plants from the nativeScotch pine, under the same treatment, andstanding alongside, although considerablyshorter, are rather stouter and quite green, sothat the beds of the one can be known from theother when seen from the distance of a mile."Closely similar facts have been observed withseedling larches.

[Hardy varieties would alone be valued or no-ticed in Europe; whilst tender varieties, requi-

ring more warmth, would generally be neglec-ted; but such occasionally arise. Thus Loudon(24/71. 'Arboretum et Fruticetum' volume 3page 1376.) describes a Cornish variety of theelm which is almost an evergreen, and of whichthe shoots are often killed by the autumnalfrosts, so that its timber is of little value. Horti-culturists know that some varieties are muchmore tender than others: thus all the varietiesof the broccoli are more tender than cabbages;but there is much difference in this respect inthe sub-varieties of the broccoli; the pink andpurple kinds are a little hardier than the whiteCape broccoli, "but they are not to be dependedon after the thermometer falls below 24 degFahr.;" the Walcheren broccoli is less tenderthan the Cape, and there are several varietieswhich will stand much severer cold than theWalcheren. (24/72. Mr. Robson in 'Journal ofHorticulture' 1861 page 23.) Cauliflowers seedmore freely in India than cabbages. (24/73. Dr.Bonavia 'Report of the Agri.-Hort. Soc. of

Oudh' 1866.) To give one instance with flowers:eleven plants raised from a hollyhock, calledthe Queen of the Whites (24/74. 'Cottage Gar-dener' 1860 April 24 page 57.) were found to bemuch more tender than various other seed-lings. It may be presumed that all tender varie-ties would succeed better under a climatewarmer than ours. With fruit-trees, it is wellknown that certain varieties, for instance of thepeach, stand forcing in a hot-house better thanothers; and this shows either pliability of orga-nisation or some constitutional difference. Thesame individual cherry-tree, when forced, hasbeen observed during successive years gradua-lly to change its period of vegetation. (24/75.'Gardener's Chronicle' 1841 page 291.) Few pe-largoniums can resist the heat of a stove, butAlba Multiflora will, as a most skilful gardenerasserts, "stand pine-apple top and bottom heatthe whole winter; without looking any moredrawn than if it had stood in a common green-house; and Blanche Fleur seems as if it had

been made on purpose for growing in winter,like many bulbs, and to rest all summer."(24/76. Mr. Beaton in 'Cottage Gardener' March20, 1860 page 377. Queen Mab will also standstove heat. See 'Gardener's Chronicle' 1845 page226.) There can hardly be a doubt that the AlbaMultiflora pelargonium must have a widelydifferent constitution from that of most othervarieties of this plant; it would probably withs-tand even an equatorial climate.

We have seen that according to Labat the vineand wheat require acclimatisation in order tosucceed in the West Indies. Similar facts havebeen observed at Madras: "two parcels of mig-nonette-seed, one direct from Europe, the othersaved at Bangalore (of which the mean tempe-rature is much below that of Madras), weresown at the same time: they both vegetatedequally favourably, but the former all died off afew days after they appeared above ground;the latter still survive, and are vigorous, healt-

hy plants." "So again, turnip and carrot seedsaved at Hyderabad are found to answer betterat Madras than seed from Europe or from theCape of Good Hope." (24/77. 'Gardener'sChronicle' 1841 page 439.) Mr. J. Scott of theCalcutta Botanic Gardens, informs me thatseeds of the sweet-pea (Lathyrus odoratus) im-ported from England produce plants, withthick, rigid stems and small leaves, which rare-ly blossom and never yield seed; plants raisedfrom French seed blossom sparingly, but all theflowers are sterile; on the other hand, plantsraised from sweet-peas grown near Darjeelingin Upper India, but originally derived fromEngland, can be successfully cultivated on theplains of India; for they flower and seed profu-sely, and their stems are lax and scandent. Insome of the foregoing cases, as Dr. Hooker hasremarked to me, the greater success may per-haps be attributed to the seeds having beenmore fully ripened under a more favourableclimate; but this view can hardly be extended

to so many cases, including plants, which, frombeing cultivated under a climate hotter thantheir native one, become fitted for a still hotterclimate. We may therefore safely conclude thatplants can to a certain extent become accusto-med to a climate either hotter or colder thantheir own; although the latter cases have beenmore frequently observed.]

We will now consider the means by which ac-climatisation may be effected, namely, throughthe appearance of varieties having a differentconstitution, and through the effects of habit. Inregard to new varieties, there is no evidencethat a change in the constitution of the offs-pring necessarily stands in any direct relationwith the nature of the climate inhabited by theparents. On the contrary, it is certain that hardyand tender varieties of the same species appearin the same country. New varieties thus spon-taneously arising become fitted to slightly dif-ferent climates in two different ways; firstly,

they may have the power, either as seedlings orwhen full-grown, of resisting intense cold, aswith the Moscow pear, or of resisting intenseheat, as with some kinds of Pelargonium, or theflowers may withstand severe frost, as with theForelle pear. Secondly, plants may becomeadapted to climates widely different from theirown, from flowering and fruiting either earlieror later in the season. In both these cases thepower of acclimatisation by man consists sim-ply in the selection and preservation of newvarieties. But without any direct intention onhis part of securing a hardier variety, acclimati-sation may be unconsciously effected by merelyraising tender plants from seed, and by occa-sionally attempting their cultivation furtherand further northwards, as in the case of maize,the orange and the peach.

How much influence ought to be attributed toinherited habit or custom in the acclimatisationof animals and plants is a much more difficult

question. In many cases natural selection canhardly have failed to have come into play andcomplicated the result. It is notorious thatmountain sheep resist severe weather andstorms of snow which would destroy lowlandbreeds; but then mountain sheep have beenthus exposed from time immemorial, and alldelicate individuals will have been destroyed,and the hardiest preserved. So with the Arrin-dy silk-moths of China and India; who can tellhow far natural selection may have taken ashare in the formation of the two races, whichare now fitted for such widely different clima-tes? It seems at first probable that the manyfruit-trees which are so well fitted for the hotsummers and cold winters of North America,in contrast with their poor success under ourclimate, have become adapted through habit;but when we reflect on the multitude of seed-lings annually raised in that country, and thatnone would succeed unless born with a fittingconstitution, it is possible that mere habit may

have done nothing towards their acclimatisa-tion. On the other hand, when we hear thatMerino sheep, bred during no great number ofgenerations at the Cape of Good Hope—thatsome European plants raised during only a fewgenerations in the cooler parts of India, withs-tand the hotter parts of that country much bet-ter than the sheep or seeds imported directlyfrom England, we must attribute some influen-ce to habit. We are led to the same conclusionwhen we hear from Naudin (24/78. Quoted byAsa Gray in 'Am. Journ. of Sc.' 2nd series Ja-nuary 1865 page 106.) that the races of melons,squashes, and gourds, which have long beencultivated in Northern Europe, are comparati-vely more precocious, and need much less heatfor maturing their fruit, than the varieties of thesame species recently brought from tropicalregions. In the reciprocal conversion of summerand winter wheat, barley, and vetches into eachother, habit produces a marked effect in thecourse of a very few generations. The same

thing apparently occurs with the varieties ofmaize, which, when carried from the SouthernStates of America, or into Germany, soon be-came accustomed to their new homes. Withvine-plants taken to the West Indies from Ma-deira, which are said to succeed better thanplants brought directly from France, we havesome degree of acclimatisation in the indivi-dual, independently of the production of newvarieties by seed.

The common experience of agriculturists is ofsome value, and they often advise persons to becautious in trying the productions of one coun-try in another. The ancient agricultural writersof China recommend the preservation and cul-tivation of the varieties peculiar to each coun-try. During the classical period, Columella wro-te, "Vernaculum pecus peregrino longe praes-tantius est." (24/79. For China see 'Memoire surles Chinois' tome 11 1786 page 60. Columella is

quoted by Carlier in 'Journal de Physique' tome24 1784.)

I am aware that the attempt to acclimatise eit-her animals or plants has been called a vainchimera. No doubt the attempt in most casesdeserves to be thus called, if made independen-tly of the production of new varieties endowedwith a different constitution. With plants pro-pagated by buds, habit rarely produces anyeffect; it apparently acts only through successi-ve seminal generations. The laurel, bay, laures-tinus, etc., and the Jerusalem artichoke, whichare propagated by cuttings or tubers, are pro-bably now as tender in England as when firstintroduced; and this appears to be the case withthe potato, which until recently was seldommultiplied by seed. With plants propagated byseed, and with animals, there will be little or noacclimatisation unless the hardier individualsare either intentionally or unconsciously pre-served. The kidney-bean has often been advan-

ced as an instance of a plant which has not be-come hardier since its first introduction intoBritain. We hear, however, on excellent authori-ty (24/80. Messrs. Hardy and Son in 'Garde-ner's Chronicle' 1856 page 589.) that some veryfine seed, imported from abroad, producedplants "which blossomed most profusely, butwere nearly all but abortive, whilst plantsgrown alongside from English seed poddedabundantly;" and this apparently shows somedegree of acclimatisation in our English plants.We have also seen that seedlings of the kidney-bean occasionally appear with a marked powerof resisting frost; but no one, as far as I canhear, has ever separated such hardy seedlings,so as to prevent accidental crossing, and thengathered their seed, and repeated the processyear after year. It may, however, be objectedwith truth that natural selection ought to havehad a decided effect on the hardiness of ourkidney- beans; for the tenderest individualsmust have been killed during every severe

spring, and the hardier preserved. But it shouldbe borne in mind that the result of increasedhardiness would simply be that gardeners, whoare always anxious for as early a crop as possi-ble, would sow their seed a few days earlierthan formerly. Now, as the period of sowingdepends much on the soil and elevation of eachdistrict, and varies with the season; and as newvarieties have often been imported fromabroad, can we feel sure that our kidney-beansare not somewhat hardier? I have not been able,by searching old horticultural works, to answerthis question satisfactorily.

On the whole the facts now given show that,though habit does something towards acclima-tisation, yet that the appearance of constitutio-nally different individuals is a far more effecti-ve agent. As no single instance has been recor-ded either with animals or plants of hardierindividuals having been long and steadily se-lected, though such selection is admitted to be

indispensable for the improvement of any othercharacter, it is not surprising that man has donelittle in the acclimatisation of domesticatedanimals and cultivated plants. We need not,however, doubt that under nature new racesand new species would become adapted to wi-dely different climates, by variation, aided byhabit, and regulated by natural selection.

[ARRESTS OF DEVELOPMENT: RUDI-MENTARY AND ABORTED ORGANS.

Modifications of structure from arrested deve-lopment, so great or so serious as to deserve tobe called monstrosities, are not infrequent withdomesticated animals, but, as they differ muchfrom any normal structure, they require only apassing notice. Thus the whole head may berepresented by a soft nipple-like projection, andthe limbs by mere papillae. These rudiments oflimbs are sometimes inherited, as has been ob-served in a dog. (24/81. Isid. Geoffroy Saint-

Hilaire 'Hist. Nat. des Anomalies' 1836 tome 2pages 210, 223, 224, 395; 'Philosoph. Transact.'1775 page 313.)

Many lesser anomalies appear to be due toarrested development. What the cause of thearrest may be, we seldom know, except in thecase of direct injury to the embryo. That thecause does not generally act at an extremelyearly embryonic period we may infer from theaffected organ seldom being wholly aborted,—a rudiment being generally preserved. The ex-ternal ears are represented by mere vestiges ina Chinese breed of sheep; and in another breed,the tail is reduced "to a little button, suffocatedin a manner, by fat." (24/82. Pallas quoted byYouatt on 'Sheep' page 25.) In tailless dogs andcats a stump is left. In certain breeds of fowlsthe comb and wattles are reduced to rudiments;in the Cochin-China breed scarcely more thanrudiments of spurs exist. With polled Suffolkcattle, "rudiments of horns can often be felt at

an early age" (24/83. Youatt on 'Cattle' 1834page 174.); and with species in a state of nature,the relatively great development of rudimenta-ry organs at an early period of life is highlycharacteristic of such organs. With hornlessbreeds of cattle and sheep, another and singu-lar kind of rudiment has been observed, name-ly, minute dangling horns attached to the skinalone, and which are often shed and growagain. With hornless goats, according to Des-marest (24/84. 'Encyclop. Method.' 1820 page483: see page 500, on the Indian zebu casting itshorns. Similar cases in European cattle weregiven in the third chapter.), the bony protube-rance which properly supports the horn existsas a mere rudiment.

With cultivated plants it is far from rare to findthe petals, stamens, and pistils represented byrudiments, like those observed in natural spe-cies. So it is with the whole seed in many fruits;thus, near Astrakhan there is a grape with mere

traces of seeds, "so small and lying so near thestalk that they are not perceived in eating thegrape." (24/85. Pallas 'Travels' English Translat.volume 1 page 243.) In certain varieties of thegourd, the tendrils, according to Naudin, arerepresented by rudiments or by various mons-trous growths. In the broccoli and cauliflowerthe greater number of the flowers are incapableof expansion, and include rudimentary organs.In the Feather hyacinth (Muscari comosum) inits natural state the upper and central flowersare brightly coloured but rudimentary; undercultivation the tendency to abortion travelsdownwards and outwards, and all the flowersbecome rudimentary; but the abortive stamensand pistils are not so small in the lower as inthe upper flowers. In the Viburnum opulus, onthe other hand, the outer flowers naturallyhave their organs of fructification in a rudimen-tary state, and the corolla is of large size; undercultivation, the change spreads to the centre,and all the flowers become affected. In the

compositae, the so- called doubling of the flo-wers consists in the greater development of thecorolla of the central florets, generally accom-panied with some degree of sterility; and it hasbeen observed (24/86. Mr. Beaton in 'Journal ofHorticulture' May 21, 1861 page 133.) that theprogressive doubling invariably spreads fromthe circumference to the centre,—that is, fromthe ray florets, which so often include rudimen-tary organs, to those of the disc. I may add, asbearing on this subject, that with Asters, seedstaken from the florets of the circumference havebeen found to yield the greatest number ofdouble flowers. (24/87. Lecoq 'De la Feconda-tion' 1862 page 233.) In the above cases we havea natural tendency in certain parts to be rudi-mentary, and this under culture spreads eitherto, or from, the axis of the plant. It deservesnotice, as showing how the same laws governthe changes which natural species and artificialvarieties undergo, that in the species of Cart-hamus, one of the Compositae, a tendency to

the abortion of the pappus may be traced ex-tending from the circumference to the centre ofthe disc as in the so-called doubling of the flo-wers in the members of the same family. Thus,according to A. de Jussieu (24/88. 'Annales duMuseum' tome 6 page 319.), the abortion is onlypartial in Carthamus creticus, but more exten-ded in C. lanatus; for in this species only two orthree of the central seeds are furnished with apappus, the surrounding seeds being eitherquite naked or furnished with a few hairs; andlastly in C. tinctorius, even the central seeds aredestitute of pappus, and the abortion is comple-te.

With animals and plants under domestication,when an organ disappears, leaving only a ru-diment, the loss has generally been sudden, aswith hornless and tailless breeds; and such ca-ses may be ranked as inherited monstrosities.But in some few cases the loss has been gra-dual, and has been effected partly by selection,

as with the rudimentary combs and wattles ofcertain fowls. We have also seen that the wingsof some domesticated birds have been slightlyreduced by disuse, and the great reduction ofthe wings in certain silk-moths, with mere ru-diments left, has probably been aided by dis-use.]

With species in a state of nature, rudimentaryorgans are extremely common. Such organs aregenerally variable, as several naturalists haveobserved; for, being useless, they are not regu-lated by natural selection, and they are more orless liable to reversion. The same rule certainlyholds good with parts which have become ru-dimentary under domestication. We do notknow through what steps under nature rudi-mentary organs have passed in being reducedto their present condition; but we so incessantlysee in species of the same group the finest gra-dations between an organ in a rudimentary andperfect state, that we are led to believe that the

passage must have been extremely gradual. Itmay be doubted whether a change of structureso abrupt as the sudden loss of an organ wouldever be of service to a species in a state of natu-re; for the conditions to which all organisms areclosely adapted usually change very slowly.Even if an organ did suddenly disappear insome one individual by an arrest of develop-ment, intercrossing with the other individualsof the same species would tend to cause its par-tial reappearance; so that its final reductioncould only be effected by some other means.The most probable view is, that a part which isnow rudimentary, was formerly, owing tochanged habits of life, used less and less, beingat the same time reduced in size by disuse, un-til at last it became quite useless and super-fluous. But as most parts or organs are notbrought into action during an early period oflife, disuse or decreased action will not lead totheir reduction until the organism arrives at asomewhat advanced age; and from the princi-

ple of inheritance at corresponding ages thereduction will be transmitted to the offspring atthe same advanced stage of growth. The part ororgan will thus retain its full size in the embr-yo, as we know to be the case with most rudi-ments. As soon as a part becomes useless,another principle, that of economy of growth,will come into play, as it would be an advanta-ge to an organism exposed to severe competi-tion to save the development of any uselesspart; and individuals having the part less deve-loped will have a slight advantage over others.But, as Mr. Mivart has justly remarked, as soonas a part is much reduced, the saving from itsfurther reduction will be utterly insignificant;so that this cannot be effected by natural selec-tion. This manifestly holds good if the part beformed of mere cellular tissue, entailing littleexpenditure of nutriment. How then can thefurther reduction of an already somewhat re-duced part be effected? That this has occurredrepeatedly under Nature is shown by the many

gradations which exist between organs in aperfect state and the merest vestiges of them.Mr. Romanes (24/89. I suggested in 'Nature'(volume 8 pages 432, 505) that with organismssubjected to unfavourable conditions all theparts would tend towards reduction, and thatunder such circumstances any part which wasnot kept up to its standard size by natural selec-tion would, owing to intercrossing, slowly butsteadily decrease. In three subsequent commu-nications to 'Nature' (March 12, April 9, andJuly 2, 1874), Mr. Romanes gives his improvedview.) has, I think, thrown much light on thisdifficult problem. His view, as far as it can begiven in a few words, is as follows: all parts aresomewhat variable and fluctuate in size roundan average point. Now, when a part has alrea-dy begun from any cause to decrease, it is veryimprobable that the variations should be asgreat in the direction of increase as of diminu-tion; for the previous reduction shows that cir-cumstances have not been favourable for its

development; whilst there is nothing to checkvariations in the opposite direction. If this beso, the long continued crossing of many indivi-duals furnished with an organ which fluctuatesin a greater degree towards decrease than to-wards increase, will slowly but steadily lead toits diminution. With respect to the completeand absolute abortion of a part, a distinct prin-ciple, which will be discussed in the chapter onpangenesis, probably comes into action.

With animals and plants reared by man there isno severe or recurrent struggle for existence,and the principle of economy will not comeinto action, so that the reduction of an organwill not thus be aided. So far, indeed, is thisfrom being the case, that in some few instancesorgans, which are naturally rudimentary in theparent-species, become partially redeveloped inthe domesticated descendants. Thus cows, likemost other ruminants, properly have four acti-ve and two rudimentary mamma; but in our

domesticated animals, the latter occasionallybecome considerably developed and yield milk.The atrophied mammae, which, in male domes-ticated animals, including man, have in somerare cases grown to full size and secreted milk,perhaps offer an analogous case. The hind feetof dogs naturally include rudiments of a fifthtoe, and in certain large breeds these toes,though still rudimentary, become considerablydeveloped and are furnished with claws. In thecommon Hen, the spurs and comb are rudi-mentary, but in certain breeds these become,independently of age or disease of the ovaria,well developed. The stallion has canine teeth,but the mare has only traces of the alveoli,which, as I am informed by the eminent veteri-narian Mr. G.T. Brown, frequently contain mi-nute irregular nodules of bone. These nodules,however, sometimes become developed intoimperfect teeth, protruding through the gumsand coated with enamel; and occasionally theygrow to a fourth or even a third of the length of

the canines in the stallion. With plants I do notknow whether the redevelopment of rudimen-tary organs occurs more frequently under cul-ture than under nature. Perhaps the pear-treemay be a case in point, for when wild it bearsthorns, which consist of branches in a rudimen-tary condition and serve as a protection, but,when the tree is cultivated, they are reconver-ted into branches.

CHAPTER 2.XXV.LAWS OF VARIATION, continued.—CORRELATED VARIABILITY.

EXPLANATION OF TERM CORRELA-TION. CONNECTED WITH DEVELOPMENT.MODIFICATIONS CORRELATED WITH THEINCREASED OR DECREASED SIZE OFPARTS. CORRELATED VARIATION OFHOMOLOGOUS PARTS. FEATHERED FEETIN BIRDS ASSUMING THE STRUCTURE OF

THE WINGS. CORRELATION BETWEEN THEHEAD AND THE EXTREMITIES. BETWEENTHE SKIN AND DERMAL APPENDAGES.BETWEEN THE ORGANS OF SIGHT ANDHEARING. CORRELATED MODIFICATIONSIN THE ORGANS OF PLANTS. CORRELA-TED MONSTROSITIES. CORRELATIONBETWEEN THE SKULL AND EARS. SKULLAND CREST OF FEATHERS. SKULL ANDHORNS. CORRELATION OF GROWTHCOMPLICATED BY THE ACCUMULATEDEFFECTS OF NATURAL SELECTION. CO-LOUR AS CORRELATED WITH CONSTITU-TIONAL PECULIARITIES.

All parts of the organisation are to a certainextent connected together; but the connectionmay be so slight that it hardly exists, as withcompound animals or the buds on the sametree. Even in the higher animals various partsare not at all closely related; for one part maybe wholly suppressed or rendered monstrouswithout any other part of the body being affec-

ted. But in some cases, when one part varies,certain other parts always, or nearly always,simultaneously vary; they are then subject tothe law of correlated variation. The whole bodyis admirably co-ordinated for the peculiarhabits of life of each organic being, and may besaid, as the Duke of Argyll insists in his 'Reignof Law' to be correlated for this purpose. Again,in large groups of animals certain structuresalways co-exist: for instance, a peculiar form ofstomach with teeth of peculiar form, and suchstructures may in one sense be said to be corre-lated. But these cases have no necessary con-nection with the law to be discussed in the pre-sent chapter; for we do not know that the initialor primary variations of the several parts werein any way related: slight modifications or in-dividual differences may have been preserved,first in one and then in another part, until thefinal and perfectly co-adapted structure wasacquired; but to this subject I shall presentlyrecur. Again, in many groups of animals the

males alone are furnished with weapons, or areornamented with gay colours; and these cha-racters manifestly stand in some sort of correla-tion with the male reproductive organs, forwhen the latter are destroyed these charactersdisappear. But it was shown in the twelfthchapter that the very same peculiarity may be-come attached at any age to either sex, and af-terwards be exclusively transmitted to the samesex at a corresponding age. In these cases wehave inheritance limited by both sex and age;but we have no reason for supposing that theoriginal cause of the variation was necessarilyconnected with the reproductive organs, orwith the age of the affected being.

In cases of true correlated variation, we aresometimes able to see the nature of the connec-tion; but in most cases it is hidden from us, andcertainly differs in different cases. We can sel-dom say which of two correlated parts firstvaries, and induces a change in the other; or

whether the two are the effects of some com-mon cause. Correlated variation is an importantsubject for us; for when one part is modifiedthrough continued selection, either by man orunder nature, other parts of the organisationwill be unavoidably modified. From this corre-lation it apparently follows that with our do-mesticated animals and plants, varieties rarelyor never differ from one another by a singlecharacter alone.

One of the simplest cases of correlation is that amodification which arises during an early stageof growth tends to influence the subsequentdevelopment of the same part, as well as ofother and intimately connected parts. IsidoreGeoffroy Saint-Hilaire states (25/1. 'Hist. desAnomalies' tome 3 page 392. Prof. Huxley ap-plies the same principle in accounting for theremarkable, though normal, differences in thearrangement of the nervous system in the Mo-llusca, in his paper on the Morphology of the

Cephalous Mollusca in 'Phil. Transact.' 1853page 56.) that this may constantly be observedwith monstrosities in the animal kingdom; andMoquin-Tandon (25/2. 'Elements de Teratolo-gie Veg.' 1841 page 13.) remarks, that, as withplants the axis cannot become monstrous wit-hout in some way affecting the organs subse-quently produced from it, so axial anomaliesare almost always accompanied by deviationsof structure in the appended parts. We shallpresently see that with short-muzzled races ofthe dog certain histological changes in the basalelements of the bones arrest their developmentand shorten them, and this affects the positionof the subsequently developed molar teeth. It isprobable that certain modifications in the lar-vae of insects would affect the structure of themature insects. But we must be careful not toextend this view too far, for during the normalcourse of development, certain species passthrough an extraordinary course of change,

whilst other and closely allied species arrive atmaturity with little change of structure.

Another simple case of correlation is that withthe increased or decreased dimensions of thewhole body, or of any particular part, certainorgans are increased or diminished in number,or are otherwise modified. Thus pigeon fan-ciers have gone on selecting pouters for lengthof body, and we have seen that their vertebraeare generally increased not only in size but innumber, and their ribs in breadth. Tumblershave been selected for their small bodies, andtheir ribs and primary wing-feathers are gene-rally lessened in number. Fantails have beenselected for their large widely-expanded tails,with numerous tail-feathers, and the caudalvertebrae are increased in size and number.Carriers have been selected for length of beak,and their tongues have become longer, but notin strict accordance with the length of beak. Inthis latter breed and in others having large feet,

the number of the scutellae on the toes is grea-ter than in the breeds with small feet. Manysimilar cases could be given. In Germany it hasbeen observed that the period of gestation islonger in large than in small breeds of cattle.With our highly-improved breeds of all kinds,the periods of maturity and of reproductionhave advanced with respect to the age of theanimal; and, in correspondence with this, theteeth are now developed earlier than formerly,so that, to the surprise of agriculturists, the an-cient rules for judging of the age of an animalby the state of its teeth are no longer trustwort-hy. (25/3. Prof. J.B. Simonds on the Age of theOx, Sheep, etc. quoted in 'Gardener's Chronicle'1854 page 588.)

CORRELATED VARIATION OF HOMO-LOGOUS PARTS.

Parts which are homologous tend to vary in thesame manner; and this is what might have been

expected, for such parts are identical in formand structure during an early period of embr-yonic development, and are exposed in the eggor womb to similar conditions. The symmetry,in most kinds of animals, of the correspondingor homologous organs on the right and left si-des of the body, is the simplest case in point;but this symmetry sometimes fails, as with rab-bits having only one ear, or stags with onehorn, or with many-horned sheep which some-times carry an additional horn on one side oftheir heads. With flowers which have regularcorollas, all the petals generally vary in the sa-me manner, as we see in the complicated andsymmetrical pattern, on the flowers, for instan-ce, of the Chinese pink; but with irregular flo-wers, though the petals are of course homolo-gous, this symmetry often fails, as with the va-rieties of the Antirrhinum or snapdragon, orthat variety of the kidney-bean (Phaseolus)which has a white standard-petal.

In the Vertebrata the front and hind limbs arehomologous, and they tend to vary in the samemanner, as we see in long and short legged, orin thick and thin legged races of the horse anddog. Isidore Geoffroy (25/4. 'Hist. des Anoma-lies' tome 1 page 674.) has remarked on the ten-dency of supernumerary digits in man to ap-pear, not only on the right and left sides, but onthe upper and lower extremities. Meckel hasinsisted (25/5. Quoted by Isid. Geoffroy ibidtome 1 page 635.) that, when the muscles of thearm depart in number or arrangement fromtheir proper type, they almost always imitatethose of the leg; and so conversely the varyingmuscles of the leg imitate the normal musclesof the arm.

In several distinct breeds of the pigeon andfowl, the legs and the two outer toes are heavi-ly feathered, so that in the trumpeter pigeonthey appear like little wings. In the feather-legged bantam the "boots" or feathers, which

grow from the outside of the leg and generallyfrom the two outer toes, have, according to theexcellent authority of Mr. Hewitt (25/6. 'ThePoultry Book' by W.B. Tegetmeier 1866 page250.), been seen to exceed the wing-feathers inlength, and in one case were actually nine and ahalf inches long! As Mr. Blyth has remarked tome, these leg-feathers resemble the primarywing- feathers, and are totally unlike the finedown which naturally grows on the legs of so-me birds, such as grouse and owls. Hence itmay be suspected that excess of food has firstgiven redundancy to the plumage, and thenthat the law of homologous variation has led tothe development of feathers on the legs, in aposition corresponding with those on the wing,namely, on the outside of the tarsi and toes. Iam strengthened in this belief by the followingcurious case of correlation, which for a longtime seemed to me utterly inexplicable, namely,that in pigeons of any breed, if the legs are feat-hered, the two outer toes are partially connec-

ted by skin. These two outer toes correspondwith our third and fourth toes. (25/7. Natura-lists differ with respect to the homologies of thedigits of birds; but several uphold the viewabove advanced. See on this subject Dr. E.S.Morse in 'Annals of the Lyceum of Nat. Hist. ofNew York' volume 10 1872 page 16.) Now, inthe wing of the pigeon or of any other bird, thefirst and fifth digits are aborted; the second isrudimentary and carries the so-called "bastard-wing;" whilst the third and fourth digits arecompletely united and enclosed by skin, toget-her forming the extremity of the wing. So thatin feather-footed pigeons, not only does theexterior surface support a row of long feathers,like wing-feathers, but the very same digitswhich in the wing are completely united byskin become partially united by skin in the feet;and thus by the law of the correlated variationof homologous parts we can understand thecurious connection of feathered legs and mem-brane between the two outer toes.

Andrew Knight (24/8. A. Walker on 'Interma-rriage' 1838 page 160.) has remarked that theface or head and the limbs usually vary toget-her in general proportions. Compare, for ins-tance, the limbs of a dray and race horse, or of agreyhound and mastiff. What a monster agreyhound would appear with the head of amastiff! The modern bulldog, however, has finelimbs, but this is a recently-selected character.From the measurements given in the sixthchapter, we see that in several breeds of thepigeon the length of the beak and the size of thefeet are correlated. The view which, as beforeexplained, seems the most probable is, that dis-use in all cases tends to diminish the feet, thebeak becoming at the same time shorterthrough correlation; but that in some fewbreeds in which length of beak has been a selec-ted point, the feet, notwithstanding disuse,have increased in size through correlation. Inthe following case some kind of correlation isseen to exist between the feet and beak: several

specimens have been sent to Mr. Bartlett at dif-ferent times, as hybrids between ducks andfowls, and I have seen one; these were, asmight be expected, ordinary ducks in a semi-monstrous condition, and in all of them theswimming-web between the toes was quitedeficient or much reduced, and in all the beakwas narrow and ill-shaped.

With the increased length of the beak in pi-geons, not only the tongue increases in length,but likewise the orifice of the nostrils. But theincreased length of the orifice of the nostrilsperhaps stands in closer correlation with thedevelopment of the corrugated skin or wattle atthe base of the beak, for when there is muchwattle round the eyes, the eyelids are greatlyincreased or even doubled in length.

There is apparently some correlation even incolour between the head and the extremities.Thus with horses a large white star or blaze onthe forehead is generally accompanied by white

feet. (25/9. 'The Farrier and Naturalist' volume1 1828 page 456. A gentleman who has atten-ded to this point, tells me that about three-fourths of white-faced horses have white legs.)With white rabbits and cattle, dark marks oftenco-exist on the tips of the ears and on the feet.In black and tan dogs of different breeds, tan-coloured spots over the eyes and tan-colouredfeet almost invariably go together. These lattercases of connected colouring may be due eitherto reversion or to analogous variation,—subjects to which I shall hereafter return,—butthis does not necessarily determine the ques-tion of their original correlation. Mr. H.W. Jack-son informs me that he has observed manyhundred white-footed cats, and he finds that allare more or less conspicuously marked withwhite on the front of the neck or chest.

The lopping forwards and downwards of theimmense ears of fancy rabbits seems partly dueto the disuse of the muscles, and partly to the

weight and length of the ears, which have beenincreased by selection during many genera-tions. Now, with the increased size and chan-ged direction of the ears not only has the bonyauditory meatus become changed in outline,direction, and greatly in size, but the wholeskull has been slightly modified. This could beclearly seen in "half-lops"—that is, in rabbitswith only one ear lopping forward— for theopposite sides of their skulls were not strictlysymmetrical. This seems to me a curious ins-tance of correlation, between hard bones andorgans so soft and flexible, as well as so unim-portant under a physiological point of view, asthe external ears. The result no doubt is largelydue to mere mechanical action, that is, to theweight of the ears, on the same principle thatthe skull of a human infant is easily modifiedby pressure.

The skin and the appendages of hair, feathers,hoofs, horns, and teeth, are homologous over

the whole body. Every one knows that the co-lour of the skin and that of the hair usually va-ry together; so that Virgil advises the shepherdto look whether the mouth and tongue of theram are black, lest the lambs should not be pu-rely white. The colour of the skin and hair, andthe odour emitted by the glands of the skin, aresaid (25/10. Godron 'Sur l'Espece' tome 2 page217.) to be connected, even in the same race ofmen. Generally the hair varies in the same wayall over the body in length, fineness, and curli-ness. The same rule holds good with feathers,as we see with the laced and frizzled breedsboth of fowls and pigeons. In the common cockthe feathers on the neck and loins are always ofa particular shape, called hackles: now in thePolish breed, both sexes are characterised by atuft of feathers on the head, and through corre-lation these feathers in the male always assumethe form of hackles. The wing and tail-feathers,though arising from parts not homologous,vary in length together; so that long or short

winged pigeons generally have long or shorttails. The case of the Jacobin-pigeon is morecurious, for the wing and tail feathers are re-markably long; and this apparently has arisenin correlation with the elongated and reversedfeathers on the back of the neck, which formthe hood.

The hoofs and hair are homologous appenda-ges; and a careful observer, namely Azara(25/11. 'Quadrupedes du Paraguay' tome 2page 333.), states that in Paraguay horses ofvarious colours are often born with their haircurled and twisted like that on the head of anegro. This peculiarity is strongly inherited.But what is remarkable is that the hoofs of the-se horses "are absolutely like those of a mule."The hair also of their manes and tails is inva-riably much shorter than usual, being onlyfrom four to twelve inches in length; so thatcurliness and shortness of the hair are here, aswith the negro, apparently correlated.

With respect to the horns of sheep, Youatt(25/12. 'On Sheep' page 142.) remarks that"multiplicity of horns is not found in any breedof much value; it is generally accompanied bygreat length and coarseness of the fleece." Seve-ral tropical breeds of sheep which are clothedwith hair instead of wool, have horns almostlike those of a goat. Sturm (25/13. 'Ueber Ra-cen, Kreuzungen' etc. 1825 s. 24.) expressly de-clares that in different races the more the woolis curled the more the horns are spirally twis-ted. We have seen in the third chapter, whereother analogous facts have been given, that theparent of the Mauchamp breed, so famous forits fleece, had peculiarly shaped horns. Theinhabitants of Angora assert (25/14. Quotedfrom Conolly in 'The Indian Field' February1859 volume 2 page 266.) that "only the whitegoats which have horns wear the fleece in thelong curly locks that are so much admired; tho-se which are not horned having a comparative-ly close coat." From these cases we may infer

that the hair or wool and the horns tend to varyin a correlated manner. (25/15. In the thirdchapter I have said that "the hair and horns areso closely related to each other, that they areapt to vary together." Dr. Wilckens ("Darwin'sTheorie" 'Jahrbuch der Deutschen Viehzucht'1866 1. Heft) translates my words into "lang-und grobhaarige Thiere sollen geneigter sein,lange und viele Horner zu bekommen" and hethen justly disputes this proposition; but what Ihave really said, in accordance with the aut-horities just quoted, may, I think, be trusted.)Those who have tried hydropathy are awarethat the frequent application of cold water sti-mulates the skin; and whatever stimulates theskin tends to increase the growth of the hair, asis well shown in the abnormal growth of hairnear old inflamed surfaces. Now, ProfessorLow (25/16. 'Domesticated Animals of the Bri-tish Islands' pages 307, 368. Dr. Wilckens ar-gues ('Landwirth. Wochenblatt' Nr. 10 1869) tothe same effect with respect to domestic ani-

mals in Germany.) is convinced that with thedifferent races of British cattle thick skin andlong hair depend on the humidity of the clima-te which they inhabit. We can thus see how ahumid climate might act on the horns—in thefirst place directly on the skin and hair, andsecondly by correlation on the horns. The pre-sence or absence of horns, moreover, both inthe case of sheep and cattle, acts, as will presen-tly be shown, by some sort of correlation on theskull.

With respect to hair and teeth, Mr. Yarrell(25/17. 'Proceedings Zoolog. Soc.' 1833 page113.) found many of the teeth deficient in threehairless "Egyptian dogs," and in a hairless te-rrier. The incisors, canines, and the premolarssuffered most, but in one case all the teeth, ex-cept the large tubercular molar on each side,were deficient. With man several striking caseshave been recorded (25/18. Sedgwick 'Brit. andForeign Medico-Chirurg. Review' April 1863

page 453.) of inherited baldness with inheriteddeficiency, either complete or partial, of theteeth. I may give an analogous case, communi-cated to me by Mr. W. Wedderburn, of a Hin-doo family in Scinde, in which ten men, in thecourse of four generations, were furnished, inboth jaws taken together, with only four smalland weak incisor teeth and with eight posteriormolars. The men thus affected have very littlehair on the body, and become bald early in life.They also suffer much during hot weather fromexcessive dryness of the skin. It is remarkablethat no instance has occurred of a daughterbeing thus affected; and this fact reminds ushow much more liable men are in England tobecome bald than women. Though the daugh-ters in the above family are never affected, theytransmit the tendency to their sons; and no casehas occurred of a son transmitting it to his sons.The affection thus appears only in alternategenerations, or after longer intervals. There is asimilar connection between hair and teeth, ac-

cording to Mr. Sedgwick, in those rare cases inwhich the hair has been renewed in old age, forthis has "usually been accompanied by a rene-wal of the teeth." I have remarked in a formerpart of this volume that the great reduction inthe size of the tusks in domestic boars probablystands in close relation with their diminishedbristles, due to a certain amount of protection;and that the reappearance of the tusks in boars,which have become feral and are fully exposedto the weather, probably depends on the reap-pearance of the bristles. I may add, though notstrictly connected with our present point, thatan agriculturist (25/19. 'Gardener's Chronicle'1849 page 205.) asserts that "pigs with little hairon their bodies are most liable to lose their tails,showing a weakness of the tegumental structu-re. It may be prevented by crossing with a morehairy breed."

In the previous cases deficient hair, and teethdeficient in number or size, are apparently con-

nected. In the following cases abnormally re-dundant hair, and teeth either deficient or re-dundant, are likewise connected. Mr. Crawfurd(25/20. 'Embassy to the Court of Ava' volume 1page 320.) saw at the Burmese Court a man,thirty years old, with his whole body, exceptthe hands and feet, covered with straight silkyhair, which on the shoulders and spine was fiveinches in length. At birth the ears alone werecovered. He did not arrive at puberty, or shedhis milk teeth, until twenty years old; and atthis period he acquired five teeth in the upperjaw, namely, four incisors and one canine, andfour incisor teeth in the lower jaw; all the teethwere small. This man had a daughter who wasborn with hair within her ears; and the hairsoon extended over her body. When CaptainYule (25/21. 'Narrative of a Mission to theCourt of Ava in 1855' page 94.) visited theCourt, he found this girl grown up; and shepresented a strange appearance with even hernose densely covered with soft hair. Like her

father, she was furnished with incisor teethalone. The King had with difficulty bribed aman to marry her, and of her two children, one,a boy fourteen months old, had hair growingout of his ears, with a beard and moustache.This strange peculiarity has, therefore, beeninherited for three generations, with the molarteeth deficient in the grandfather and mother;whether these teeth would likewise fail in theinfant could not then be told.

A parallel case of a man fifty-five years old, andof his son, with their faces covered with hair,has recently occurred in Russia. Dr. Alex.Brandt has sent me an account of this case, to-gether with specimens of the extremely finehair from the cheeks. The man is deficient inteeth, possessing only four incisors in the lowerand two in the upper jaw. His son, about threeyears old, has no teeth except four lower inci-sors. The case, as Dr. Brandt remarks in his let-ter, no doubt is due to an arrest of development

in the hair and teeth. We here see how inde-pendent of the ordinary conditions of existencesuch arrests must be, for the lives of a Russianpeasant and of a native of Burmah are as diffe-rent as possible. (25/22. I owe to the kindnessof M. Chauman, of St. Petersburg, excellentphotographs of this man and his son, both ofwhom have since been exhibited in Paris andLondon.)

Here is another and somewhat different casecommunicated to me by Mr. Wallace on theauthority of Dr. Purland, a dentist: Julia Pastra-na, a Spanish dancer, was a remarkably finewoman, but she had a thick masculine beardand a hairy forehead; she was photographed,and her stuffed skin was exhibited as a show;but what concerns us is, that she had in boththe upper and lower jaw an irregular double setof teeth, one row being placed within the other,of which Dr. Purland took a cast. From the re-dundancy of teeth her mouth projected, and

her face had a gorilla-like appearance. Thesecases and those of the hairless dogs forcibly callto mind the fact, that the two orders of mam-mals—namely, the Edentata and Cetacea—which are the most abnormal in their dermalcovering, are likewise the most abnormal eitherby deficiency or redundancy of teeth.

The organs of sight and hearing are generallyadmitted to be homologous with one anotherand with various dermal appendages; hencethese parts are liable to be abnormally affectedin conjunction. Mr. White Cowper says "that inall cases of double microphthalmia broughtunder his notice he has at the same time metwith defective development of the dental sys-tem." Certain forms of blindness seem to beassociated with the colour of the hair; a manwith black hair and a woman with light-coloured hair, both of sound constitution, ma-rried and had nine children, all of whom wereborn blind; of these children, five "with dark

hair and brown iris were afflicted with amau-rosis; the four others, with light-coloured hairand blue iris, had amaurosis and cataract con-joined." Several cases could be given, showingthat some relation exists between various affec-tions of the eyes and ears; thus Liebreich statesthat out of 241 deaf-mutes in Berlin, no lessthan fourteen suffered from the rare diseasecalled pigmentary retinitis. Mr. White Cowperand Dr. Earle have remarked that inability todistinguish different colours, or colour-blindness, "is often associated with a corres-ponding inability to distinguish musicalsounds." (25/23. These statements are takenfrom Mr. Sedgwick in the 'Medico-Chirurg.Review' July 1861 page 198; April 1863 pages455 and 458. Liebreich is quoted by ProfessorDevay in his 'Mariages Consanguins' 1862 page116.)

Here is a more curious case: white cats, if theyhave blue eyes, are almost always deaf. I for-

merly thought that the rule was invariable, butI have heard of a few authentic exceptions. Thefirst two notices were published in 1829 andrelate to English and Persian cats: of the latter,the Rev. W.T. Bree possessed a female, and hestates, "that of the offspring produced at oneand the same birth, such as, like the mother,were entirely white (with blue eyes) were, likeher, invariably deaf; while those that had theleast speck of colour on their fur, as invariablypossessed the usual faculty of hearing." (25/24.Loudon's 'Mag. of Nat. Hist.' volume 1 1829pages 66, 178. See also Dr. P. Lucas 'L'Hered.Nat.' tome 1 page 428 on the inheritance ofdeafness in cats. Mr. Lawson Tait states ('Natu-re' 1873 page 323) that only male cats are thusaffected; but this must be a hasty generalisa-tion. The first case recorded in England by Mr.Bree related to a female, and Mr. Fox informsme that he has bred kittens from a white femalewith blue eyes, which was completely deaf; hehas also observed other females in the same

condition.) The Rev. W. Darwin Fox informsme that he has seen more than a dozen instan-ces of this correlation in English, Persian, andDanish cats; but he adds "that, if one eye, as Ihave several times observed, be not blue, thecat hears. On the other hand, I have never seena white cat with eyes of the common colourthat was deaf." In France Dr. Sichel (25/25.'Annales des Sc. Nat.' Zoolog. 3rd series 1847tome 8 page 239.) has observed during twentyyears similar facts; he adds the remarkable caseof the iris beginning, at the end of four months,to grow dark-coloured, and then the cat firstbegan to hear.

This case of correlation in cats has struck manypersons as marvellous. There is nothing unu-sual in the relation between blue eyes and whi-te fur; and we have already seen that the or-gans of sight and hearing are often simulta-neously affected. In the present instance thecause probably lies in a slight arrest of deve-

lopment in the nervous system in connectionwith the sense-organs. Kittens during the firstnine days, whilst their eyes are closed, appearto be completely deaf; I have made a greatclanging noise with a poker and shovel close totheir heads, both when they were asleep andawake, without producing any effect. The trialmust not be made by shouting close to theirears, for they are, even when asleep, extremelysensitive to a breath of air. Now, as long as theeyes continue closed, the iris is no doubt blue,for in all the kittens which I have seen this co-lour remains for some time after the eyelidsopen. Hence, if we suppose the development ofthe organs of sight and hearing to be arrested atthe stage of the closed eyelids, the eyes wouldremain permanently blue and the ears wouldbe incapable of perceiving sound; and weshould thus understand this curious case. As,however, the colour of the fur is determinedlong before birth, and as the blueness of theeyes and the whiteness of the fur are obviously

connected, we must believe that some primarycause acts at a much earlier period.

The instances of correlated variability hithertogiven have been chiefly drawn from the animalkingdom, and we will now turn to plants. Lea-ves, sepals, petals, stamens, and pistils are allhomologous. In double flowers we see that thestamens and pistils vary in the same manner,and assume the form and colour of the petals.In the double columbine (Aquilegia vulgaris),the successive whorls of stamens are convertedinto cornucopias, which are enclosed withinone another and resemble the true petals. Inhose-in-hose flowers the sepals mock the pe-tals. In some cases the flowers and leaves varytogether in tint: in all the varieties of the com-mon pea, which have purple flowers, a purplemark may be seen on the stipules.

M. Faivre states that with the varieties of Pri-mula sinensis the colour of the flower is eviden-tly correlated with the colour of the under side

of the leaves; and he adds that the varietieswith fimbriated flowers almost always havevoluminous, balloon-like calyces. (25/26. 'Re-vue des Cours Scientifiques' June 5, 1869 page430.) With other plants the leaves and fruit orseeds vary together in colour, as in a curiouspale-leaved variety of the sycamore, which hasrecently been described in France (25/27. 'Gar-dener's Chronicle' 1864 page 1202.), and as inthe purple-leaved hazel, in which the leaves,the husk of the nut, and the pellicle round thekernel are all coloured purple. (25/28. Verlotgives several other instances 'Des Varietes' 1865page 72.) Pomologists can predict to a certainextent, from the size and appearance of the lea-ves of their seedlings, the probable nature ofthe fruit; for, as Van Mons remarks (25/29. 'Ar-bres Fruitiers' 1836 tome 2 pages 204, 226.) va-riations in the leaves are generally accompa-nied by some modification in the flower, andconsequently in the fruit. In the Serpent melon,which has a narrow tortuous fruit above a yard

in length, the stem of the plant, the peduncle ofthe female flower, and the middle lobe of theleaf, are all elongated in a remarkable manner.On the other hand, several varieties of Cucurbi-ta, which have dwarfed stems, all produce, asNaudin remarks, leaves of the same peculiarshape. Mr. G. Maw informs me that all the va-rieties of the scarlet Pelargoniums which havecontracted or imperfect leaves have contractedflowers: the difference between "Brilliant" andits parent "Tom Thumb" is a good instance ofthis. It may be suspected that the curious casedescribed by Risso (25/30. 'Annales du Mu-seum' tome 20 page 188.), of a variety of theOrange which produces on the young shootsrounded leaves with winged petioles, and af-terwards elongated leaves on long but winglesspetioles, is connected with the remarkablechange in form and nature which the fruit un-dergoes during its development.

In the following instance we have the colourand the form of the petals apparently correla-ted, and both dependent on the nature of theseason. An observer, skilled in the subject, wri-tes (25/31. 'Gardener's Chronicle' 1843 page877.), "I noticed, during the year 1842, that eve-ry Dahlia of which the colour had any tendencyto scarlet, was deeply notched—indeed, to sogreat an extent as to give the petals the appea-rance of a saw; the indentures were, in someinstances, more than a quarter of an inch deep."Again, Dahlias which have their petals tippedwith a different colour from the rest of the flo-wer are very inconstant, and during certainyears some, or even all the flowers, becomeuniformly coloured; and it has been observedwith several varieties (25/32. Ibid 1845 page102.) that when this happens the petals growmuch elongated and lose their proper shape.This, however, may be due to reversion, both incolour and form, to the aboriginal species.

In this discussion on correlation, we have hit-herto treated of cases in which we can partlyunderstand the bond of connection; but I willnow give cases in which we cannot even con-jecture, or can only very obscurely see, the na-ture of the bond. Isidore Geoffroy Saint-Hilaire,in his work on Monstrosities, insists (25/33.'Hist. des Anomalies' tome 3 page 402. See alsoM. Camille Dareste 'Recherches sur les Condi-tions' etc. 1863 pages 16, 48.), "que certainesanomalies coexistent rarement entr'elles, d'au-tres frequemment, d'autres enfin presque cons-tamment, malgre la difference tres-grande deleur nature, et quoiqu'elles puissent paraitreCOMPLETEMENT INDEPENDANTES lesunes des autres." We see something analogousin certain diseases: thus in a rare affection ofthe renal capsules (of which the functions areunknown), the skin becomes bronzed; and inhereditary syphilis, as I hear from Sir J. Paget,both the milk and the second teeth assume apeculiar and characteristic form. Professor Ro-

lleston, also, informs me that the incisor teethare sometimes furnished with a vascular rim incorrelation with intra-pulmonary deposition oftubercles. In other cases of phthisis and of cya-nosis the nails and finger- ends become clubbedlike acorns. I believe that no explanation hasbeen offered of these and of many other casesof correlated disease.

What can be more curious and less intelligiblethan the fact previously given, on the authorityof Mr. Tegetmeier, that young pigeons of allbreeds, which when mature have white, ye-llow, silver-blue, or dun-coloured plumage,come out of the egg almost naked; whereaspigeons of other colours when first born areclothed with plenty of down? White Pea-fowls,as has been observed both in England andFrance (25/34. Rev. E.S. Dixon 'OrnamentalPoultry' 1848 page 111; Isidore Geoffroy 'Hist.Anomalies' tome 1 page 211.), and as I havemyself seen, are inferior in size to the common

coloured kind; and this cannot be accounted forby the belief that albinism is always accompa-nied by constitutional weakness; for white oralbino moles are generally larger than thecommon kind.

To turn to more important characters: the niatacattle of the Pampas are remarkable from theirshort foreheads, upturned muzzles, and curvedlower jaws. In the skull the nasal and premaxi-llary bones are much shortened, the maxillariesare excluded from any junction with the nasals,and all the bones are slightly modified, even tothe plane of the occiput. From the analogouscase of the dog, hereafter to be given, it is pro-bable that the shortening of the nasal and ad-joining bones is the proximate cause of the ot-her modifications in the skull, including theupward curvature of the lower jaw, though wecannot follow out the steps by which thesechanges have been effected.

Polish fowls have a large tuft of feathers ontheir heads; and their skulls are perforated bynumerous holes, so that a pin can be driveninto the brain without touching any bone. Thatthis deficiency of bone is in some way connec-ted with the tuft of feathers is clear from tuftedducks and geese likewise having perforatedskulls. The case would probably be consideredby some authors as one of balancement or com-pensation. In the chapter on Fowls, I haveshown that with Polish fowls the tuft of feat-hers was probably at first small; by continuedselection it became larger, and then rested on afibrous mass; and finally, as it became still lar-ger, the skull itself became more and more pro-tuberant until it acquired its present extraordi-nary structure. Through correlation with theprotuberance of the skull, the shape and eventhe relative connection of the premaxillary andnasal bones, the shape of the orifice of the nos-trils, the breadth of the frontal bone, the shapeof the post-lateral processes of the frontal and

squamosal bones, and the direction of the bonycavity of the ear, have all been modified. Theinternal configuration of the skull and the who-le shape of the brain have likewise been alteredin a truly marvellous manner.

After this case of the Polish fowl it would besuperfluous to do more than refer to the detailspreviously given on the manner in which thechanged form of the comb has affected theskull, in various breeds of the fowl, causing bycorrelation crests, protuberances, and depres-sions on its surface.

With our cattle and sheep the horns stand inclose connection with the size of the skull, andwith the shape of the frontal bones; thus Cline(25/35. 'On the Breeding of Domestic Animals'1829 page 6.) found that the skull of a hornedram weighed five times as much as that of ahornless ram of the same age. When cattle be-come hornless, the frontal bones are "materiallydiminished in breadth towards the poll;" and

the cavities between the bony plates "are not sodeep, nor do they extend beyond the frontals."(25/36. Youatt on 'Cattle' 1834 page 283.)

It may be well here to pause and observe howthe effects of correlated variability, of the in-creased use of parts, and of the accumulation ofso- called spontaneous variations through na-tural selection, are in many cases inextricablycommingled. We may borrow an illustrationfrom Mr. Herbert Spencer, who remarks that,when the Irish elk acquired its gigantic horns,weighing above one hundred pounds, nume-rous co-ordinated changes of structure wouldhave been indispensable,—namely, a thickenedskull to carry the horns; strengthened cervicalvertebrae, with strengthened ligaments; enlar-ged dorsal vertebrae to support the neck, withpowerful fore-legs and feet; all these partsbeing supplied with proper muscles, blood-vessels, and nerves. How then could these ad-mirably co-ordinated modifications of structure

have been acquired? According to the doctrinewhich I maintain, the horns of the male elk we-re slowly gained through sexual selection,—that is, by the best-armed males conquering theworse-armed, and leaving a greater number ofdescendants. But it is not at all necessary thatthe several parts of the body should have si-multaneously varied. Each stag presents indi-vidual characteristics, and in the same districtthose which had slightly heavier horns, orstronger necks, or stronger bodies, or were themost courageous, would secure the greaternumber of does, and consequently have a grea-ter number of offspring. The offspring wouldinherit, in a greater or less degree, these samequalities, would occasionally intercross withone another, or with other individuals varyingin some favourable manner; and of their offs-pring, those which were the best endowed inany respect would continue multiplying; andso onwards, always progressing, sometimes inone direction, and sometimes in another, to-

wards the excellently co-ordinated structure ofthe male elk. To make this clear, let us reflect onthe probable steps, as shown in the twentiethchapter, by which our race and dray horseshave arrived at their present state of excellence;if we could view the whole series of interme-diate forms between one of these animals andan early unimproved progenitor, we shouldbehold a vast number of animals, not equallyimproved in each generation throughout theirentire structure, but sometimes a little more inone point, and sometimes in another, yet on thewhole gradually approaching in character toour present race or dray horses, which are soadmirably fitted in the one case for fleetnessand in the other for draught.

Although natural selection would thus (25/37.Mr. Herbert Spencer 'Principles of Biology' 1864volume 1 pages 452, 468 takes a different view;and in one place remarks: "We have seen rea-son to think that, as fast as essential faculties

multiply, and as fast as the number of organsthat co-operate in any given function increases,indirect equilibration through natural selectionbecomes less and less capable of producingspecific adaptations; and remains fully capableonly of maintaining the general fitness of cons-titution to conditions." This view that naturalselection can do little in modifying the higheranimals surprises me, seeing that man's selec-tion has undoubtedly effected much with ourdomesticated quadrupeds and birds.) tend togive to the male elk its present structure, yet itis probable that the inherited effects of use, andof the mutual action of part on part, have beenequally or more important. As the horns gra-dually increased in weight the muscles of theneck, with the bones to which they are atta-ched, would increase in size and strength; andthese parts would react on the body and legs.Nor must we overlook the fact that certainparts of the skull and the extremities would,judging by analogy, tend from the first to vary

in a correlated manner. The increased weight ofthe horns would also act directly on the skull,in the same manner as when one bone is remo-ved in the leg of a dog, the other bone, whichhas to carry the whole weight of the body, in-creases in thickness. But from the fact givenwith respect to horned and hornless cattle, it isprobable that the horns and skull would im-mediately act on each other through the princi-ple of correlation. Lastly, the growth and sub-sequent wear and tear of the augmented mus-cles and bones would require an increasedsupply of blood, and consequently increasedsupply of food; and this again would requireincreased powers of mastication, digestion,respiration, and excretion.

COLOUR AS CORRELATED WITH CONS-TITUTIONAL PECULIARITIES.

It is an old belief that with man there is a con-nection between complexions and constitution;

and I find that some of the best authorities be-lieve in this to the present day. (25/38. Dr.Prosper Lucas apparently disbelieves in anysuch connection; 'L'Hered. Nat.' tome 2 pages88-94.) Thus Dr. Beddoe by his tables shows(25/39. 'British Medical Journal' 1862 page 433.)that a relation exists between liability to con-sumption and the colour of the hair, eyes, andskin. It has been affirmed (25/40. Boudin 'Geo-graph. Medicale' tome 1 page 406.) that, in theFrench army which invaded Russia, soldiershaving a dark complexion from the southernparts of Europe, withstood the intense coldbetter than those with lighter complexions fromthe north; but no doubt such statements areliable to error.

In the second chapter on Selection I have givenseveral cases proving that with animals andplants differences in colour are correlated withconstitutional differences, as shown by greateror less immunity from certain diseases, from

the attacks of parasitic plants and animals, fromscorching by the sun, and from the action ofcertain poisons. When all the individuals of anyone variety possess an immunity of this nature,we do not know that it stands in any sort ofcorrelation with their colour; but when severalsimilarly coloured varieties of the same speciesare thus characterised, whilst other colouredvarieties are not thus favoured, we must belie-ve in the existence of a correlation of this kind.Thus, in the United States purple- fruitedplums of many kinds are far more affected by acertain disease than green or yellow-fruitedvarieties. On the other hand, yellow-fleshedpeaches of various kinds suffer from anotherdisease much more than the white-fleshed va-rieties. In the Mauritius red sugar-canes aremuch less affected by a particular disease thanthe white canes. White onions and verbenas arethe most liable to mildew; and in Spain thegreen-fruited grapes suffered from the vine-disease more than other coloured varieties.

Dark-coloured pelargoniums and verbenas aremore scorched by the sun than varieties of ot-her colours. Red wheats are believed to be har-dier than white; and red-flowered hyacinthswere more injured during one particular winterin Holland than other coloured varieties. Withanimals, white terriers suffer most from thedistemper, white chickens from a parasiticworm in their tracheae, white pigs from scor-ching by the sun, and white cattle from flies;but the caterpillars of the silk-moth which yieldwhite cocoons suffered in France less from thedeadly parasitic fungus than those producingyellow silk.

The cases of immunity from the action of cer-tain vegetable poisons, in connexion with co-lour, are more interesting, and are at presentwholly inexplicable. I have already given a re-markable instance, on the authority of Profes-sor Wyman, of all the hogs, excepting those of ablack colour, suffering severely in Virginia

from eating the root of the Lachnanthes tincto-ria. According to Spinola and others (25/41.This fact and the following cases, when notstated to the contrary, are taken from a verycurious paper by Prof. Heusinger in 'Wo-chenschrift fur Heilkunde' May 1846 s. 277.Settegast 'Die Thierzucht' 1868 page 39 saysthat white or white-spotted sheep suffer likepigs, or even die from eating buckwheat; whilstblack or dark-woolled individuals are not in theleast affected.), buckwheat (Po1ygonum fago-pyrum), when in flower, is highly injurious towhite or white-spotted pigs, if they are exposedto the heat of the sun, but is quite innocuous toblack pigs. According to two accounts, the Hy-pericum crispum in Sicily is poisonous to whitesheep alone; their heads swell, their wool fallsoff, and they often die; but this plant, accordingto Lecce, is poisonous only when it grows inswamps; nor is this improbable, as we knowhow readily the poisonous principle in plants is

influenced by the conditions under which theygrow.

Three accounts have been published in EasternPrussia, of white and white- spotted horsesbeing greatly injured by eating mildewed andhoneydewed vetches; every spot of skin bea-ring white hairs becoming inflamed and gan-grenous. The Rev. J. Rodwell informs me thathis father turned out about fifteen cart-horsesinto a field of tares which in parts swarmedwith black aphides, and which no doubt werehoneydewed, and probably mildewed; the hor-ses, with two exceptions, were chestnuts andbays with white marks on their faces and pas-terns, and the white parts alone swelled andbecame angry scabs. The two bay horses withno white marks entirely escaped all injury. InGuernsey, when horses eat fool's parsley (Aet-husa cynapium) they are sometimes violentlypurged; and this plant "has a peculiar effect onthe nose and lips, causing deep cracks and ul-

cers, particularly on horses with white muzz-les." (25/42. Mr. Mogford in the 'Veterinarian'quoted in 'The Field' January 22, 1861 page545.) With cattle, independently of the action ofany poison, cases have been published byYouatt and Erdt of cutaneous diseases withmuch constitutional disturbance (in one instan-ce after exposure to a hot sun) affecting everysingle point which bore a white hair, but com-pletely passing over other parts of the body.Similar cases have been observed with horses.(25/43. 'Edinburgh Veterinary Journal' October1860 page 347.)

We thus see that not only do those parts of theskin which bear white hair differ in a remarka-ble manner from those bearing hair of any ot-her colour, but that some great constitutionaldifference must be correlated with the colour ofthe hair; for in the above-mentioned cases, ve-getable poisons caused fever, swelling of thehead, as well as other symptoms, and even

death, to all the white, or white-spotted ani-mals.

CHAPTER 2.XXVI.LAWS OF VARIATION, continued.—SUMMARY.

THE FUSION OF HOMOLOGOUS PARTS.THE VARIABILITY OF MULTIPLE ANDHOMOLOGOUS PARTS. COMPENSATIONOF GROWTH. MECHANICAL PRESSURE.RELATIVE POSITION OF FLOWERS WITHRESPECT TO THE AXIS, AND OF SEEDS INTHE OVARY, AS INDUCING VARIATION.ANALOGOUS OR PARALLEL VARIETIES.SUMMARY OF THE THREE LAST CHAP-TERS.

THE FUSION OF HOMOLOGOUS PARTS.

Geoffroy Saint-Hilaire formerly propoundedwhat he called la loi de l'affinite de soi pour soi,which has been discussed and illustrated by hisson, Isidore, with respect to monsters in theanimal kingdom (26/1. 'Hist. des Anomalies'1832 tome 1 pages 22, 537-556; tome 3 page462.), and by Moquin-Tandon, with respect tomonstrous plants.

This law seems to imply that homologous partsactually attract one another and then unite. Nodoubt there are many wonderful cases, inwhich such parts become intimately fused to-gether. This is perhaps best seen in monsterswith two heads, which are united, summit tosummit, or face to face, or Janus-like, back toback, or obliquely side to side. In one instanceof two heads united almost face to face, but alittle obliquely, four ears were developed, andon one side a perfect face, which was manifes-tly formed by the fusion of two half-faces.Whenever two bodies or two heads are united,

each bone, muscle, vessel, and nerve on the lineof junction appears as if it had sought out itsfellow, and had become completely fused withit. Lereboullet (26/2. 'Comptes Rendus' 1855pages 855, 1039.), who carefully studied thedevelopment of double monsters in fishes, ob-served in fifteen instances the steps by whichtwo heads gradually became united into one. Inall such cases it is now thought by the greaternumber of capable judges that the homologousparts do not attract each other, but that in thewords of Mr. Lowne (26/3. 'Catalogue of theTeratological Series in the Museum of the R.Coll. of Surgeons' 1872 page 16.): "As uniontakes place before the differentiation of distinctorgans occurs, these are formed in continuitywith each other." He adds that organs alreadydifferentiated probably in no case become uni-ted to homologous ones. M. Dareste does notspeak (26/4. 'Archives de Zoolog. Exper.' Ja-nuary 1874 page 78.) quite decisively againstthe law of soi pour soi, but concludes by sa-

ying, "On se rend parfaitement compte de laformation des monstres, si l'on admet que lesembryons qui se soudent appartiennent a unmeme oeuf; qu'ils s'unissent en meme tempsqu'ils se forment, et que la soudure ne se pro-duit que pendant la premiere periode de la vieembryonnaire, celle ou les organes ne sont en-core constitues que par des blastemes homoge-nes."

By whatever means the abnormal fusion ofhomologous parts is effected, such cases throwlight on the frequent presence of organs whichare double during an embryonic period (andthroughout life in other and lower members ofthe same class) but which afterwards unite by anormal process into a single medial organ. Inthe vegetable kingdom Moquin-Tandon (26/5.'Teratologie Veg.' 1841 livre 3.) gives a long listof cases, showing how frequently homologousparts, such as leaves, petals, stamens, and pis-tils, flowers, and aggregates of homologous

parts, such as buds, as well as fruit, becomeblended, both normally and abnormally, withperfect symmetry into one another.

THE VARIABILITY OF MULTIPLE ANDHOMOLOGOUS PARTS.

Isidore Geoffroy (26/6. 'Hist. des Anomalies'tome 3 pages 4, 5, 6.) insists that, when any partor organ is repeated many times in the sameanimal, it is particularly liable to vary both innumber and structure. With respect to number,the proposition may, I think, be considered asfully established; but the evidence is chieflyderived from organic beings living under theirnatural conditions, with which we are not hereconcerned. Whenever such parts as the verte-brae or teeth, the rays in the fins of fishes, orthe feathers in the tails of birds, or petals, sta-mens, pistils, or seeds, are very numerous, thenumber is generally variable. With respect tothe structure of multiple parts, the evidence of

variability is not so decisive; but the fact, as faras it may be trusted, probably depends on mul-tiple parts being of less physiological importan-ce than single parts; consequently their structu-re has been less rigorously guarded by naturalselection.

COMPENSATION OF GROWTH, OR BA-LANCEMENT.

This law, as applied to natural species, waspropounded by Goethe and Geoffroy Saint-Hilaire at nearly the same time. It implies that,when much organised matter is used in buil-ding up some one part, other parts are starvedand become reduced. Several authors, especia-lly botanists, believe in this law; others reject it.As far as I can judge, it occasionally holdsgood; but its importance has probably beenexaggerated. It is scarcely possible to distin-guish between the supposed effects of suchcompensation, and the effects of long-

continued selection which may lead to theaugmentation of one part, and simultaneouslyto the diminution of another. Anyhow, therecan be no doubt that an organ may be greatlyincreased without any corresponding diminu-tion of an adjoining part. To recur to our formerillustration of the Irish elk, it may be askedwhat part has suffered in consequence of theimmense development of the horns?

It has already been observed that the strugglefor existence does not bear hard on our domes-ticated productions, and consequently the prin-ciple of economy of growth will seldom comeinto play, so that we ought not to expect to findwith them frequent evidence of compensation.We have, however, some such cases. Moquin-Tandon describes a monstrous bean (26/7. 'Te-ratologie Veg.' page 156. See also my book on'The Movements and Habits of ClimbingPlants' 2nd edition 1875 page 202.), in whichthe stipules were enormously developed, and

the leaflets apparently in consequence comple-tely aborted; this case is interesting, as it repre-sents the natural condition of Lathyrus aphaca,with its stipules of great size, and its leavesreduced to mere threads, which act as tendrils.De Candolle (26/8. 'Memoires du Museum' etc.tome 8 page 178.) has remarked that the varie-ties of Raphanus sativus which have smallroots yield numerous seed containing much oil,whilst those with large roots are not productivein oil; and so it is with Brassica asperifolia. Thevarieties of Cucurbita pepo which bear largefruit yield a small crop, according to Naudin;whilst those producing small fruit yield a vastnumber. Lastly, I have endeavoured to show inthe eighteenth chapter that with many cultiva-ted plants unnatural treatment checks the fulland proper action of the reproductive organs,and they are thus rendered more or less sterile;consequently, in the way of compensation, thefruit becomes greatly enlarged, and, in double

flowers, the petals are greatly increased innumber.

With animals, it has been found difficult to pro-duce cows which yield much milk, and are af-terwards capable of fattening well. With fowlswhich have large top-knots and beards thecomb and wattles are generally much reducedin size; though there are exceptions to this rule.Perhaps the entire absence of the oil-gland infantail pigeons may be connected with thegreat development of their tails.

MECHANICAL PRESSURE AS A CAUSEOF MODIFICATIONS.

In some few cases there is reason to believe thatmere mechanical pressure has affected certainstructures. Vrolik and Weber (26/9. Prichard'Phys. Hist. of Mankind' 1851 volume 1 page324.) maintain that the shape of the humanhead is influenced by the shape of the mother'spelvis. The kidneys in different birds differ

much in form, and St. Ange (26/10. 'Annalesdes Sc. Nat.' 1st series tome 19 page 327.) belie-ves that this is determined by the form of thepelvis, which again, no doubt, stands in closerelation with their power of locomotion. Insnakes, the viscera are curiously displaced, incomparison with their position in other verte-brates; and this has been attributed by someauthors to the elongation of their bodies; buthere, as in so many previous cases, it is impos-sible to disentangle a direct result of this kindfrom that consequent on natural selection. Go-dron has argued (26/11. 'Comptes Rendus' De-cember 1864 page 1039.) that the abortion of thespur on the inner side of the flowers in Coryda-lis, is caused by the buds at a very early periodof growth whilst underground being closelypressed against one another and against thestem. Some botanists believe that the singulardifference in the shape both of the seed andcorolla, in the interior and exterior florets incertain Compositous and Umbelliferous plants,

is due to the pressure to which the inner floretsare subjected; but this conclusion is doubtful.

The facts just given do not relate to domestica-ted productions, and therefore do not strictlyconcern us. But here is a more appropriate case:H. Muller (26/12. "Ueber fotale Rachites"'Wurzburger Medicin. Zeitschrift' 1860 b. 1 s.265.) has shown that in shortfaced races of thedog some of the molar teeth are placed in aslightly different position to that which theyoccupy in other dogs, especially in thosehaving elongated muzzles; and as he remarks,any inherited change in the arrangement of theteeth deserves notice, considering their classifi-catory importance. This difference in position isdue to the shortening of certain facial bonesand the consequent want of space; and theshortening results from a peculiar and abnor-mal state of the embryonal cartilages of the bo-nes.

[RELATIVE POSITION OF FLOWERSWITH RESPECT TO THE AXIS, AND OFSEEDS IN THE OVARY, AS INDUCING VA-RIATION.

In the thirteenth chapter various peloric flowerswere described, and their production wasshown to be due either to arrested develop-ment, or to reversion to a primordial condition.Moquin-Tandon has remarked that the flowerswhich stand on the summit of the main stem orof a lateral branch are more liable to becomepeloric than those on the sides (26/13. 'Terato-logie Veg.' page 192.); and he adduces, amongstother instances, that of Teucrium campanula-tum. In another Labiate plant grown by me,viz., the Galeobdolon luteum, the peloric flo-wers were always produced on the summit ofthe stem, where flowers are not usually borne.In Pelargonium, a SINGLE flower in the truss isfrequently peloric, and when this occurs I haveduring several years invariably observed it tobe the central flower. This is of such frequent

occurrence that one observer (26/14. 'Journal ofHorticulture' July 2, 1861 page 253.) gives thenames of ten varieties flowering at the sametime, in every one of which the central flowerwas peloric. Occasionally more than one flowerin the truss is peloric, and then of course theadditional ones must be lateral. These flowersare interesting as showing how the wholestructure is correlated. In the common Pelargo-nium the upper sepal is produced into a necta-ry which coheres with the flower-peduncle; thetwo upper petals differ a little in shape fromthe three lower ones, and are marked with darkshades of colour; the stamens are graduated inlength and upturned. In the peloric flowers, thenectary aborts; all the petals become alike bothin shape and colour; the stamens are generallyreduced in number and become straight, so thatthe whole flower resembles that of the alliedgenus Erodium. The correlation between thesechanges is well shown when one of the twoupper petals alone loses its dark mark, for in

this case the nectary does not entirely abort, butis usually much reduced in length. (26/15. Itwould be worth trial to fertilise with the samepollen the central and lateral flowers of the pe-largonium, or of other highly cultivated plants,protecting them of course from insects: then tosow the seed separately, and observe whetherthe one or the other lot of seedlings varied themost.)

Morren has described (26/16. Quoted in 'Jour-nal of Horticulture' February 24, 1863 page152.) a marvellous flask-shaped flower of theCalceolaria, nearly four inches in length, whichwas almost completely peloric; it grew on thesummit of the plant, with a normal flower oneach side; Prof. Westwood also has described(26/17. 'Gardener's Chronicle' 1866 page 612.For the Phalaenopsis see ibid 1867 page 211.)three similar peloric flowers, which all occu-pied a central position on the flower-branches.In the Orchideous genus, Phalaenopsis, the

terminal flower has been seen to become pelo-ric.

In a Laburnum-tree I observed that about afourth part of the racemes produced terminalflowers which had lost their papilionaceousstructure. These were produced after almost allthe other flowers on the same racemes had wit-hered. The most perfectly pelorised exampleshad six petals, each marked with black striaelike those on the standard-petal. The keel see-med to resist the change more than the otherpetals. Dutrochet has described (26/18. 'Me-moires…des Vegetaux' 1837 tome 2 page 170.)an exactly similar case in France, and I believethese are the only two instances of pelorism inthe laburnum which have been recorded. Du-trochet remarks that the racemes on this tree donot properly produce a terminal flower, so that(as in the case of the Galeobdolon) their posi-tion as well as structure are both anomalies,which no doubt are in some manner related.

Dr. Masters has briefly described another le-guminous plant (26/19. 'Journal of Horticultu-re' July 23, 1861 page 311.), namely, a species ofclover, in which the uppermost and centralflowers were regular or had lost their papilio-naceous structure. In some of these plants theflower-heads were also proliferous.

Lastly, Linaria produces two kinds of peloricflowers, one having simple petals, and the ot-her having them all spurred. The two forms, asNaudin remarks (26/20. 'Nouvelles Archivesdu Museum' tome 1 page 137.), not rarely occuron the same plant, but in this case the spurredform almost invariably stands on the summit ofthe spike.

The tendency in the terminal or central flowerto become peloric more frequently than theother flowers, probably results from "the budwhich stands on the end of a shoot receivingthe most sap; it grows out into a stronger shootthan those situated lower down." (26/21. Hugo

von Mohl 'The Vegetable Cell' English transla-tion 1852 page 76.) I have discussed the connec-tion between pelorism and a central position,partly because some few plants are knownnormally to produce a terminal flower differentin structure from the lateral ones; but chiefly onaccount of the following case, in which we see atendency to variability or to reversion connec-ted with the same position. A great judge ofAuriculas (26/22. The Rev. H.H. Dombrain in'Journal of Horticulture' 1861 June 4 page 174;and June 25 page 234; 1862 April 29 page 83.)states that when one throws up a side bloom itis pretty sure to keep its character; but that if itgrows from the centre or heart of the plant,whatever the colour of the edging ought to be,"it is just as likely to come in any other class asin the one to which it properly belongs." This isso notorious a fact, that some florists regularlypinch off the central trusses of flowers. Whet-her in the highly improved varieties the depar-ture of the central trusses from their proper

type is due to reversion, I do not know. Mr.Dombrain insists that, whatever may be thecommonest kind of imperfection in each varie-ty, this is generally exaggerated in the centraltruss. Thus one variety "sometimes has the faultof producing a little green floret in the centre ofthe flower," and in central blooms these becomeexcessive in size. In some central blooms, sentto me by Mr. Dombrain, all the organs of theflower were rudimentary in structure, of minu-te size, and of a green colour, so that by a littlefurther change all would have been convertedinto small leaves. In this case we clearly see atendency to prolification—a term which I mayexplain, for those who have never attended tobotany, to mean the production of a branch orflower, or head of flowers, out of another flo-wer. Now Dr. Masters (26/23. 'Transact. Linn.Soc.' volume 23 1861 page 360.) states that thecentral or uppermost flower on a plant is gene-rally the most liable to prolification. Thus, inthe varieties of the Auricula, the loss of their

proper character and a tendency to prolifica-tion, also a tendency to prolification with pelo-rism, are all connected together, and are dueeither to arrested development, or to reversionto a former condition.

The following is a more interesting case; Metz-ger (26/24. 'Die Getreidearten' 1845 s. 208, 209.)cultivated in Germany several kinds of maizebrought from the hotter parts of America, andhe found, as previously described, that in twoor three generations the grains became greatlychanged in form, size, and colour; and withrespect to two races he expressly states that inthe first generation, whilst the lower grains oneach head retained their proper character, theuppermost grains already began to assume thatcharacter which in the third generation all thegrains acquired. As we do not know the abori-ginal parent of the maize, we cannot tell whet-her these changes are in any way connectedwith reversion.

In the two following cases, reversion comesinto play and is determined by the position ofthe seed in the capsule. The Blue Imperial peais the offspring of the Blue Prussian, and haslarger seed and broader pods than its parent.Now Mr. Masters, of Canterbury, a careful ob-server and a raiser of new varieties of the pea,states (26/25. 'Gardener's Chronicle' 1850 page198.) that the Blue Imperial always has a strongtendency to revert to its parent-stock, and thereversion "occurs in this manner: the last (oruppermost) pea in the pod is frequently muchsmaller than the rest; and if these small peas arecarefully collected and sown separately, verymany more, in proportion, will revert to theirorigin, than those taken from the other parts ofthe pod." Again, M. Chate (26/26. Quoted in'Gardener's Chronicle' 1866 page 74.) says thatin raising seedling stocks he succeeds in gettingeighty per cent to bear double flowers, by lea-ving only a few of the secondary branches toseed; but in addition to this, "at the time of ex-

tracting the seeds, the upper portion of the podis separated and placed aside, because it hasbeen ascertained that the plants coming fromthe seeds situated in this portion of the pod,give eighty per cent of single flowers." Now theproduction of single-flowering plants from theseed of double-flowering plants is clearly a caseof reversion. These latter facts, as well as theconnection between a central position and pelo-rism and prolification, show in an interestingmanner how small a difference—namely, a lit-tle greater or less freedom in the flow of saptowards one part of the plant—determines im-portant changes of structure.]

ANALOGOUS OR PARALLEL VARIA-TION.

By this term I mean that similar characters oc-casionally make their appearance in the severalvarieties or races descended from the same spe-cies, and more rarely in the offspring of widely

distinct species. We are here concerned, not ashitherto with the causes of variation, but withthe results; but this discussion could not havebeen more conveniently introduced elsewhere.The cases of analogous variation, as far as theirorigin is concerned, may be grouped, disregar-ding minor subdivisions, under two mainheads; firstly, those due to unknown causesacting on similarly constituted organisms, andwhich consequently have varied in a similarmanner; and secondly, those due to the reap-pearance of characters which were possessedby a more or less remote progenitor. But thesetwo main divisions can often be separated onlyconjecturally, and graduate, as we shall presen-tly see, into each other.

[Under the first head of analogous variations,not due to reversion, we have the many cases oftrees belonging to quite different orders whichhave produced pendulous and fastigiate varie-ties. The beech, hazel, and barberry have given

rise to purple-leaved varieties; and, as Bern-hardi remarks (26/27. 'Ueber den Begriff derPflanzenart' 1834 s. 14.), a multitude of plants,as distinct as possible, have yielded varietieswith deeply-cut or laciniated leaves. Varietiesdescended from three distinct species of Brassi-ca have their stems, or so-called roots, enlargedinto globular masses. The nectarine is the offs-pring of the peach; and the varieties of peachesand nectarines offer a remarkable parallelism inthe fruit being white, red, or yellow fleshed—inbeing clingstones or freestones—in the flowersbeing large or small—in the leaves being serra-ted or crenated, furnished with globose or reni-form glands, or quite destitute of glands. Itshould be remarked that each variety of thenectarine has not derived its character from acorresponding variety of the peach. The severalvarieties also of a closely allied genus, namelythe apricot, differ from one another in nearlythe same parallel manner. There is no reason tobelieve that any of these varieties have merely

reacquired long- lost characters; and in most ofthem this certainly is not the case.

Three species of Cucurbita have yielded a mul-titude of races which correspond so closely incharacter that, as Naudin insists, they may bearranged in almost strictly parallel series. Seve-ral varieties of the melon are interesting fromresembling, in important characters, other spe-cies, either of the same genus or of allied gene-ra; thus, one variety has fruit so like, both ex-ternally and internally, the fruit of a perfectlydistinct species, namely, the cucumber, as hard-ly to be distinguished from it; another has longcylindrical fruit twisting about like a serpent; inanother the seeds adhere to portions of thepulp; in another the fruit, when ripe, suddenlycracks and falls into pieces; and all these highlyremarkable peculiarities are characteristic ofspecies belonging to allied genera. We canhardly account for the appearance of so manyunusual characters by reversion to a single an-

cient form; but we must believe that all themembers of the family have inherited a nearlysimilar constitution from an early progenitor.Our cereal and many other plants offer similarcases.

With animals we have fewer cases of analogousvariation, independently of direct reversion.We see something of the kind in the resemblan-ce between the short-muzzled races of the dog,such as the pug and bull-dog; in feather- footedraces of the fowl, pigeon, and canary-bird; inhorses of the most different races presentingthe same range of colour; in all black-and-tandogs having tan-coloured eye-spots and feet,but in this latter case reversion may possiblyhave played a part. Low has remarked (26/28.'Domesticated Animals' 1845 page 351.) thatseveral breeds of cattle are "sheeted,"—that is,have a broad band of white passing round theirbodies like a sheet; this character is stronglyinherited, and sometimes originates from a

cross; it may be the first step in reversion to anearly type, for, as was shown in the third chap-ter, white cattle with dark ears, dark feet andtip of tail, formerly existed, and now exist inferal or semi-feral condition in several quartersof the world.

Under our second main division, namely, ofanalogous variations due to reversion, the bestcases are afforded by pigeons. In all the mostdistinct breeds, sub-varieties occasionally ap-pear coloured exactly like the parent rock-pigeon, with black wing-bars, white loins, ban-ded tail, etc.; and no one can doubt that thesecharacters are due to reversion. So with minordetails; turbits properly have white tails, butoccasionally a bird is born with a dark-colouredand banded tail; pouters properly have theirprimary wing- feathers white, but not rarely a"sword-flighted" bird appears, that is, one withthe few first primaries dark-coloured; and inthese cases we have characters proper to the

rock-pigeon, but new to the breed, evidentlyappearing from reversion. In some domesticvarieties the wing-bars, instead of being simplyblack, as in the rock-pigeon, are beautifullyedged with different zones of colour, and theythen present a striking analogy with the wing-bars in certain natural species of the same fami-ly, such as Phaps chalcoptera; and this mayprobably be accounted for by all the species ofthe family being descended from the same re-mote progenitor and having a tendency to varyin the same manner. Thus, also, we can perhapsunderstand the fact of some Laugher-pigeonscooing almost like turtle-doves, and for severalraces having peculiarities in their flight, sincecertain natural species (viz., C. torquatrix andpalumbus), display singular vagaries in thisrespect. In other cases a race, instead of imita-ting a distinct species, resembles some otherrace; thus, certain runts tremble and slightlyelevate their tails, like fantails; and turbits infla-

te the upper part of their oesophagus, like pou-ter- pigeons.

It is a common circumstance to find certaincoloured marks persistently characterising allthe species of a genus, but differing much intint; and the same thing occurs with the varie-ties of the pigeon: thus, instead of the generalplumage being blue, with the wing-bars black,there are snow-white varieties with red bars,and black varieties with white bars; in othervarieties the wing-bars, as we have seen, areelegantly zoned with different tints. The Spotpigeon is characterised by the whole plumagebeing white, excepting a spot on the foreheadand the tail; but these parts may be red, yellow,or black. In the rock-pigeon and in many varie-ties the tail is blue, with the outer edges of theouter feathers white; but in the sub-variety ofthe monk-pigeon we have a reversed style ofcoloration, for the tail is white, except the outeredges of the outer feathers, which are black.

(26/29. Bechstein 'Naturgeschichte Deutsch-lands' b. 4 1795 s. 31.)

With some species of birds, for instance withgulls, certain coloured parts appear as if almostwashed out, and I have observed exactly thesame appearance in the terminal dark tail-barin certain pigeons, and in the whole plumage ofcertain varieties of the duck. Analogous facts inthe vegetable kingdom could be given.

Many sub-varieties of the pigeon have reversedand somewhat lengthened feathers on the backpart of their heads, and this is certainly not dueto reversion to the parent-species, which showsno trace of such structure: but when we re-member that sub-varieties of the fowl, turkey,canary-bird, duck, and goose, all have eithertopknots or reversed feathers on their heads;and when we remember that scarcely a singlelarge natural group of birds can be named, inwhich some members have not a tuft of feat-hers on their heads, we may suspect that rever-

sion to some extremely remote form has comeinto action.

Several breeds of the fowl have either spangledor pencilled feathers; and these cannot be deri-ved from the parent-species, the Gallus banki-va; though of course it is possible that one earlyprogenitor of this species may have been span-gled, and another pencilled. But, as many galli-naceous birds are either spangled or pencilled,it is a more probable view that the several do-mestic breeds of the fowl have acquired thiskind of plumage from all the members of thefamily inheriting a tendency to vary in a likemanner. The same principle may account forthe ewes in certain breeds of sheep being horn-less, like the females of some other hollow-horned ruminants; it may account for certaindomestic cats having slightly-tufted ears, likethose of the lynx; and for the skulls of domesticrabbits often differing from one another in the

same characters by which the skulls of the va-rious species of the genus Lepus differ.

I will only allude to one other case, alreadydiscussed. Now that we know that the wildparent of the ass commonly has striped legs, wemay feel confident that the occasional appea-rance of stripes on the legs of the domestic assis due to reversion; but this will not account forthe lower end of the shoulder- stripe being so-metimes angularly bent or slightly forked. So,again, when we see dun and other colouredhorses with stripes on the spine, shoulders, andlegs, we are led, from reasons formerly given,to believe that they reappear through reversionto the wild parent-horse. But when horses havetwo or three shoulder-stripes, with one of themoccasionally forked at the lower end, or whenthey have stripes on their faces, or are faintlystriped as foals over nearly their whole bodies,with the stripes angularly bent one under theother on the forehead, or irregularly branched

in other parts, it would be rash to attribute suchdiversified characters to the reappearance ofthose proper to the aboriginal wild horse. Asthree African species of the genus are muchstriped, and as we have seen that the crossingof the unstriped species often leads to the hy-brid offspring being conspicuously striped—bearing also in mind that the act of crossingcertainly causes the reappearance of long-lostcharacters—it is a more probable view that theabove-specified stripes are due to reversion, notto the immediate wild parent-horse, but to thestriped progenitor of the whole genus.]

I have discussed this subject of analogous va-riation at considerable length, because it is wellknown that the varieties of one species frequen-tly resemble distinct species—a fact in perfectharmony with the foregoing cases, and explica-ble on the theory of descent. Secondly, becausethese facts are important from showing, as re-marked in a former chapter, that each trifling

variation is governed by law, and is determi-ned in a much higher degree by the nature ofthe organisation, than by the nature of the con-ditions to which the varying being has beenexposed. Thirdly, because these facts are to acertain extent related to a more general law,namely, that which Mr. B.D. Walsh (26/30.'Proc. Entomolog. Soc. of Philadelphia' October1863 page 213.) has called the "Law of EQUA-BLE VARIABILITY," or, as he explains it, "ifany given character is very variable in one spe-cies of a group, it will tend to be variable inallied species; and if any given character is per-fectly constant in one species of a group, it willtend to be constant in allied species."

This leads me to recall a discussion in the chap-ter on Selection, in which it was shown thatwith domestic races, which are now under-going rapid improvement, those parts or cha-racters vary the most, which are the most va-lued. This naturally follows from recently selec-

ted characters continually tending to revert totheir former less improved standard, and fromtheir being still acted on by the same agencies,whatever these may be, which first caused thecharacters in question to vary. The same prin-ciple is applicable to natural species, for, asstated in my 'Origin of Species' generic charac-ters are less variable than specific characters;and the latter are those which have been modi-fied by variation and natural selection, since theperiod when all the species belonging to thegenus branched off from a common progenitor,whilst generic characters are those which haveremained unaltered from a much more remoteepoch, and accordingly are now less variable.This statement makes a near approach to Mr.Walsh's law of Equable Variability. Secondarysexual characters, it may be added, rarely serveto characterise distinct genera, for they usuallydiffer much in the species of the same genus,and they are highly variable in the individualsof the same species; we have also seen in the

earlier chapters of this work how variable se-condary sexual characters become under do-mestication.

SUMMARY OF THE THREE PREVIOUSCHAPTERS ON THE LAWS OF VARIATION.

In the twenty-third chapter we saw that chan-ged conditions occasionally, or even often, actin a definite manner on the organisation, sothat all, or nearly all, the individuals thus expo-sed become modified in the same manner. Buta far more frequent result of changed condi-tions, whether acting directly on the organisa-tion or indirectly through the reproductive sys-tem, is indefinite and fluctuating variability. Inthe three last chapters, some of the laws bywhich such variability is regulated have beendiscussed.

Increased use adds to the size of muscles, to-gether with the blood-vessels, nerves, liga-ments, the crests of bone and the whole bones,

to which they are attached. Increased functio-nal activity increases the size of various glands,and strengthens the sense-organs. Increasedand intermittent pressure thickens the epider-mis. A change in the nature of the food some-times modifies the coats of the stomach, andaugments or decreases the length of the intesti-nes. Continued disuse, on the other hand, wea-kens and diminishes all parts of the organisa-tion. Animals which during many generationshave taken but little exercise, have their lungsreduced in size, and as a consequence the bonyfabric of the chest and the whole form of thebody become modified. With our ancientlydomesticated birds, the wings have been littleused, and they are slightly reduced; with theirdecrease, the crest of the sternum, the scapulae,coracoids, and furculum, have all been redu-ced.

With domesticated animals, the reduction of apart from disuse is never carried so far that a

mere rudiment is left; whereas we have reasonto believe that this has often occurred undernature; the effects of disuse in this latter casebeing aided by economy of growth, togetherwith the intercrossing of many varying indivi-duals. The cause of this difference between or-ganisms in a state of nature, and under domes-tication, probably is that in the latter case therehas not been time sufficient for any very greatchange, and that the principle of economy ofgrowth does not come into action. On the con-trary, structures which are rudimentary in theparent-species, sometimes become partiallyredeveloped in our domesticated productions.Such rudiments as occasionally make their ap-pearance under domestication, seem always tobe the result of a sudden arrest of development;nevertheless they are of interest, as showingthat rudiments are the relics of organs onceperfectly developed.

Corporeal, periodical, and mental habits,though the latter have been almost passed overin this work, become changed under domesti-cation, and the changes are often inherited.Such changed habits in an organic being, espe-cially when living a free life, would often leadto the augmented or diminished use of variousorgans, and consequently to their modification.From long-continued habit, and more especia-lly from the occasional birth of individuals witha slightly different constitution, domestic ani-mals and cultivated plants become to a certainextent acclimatised or adapted to a climate dif-ferent from that proper to the parent-species.

Through the principle of correlated variability,taken in its widest sense, when one part variesother parts vary, either simultaneously, or oneafter the other. Thus, an organ modified duringan early embryonic period affects other partssubsequently developed. When an organ, suchas the beak, increases or decreases in length,

adjoining or correlated parts, as the tongue andthe orifice of the nostrils, tend to vary in thesame manner. When the whole body increasesor decreases in size, various parts become mo-dified; thus, with pigeons the ribs increase ordecrease in number and breadth. Homologousparts which are identical during their early de-velopment and are exposed to similar condi-tions, tend to vary in the same or in some con-nected manner,—as in the case of the right andleft sides of the body, and of the front and hindlimbs. So it is with the organs of sight and hea-ring; for instance, white cats with blue eyes arealmost always deaf. There is a manifest relationthroughout the body between the skin and va-rious dermal appendages, such as hair, feat-hers, hoofs, horns, and teeth. In Paraguay, hor-ses with curly hair have hoofs like those of amule; the wool and the horns of sheep oftenvary together; hairless dogs are deficient intheir teeth; men with redundant hair have ab-normal teeth, either by deficiency or excess.

Birds with long wing-feathers usually havelong tail-feathers. When long feathers growfrom the outside of the legs and toes of pigeons,the two outer toes are connected by membrane;for the whole leg tends to assume the structureof the wing. There is a manifest relation bet-ween a crest of feathers on the head and a mar-vellous amount of change in the skull of va-rious fowls; and in a lesser degree, between thegreatly elongated, lopping ears of rabbits andthe structure of their skulls. With plants, theleaves, various parts of the flower, and thefruit, often vary together to a correlated man-ner.

In some cases we find correlation without beingable even to conjecture what is the nature of theconnection, as with various monstrosities anddiseases. This is likewise the case with the co-lour of the adult pigeon, in connection with thepresence of down on the young bird. Nume-rous curious instances have been given of pecu-

liarities of constitution, in correlation with co-lour, as shown by the immunity of individualsof one colour from certain diseases, from theattacks of parasites and from the action of cer-tain vegetable poisons.

Correlation is an important subject; for withspecies, and in a lesser degree with domesticraces, we continually find that certain partshave been greatly modified to serve some use-ful purpose; but we almost invariably find thatother parts have likewise been more or lessmodified, without our being able to discoverany advantage in the change. No doubt greatcaution is necessary with respect to this latterpoint, for it is difficult to overrate our ignoran-ce on the use of various parts of the organisa-tion; but from what we have seen, we may be-lieve that many modifications are of no directservice, having arisen in correlation with otherand useful changes.

Homologous parts during their early develop-ment often become fused together. Multipleand homologous organs are especially liable tovary in number and probably in form. As thesupply of organised matter is not unlimited, theprinciple of compensation sometimes comesinto action; so that, when one part is greatlydeveloped, adjoining parts are apt to be redu-ced; but this principle is probably of much lessimportance than the more general one of theeconomy of growth. Through mere mechanicalpressure hard parts occasionally affect adjoi-ning parts. With plants the position of the flo-wers on the axis, and of the seeds in the ovary,sometimes leads, through a more or less freeflow of sap, to changes of structure; but suchchanges are often due to reversion. Modifica-tions, in whatever manner caused, will be to acertain extent regulated by that co-ordinatingpower, or so-called nisus formativus, which isin fact a remnant of that simple form of repro-duction, displayed by many lowly organised

beings in their power of fissiparous generationand budding. Finally, the effects of the lawswhich directly or indirectly govern variability,may be largely regulated by man's selection,and will so far be determined by natural selec-tion that changes advantageous to any race willbe favoured, and disadvantageous changes willbe checked.

Domestic races descended from the same spe-cies, or from two or more allied species, areliable to revert to characters derived from theircommon progenitor; and, as they inherit a so-mewhat similar constitution, they are liable tovary in the same manner. From these two cau-ses analogous varieties often arise. When wereflect on the several foregoing laws, imperfec-tly as we understand them, and when we bearin mind how much remains to be discovered,we need not be surprised at the intricate and tous unintelligible manner in which our domestic

productions have varied, and still go on var-ying.

CHAPTER 2.XXVII.

PROVISIONAL HYPOTHESIS OF PAN-GENESIS.

PRELIMINARY REMARKS. FIRST PART:THE FACTS TO BE CONNECTED UNDER ASINGLE POINT OF VIEW, NAMELY, THEVARIOUS KINDS OF REPRODUCTION. RE-GROWTH OF AMPUTATED PARTS. GRAFT-HYBRIDS. THE DIRECT ACTION OF THEMALE ELEMENT ON THE FEMALE. DEVE-LOPMENT. THE FUNCTIONAL INDEPEN-DENCE OF THE UNITS OF THE BODY. VA-RIABILITY. INHERITANCE. REVERSION.

SECOND PART: STATEMENT OF THEHYPOTHESIS. HOW FAR THE NECESSARY

ASSUMPTIONS ARE IMPROBABLE. EX-PLANATION BY AID OF THE HYPOTHESISOF THE SEVERAL CLASSES OF FACTS SPE-CIFIED IN THE FIRST PART. CONCLUSION.

In the previous chapters large classes of facts,such as those bearing on bud- variation, thevarious forms of inheritance, the causes andlaws of variation, have been discussed; and it isobvious that these subjects, as well as the seve-ral modes of reproduction, stand in some sortof relation to one another. I have been led, orrather forced, to form a view which to a certainextent connects these facts by a tangible met-hod. Every one would wish to explain to him-self, even in an imperfect manner, how it ispossible for a character possessed by some re-mote ancestor suddenly to reappear in the offs-pring; how the effects of increased or decreaseduse of a limb can be transmitted to the child;how the male sexual element can act not solelyon the ovules, but occasionally on the mother-form; how a hybrid can be produced by the

union of the cellular tissue of two plants inde-pendently of the organs of generation; how alimb can be reproduced on the exact line ofamputation, with neither too much nor too lit-tle added; how the same organism may be pro-duced by such widely different processes, asbudding and true seminal generation; and, las-tly, how of two allied forms, one passes in thecourse of its development through the mostcomplex metamorphoses, and the other doesnot do so, though when mature both are alikein every detail of structure. I am aware that myview is merely a provisional hypothesis or spe-culation; but until a better one be advanced, itwill serve to bring together a multitude of factswhich are at present left disconnected by anyefficient cause. As Whewell, the historian of theinductive sciences, remarks:—"Hypothesesmay often be of service to science, when theyinvolve a certain portion of incompleteness,and even of error." Under this point of view Iventure to advance the hypothesis of Pangene-

sis, which implies that every separate part ofthe whole organisation reproduces itself. Sothat ovules, spermatozoa, and pollen-grains,—the fertilised egg or seed, as well as buds,—include and consist of a multitude of germsthrown off from each separate part or unit.(27/1. This hypothesis has been severely critici-sed by many writers, and it will be fair to givereferences to the more important articles. Thebest essay which I have seen is by Prof. Delpi-no, entitled 'Sulla Darwiniana Teoria della Pan-genesi, 1869' of which a translation appeared in'Scientific Opinion' September 29, 1869 and thesucceeding numbers. He rejects the hypothesis,but criticises it fairly, and I have found his criti-cisms very useful. Mr. Mivart ('Genesis of Spe-cies' 1871 chapter 10.) follows Delpino, butadds no new objections of any weight. Dr. Bas-tian ('The Beginnings of Life' 1872 volume 2page 98) says that the hypothesis "looks like arelic of the old rather than a fitting appanage ofthe new evolution philosophy." He shows that I

ought not to have used the term "pangenesis,"as it had been previously used by Dr. Gros inanother sense. Dr. Lionel Beale ('Nature' May11, 1871 page 26) sneers at the whole doctrinewith much acerbity and some justice. Prof. Wi-gand ('Schriften der Gesell. der gesammt. Na-turwissen. zu Marburg' b. 9 1870) considers thehypothesis as unscientific and worthless. Mr.G.H. Lewes ('Fortnightly Review' November 1,1868 page 503) seems to consider that it may beuseful: he makes many good criticisms in a per-fectly fair spirit. Mr. F. Galton, after describinghis valuable experiments ('Proc. Royal Soc.'volume 19 page 393) on the intertransfusion ofthe blood of distinct varieties of the rabbit, con-cludes by saying that in his opinion the resultsnegative beyond all doubt the doctrine of Pan-genesis. He informs me that subsequently tothe publication of his paper he continued hisexperiments on a still larger scale for two moregenerations, without any sign of mongrelismshowing itself in the very numerous offspring. I

certainly should have expected that gemmuleswould have been present in the blood, but thisis no necessary part of the hypothesis, whichmanifestly applies to plants and the lowestanimals. Mr. Galton, in a letter to 'Nature'(April 27, 1871 page 502), also criticises variousincorrect expressions used by me. On the otherhand, several writers have spoken favourablyof the hypothesis, but there would be no use ingiving references to their articles. I may, howe-ver, refer to Dr. Ross' work, 'The Graft Theoryof Disease; being an application of Mr. Dar-win's hypothesis of Pangenesis' 1872 as he gi-ves several original and ingenious discussions.)

In the First Part I will enumerate as briefly as Ican the groups of facts which seem to demandconnection; but certain subjects, not hithertodiscussed, must be treated at disproportionatelength. In the Second Part the hypothesis willbe given; and after considering how far thenecessary assumptions are in themselves im-

probable, we shall see whether it serves tobring under a single point of view the variousfacts.

PART I.

Reproduction may be divided into two mainclasses, namely, sexual and asexual. The latteris effected in many ways—by the formation ofbuds of various kinds, and by fissiparous gene-ration, that is by spontaneous or artificial divi-sion. It is notorious that some of the lower ani-mals, when cut into many pieces, reproduce somany perfect individuals: Lyonnet cut a Nais orfreshwater worm into nearly forty pieces, andthese all reproduced perfect animals. (27/2.Quoted by Paget 'Lectures on Pathology' 1853page 159.) It is probable that segmentationcould be carried much further in some of theprotozoa; and with some of the lowest plantseach cell will reproduce the parent-form. Jo-hannes Muller thought that there was an im-

portant distinction between gemmation andfission; for in the latter case the divided por-tion, however small, is more fully developedthan a bud, which also is a younger formation;but most physiologists are now convinced thatthe two processes are essentially alike. (27/3.Dr. Lachmann also observes ('Annals and Mag.of Nat. History' 2nd series volume 19 1857 page231) with respect to infusoria, that "fissationand gemmation pass into each other almostimperceptibly." Again, Mr. W.C. Minor ('An-nals and Mag. of Nat. Hist.' 3rd series volume11 page 328) shows that with Annelids the dis-tinction that has been made between fissionand budding is not a fundamental one. See alsoProfessor Clark's work 'Mind in Nature' NewYork 1865 pages 62, 94.) Prof. Huxley remarks,"fission is little more than a peculiar mode ofbudding," and Prof. H.J. Clark shows in detailthat there is sometimes "a compromise betweenself-division and budding." When a limb isamputated, or when the whole body is bisec-

ted, the cut extremities are said to bud forth(27/4. See Bonnet 'Oeuvres d'Hist. Nat.' tome 51781 page 339 for remarks on the budding-outof the amputated limbs of Salamanders.); andas the papilla, which is first formed, consists ofundeveloped cellular tissue like that formingan ordinary bud, the expression is apparentlycorrect. We see the connection of the two pro-cesses in another way; for Trembley observedwith the hydra, that the reproduction of thehead after amputation was checked as soon asthe animal put forth reproductive gemmae.(27/5. Paget 'Lectures on Pathology' 1853 page158.)

Between the production, by fissiparous genera-tion, of two or more complete individuals, andthe repair of even a very slight injury, there isso perfect a gradation, that it is impossible todoubt that the two processes are connected. Asat each stage of growth an amputated part isreplaced by one in the same state of develop-

ment, we must also follow Sir J. Paget in admit-ting, "that the powers of development from theembryo, are identical with those exercised forthe restoration from injuries: in other words,that the powers are the same by which perfec-tion is first achieved, and by which, when lost,it is recovered." (27/6. Ibid pages 152, 164.) Fi-nally, we may conclude that the several formsof budding, fissiparous generation, the repair ofinjuries, and development, are all essentiallythe results of one and the same power.

SEXUAL GENERATION.

The union of the two sexual elements seems atfirst sight to make a broad distinction betweensexual and asexual generation. But the conjuga-tion of algae, by which process the contents oftwo cells unite into a single mass capable ofdevelopment, apparently gives us the first steptowards sexual union: and Pringsheim, in hismemoir on the pairing of Zoospores (27/7.

Translated in 'Annals and Mag. of Nat. Hist.'April 1870 page 272.), shows that conjugationgraduates into true sexual reproduction. Mo-reover, the now well-ascertained cases of Part-henogenesis prove that the distinction betweensexual and asexual generation is not nearly sogreat as was formerly thought; for ova occasio-nally, and even in some cases frequently, be-come developed into perfect beings, withoutthe concourse of the male. With most of thelower animals and even with mammals, theova show a trace of parthenogenetic power, forwithout being fertilised they pass through thefirst stages of segmentation. (27/8. Bischoff asquoted by von Siebold "Ueber Parthenogenesis"'Sitzung der math. phys. Classe.' Munich No-vember 4, 1871 page 240. See also Quatrefages'Annales des Sc. Nat. Zoolog.' 3rd series 1850page 138.) Nor can pseudova which do notneed fertilisation, be distinguished from trueova, as was first shown by Sir J. Lubbock, andis now admitted by Siebold. So, again, the

germ-balls in the larvae of Cecidomyia are saidby Leuckart (27/9. 'On the Asexual Reproduc-tion of Cecidomyide Larvae' translated in 'An-nals and Mag. of Nat. Hist.' March 1866 pages167, 171.) to be formed within the ovarium, butthey do not require to be fertilised. It shouldalso be observed that in sexual generation, theovules and the male element have equal powerof transmitting every single character possessedby either parent to their offspring. We see thisclearly when hybrids are paired inter se, for thecharacters of both grandparents often appear inthe progeny, either perfectly or by segments. Itis an error to suppose that the male transmitscertain characters and the female other charac-ters; although no doubt, from unknown causes,one sex sometimes has a much stronger powerof transmission than the other.

It has, however, been maintained by some aut-hors that a bud differs essentially from a fertili-sed germ, in always reproducing the perfect

character of the parent-stock; whilst fertilisedgerms give birth to variable beings. But there isno such broad distinction as this. In the ele-venth chapter numerous cases were advancedshowing that buds occasionally grow intoplants having quite new characters; and thevarieties thus produced can be propagated for alength of time by buds, and occasionally byseed. Nevertheless, it must be admitted thatbeings produced sexually are much more liableto vary than those produced asexually; and ofthis fact a partial explanation will hereafter beattempted. The variability in both cases is de-termined by the same general causes, and isgoverned by the same laws. Hence new varie-ties arising from buds cannot be distinguishedfrom those arising from seed. Although bud-varieties usually retain their character duringsuccessive bud-generations, yet they occasiona-lly revert, even after a long series of bud-generations, to their former character. This ten-dency to reversion in buds, is one of the most

remarkable of the several points of agreementbetween the offspring from bud and seminalreproduction.

But there is one difference between organismsproduced sexually and asexually, which is verygeneral. The former pass in the course of theirdevelopment from a very low stage to theirhighest stage, as we see in the metamorphosesof insects and of many other animals, and inthe concealed metamorphoses of the vertebrata.Animals propagated asexually by buds or fis-sion, on the other hand, commence their deve-lopment at that stage at which the budding orself- dividing animal may happen to be, andtherefore do not pass through some of the lo-wer developmental stages. (27/10. Prof. All-man speaks ('Transact. R. Soc. of Edinburgh'volume 26 1870 page 102) decisively on thishead with respect to the Hydroida: he says, "Itis a universal law in the succession of zooids,that no retrogression ever takes place in the

series.") Afterwards, they often advance in or-ganisation, as we see in the many cases of "al-ternate generation." In thus speaking of alterna-te generation, I follow those naturalists wholook at this process as essentially one of inter-nal budding or of fissiparous generation. Someof the lower plants, however, such as mossesand certain algae, according to Dr. L. Radlkofer(27/11. 'Annals and Mag. of Nat. Hist.' 2ndseries volume 20 1857 pages 153-455), whenpropagated asexually, do undergo a retrogres-sive metamorphosis. As far as the final cause isconcerned, we can to a certain extent unders-tand why beings propagated by buds shouldnot pass through all the early stages of deve-lopment; for with each organism the structureacquired at each stage must be adapted to itspeculiar habits; and if there are places for thesupport of many individuals at some one stage,the simplest plan will be that they should bemultiplied at this stage, and not that theyshould first retrograde in their development to

an earlier or simpler structure, which might notbe fitted for the then surrounding conditions.

From the several foregoing considerations wemay conclude that the difference betweensexual and asexual generation is not nearly sogreat as at first appears; the chief differencebeing that an ovule cannot continue to live andto be fully developed unless it unites with themale element; but even this difference is farfrom invariable, as shown by the many cases ofparthenogenesis. We are therefore naturally ledto inquire what the final cause can be of thenecessity in ordinary generation for the con-course of the two sexual elements.

Seeds and ova are often highly serviceable asthe means of disseminating plants and animals,and of preserving them during one or moreseasons in a dormant state; but unimpregnatedseeds or ova, and detached buds, would beequally serviceable for both purposes. We can,however, indicate two important advantages

gained by the concourse of the two sexes, orrather of two individuals belonging to oppositesexes; for, as I have shown in a former chapter,the structure of every organism appears to beespecially adapted for the concurrence, at leastoccasionally, of two individuals. When speciesare rendered highly variable by changed condi-tions of life, the free intercrossing of the var-ying individuals tends to keep each form fittedfor its proper place in nature; and crossing canbe effected only by sexual generation; butwhether the end thus gained is of sufficientimportance to account for the first origin ofsexual intercourse is extremely doubtful. Se-condly, I have shown from a large body offacts, that, as a slight change in the conditionsof life is beneficial to each creature, so, in ananalogous manner, is the change effected in thegerm by sexual union with a distinct indivi-dual; and I have been led, from observing themany widely-extended provisions throughoutnature for this purpose, and from the greater

vigour of crossed organisms of all kinds, asproved by direct experiments, as well as fromthe evil effects of close interbreeding when longcontinued, to believe that the advantage thusgained is very great.

Why the germ, which before impregnation un-dergoes a certain amount of development, cea-ses to progress and perishes, unless it be actedon by the male element; and why converselythe male element, which in the case of someinsects is enabled to keep alive for four or fiveyears, and in the case of some plants for severalyears, likewise perishes, unless it acts on orunites with the germ, are questions which can-not be answered with certainty. It is, however,probable that both sexual elements perish, un-less brought into union, simply from includingtoo little formative matter for independent de-velopment. Quatrefages has shown in the caseof the Teredo (27/12. 'Annales des Sc. Nat.' 3rdseries 1850 tome 13.), as did formerly Prevost

and Dumas with other animals, that more thanone spermatozoon is requisite to fertilise anovum. This has likewise been shown by New-port (27/13. 'Transact. Phil. Soc.' 1851 pages196, 208, 210; 1853 pages 245, 247.), who provedby numerous experiments, that, when a verysmall number of spermatozoa are applied tothe ova of Batrachians, they are only partiallyimpregnated, and an embryo is never fully de-veloped. The rate also of the segmentation ofthe ovum is determined by the number of thespermatozoa. With respect to plants, nearly thesame results were obtained by Kolreuter andGartner. This last careful observer, after makingsuccessive trials on a Malva with more andmore pollen- grains, found (27/14. 'Beitrage zurKenntniss' etc. 1844 s. 345.), that even thirtygrains did not fertilise a single seed; but whenforty grains were applied to the stigma, a fewseeds of small size were formed. In the case ofMirabilis the pollen grains are extraordinarilylarge, and the ovarium contains only a single

ovule; and these circumstances led Naudin(27/15. 'Nouvelles Archives du Museum' tome1 page 27.) to make the following experiments:a flower was fertilised by three grains and suc-ceeded perfectly; twelve flowers were fertilisedby two grains, and seventeen flowers by a sin-gle grain, and of these one flower alone in eachlot perfected its seed: and it deserves especialnotice that the plants produced by these twoseeds never attained their proper dimensions,and bore flowers of remarkably small size.From these facts we clearly see that the quanti-ty of the peculiar formative matter which iscontained within the spermatozoa and pollen-grains is an all-important element in the act offertilisation, not only for the full developmentof the seed, but for the vigour of the plant pro-duced from such seed. We see something of thesame kind in certain cases of parthenogenesis,that is, when the male element is wholly exclu-ded; for M. Jourdan (27/16. As quoted by Sir J.Lubbock in 'Nat. Hist. Review' 1862 page 345.

Weijenbergh also raised ('Nature' December 21,1871 page 149) two successive generations fromunimpregnated females of another lepidopte-rous insect, Liparis dispar. These females didnot produce at most one-twentieth of their fullcomplement of eggs, and many of the eggs we-re worthless. Moreover the caterpillars raisedfrom these unfertilised eggs "possessed far lessvitality" than those from fertilised eggs. In thethird parthenogenetic generation not a singleegg yielded a caterpillar.) found that, out ofabout 58,000 eggs laid by unimpregnated silk-moths, many passed through their early embr-yonic stages, showing that they were capable ofself-development, but only twenty-nine out ofthe whole number produced caterpillars. Thesame principle of quantity seems to hold goodeven in artificial fissiparous reproduction, forHackel (27/17. 'Entwickelungsgeschichte derSiphonophora' 1869 page 73.) found that bycutting the segmented and fertilised ova or lar-va of Siphonophorae (jelly- fishes) into pieces,

the smaller the pieces were, the slower was therate of development, and the larvae thus pro-duced were by so much the more imperfect andinclined to monstrosity. It seems, therefore,probable that with the separate sexual elementsdeficient quantity of formative matter is themain cause of their not having the capacity forprolonged existence and development, unlessthey combine and thus increase each other'sbulk. The belief that it is the function of thespermatozoa to communicate life to the ovuleseems a strange one, seeing that the unimpreg-nated ovule is already alive and generally un-dergoes a certain amount of independent deve-lopment. Sexual and asexual reproduction arethus seen not to differ essentially; and we havealready shown that asexual reproduction, thepower of regrowth and development are allparts of one and the same great law.

REGROWTH OF AMPUTATED PARTS.

This subject deserves a little further discussion.A multitude of the lower animals and somevertebrates possess this wonderful power. Forinstance, Spallanzani cut off the legs and tail ofthe same salamander six times successively,and Bonnet (27/18. Spallanzani 'An Essay onAnimal Reproduction' translated by Dr. Maty1769 page 79. Bonnet 'Oeuvres d'Hist. Nat.'tome 5 part 1 4to. edition 1781 pages 343, 350.)did so eight times; and on each occasion thelimbs were reproduced on the exact line of am-putation, with no part deficient or in excess. Anallied animal, the axolotl, had a limb bitten off,which was reproduced in an abnormal condi-tion, but when this was amputated it was re-placed by a perfect limb. (27/19. Vulpian asquoted by Prof. Faivre 'La Variabilite des Espe-ces' 1868 page 112.) The new limbs in these ca-ses bud forth, and are developed in the samemanner as during the regular development of ayoung animal. For instance, with the Amblys-toma lurida, three toes are first developed, then

the fourth, and on the hind-feet the fifth, and soit is with a reproduced limb. (27/20. Dr. P. Hoy'The American Naturalist' September 1871 page579.)

The power of regrowth is generally much grea-ter during the youth of an animal or during theearlier stages of its development than duringmaturity. The larvae or tadpoles of the Batra-chians are capable of reproducing lost mem-bers, but not so the adults. (27/21. Dr. Guntherin Owen 'Anatomy of Vertebrates' volume 11866 page 567. Spallanzani has made similarobservations.) Mature insects have no power ofregrowth, excepting in one order, whilst thelarvae of many kinds have this power. Animalslow in the scale are able, as a general rule, toreproduce lost parts far more easily than thosewhich are more highly organised. The myria-pods offer a good illustration of this rule; butthere are some strange exceptions to it—thusNemerteans, though lowly organised, are said

to exhibit little power of regrowth. With thehigher vertebrata, such as birds and mammals,the power is extremely limited. (27/22. Athrush was exhibited before the British Associa-tion at Hull in 1853 which had lost its tarsus,and this member, it was asserted, had beenthrice reproduced; having been lost, I presume,each time by disease. Sir J. Paget informs methat he feels some doubt about the facts recor-ded by Sir J. Simpson ('Monthly Journal of Me-dical Science' Edinburgh 1848 new series volu-me 2 page 890) of the regrowth of limbs in thewomb in the case of man.)

In the case of those animals which may be bi-sected or chopped into pieces, and of whichevery fragment will reproduce the whole, thepower of regrowth must be diffused throug-hout the whole body. Nevertheless there seemsto be much truth in the view maintained byProf. Lessona (27/23. 'Atti della Soc. Ital. di Sc.Nat.' volume 11 1869 page 493.), that this capa-

city is generally a localised and special one,serving to replace parts which are eminentlyliable to be lost in each particular animal. Themost striking case in favour of this view, is thatthe terrestrial salamander, according to Lesso-na, cannot reproduce lost parts, whilst anotherspecies of the same genus, the aquatic sala-mander, has extraordinary powers of regrowth,as we have just seen; and this animal is eminen-tly liable to have its limbs, tail, eyes and jawsbitten off by other tritons. (27/24. Lessona sta-tes that this is so in the paper just referred to.See also 'The American Naturalist' September1871 page 579.) Even with the aquatic salaman-der the capacity is to a certain extent localised,for when M. Philipeaux (27/25. 'Comptes Ren-dus' October 1, 1866 and June 1867.) extirpatedthe entire fore limb together with the scapula,the power of regrowth was completely lost. It isalso a remarkable fact, standing in oppositionto a very general rule, that the young of theaquatic salamander do not possess the power

of repairing their limbs in an equal degree withthe adults (27/26. Bonnet 'Oeuvres Hist. Nat.'volume 5 page 294, as quoted by Prof. Rolles-ton in his remarkable address to the 36th an-nual meeting of the British Medical Associa-tion.) but I do not know that they are more ac-tive, or can otherwise better escape the loss oftheir limbs, than the adults. The walking-stickinsect, Diapheromera femorata, like other in-sects of the same order, can reproduce its legsin the mature state, and these from their greatlength must be liable to be lost: but the capacityis localised (as in the case of the salamander),for Dr. Scudder found (27/27. 'Proc. BostonSoc. of Nat. Hist.' volume 12 1868-69 page 1.),that if the limb was removed within the tro-chanto-femoral articulation, it was never rene-wed. When a crab is seized by one of its legs,this is thrown off at the basal joint, being after-wards replaced by a new leg; and it is generallyadmitted that this is a special provision for thesafety of the animal. Lastly, with gasteropod

molluscs, which are well known to have thepower of reproducing their heads, Lessonashows that they are very liable to have theirheads bitten off by fishes; the rest of the bodybeing protected by the shell. Even with plantswe see something of the same kind, for non-deciduous leaves and young stems have nopower of regrowth, these parts being easilyreplaced by growth from new buds; whilst thebark and subjacent tissues of the trunks of treeshave great power of regrowth, probably onaccount of their increase in diameter, and oftheir liability to injury from being gnawed byanimals.

GRAFT-HYBRIDS.

It is well known from innumerable trials madein all parts of the world, that buds may be in-serted into a stock, and that the plants thus rai-sed are not affected in a greater degree than canbe accounted for by changed nutrition. Nor do

the seedlings raised from such inserted budspartake of the character of the stock, thoughthey are more liable to vary than are seedlingsfrom the same variety growing on its ownroots. A bud, also, may sport into a new andstrongly-marked variety without any other budon the same plant being in the least degree af-fected. We may therefore infer, in accordancewith the common view, that each bud is a dis-tinct individual, and that its formative elementsdo not spread beyond the parts subsequentlydeveloped from it. Nevertheless, we have seenin the abstract on graft-hybridisation in theeleventh chapter that buds certainly includeformative matter, which can occasionally com-bine with that included in the tissues of a dis-tinct variety or species; a plant intermediatebetween the two parent-forms being thus pro-duced. In the case of the potato we have seenthat the tubers produced from a bud of onekind inserted into another are intermediate incolour, size, shape and state of surface; that the

stems, foliage, and even certain constitutionalpeculiarities, such as precocity, are likewiseintermediate. With these well- established ca-ses, the evidence that graft-hybrids have alsobeen produced with the laburnum, orange,vine, rose, etc., seems sufficient. But we do notknow under what conditions this rare form ofreproduction is possible. From these severalcases we learn the important fact that formativeelements capable of blending with those of adistinct individual (and this is the chief charac-teristic of sexual generation), are not confinedto the reproductive organs, but are present inthe buds and cellular tissue of plants; and thisis a fact of the highest physiological importan-ce.

DIRECT ACTION OF THE MALE ELE-MENT ON THE FEMALE.

In the eleventh chapter, abundant proofs weregiven that foreign pollen occasionally affects in

a direct manner the mother-plant. Thus, whenGallesio fertilised an orange-flower with pollenfrom the lemon, the fruit bore stripes of perfec-tly characterised lemon-peel. With peas, severalobservers have seen the colour of the seed-coatsand even of the pod directly affected by thepollen of a distinct variety. So it has been withthe fruit of the apple, which consists of the mo-dified calyx and upper part of the flower-stalk.In ordinary cases these parts are wholly formedby the mother-plant. We here see that the for-mative elements included within the male ele-ment or pollen of one variety can affect andhybridise, not the part which they are properlyadapted to affect, namely, the ovules, but thepartially-developed tissues of a distinct varietyor species. We are thus brought half-way to-wards a graft- hybrid, in which the formativeelements included within the tissues of oneindividual combine with those included in thetissues of a distinct variety or species, thus gi-

ving rise to a new and intermediate form, inde-pendently of the male or female sexual organs.

With animals which do not breed until nearlymature, and of which all the parts are then fullydeveloped, it is hardly possible that the maleelement should directly affect the female. Butwe have the analogous and perfectly well-ascertained case of the male element affecting(as with the quagga and Lord Morton's mare)the female or her ova, in such a manner thatwhen she is impregnated by another male heroffspring are affected and hybridised by thefirst male. The explanation would be simple ifthe spermatozoa could keep alive within thebody of the female during the long intervalwhich has sometimes elapsed between the twoacts of impregnation; but no one will supposethat this is possible with the higher animals.

DEVELOPMENT.

The fertilised germ reaches maturity by a vastnumber of changes: these are either slight andslowly effected, as when the child grows intothe man, or are great and sudden, as with themetamorphoses of most insects. Between theseextremes we have every gradation, even withinthe same class; thus, as Sir J. Lubbock hasshown (27/28. 'Transact. Linn. Soc.' volume 241863 page 62.) there is an Ephemerous insectwhich moults above twenty times, undergoingeach time a slight but decided change of struc-ture; and these changes, as he further remarks,probably reveal to us the normal stages of de-velopment, which are concealed and hurriedthrough or suppressed in most other insects. Inordinary metamorphoses, the parts and organsappear to become changed into the correspon-ding parts in the next stage of development;but there is another form of development,which has been called by Professor Owen me-tagenesis. In this case "the new parts are notmoulded upon the inner surface of the old

ones. The plastic force has changed its course ofoperation. The outer case, and all that gaveform and character to the precedent individual,perish and are cast off; they are not changedinto the corresponding parts of the new indivi-dual. These are due to a new and distinct deve-lopmental process," etc. (27/29. 'Parthenogene-sis' 1849 pages 25, 26. Prof. Huxley has someexcellent remarks ('Medical Times' 1856 page637) on this subject in reference to the deve-lopment of star-fishes, and shows how curious-ly metamorphosis graduates into gemmation orzoid-formation, which is in fact the same asmetagenesis.) Metamorphosis, however, gra-duates so insensibly, into metagenesis, that thetwo processes cannot be distinctly separated.For instance, in the last change which Cirripe-des undergo, the alimentary canal and someother organs are moulded on pre-existing parts;but the eyes of the old and the young animalare developed in entirely different parts of thebody; the tips of the mature limbs are formed

within the larval limbs, and may be said to bemetamorphosed from them; but their basalportions and the whole thorax are developed ina plane at right angles to the larval limbs andthorax; and this may be called metagenesis. Themetagenetic process is carried to an extremepoint in the development of some Echino-derms, for the animal in the second stage ofdevelopment is formed almost like a bud wit-hin the animal of the first stage, the latter beingthen cast off like an old vestment, yet someti-mes maintaining for a short period an inde-pendent vitality. (27/30. Prof. J. Reay Greene inGunther's 'Record of Zoolog. Lit.' 1865 page625.) If, instead of a single individual, severalwere to be thus developed metageneticallywithin a pre- existing form, the process wouldbe called one of alternate generation. Theyoung thus developed may either closely re-semble the encasing parent-form, as with thelarvae of Cecidomyia, or may differ to an asto-nishing degree, as with many parasitic worms

and jelly-fishes; but this does not make anyessential difference in the process, any morethan the greatness or abruptness of the changein the metamorphoses of insects.

The whole question of development is of greatimportance for our present subject. When anorgan, the eye, for instance, is metageneticallyformed in a part of the body where during theprevious stage of development no eye existed,we must look at it as a new and independentgrowth. The absolute independence of new andold structures, although corresponding instructure and function, is still more obviouswhen several individuals are formed within aprevious form, as in the cases of alternate gene-ration. The same important principle probablycomes largely into play even in the case of ap-parently continuous growth, as we shall seewhen we consider the inheritance of modifica-tions at corresponding ages.

We are led to the same conclusion, namely, theindependence of parts successively developed,by another and quite distinct group of facts. Itis well known that many animals belonging tothe same order, and therefore not differing wi-dely from each other, pass through an extreme-ly different course of development. Thus cer-tain beetles, not in any way remarkably diffe-rent from others of the same order, undergowhat has been called a hyper-metamorphosis—that is, they pass through an early stage whollydifferent from the ordinary grub-like larva. Inthe same sub-order of crabs, namely, the Ma-croura, as Fritz Muller remarks, the river cray-fish is hatched under the same form which itever afterwards retains; the young lobster hasdivided legs, like a Mysis; the Palaemon ap-pears under the form of a Zoea, and Peneusunder the Nauplius- form; and how wonderfu-lly these larval forms differ from one another, isknown to every naturalist. (27/31. Fritz Muller'Fur Darwin' 1864 s. 65, 71. The highest authori-

ty on crustaceans, Prof. Milne-Edwards, insists('Annal. des Sci. Nat.' 2nd series Zoolog. tome 3page 322) on the difference in the metamorp-hosis of closely-allied genera.) Some other crus-taceans, as the same author observes, start fromthe same point and arrive at nearly the sameend, but in the middle of their development arewidely different from one another. Still morestriking cases could be given with respect to theEchinodermata. With the Medusae or jelly-fishes Professor Allman observes, "The classifi-cation of the Hydroida would be a comparati-vely simple task if, as has been erroneouslyasserted, generically-identical medusoids al-ways arose from generically-identical poly-poids; and, on the other hand, that generically-identical polypoids always gave origin to gene-rically-identical medusoids." So again, Dr.Strethill Wright remarks, "In the life-history ofthe Hydroidae any phase, planuloid, polypoid,or medusoid, may be absent." (27/32. Prof.Allman 'Annals and Mag. of Nat. Hist.' 3rd

series volume 13 1864 page 348; Dr. S. Wrightibid volume 8 1861 page 127. See also page 358for analogous statements by Sars.)

According to the belief now generally acceptedby our best naturalists, all the members of thesame order or class, for instance, the Medusaeor the Macrourous crustaceans, are descendedfrom a common progenitor. During their des-cent they have diverged much in structure, buthave retained much in common; and this hasoccurred, though they have passed throughand still pass through marvellously differentmetamorphoses. This fact well illustrates howindependent each structure is from that whichprecedes and that which follows it in the courseof development.

THE FUNCTIONAL INDEPENDENCE OFTHE ELEMENTS OR UNITS OF THE BODY.

Physiologists agree that the whole organismconsists of a multitude of elemental parts,

which are to a great extent independent of oneanother. Each organ, says Claude Bernard(27/33. 'Tissus Vivants' 1866 page 22.), has itsproper life, its autonomy; it can develop andreproduce itself independently of the adjoiningtissues. A great German authority, Virchow(27/34. 'Cellular Pathology' translated by Dr.Chance 1860 pages 14, 18, 83, 460.), asserts stillmore emphatically that each system consists ofan "enormous mass of minute centres of ac-tion…Every element has its own special action,and even though it derive its stimulus to activi-ty from other parts, yet alone effects the actualperformance of duties…Every single epithelialand muscular fibre- cell leads a sort of parasiti-cal existence in relation to the rest of the bo-dy…Every single bone-corpuscle really posses-ses conditions of nutrition peculiar to itself."Each element, as Sir J. Paget remarks, lives itsappointed time and then dies, and is replacedafter being cast off or absorbed. (27/35. Paget'Surgical Pathology' volume 1 1853 pages 12-

14.) I presume that no physiologist doubts that,for instance, each bone-corpuscle of the fingerdiffers from the corresponding corpuscle in thecorresponding joint of the toe; and there canhardly be a doubt that even those on the co-rresponding sides of the body differ, thoughalmost identical in nature. This near approachto identity is curiously shown in many diseasesin which the same exact points on the right andleft sides of the body are similarly affected; thusSir J. Paget (27/36. Ibid page 19.) gives a dra-wing of a diseased pelvis, in which the bonehas grown into a most complicated pattern, but"there is not one spot or line on one side whichis not represented, as exactly as it would be in amirror, on the other."

Many facts support this view of the indepen-dent life of each minute element of the body.Virchow insists that a single bone-corpuscle ora single cell in the skin may become diseased.The spur of a cock, after being inserted into the

ear of an ox, lived for eight years, and acquireda weight of 396 grammes (nearly fourteen oun-ces), and the astonishing length of twenty-fourcentimetres, or about nine inches; so that thehead of the ox appeared to bear three horns.(27/37. See Prof. Mantegazza's interesting work'Degli innesti Animali' etc. Milano 1865 page 51tab. 3.) The tail of a pig has been grafted intothe middle of its back, and reacquired sensibili-ty. Dr. Ollier (27/38. 'De la Production Artificie-lle des Os' page 8.) inserted a piece of perios-teum from the bone of a young dog under theskin of a rabbit, and true bone was developed.A multitude of similar facts could be given. Thefrequent presence of hairs and of perfectly de-veloped teeth, even teeth of the second denti-tion, in ovarian tumours (27/39. Isidore Geof-froy Saint-Hilaire 'Hist. des Anomalies' tome 2pages 549, 560, 562; Virchow ibid page 484.Lawson Tait 'The Pathology of Diseases of theOvaries' 1874 pages 61, 62.), are facts leading tothe same conclusion. Mr. Lawson Tait refers to

a tumour in which "over 300 teeth were found,resembling in many respects milk-teeth;" and toanother tumour, "full of hair which had grownand been shed from one little spot of skin notbigger than the tip of my little finger. Theamount of hair in the sac, had it grown from asimilarly sized area of the scalp, would havetaken almost a lifetime to grow and be shed."

Whether each of the innumerable autonomouselements of the body is a cell or the modifiedproduct of a cell, is a more doubtful question,even if so wide a definition be given to theterm, as to include cell-like bodies withoutwalls and without nuclei. (27/40. For the mostrecent classification of cells, see Ernst Hackel'Generelle Morpholog.' b. 2 1866 s. 275.) Thedoctrine of omnis cellula e cellula is admittedfor plants, and widely prevails with respect toanimals. (27/41. Dr. W. Turner 'The PresentAspect of Cellular Pathology' 'Edinburgh Me-dical Journal' April 1863.) Thus Virchow, the

great supporter of the cellular theory, whilstallowing that difficulties exist, maintains thatevery atom of tissue is derived from cells, andthese from pre-existing cells, and these primari-ly from the egg, which he regards as a greatcell. That cells, still retaining the same nature,increase by self-division or proliferation, is ad-mitted by every one. But when an organismundergoes great changes of structure duringdevelopment, the cells, which at each stage aresupposed to be directly derived from previous-ly existing cells, must likewise be greatly chan-ged in nature; this change is attributed by thesupporters of the cellular doctrine to some in-herent power which the cells possess, and notto any external agency. Others maintain thatcells and tissues of all kinds may be formed,independently of pre-existing cells, from plasticlymph or blastema. Whichever view may becorrect, every one admits that the body consistsof a multitude of organic units, all of whichpossess their own proper attributes, and are to

a certain extent independent of all others. Hen-ce it will be convenient to use indifferently theterms cells or organic units, or simply units.

VARIABILITY AND INHERITANCE.

We have seen in the twenty-second chapter thatvariability is not a principle co-ordinate withlife or reproduction, but results from specialcauses, generally from changed conditions ac-ting during successive generations. The fluctua-ting variability thus induced is apparently duein part to the sexual system being easily affec-ted, so that it is often rendered impotent; andwhen not so seriously affected, it often fails inits proper function of transmitting truly thecharacters of the parents to the offspring. Butvariability is not necessarily connected with thesexual system, as we see in the cases of bud-variation. Although we are seldom able to tracethe nature of the connection, many deviationsof structure no doubt result from changed con-

ditions acting directly on the organisation, in-dependently of the reproductive system. Insome instances we may feel sure of this, whenall, or nearly all the individuals which havebeen similarly exposed are similarly and defini-tely affected, of which several instances havebeen given. But it is by no means clear why theoffspring should be affected by the exposure ofthe parents to new conditions, and why it isnecessary in most cases that several generationsshould have been thus exposed.

How, again, can we explain the inherited ef-fects of the use or disuse of particular organs?The domesticated duck flies less and walksmore than the wild duck, and its limb-boneshave become diminished and increased in acorresponding manner in comparison with tho-se of the wild duck. A horse is trained to certainpaces, and the colt inherits similar consensualmovements. The domesticated rabbit becomestame from close confinement; the dog, intelli-

gent from associating with man; the retriever istaught to fetch and carry; and these mental en-dowments and bodily powers are all inherited.Nothing in the whole circuit of physiology ismore wonderful. How can the use or disuse ofa particular limb or of the brain affect a smallaggregate of reproductive cells, seated in a dis-tant part of the body, in such a manner that thebeing developed from these cells inherits thecharacters of either one or both parents? Evenan imperfect answer to this question would besatisfactory.

In the chapters devoted to inheritance it wasshown that a multitude of newly acquired cha-racters, whether injurious or beneficial, whet-her of the lowest or highest vital importance,are often faithfully transmitted—frequentlyeven when one parent alone possesses somenew peculiarity; and we may on the whole con-clude that inheritance is the rule, and non-inheritance the anomaly. In some instances a

character is not inherited, from the conditionsof life being directly opposed to its develop-ment; in many instances, from the conditionsincessantly inducing fresh variability, as withgrafted fruit-trees and highly-cultivated flo-wers. In the remaining cases the failure may beattributed to reversion, by which the child re-sembles its grandparents or more remote pro-genitors, instead of its parents.

Inheritance is governed by various laws. Cha-racters which first appear at any particular agetend to reappear at a corresponding age. Theyoften become associated with certain seasons ofthe year, and reappear in the offspring at a co-rresponding season. If they appear rather latein life in one sex, they tend to reappear exclusi-vely in the same sex at the same period of life.

The principle of reversion, recently alluded to,is one of the most wonderful of the attributes ofInheritance. It proves to us that the transmis-sion of a character and its development, which

ordinarily go together and thus escape discri-mination, are distinct powers; and these po-wers in some cases are even antagonistic, foreach acts alternately in successive generations.Reversion is not a rare event, depending onsome unusual or favourable combination ofcircumstances, but occurs so regularly withcrossed animals and plants, and so frequentlywith uncrossed breeds, that it is evidently anessential part of the principle of inheritance. Weknow that changed conditions have the powerof evoking long-lost characters, as in the case ofanimals becoming feral. The act of crossing initself possesses this power in a high degree.What can be more wonderful than that charac-ters, which have disappeared during scores, orhundreds, or even thousands of generations,should suddenly reappear perfectly developed,as in the case of pigeons and fowls, both whenpurely bred and especially when crossed; or aswith the zebrine stripes on dun-coloured hor-ses, and other such cases? Many monstrosities

come under this same head, as when rudimen-tary organs are redeveloped, or when an organwhich we must believe was possessed by anearly progenitor of the species, but of which noteven a rudiment is left, suddenly reappears, aswith the fifth stamen in some Scrophulariaceae.We have already seen that reversion acts inbud- reproduction; and we know that it occa-sionally acts during the growth of the sameindividual animal, especially, but not exclusive-ly, if of crossed parentage,—as in the rare casesdescribed of fowls, pigeons, cattle, and rabbits,which have reverted to the colours of one oftheir parents or ancestors as they advanced inyears.

We are led to believe, as formerly explained,that every character which occasionally reap-pears is present in a latent form in each genera-tion, in nearly the same manner as in male andfemale animals the secondary characters of theopposite sex lie latent and ready to be evolved

when the reproductive organs are injured. Thiscomparison of the secondary sexual characterswhich lie latent in both sexes, with other latentcharacters, is the more appropriate from thecase recorded of a Hen, which assumed someof the masculine characters, not of her own ra-ce, but of an early progenitor; she thus exhibi-ted at the same time the redevelopment of la-tent characters of both kinds. In every livingcreature we may feel assured that a host oflong-lost characters lie ready to be evolved un-der proper conditions. How can we make inte-lligible and connect with other facts, this won-derful and common capacity of reversion,—thispower of calling back to life long-lost charac-ters?

PART II.

I have now enumerated the chief facts whichevery one would desire to see connected bysome intelligible bond. This can be done, if we

make the following assumptions, and muchmay be advanced in favour of the chief one.The secondary assumptions can likewise besupported by various physiological considera-tions. It is universally admitted that the cells orunits of the body increase by self-division orproliferation, retaining the same nature, andthat they ultimately become converted into thevarious tissues and substances of the body. Butbesides this means of increase I assume that theunits throw off minute granules which are dis-persed throughout the whole system; that the-se, when supplied with proper nutriment, mul-tiply by self-division, and are ultimately deve-loped into units like those from which theywere originally derived. These granules may becalled gemmules. They are collected from allparts of the system to constitute the sexual ele-ments, and their development in the next gene-ration forms a new being; but they are likewisecapable of transmission in a dormant state tofuture generations and may then be developed.

Their development depends on their unionwith other partially developed or nascent cellswhich precede them in the regular course ofgrowth. Why I use the term union, will be seenwhen we discuss the direct action of pollen onthe tissues of the mother-plant. Gemmules aresupposed to be thrown off by every unit, notonly during the adult state, but during eachstage of development of every organism; butnot necessarily during the continued existenceof the same unit. Lastly, I assume that thegemmules in their dormant state have a mutualaffinity for each other, leading to their aggrega-tion into buds or into the sexual elements. Hen-ce, it is not the reproductive organs or budswhich generate new organisms, but the units ofwhich each individual is composed. These as-sumptions constitute the provisional hypot-hesis which I have called Pangenesis. Views inmany respects similar have been propoundedby various authors. (27/42. Mr. G.H. Lewes('Fortnightly Review' November 1, 1868 page

506) remarks on the number of writers whohave advanced nearly similar views. More thantwo thousand years ago Aristotle combated aview of this kind, which, as I hear from Dr. W.Ogle, was held by Hippocrates and others. Ray,in his 'Wisdom of God' (2nd edition 1692 page68), says that "every part of the body seems toclub and contribute to the seed." The "organicmolecules" of Buffon ('Hist. Nat. Gen.' editionof 1749 tome 2 pages 54, 62, 329, 333, 420, 425)appear at first sight to be the same as the gem-mules of my hypothesis, but they are essentia-lly different. Bonnet ('Oeuvres d'Hist. Nat.' to-me 5 part 1 1781 4to edition page 334) speaks ofthe limbs having germs adapted for the repara-tion of all possible losses; but whether thesegerms are supposed to be the same with thosewithin buds and the sexual organs is not clear.Prof. Owen says ('Anatomy of Vertebrates' vo-lume 3 1868 page 813) that he fails to see anyfundamental difference between the viewswhich he propounded in his 'Parthenogenesis'

(1849 pages 5- 8), and which he now considersas erroneous, and my hypothesis of pangenesis:but a reviewer ('Journal of Anat. and Phys.'May 1869 page 441) shows how different theyreally are. I formerly thought that the "physio-logical units" of Herbert Spencer ('Principles ofBiology' volume 1 chapters 4 and 8 1863-64)were the same as my gemmules, but I nowknow that this is not the case. Lastly, it appearsfrom a review of the present work by Prof.Mantegazza ('Nuova Antologia, Maggio' 1868),that he (in his 'Elementi di Igiene' Ediz. 3 page540) clearly foresaw the doctrine of pangene-sis.)

Before proceeding to show, firstly, how far the-se assumptions are in themselves probable, andsecondly, how far they connect and explain thevarious groups of facts with which we are con-cerned, it may be useful to give an illustration,as simple as possible, of the hypothesis. If oneof the Protozoa be formed, as it appears under

the microscope, of a small mass of homoge-neous gelatinous matter, a minute particle orgemmule thrown off from any part and nouris-hed under favourable circumstances wouldreproduce the whole; but if the upper and lo-wer surfaces were to differ in texture from eachother and from the central portion, then allthree parts would have to throw off gemmules,which when aggregated by mutual affinitywould form either buds or the sexual elements,and would ultimately be developed into a simi-lar organism. Precisely the same view may beextended to one of the higher animals; althoughin this case many thousand gemmules must bethrown off from the various parts of the bodyat each stage of development; these gemmulesbeing developed in union with pre-existingnascent cells in due order of succession.

Physiologists maintain, as we have seen, thateach unit of the body, though to a large extentdependent on others, is likewise to a certain

extent independent or autonomous, and has thepower of increasing by self-division. I go onestep further, and assume that each unit casts offfree gemmules which are dispersed throughoutthe system, and are capable under proper con-ditions of being developed into similar units.Nor can this assumption be considered as gra-tuitous and improbable. It is manifest that thesexual elements and buds include formativematter of some kind, capable of development;and we now know from the production ofgraft-hybrids that similar matter is dispersedthroughout the tissues of plants, and can com-bine with that of another and distinct plant,giving rise to a new being, intermediate in cha-racter. We know also that the male element canact directly on the partially developed tissuesof the mother-plant, and on the future progenyof female animals. The formative matter whichis thus dispersed throughout the tissues ofplants, and which is capable of being develo-ped into each unit or part, must be generated

there by some means; and my chief assumptionis that this matter consists of minute particles orgemmules cast off from each unit or cell.(27/43. Mr. Lowne has observed ('Journal ofQueckett Microscopical Club' September 23,1870) certain remarkable changes in the tissuesof the larva of a fly, which makes him believe"it possible that organs and organisms are so-metimes developed by the aggregation of ex-cessively minute gemmules, such as thosewhich Mr. Darwin's hypothesis demands.")

But I have further to assume that the gemmulesin their undeveloped state are capable of large-ly multiplying themselves by self-division, likeindependent organisms. Delpino insists that to"admit of multiplication by fissiparity in cor-puscles, analogous to seeds or buds…is repug-nant to all analogy." But this seems a strangeobjection, as Thuret (27/44. 'Annales des Sc.Nat.' 3rd series Bot. tome 14 1850 page 244.) hasseen the zoospore of an alga divide itself, and

each half germinated. Haeckel divided thesegmented ovum of a siphonophora into manypieces, and these were developed. Nor does theextreme minuteness of the gemmules, whichcan hardly differ much in nature from the lo-west and simplest organisms, render it impro-bable that they should grow and multiply. Agreat authority, Dr. Beale (27/45. 'DiseaseGerms' page 20.), says "that minute yeast cellsare capable of throwing off buds or gemmules,much less than the 1/100000 of an inch in dia-meter;" and these he thinks are "capable of sub-division practically ad infinitum."

A particle of small-pox matter, so minute as tobe borne by the wind, must multiply itself ma-ny thousandfold in a person thus inoculated;and so with the contagious matter of scarletfever. (27/46. See some very interesting paperson this subject by Dr. Beale in 'Medical Timesand Gazette' September 9, 1865 pages 273, 330.)It has recently been ascertained (27/47. Third

Report of the R. Comm. on the Cattle Plague asquoted in 'Gardener's Chronicle' 1866 page446.) that a minute portion of the mucous dis-charge from an animal affected with rinderpest,if placed in the blood of a healthy ox, increasesso fast that in a short space of time "the wholemass of blood, weighing many pounds, is infec-ted, and every small particle of that blood con-tains enough poison to give, within less thanforty-eight hours, the disease to another ani-mal."

The retention of free and undeveloped gemmu-les in the same body from early youth to oldage will appear improbable, but we shouldremember how long seeds lie dormant in theearth and buds in the bark of a tree. Theirtransmission from generation to generation willappear still more improbable; but here againwe should remember that many rudimentaryand useless organs have been transmitted du-ring an indefinite number of generations. We

shall presently see how well the long-continuedtransmission of undeveloped gemmules ex-plains many facts.

As each unit, or group of similar units, throug-hout the body, casts off its gemmules, and as allare contained within the smallest ovule, andwithin each spermatozoon or pollen-grain, andas some animals and plants produce an asto-nishing number of pollen-grains and ovules(27/48. Mr. F. Buckland found 6,867,840 eggs ina cod-fish ('Land and Water' 1868 page 62). AnAscaris produces about 64,000,000 eggs (Car-penter's 'Comp. Phys.' 1854 page 590). Mr. J.Scott, of the Royal Botanic Garden of Edin-burgh, calculated, in the same manner as I havedone for some British Orchids ('Fertilisation ofOrchids' page 344), the number of seeds in acapsule of an Acropera and found the numberto be 371,250. Now this plant produces severalflowers on a raceme, and many racemes duringa season. In an allied genus, Gongora, Mr. Scott

has seen twenty capsules produced on a singleraceme; ten such racemes on the Acroperawould yield above seventy-four millions ofseed.), the number and minuteness of the gem-mules must be something inconceivable. Butconsidering how minute the molecules are, andhow many go to the formation of the smallestgranule of any ordinary substance, this difficul-ty with respect to the gemmules is not insupe-rable. From the data arrived at by Sir W. Thom-son, my son George finds that a cube of1/10000 of an inch of glass or water must con-sist of between 16 million millions, and 131thousand million million molecules. No doubtthe molecules of which an organism is formedare larger, from being more complex, than tho-se of an inorganic substance, and probably ma-ny molecules go to the formation of a gemmule;but when we bear in mind that a cube of1/10000 of an inch is much smaller than anypollen-grain, ovule or bud, we can see what a

vast number of gemmules one of these bodiesmight contain.

The gemmules derived from each part or organmust be thoroughly dispersed throughout thewhole system. We know, for instance, that evena minute fragment of a leaf of a Begonia willreproduce the whole plant; and that if a fresh-water worm is chopped into small pieces, eachwill reproduce the whole animal. Consideringalso the minuteness of the gemmules and thepermeability of all organic tissues, the thoroughdispersion of the gemmules is not surprising.That matter may be readily transferred withoutthe aid of vessels from part to part of the body,we have a good instance in a case recorded bySir J. Paget of a lady, whose hair lost its colourat each successive attack of neuralgia and reco-vered it again in the course of a few days. Withplants, however, and probably with compoundanimals, such as corals, the gemmules do notordinarily spread from bud to bud, but are con-

fined to the parts developed from each separatebud; and of this fact no explanation can be gi-ven.

The assumed elective affinity of each gemmulefor that particular cell which precedes it in dueorder of development is supported by manyanalogies. In all ordinary cases of sexual repro-duction, the male and female elements certainlyhave a mutual affinity for each other: thus, it isbelieved that about ten thousand species ofCompositae exist, and there can be no doubtthat if the pollen of all these species could besimultaneously or successively placed on thestigma of any one species, this one would electwith unerring certainty its own pollen. Thiselective capacity is all the more wonderful, as itmust have been acquired since the many spe-cies of this great group of plants branched offfrom a common progenitor. On any view of thenature of sexual reproduction, the formativematter of each part contained within the ovules

and the male element act on each other by somelaw of special affinity, so that correspondingparts affect one another; thus, a calf producedfrom a short-horned cow by a long-horned bullhas its horns affected by the union of the twoforms, and the offspring from two birds withdifferently coloured tails have their tails affec-ted.

The various tissues of the body plainly show, asmany physiologists have insisted (27/49. Paget'Lectures on Pathology' page 27; Virchow 'Ce-llular Pathology' translated by Dr. Chance pa-ges 123, 126, 294. Claude Bernard 'Des TissusVivants' pages 177, 210, 337; Muller 'Physiolo-gy' English translation page 290.), an affinityfor special organic substances, whether naturalor foreign to the body. We see this in the cellsof the kidneys attracting urea from the blood;in curare affecting certain nerves; Lytta vesica-toria the kidneys; and the poisonous matter ofvarious diseases, as small-pox, scarlet- fever,

hooping-cough, glanders, and hydrophobia,affecting certain definite parts of the body. Ithas also been assumed that the development ofeach gemmule depends on its union with anot-her cell or unit which has just commenced itsdevelopment, and which precedes it in dueorder of growth. That the formative matterwithin the pollen of plants, which by our hy-pothesis consists of gemmules, can unite withand modify the partially developed cells of themother-plant, we have clearly seen in the sec-tion devoted to this subject. As the tissues ofplants are formed, as far as is known, only bythe proliferation of pre-existing cells, we mustconclude that the gemmules derived from theforeign pollen do not become developed intonew and separate cells, but penetrate and mo-dify the nascent cells of the mother-plant. Thisprocess may be compared with what takes pla-ce in the act of ordinary fertilisation, duringwhich the contents of the pollen-tubes penetra-te the closed embryonic sac within the ovule,

and determine the development of the embryo.According to this view, the cells of the mother-plant may almost literally be said to be fertili-sed by the gemmules derived from the foreignpollen. In this case and in all others the propergemmules must combine in due order withpre-existing nascent cells, owing to their electi-ve affinities. A slight difference in nature bet-ween the gemmules and the nascent cellswould be far from interfering with their mutualunion and development, for we well know inthe case of ordinary reproduction that suchslight differentiation in the sexual elementsfavours in a marked manner their union andsubsequent development, as well as the vigourof the offspring thus produced.

Thus far we have been able by the aid of ourhypothesis to throw some obscure light on theproblems which have come before us; but itmust be confessed that many points remainaltogether doubtful. Thus it is useless to specu-

late at what period of development each unit ofthe body casts off its gemmules, as the wholesubject of the development of the various tis-sues is as yet far from clear. We do not knowwhether the gemmules are merely collected bysome unknown means at certain seasons withinthe reproductive organs, or whether after beingthus collected they rapidly multiply there, asthe flow of blood to these organs at each bree-ding season seems to render probable. Nor dowe know why the gemmules collect to formbuds in certain definite places, leading to thesymmetrical growth of trees and corals. Wehave no means of deciding whether the ordina-ry wear and tear of the tissues is made good bymeans of gemmules, or merely by the prolifera-tion of pre-existing cells. If the gemmules arethus consumed, as seems probable from theintimate connection between the repair of was-te, regrowth, and development, and more espe-cially from the periodical changes which manymale animals undergo in colour and structure,

then some light would be thrown on the phe-nomena of old age, with its lessened power ofreproduction and of the repair of injuries, andon the obscure subject of longevity. The fact ofcastrated animals, which do not cast off innu-merable gemmules in the act of reproduction,not being longer-lived than perfect males,seems opposed to the belief that gemmules areconsumed in the ordinary repair of wasted tis-sues; unless indeed the gemmules after beingcollected in small numbers within the repro-ductive organs are there largely multiplied.(27/50. Prof. Ray Lankester has discussed seve-ral of the points here referred to as bearing onpangenesis, in his interesting essay, 'On Com-parative Longevity in Man and the LowerAnimals' 1870 pages 33, 77, etc.)

That the same cells or units may live for a longperiod and continue multiplying without beingmodified by their union with free gemmules ofany kind, is probable from such cases as that of

the spur of a cock which grew to an enormoussize when grafted into the ear of an ox. How farunits are modified during their normal growthby absorbing peculiar nutriment from the su-rrounding tissues, independently of their unionwith gemmules of a distinct nature, is anotherdoubtful point. (27/51. Dr. Ross refers to thissubject in his 'Graft Theory of Disease' 1872page 53.) We shall appreciate this difficulty bycalling to mind what complex yet symmetricalgrowths the cells of plants yield when inocula-ted by the poison of a gall-insect. With animalsvarious polypoid excrescences and tumours aregenerally admitted (27/52. Virchow 'CellularPathology' translated by Dr. Chance 1860 pages60, 162, 245, 441, 454.) to be the direct product,through proliferation, of normal cells whichhave become abnormal. In the regular growthand repair of bones, the tissues undergo, asVirchow remarks (27/53. Ibid pages 412-426.), awhole series of permutations and substitutions."The cartilage cells may be converted by a dire-

ct transformation into marrow-cells, and conti-nue as such; or they may first be converted intoosseous and then into medullary tissue; or las-tly, they may first be converted into marrowand then into bone. So variable are the permu-tations of these tissues, in themselves so nearlyallied, and yet in their external appearance socompletely distinct." But as these tissues thuschange their nature at any age, without anyobvious change in their nutrition, we mustsuppose in accordance with our hypothesis thatgemmules derived from one kind of tissuecombine with the cells of another kind, andcause the successive modifications.

We have good reason to believe that severalgemmules are requisite for the development ofone and the same unit or cell; for we cannototherwise understand the insufficiency of asingle or even of two or three pollen-grains orspermatozoa. But we are far from knowingwhether the gemmules of all the units are free

and separate from one another, or whether so-me are from the first united into small aggrega-tes. A feather, for instance, is a complex struc-ture, and, as each separate part is liable to in-herited variations, I conclude that each feathergenerates a large number of gemmules; but it ispossible that these may be aggregated into acompound gemmule. The same remark appliesto the petals of flowers, which are sometimeshighly complex structures, with each ridge andhollow contrived for a special purpose, so thateach part must have been separately modified,and the modifications transmitted; consequen-tly, separate gemmules, according to our hy-pothesis, must have been thrown off from eachcell or unit. But, as we sometimes see half ananther or a small portion of a filament beco-ming petali-form, or parts or mere stripes of thecalyx assuming the colour and texture of thecorolla, it is probable that with petals the gem-mules of each cell are not aggregated togetherinto a compound gemmule, but are free and

separate. Even in so simple a case as that of aperfect cell, with its protoplasmic contents, nu-cleus, nucleolus, and walls, we do not knowwhether or not its development depends on acompound gemmule derived from each part.(27/54. See some good criticisms on this headby Delpino and by Mr. G.H. Lewes in the 'Fort-nightly Review' November 1, 1868 page 509.)

Having now endeavoured to show that theseveral foregoing assumptions are to a certainextent supported by analogous facts, andhaving alluded to some of the most doubtfulpoints, we will consider how far the hypothesisbrings under a single point of view the variouscases enumerated in the First Part. All theforms of reproduction graduate into one anot-her and agree in their product; for it is impossi-ble to distinguish between organisms producedfrom buds, from self-division, or from fertilisedgerms; such organisms are liable to variationsof the same nature and to reversions of the sa-

me kind; and as, according to our hypothesis,all the forms of reproduction depend on theaggregation of gemmules derived from thewhole body, we can understand this remarka-ble agreement. Parthenogenesis is no longerwonderful, and if we did not know that greatgood followed from the union of the sexualelements derived from two distinct individuals,the wonder would be that parthenogenesis didnot occur much oftener than it does. On anyordinary theory of reproduction the formationof graft-hybrids, and the action of the maleelement on the tissues of the mother-plant, aswell as on the future progeny of female ani-mals, are great anomalies; but they are intelli-gible on our hypothesis. The reproductive or-gans do not actually create the sexual elements;they merely determine the aggregation andperhaps the multiplication of the gemmules ina special manner. These organs, however, to-gether with their accessory parts, have highfunctions to perform. They adapt one or both

elements for independent temporary existence,and for mutual union. The stigmatic secretionacts on the pollen of a plant of the same speciesin a wholly different manner to what it does onthe pollen of one belonging to a distinct genusor family. The spermatophores of the Cephalo-poda are wonderfully complex structures,which were formerly mistaken for parasiticworms; and the spermatozoa of some animalspossess attributes which, if observed in an in-dependent animal, would be put down to ins-tinct guided by sense-organs,—as when thespermatozoa of an insect find their way into theminute micropyle of the egg.

The antagonism which has long been observed(27/55. Mr. Herbert Spencer ('Principles of Bio-logy' volume 2 page 430) has fully discussedthis antagonism.), with certain exceptions, bet-ween growth and the power of sexual repro-duction (27/56. The male salmon is known tobreed at a very early age. The Triton and Sire-

don, whilst retaining their larval branchiae,according to Filippi and Dumeril ('Annals andMag. of Nat. Hist.' 3rd series 1866 page 157) arecapable of reproduction. Ernst Haeckel has re-cently ('Monatsbericht Akad. Wiss. Berlin' Fe-bruary 2, 1865) observed the surprising case ofa medusa, with its reproductive organs active,which produces by budding a widely differentform of medusa; and this latter also has thepower of sexual reproduction. Krohn hasshown ('Annals and Mag. of Nat. Hist.' 3rd se-ries volume 19 1862 page 6) that certain othermedusae, whilst sexually mature, propagate bygemmae. See also Kolliker 'Morphologie undEntwickelungsgeschichte des Pennatulidens-tammes' 1872 page 12.)—between the repair ofinjuries and gemmation—and with plants, bet-ween rapid increase by buds, rhizomes, etc.,and the production of seed, is partly explainedby the gemmules not existing in sufficientnumbers for these processes to be carried onsimultaneously.

Hardly any fact in physiology is more wonder-ful than the power of regrowth; for instance,that a snail should be able to reproduce itshead, or a salamander its eyes, tail, and legs,exactly at the points where they have been cutoff. Such cases are explained by the presence ofgemmules derived from each part, and disse-minated throughout the body. I have heard theprocess compared with that of the repair of thebroken angles of a crystal by re- crystallisation;and the two processes have this much in com-mon, that in the one case the polarity of themolecules is the efficient cause, and in the otherthe affinity of the gemmules for particular nas-cent cells. But we have here to encounter twoobjections which apply not only to the re-growth of a part, or of a bisected individual,but to fissiparous generation and budding. Thefirst objection is that the part which is reprodu-ced is in the same stage of development as thatof the being which has been operated on orbisected; and in the case of buds, that the new

beings thus produced are in the same stage asthat of the budding parent. Thus a mature sa-lamander, of which the tail has been cut off,does not reproduce a larval tail; and a crab doesnot reproduce a larval leg. In the case of bud-ding it was shown in the first part of this chap-ter that the new being thus produced does notretrograde in development,—that is, does notpass through those earlier stages, which thefertilised germ has to pass through. Nevert-heless, the organisms operated on or multipl-ying themselves by buds must, by our hypot-hesis, include innumerable gemmules derivedfrom every part or unit of the earlier stages ofdevelopment; and why do not such gemmulesreproduce the amputated part or the wholebody at a corresponding early stage of deve-lopment?

The second objection, which has been insistedon by Delpino, is that the tissues, for instance,of a mature salamander or crab, of which a

limb has been removed, are already differentia-ted and have passed through their whole cour-se of development; and how can such tissues inaccordance with our hypothesis attract andcombine with the gemmules of the part whichis to be reproduced? In answer to these twoobjections we must bear in mind the evidencewhich has been advanced, showing that at leastin a large number of cases the power of re-growth is a localised faculty, acquired for thesake of repairing special injuries to which eachparticular creature is liable; and in the case ofbuds or fissiparous generation, for the sake ofquickly multiplying the organism at a period oflife when it can be supported in large numbers.These considerations lead us to believe that inall such cases a stock of nascent cells or of par-tially developed gemmules are retained for thisspecial purpose either locally or throughout thebody, ready to combine with the gemmulesderived from the cells which come next in duesuccession. If this be admitted we have a suffi-

cient answer to the above two objections.Anyhow, pangenesis seems to throw a conside-rable amount of light on the wonderful powerof regrowth.

It follows, also, from the view just given, thatthe sexual elements differ from buds in notincluding nascent cells or gemmules in a so-mewhat advanced stage of development, sothat only the gemmules belonging to the ear-liest stages are first developed. As young ani-mals and those which stand low in the scalegenerally have a much greater capacity for re-growth than older and higher animals, it wouldalso appear that they retain cells in a nascentstate, or partially developed gemmules, morereadily than do animals which have alreadypassed through a long series of developmentalchanges. I may here add that although ovulescan be detected in most or all female animals atan extremely early age, there is no reason to

doubt that gemmules derived from parts modi-fied during maturity can pass into the ovules.

With respect to hybridism, pangenesis agreeswell with most of the ascertained facts. Wemust believe, as previously shown, that severalgemmules are requisite for the development ofeach cell or unit. But from the occurrence ofparthenogenesis, more especially from thosecases in which an embryo is only partially for-med, we may infer that the female element ge-nerally includes gemmules in nearly sufficientnumber for independent development, so thatwhen united with the male element the gem-mules are superabundant. Now, when two spe-cies or races are crossed reciprocally, the offs-pring do not commonly differ, and this showsthat the sexual elements agree in power, in ac-cordance with the view that both include thesame gemmules. Hybrids and mongrels arealso generally intermediate in character bet-ween the two parent-forms, yet occasionally

they closely resemble one parent in one partand the other parent in another part, or even intheir whole structure: nor is this difficult tounderstand on the admission that the gemmu-les in the fertilised germ are superabundant innumber, and that those derived from one pa-rent may have some advantage in number, af-finity, or vigour over those derived from theother parent. Crossed forms sometimes exhibitthe colour or other characters of either parent instripes or blotches; and this occurs in the firstgeneration, or through reversion in succeedingbud and seminal generations, of which factseveral instances were given in the eleventhchapter. In these cases we must follow Naudin(27/57. See his excellent discussion on this sub-ject in 'Nouvelles Archives du Museum' tome 1page 151.) and admit that the "essence" or "ele-ment" of the two species,—terms which Ishould translate into the gemmules,—have anaffinity for their own kind, and thus separatethemselves into distinct stripes or blotches; and

reasons were given, when discussing in thefifteenth chapter the incompatibility of certaincharacters to unite, for believing in such mutualaffinity. When two forms are crossed, one is notrarely found to be prepotent in the transmis-sion of its characters over the other; and this wecan explain by again assuming that the oneform has some advantage over the other in thenumber, vigour, or affinity of its gemmules. Insome cases, however, certain characters arepresent in the one form and latent in the other;for instance, there is a latent tendency in allpigeons to become blue, and, when a blue pi-geon is crossed with one of any other colour,the blue tint is generally prepotent. The expla-nation of this form of prepotency will be ob-vious when we come to the consideration ofReversion.

When two distinct species are crossed, it is no-torious that they do not yield the full or propernumber of offspring; and we can only say on

this head that, as the development of each or-ganism depends on such nicely-balanced affini-ties between a host of gemmules and nascentcells, we need not feel at all surprised that thecommixture of gemmules derived from twodistinct species should lead to partial or com-plete failure of development. With respect tothe sterility of hybrids produced from theunion of two distinct species, it was shown inthe nineteenth chapter that this depends exclu-sively on the reproductive organs being specia-lly affected; but why these organs should bethus affected we do not know, any more thanwhy unnatural conditions of life, though com-patible with health, should cause sterility; orwhy continued close interbreeding, or the ille-gitimate unions of heterostyled plants, inducethe same result. The conclusion that the repro-ductive organs alone are affected, and not thewhole organisation, agrees perfectly with theunimpaired or even increased capacity in hy-brid plants for propagation by buds; for this

implies, according to our hypothesis, that thecells of the hybrids throw off hybridised gem-mules, which become aggregated into buds, butfail to become aggregated within the reproduc-tive organs, so as to form the sexual elements.In a similar manner many plants, when placedunder unnatural conditions, fail to produceseed, but can readily be propagated by buds.We shall presently see that pangenesis agreeswell with the strong tendency to reversion ex-hibited by all crossed animals and plants.

Each organism reaches maturity through a lon-ger or shorter course of growth and develop-ment: the former term being confined to mereincrease of size, and development to changedstructure. The changes may be small and insen-sibly slow, as when a child grows into a man,or many, abrupt, and slight, as in the meta-morphoses of certain ephemerous insects, or,again, few and strongly- marked, as with mostother insects. Each newly formed part may be

moulded within a previously existing and co-rresponding part, and in this case it will ap-pear, falsely as I believe, to be developed fromthe old part; or it may be formed within a dis-tinct part of the body, as in the extreme cases ofmetagenesis. An eye, for instance, may be deve-loped at a spot where no eye previously exis-ted. We have also seen that allied organicbeings in the course of their metamorphosessometimes attain nearly the same structure af-ter passing through widely different forms; orconversely, after passing through nearly thesame early forms, arrive at widely differentmature forms. In these cases it is very difficultto accept the common view that the first-formed cells or units possess the inherent po-wer, independently of any external agency, ofproducing new structures wholly different inform, position, and function. But all these casesbecome plain on the hypothesis of pangenesis.The units, during each stage of development,throw off gemmules, which, multiplying, are

transmitted to the offspring. In the offspring, assoon as any particular cell or unit becomes par-tially developed, it unites with (or, to speakmetaphorically, is fertilised by) the gemmule ofthe next succeeding cell, and so onwards. Butorganisms have often been subjected to chan-ged conditions of life at a certain stage of theirdevelopment, and in consequence have beenslightly modified; and the gemmules cast offfrom such modified parts will tend to reprodu-ce parts modified in the same manner. Thisprocess may be repeated until the structure ofthe part becomes greatly changed at one parti-cular stage of development, but this will notnecessarily affect other parts, whether pre-viously or subsequently formed. In this mannerwe can understand the remarkable indepen-dence of structure in the successive metamorp-hoses, and especially in the successive metage-neses of many animals. In the case, however, ofdiseases which supervene during old age, sub-sequently to the ordinary period of procreation,

and which, nevertheless, are sometimes inheri-ted, as occurs with brain and heart complaints,we must suppose that the organs were affectedat an early age and threw off at this period af-fected gemmules; but that the affection becamevisible or injurious only after the prolongedgrowth, in the strict sense of the word, of thepart. In all the changes of structure which regu-larly supervene during old age, we probablysee the effects of deteriorated growth, and notof true development.

The principle of the independent formation ofeach part, owing to the union of the propergemmules with certain nascent cells, togetherwith the superabundance of the gemmules de-rived from both parents, and the subsequentself-multiplication of the gemmules, throwslight on a widely different group of facts, whichon any ordinary view of development appearsvery strange. I allude to organs which are ab-normally transposed or multiplied. For instan-

ce, a curious case has been recorded by Dr.Elliott Coues (27/58. 'Proc. Boston Soc. of Nat.Hist.' republished in 'Scientific Opinion' No-vember 10, 1869 page 488.) of a monstrouschicken with a perfect additional RIGHT legarticulated to the LEFT side of the pelvis. Gold-fish often have supernumerary fins placed onvarious parts of their bodies. When the tail of alizard is broken off, a double tail is sometimesreproduced; and when the foot of the salaman-der was divided longitudinally by Bonnet, ad-ditional digits were occasionally formed. Va-lentin injured the caudal extremity of an embr-yo, and three days afterwards it produced ru-diments of a double pelvis and of double hind-limbs. (27/59. Todd 'Cyclop. of Anat. andPhys.' volume 4 1849-52 page 975.) When frogs,toads, etc., are born with their limbs doubled,as sometimes happens, the doubling, as Ger-vais remarks (27/60. 'Compte Rendus' Novem-ber 14, 1865 page 800.), cannot be due to thecomplete fusion of two embryos, with the ex-

ception of the limbs, for the larvae are limbless.The same argument is applicable (27/61. Aspreviously remarked by Quatrefages in his 'Me-tamorphoses de l'Homme' etc. 1862 page 129.)to certain insects produced with multiple legsor antennae, for these are metamorphosed fromapodal or antennae-less larvae. Alphonse Mil-ne-Edwards (27/62. Gunther 'Zoological Re-cord' 1864 page 279.) has described the curiouscase of a crustacean in which one eye-pedunclesupported, instead of a complete eye, only animperfect cornea, and out of the centre of this aportion of an antenna was developed. A casehas been recorded (27/63. Sedgwick 'Medico-Chirurg. Review' April 1863 page 454.) of aman who had during both dentitions a doubletooth in place of the left second incisor, and heinherited this peculiarity from his paternalgrandfather. Several cases are known (27/64.Isid. Geoffroy Saint-Hilaire 'Hist. des Anoma-lies' tome 1 1832 pages 435, 657; and tome 2page 560.) of additional teeth having been de-

veloped in the orbit of the eye, and, more espe-cially with horses, in the palate. Hairs occasio-nally appear in strange situations, as "withinthe substance of the brain." (27/65. Virchow'Cellular Pathology' 1860 page 66.) Certainbreeds of sheep bear a whole crowd of horns ontheir foreheads. As many as five spurs havebeen seen on both legs of certain Game-fowls.In the Polish fowl the male is ornamented witha topknot of hackles like those on his neck,whilst the female has a top-knot formed ofcommon feathers. In feather- footed pigeonsand fowls, feathers like those on the wing arisefrom the outer side of the legs and toes. Eventhe elemental parts of the same feather may betransposed; for in the Sebastopol goose, barbu-les are developed on the divided filaments ofthe shaft. Imperfect nails sometimes appear onthe stumps of the amputated fingers of man(27/66. Muller 'Phys.' English Translation vo-lume 1 1833 page 407. A case of this kind haslately been communicated to me.) and it is an

interesting fact that with the snake-like Sau-rians, which present a series with more andmore imperfect limbs, the terminations of thephalanges first disappear, "the nails becomingtransferred to their proximal remnants, or evento parts which are not phalanges." (27/67. Dr.Furbringer 'Die Knochen etc. bei den schlange-nahnlichen Sauriern' as reviewed in 'Journal ofAnat. and Phys.' May 1870 page 286.)

Analogous cases are of such frequent occurren-ce with plants that they do not strike us withsufficient surprise. Supernumerary petals, sta-mens, and pistils, are often produced. I haveseen a leaflet low down in the compound leafof Vicia sativa replaced by a tendril; and a ten-dril possesses many peculiar properties, suchas spontaneous movement and irritability. Thecalyx sometimes assumes, either wholly or bystripes, the colour and texture of the corolla.Stamens are so frequently converted into pe-tals, more or less completely, that such cases

are passed over as not deserving notice; but aspetals have special functions to perform, name-ly, to protect the included organs, to attractinsects, and in not a few cases to guide theirentrance by well-adapted contrivances, we canhardly account for the conversion of stamensinto petals merely by unnatural or superfluousnourishment. Again, the edge of a petal mayoccasionally be found including one of the hig-hest products of the plant, namely, pollen; forinstance, I have seen the pollen-mass of anOphrys, which is a very complex structure,developed in the edge of an upper petal. Thesegments of the calyx of the common pea havebeen observed partially converted into carpels,including ovules, and with their tips convertedinto stigmas. Mr. Salter and Dr. Maxwell Mas-ters have found pollen within the ovules of thepassion-flower and of the rose. Buds may bedeveloped in the most unnatural positions, ason the petal of a flower. Numerous analogousfacts could be given. (27/68. Moquin-Tandon

'Teratologie Veg.' 1841 pages 218, 220, 353. Forthe case of the pea see 'Gardener's Chronicle'1866 page 897. With respect to pollen withinovules see Dr. Masters in 'Science Review' Oc-tober 1873 page 369. The Rev. J.M. Berkeleydescribes a bud developed on a petal of a Clar-kia in 'Gardener's Chronicle' April 28, 1866.)

I do not know how physiologists look at suchfacts as the foregoing. According to the doctri-ne of pangenesis, the gemmules of the transpo-sed organs become developed in the wrongplace, from uniting with wrong cells or aggre-gates of cells during their nascent state; and thiswould follow from a slight modification intheir elective affinities. Nor ought we to feelmuch surprise at the affinities of cells andgemmules varying, when we remember themany curious cases given in the seventeenthchapter, of plants which absolutely refuse to befertilised by their own pollen, though abundan-tly fertile with that of any other individual of

the same species, and in some cases only withthat of a distinct species. It is manifest that thesexual elective affinities of such plants—to usethe term employed by Gartner—have beenmodified. As the cells of adjoining or homolo-gous parts will have nearly the same nature,they will be particularly liable to acquire byvariation each other's elective affinities; and wecan thus understand to a certain extent suchcases as a crowd of horns on the heads of cer-tain sheep, of several spurs on the legs of fowls,hackle-like feathers on the heads of the males ofother fowls, and with the pigeon wing-likefeathers on their legs and membrane betweentheir toes, for the leg is the homologue of thewing. As all the organs of plants are homolo-gous and spring from a common axis, it is natu-ral that they should be eminently liable totransposition. It ought to be observed thatwhen any compound part, such as an additio-nal limb or an antenna, springs from a falseposition, it is only necessary that the few first

gemmules should be wrongly attached; for the-se whilst developing would attract other gem-mules in due succession, as in the regrowth ofan amputated limb. When parts which arehomologous and similar in structure, as thevertebrae of snakes or the stamens of polyan-drous flowers, etc., are repeated many times inthe same organism, closely allied gemmulesmust be extremely numerous, as well as thepoints to which they ought to become united;and, in accordance with the foregoing views,we can to a certain extent understand Isid.Geoffroy Saint-Hilaire's law, that parts, whichare already multiple, are extremely liable tovary in number.

Variability often depends, as I have attemptedto show, on the reproductive organs being inju-riously affected by changed conditions; and inthis case the gemmules derived from the va-rious parts of the body are probably aggregatedin an irregular manner, some superfluous and

others deficient. Whether a superabundance ofgemmules would lead to the increased size ofany part cannot be told; but we can see thattheir partial deficiency, without necessarilyleading to the entire abortion of the part, mightcause considerable modifications; for in thesame manner as plants, if their own pollen beexcluded, are easily hybridised, so, in the caseof cells, if the properly succeeding gemmuleswere absent, they would probably combineeasily with other and allied gemmules, as wehave just seen with transposed parts.

In variations caused by the direct action ofchanged conditions, of which several instanceshave been given, certain parts of the body aredirectly affected by the new conditions, andconsequently throw off modified gemmules,which are transmitted to the offspring. On anyordinary view it is unintelligible how changedconditions, whether acting on the embryo, theyoung or the adult, can cause inherited modifi-

cations. It is equally or even more unintelligibleon any ordinary view, how the effects of thelong-continued use or disuse of a part, or ofchanged habits of body or mind, can be inheri-ted. A more perplexing problem can hardly beproposed; but on our view we have only tosuppose that certain cells become at last struc-turally modified; and that these throw off simi-larly modified gemmules. This may occur atany period of development, and the modifica-tion will be inherited at a corresponding pe-riod; for the modified gemmules will unite inall ordinary cases with the proper precedingcells, and will consequently be developed at thesame period at which the modification firstarose. With respect to mental habits or instincts,we are so profoundly ignorant of the relationbetween the brain and the power of thoughtthat we do not know positively whether a fixedhabit induces any change in the nervous sys-tem, though this seems highly probable; butwhen such habit or other mental attribute, or

insanity, is inherited, we must believe that so-me actual modification is transmitted (27/69.See some remarks to this effect by Sir H.Holland in his 'Medical Notes' 1839 page 32.);and this implies, according to our hypothesis,that gemmules derived from modified nerve-cells are transmitted to the offspring.

It is generally necessary that an organismshould be exposed during several generationsto changed conditions or habits, in order thatany modification thus acquired should appearin the offspring. This may be partly due to thechanges not being at first marked enough tocatch attention, but this explanation is insuffi-cient; and I can account for the fact only by theassumption, which we shall see under the headof reversion is strongly supported, that gem-mules derived from each unmodified unit orpart are transmitted in large numbers to suc-cessive generations, and that the gemmulesderived from the same unit after it has been

modified go on multiplying under the samefavourable conditions which first caused themodification, until at last they become suffi-ciently numerous to overpower and supplantthe old gemmules.

A difficulty may be here noticed; we have seenthat there is an important difference in the fre-quency, though not in the nature, of the varia-tions in plants propagated by sexual andasexual generation. As far as variability de-pends on the imperfect action of the reproduc-tive organs under changed conditions, we canat once see why plants propagated asexuallyshould be far less variable than those propaga-ted sexually. With respect to the direct action ofchanged conditions, we know that organismsproduced from buds do not pass through theearlier phases of development; they will there-fore not be exposed, at that period of life whenstructure is most readily modified, to the va-rious causes inducing variability in the same

manner as are embryos and young larvalforms; but whether this is a sufficient explana-tion I know not.

With respect to variations due to reversion,there is a similar difference between plantspropagated from buds and seeds. Many varie-ties can be propagated securely by buds, butgenerally or invariably revert to their parent-forms by seed. So, also, hybridised plants canbe multiplied to any extent by buds, but arecontinually liable to reversion by seed,—that is,to the loss of their hybrid or intermediate cha-racter. I can offer no satisfactory explanation ofthese facts. Plants with variegated leaves,phloxes with striped flowers, barberries withseedless fruit, can all be securely propagated bybuds taken from the stem or branches; but budsfrom the roots of these plants almost invariablylose their character and revert to their formercondition. This latter fact is also inexplicable,unless buds developed from the roots are as

distinct from those on the stem, as is one budon the stem from another, and we know thatthese latter behave like independent organisms.

Finally, we see that on the hypothesis of pan-genesis variability depends on at least two dis-tinct groups of causes. Firstly, the deficiency,superabundance, and transposition of gemmu-les, and the redevelopment of those which havelong been dormant; the gemmules themselvesnot having undergone any modification; andsuch changes will amply account for much fluc-tuating variability. Secondly, the direct actionof changed conditions on the organisation, andof the increased use or disuse of parts; and inthis case the gemmules from the modified unitswill be themselves modified, and, when suffi-ciently multiplied, will supplant the old gem-mules and be developed into new structures.

Turning now to the laws of Inheritance. If wesuppose a homogeneous gelatinous protozoonto vary and assume a reddish colour, a minute

separated particle would naturally, as it grewto full size, retain the same colour; and weshould have the simplest form of inheritance.(27/70. This is the view taken by Prof. Hackelin his 'Generelle Morphologie' b. 2 s. 171, whosays: "Lediglich die partielle Identitat der speci-fisch constituirten Materie im elterlichen undim kindlichen Organismus, die Theilung dieserMaterie bei der Fortpflanzung, ist die Ursacheder Erblichkeit.") Precisely the same view maybe extended to the infinitely numerous anddiversified units of which the whole body ofone of the higher animals is composed; the se-parated particles being our gemmules. We havealready sufficiently discussed by implication,the important principle of inheritance at co-rresponding ages. Inheritance as limited by sexand by the season of the year (for instance withanimals becoming white in winter) is intelligi-ble if we may believe that the elective affinitiesof the units of the body are slightly different inthe two sexes, especially at maturity, and in one

or both sexes at different seasons, so that theyunite with different gemmules. It should beremembered that, in the discussion on the ab-normal transposition of organs, we have seenreason to believe that such elective affinities arereadily modified. But I shall soon have to recurto sexual and seasonal inheritance. These seve-ral laws are therefore explicable to a large ex-tent through pangenesis, and on no other hy-pothesis which has as yet been advanced.

But it appears at first sight a fatal objection toour hypothesis that a part or organ may be re-moved during several successive generations,and if the operation be not followed by disease,the lost part reappears in the offspring. Dogsand horses formerly had their tails docked du-ring many generations without any inheritedeffect; although, as we have seen, there is somereason to believe that the tailless condition ofcertain sheep-dogs is due to such inheritance.Circumcision has been practised by the Jews

from a remote period, and in most cases theeffects of the operation are not visible in theoffspring; though some maintain that an inheri-ted effect does occasionally appear. If inheri-tance depends on the presence of disseminatedgemmules derived from all the units of the bo-dy, why does not the amputation or mutilationof a part, especially if effected on both sexes,invariably affect the offspring? The answer inaccordance with our hypothesis probably isthat gemmules multiply and are transmittedduring a long series of generations—as we seein the reappearance of zebrine stripes on thehorse—in the reappearance of muscles andother structures in man which are proper to hislowly organised progenitors, and in many ot-her such cases. Therefore the long-continuedinheritance of a part which has been removedduring many generations is no real anomaly,for gemmules formerly derived from the partare multiplied and transmitted from generationto generation.

We have as yet spoken only of the removal ofparts, when not followed by morbid action: butwhen the operation is thus followed, it is cer-tain that the deficiency is sometimes inherited.In a former chapter instances were given, as ofa cow, the loss of whose horn was followed bysuppuration, and her calves were destitute of ahorn on the same side of their heads. But theevidence which admits of no doubt is that gi-ven by Brown-Sequard with respect to guinea-pigs, which after their sciatic nerves had beendivided, gnawed off their own gangrenoustoes, and the toes of their offspring were defi-cient in at least thirteen instances on the corres-ponding feet. The inheritance of the lost part inseveral of these cases is all the more remarkableas only one parent was affected; but we knowthat a congenital deficiency is often transmittedfrom one parent alone—for instance, the offs-pring of hornless cattle of either sex, whencrossed with perfect animals, are often horn-less. How, then, in accordance with our hypot-

hesis can we account for mutilations being so-metimes strongly inherited, if they are followedby diseased action? The answer probably is thatall the gemmules of the mutilated or amputatedpart are gradually attracted to the diseased sur-face during the reparative process, and are the-re destroyed by the morbid action.

A few words must be added on the completeabortion of organs. When a part becomes dimi-nished by disuse prolonged during many gene-rations, the principle of economy of growth,together with intercrossing, will tend to reduceit still further as previously explained, but thiswill not account for the complete or almostcomplete obliteration of, for instance, a minutepapilla of cellular tissue representing a pistil, orof a microscopically minute nodule of bonerepresenting a tooth. In certain cases of sup-pression not yet completed, in which a rudi-ment occasionally reappears through reversion,dispersed gemmules derived from this part

must, according to our view, still exist; we musttherefore suppose that the cells, in union withwhich the rudiment was formerly developed,fail in their affinity for such gemmules, exceptin the occasional cases of reversion. But whenthe abortion is complete and final, the gemmu-les themselves no doubt perish; nor is this inany way improbable, for, though a vast numberof active and long-dormant gemmules are nou-rished in each living creature, yet there must besome limit to their number; and it appears na-tural that gemmules derived from reduced anduseless parts would be more liable to perishthan those freshly derived from other partswhich are still in full functional activity.

The last subject that need be discussed, namely,Reversion, rests on the principle that transmis-sion and development, though generally actingin conjunction, are distinct powers; and thetransmission of gemmules with their subse-quent development shows us how this is possi-

ble. We plainly see the distinction in the manycases in which a grandfather transmits to hisgrandson, through his daughter, characterswhich she does not, or cannot, possess. Butbefore proceeding, it will be advisable to say afew words about latent or dormant characters.Most, or perhaps all, of the secondary charac-ters, which appertain to one sex, lie dormant inthe other sex; that is, gemmules capable of de-velopment into the secondary male sexual cha-racters are included within the female; andconversely female characters in the male: wehave evidence of this in certain masculine cha-racters, both corporeal and mental, appearingin the female, when her ovaria are diseased orwhen they fail to act from old age. In like man-ner female characters appear in castrated ma-les, as in the shape of the horns of the ox, and inthe absence of horns in castrated stags. Even aslight change in the conditions of life due toconfinement sometimes suffices to prevent thedevelopment of masculine characters in male

animals, although their reproductive organs arenot permanently injured. In the many cases inwhich masculine characters are periodicallyrenewed, these are latent at other seasons; in-heritance as limited by sex and season beinghere combined. Again, masculine charactersgenerally lie dormant in male animals untilthey arrive at the proper age for reproduction.The curious case formerly given of a Hen whichassumed the masculine characters, not of herown breed but of a remote progenitor, illustra-tes the close connection between latent sexualcharacters and ordinary reversion.

With those animals and plants which habituallyproduce several forms, as with certain butter-flies described by Mr. Wallace, in which threefemale forms and one male form co-exist, or, aswith the trimorphic species of Lythrum andOxalis, gemmules capable of reproducing thesedifferent forms must be latent in each indivi-dual.

Insects are occasionally produced with one sideor one quarter of their bodies like that of themale, with the other half or three-quarters likethat of the female. In such cases the two sidesare sometimes wonderfully different in structu-re, and are separated from each other by asharp line. As gemmules derived from everypart are present in each individual of bothsexes, it must be the elective affinities of thenascent cells which in these cases differ abnor-mally on the two sides of the body. Almost thesame principle comes into play with those ani-mals, for instance, certain gasteropods and Ve-rruca amongst cirripedes, which normally havethe two sides of the body constructed on a verydifferent plan; and yet a nearly equal numberof individuals have either side modified in thesame remarkable manner.

Reversion, in the ordinary sense of the word,acts so incessantly, that it evidently forms anessential part of the general law of inheritance.

It occurs with beings, however propagated,whether by buds or seminal generation, andsometimes may be observed with advancingage even in the same individual. The tendencyto reversion is often induced by a change ofconditions, and in the plainest manner by cros-sing. Crossed forms of the first generation aregenerally nearly intermediate in character bet-ween their two parents; but in the next genera-tion the offspring commonly revert to one orboth of their grandparents, and occasionally tomore remote ancestors. How can we accountfor these facts? Each unit in a hybrid mustthrow off, according to the doctrine of pange-nesis, an abundance of hybridised gemmules,for crossed plants can be readily and largelypropagated by buds; but by the same hypot-hesis dormant gemmules derived from bothpure parent-forms are likewise present; and asthese gemmules retain their normal condition,they would, it is probable, be enabled to multi-ply largely during the lifetime of each hybrid.

Consequently the sexual elements of a hybridwill include both pure and hybridised gemmu-les; and when two hybrids pair, the combina-tion of pure gemmules derived from the onehybrid with the pure gemmules of the sameparts derived from the other, would necessarilylead to complete reversion of character; and itis, perhaps, not too bold a supposition thatunmodified and undeteriorated gemmules ofthe same nature would be especially apt tocombine. Pure gemmules in combination withhybridised gemmules would lead to partialreversion. And lastly, hybridised gemmulesderived from both parent-hybrids would sim-ply reproduce the original hybrid form. (27/71.In these remarks I, in fact, follow Naudin, whospeaks of the elements or essences of the twospecies which are crossed. See his excellentmemoir in the 'Nouvelles Archives du Mu-seum' tome 1 page 151.) All these cases anddegrees of reversion incessantly occur.

It was shown in the fifteenth chapter that cer-tain characters are antagonistic to each other ordo not readily blend; hence, when two animalswith antagonistic characters are crossed, itmight well happen that a sufficiency of gem-mules in the male alone for the reproduction ofhis peculiar characters, and in the female alonefor the reproduction of her peculiar characters,would not be present; and in this case dormantgemmules derived from the same part in someremote progenitor might easily gain the ascen-dancy, and cause the reappearance of the long-lost character. For instance, when black andwhite pigeons, or black and white fowls, arecrossed,—colours which do not readilyblend,—blue plumage in the one case, eviden-tly derived from the rock-pigeon, and red plu-mage in the other case, derived from the wildjungle-cock, occasionally reappear. With un-crossed breeds the same result follows, underconditions which favour the multiplication anddevelopment of certain dormant gemmules, as

when animals become feral and revert to theirpristine character. A certain number of gemmu-les being requisite for the development of eachcharacter, as is known to be the case from seve-ral spermatozoa or pollen- grains being neces-sary for fertilisation, and time favouring theirmultiplication, will perhaps account for thecurious cases, insisted on by Mr. Sedgwick, ofcertain diseases which regularly appear in al-ternate generations. This likewise holds good,more or less strictly, with other weakly inheri-ted modifications. Hence, as I have heard itremarked, certain diseases appear to gainstrength by the intermission of a generation.The transmission of dormant gemmules duringmany successive generations is hardly in itselfmore improbable, as previously remarked, thanthe retention during many ages of rudimentaryorgans, or even only of a tendency to the pro-duction of a rudiment; but there is no reason tosuppose that dormant gemmules can be trans-mitted and propagated for ever. Excessively

minute and numerous as they are believed tobe, an infinite number derived, during a longcourse of modification and descent, from eachunit of each progenitor, could not be supportedor nourished by the organism. But it does notseem improbable that certain gemmules, underfavourable conditions, should be retained andgo on multiplying for a much longer periodthan others. Finally, on the view here given, wecertainly gain some insight into the wonderfulfact that the child may depart from the type ofboth its parents, and resemble its grandparents,or ancestors removed by many hundreds ofgenerations.

CONCLUSION.

The hypothesis of Pangenesis, as applied to theseveral great classes of facts just discussed, nodoubt is extremely complex, but so are thefacts. The chief assumption is that all the unitsof the body, besides having the universally

admitted power of growing by self-division,throw off minute gemmules which are disper-sed through the system. Nor can this assump-tion be considered as too bold, for we knowfrom the cases of graft-hybridisation that for-mative matter of some kind is present in thetissues of plants, which is capable of combiningwith that included in another individual, and ofreproducing every unit of the whole organism.But we have further to assume that the gemmu-les grow, multiply, and aggregate themselvesinto buds and the sexual elements; their deve-lopment depending on their union with othernascent cells or units. They are also believed tobe capable of transmission in a dormant state,like seeds in the ground, to successive genera-tions.

In a highly-organised animal, the gemmulesthrown off from each different unit throughoutthe body must be inconceivably numerous andminute. Each unit of each part, as it changes

during development, and we know that someinsects undergo at least twenty metamorp-hoses, must throw off its gemmules. But thesame cells may long continue to increase byself-division, and even become modified byabsorbing peculiar nutriment, without necessa-rily throwing off modified gemmules. All or-ganic beings, moreover, include many dormantgemmules derived from their grandparents andmore remote progenitors, but not from all theirprogenitors. These almost infinitely numerousand minute gemmules are contained withineach bud, ovule, spermatozoon, and pollen-grain. Such an admission will be declared im-possible; but number and size are only relativedifficulties. Independent organisms exist whichare barely visible under the highest powers ofthe microscope, and their germs must be exces-sively minute. Particles of infectious matter, sosmall as to be wafted by the wind or to adhereto smooth paper, will multiply so rapidly as toinfect within a short time the whole body of a

large animal. We should also reflect on the ad-mitted number and minuteness of the molecu-les composing a particle of ordinary matter.The difficulty, therefore, which at first appearsinsurmountable, of believing in the existence ofgemmules so numerous and small as they mustbe according to our hypothesis, has no greatweight.

The units of the body are generally admitted byphysiologists to be autonomous. I go one stepfurther and assume that they throw off repro-ductive gemmules. Thus an organism does notgenerate its kind as a whole, but each separateunit generates its kind. It has often been said bynaturalists that each cell of a plant has the po-tential capacity of reproducing the whole plant;but it has this power only in virtue of contai-ning gemmules derived from every part. Whena cell or unit is from some cause modified, thegemmules derived from it will be in like man-ner modified. If our hypothesis be provisiona-

lly accepted, we must look at all the forms ofasexual reproduction, whether occurring atmaturity or during youth, as fundamentally thesame, and dependent on the mutual aggrega-tion and multiplication of the gemmules. Theregrowth of an amputated limb and the healingof a wound is the same process partially carriedout. Buds apparently include nascent cells, be-longing to that stage of development at whichthe budding occurs, and these cells are ready tounite with the gemmules derived from the nextsucceeding cells. The sexual elements, on theother hand, do not include such nascent cells;and the male and female elements taken sepa-rately do not contain a sufficient number ofgemmules for independent development, ex-cept in the cases of parthenogenesis. The deve-lopment of each being, including all the formsof metamorphosis and metagenesis, dependson the presence of gemmules thrown off at eachperiod of life, and on their development, at acorresponding period, in union with preceding

cells. Such cells may be said to be fertilised bythe gemmules which come next in due order ofdevelopment. Thus the act of ordinary impreg-nation and the development of each part ineach being are closely analogous processes. Thechild, strictly speaking, does not grow into theman, but includes germs which slowly andsuccessively become developed and form theman. In the child, as well as in the adult, eachpart generates the same part. Inheritance mustbe looked at as merely a form of growth, likethe self- division of a lowly-organised unicellu-lar organism. Reversion depends on the trans-mission from the forefather to his descendantsof dormant gemmules, which occasionally be-come developed under certain known or unk-nown conditions. Each animal and plant maybe compared with a bed of soil full of seeds,some of which soon germinate, some lie dor-mant for a period, whilst others perish. Whenwe hear it said that a man carries in his consti-tution the seeds of an inherited disease, there is

much truth in the expression. No other attempt,as far as I am aware, has been made, imperfectas this confessedly is, to connect under onepoint of view these several grand classes offacts. An organic being is a microcosm—a littleuniverse, formed of a host of self-propagatingorganisms, inconceivably minute and nume-rous as the stars in heaven.

CHAPTER 2.XXVIII.

CONCLUDING REMARKS.

DOMESTICATION. NATURE AND CAU-SES OF VARIABILITY. SELECTION. DIVER-GENCE AND DISTINCTNESS OF CHARAC-TER. EXTINCTION OF RACES. CIRCUMS-TANCES FAVOURABLE TO SELECTION BYMAN. ANTIQUITY OF CERTAIN RACES. THEQUESTION WHETHER EACH PARTICULAR

VARIATION HAS BEEN SPECIALLYPREORDAINED.

As summaries have been added to nearly allthe chapters, and as, in the chapter on pangene-sis, various subjects, such as the forms of re-production, inheritance, reversion, the causesand laws of variability, etc., have been recentlydiscussed, I will here only make a few generalremarks on the more important conclusionswhich may be deduced from the multifariousdetails given throughout this work.

Savages in all parts of the world easily succeedin taming wild animals; and those inhabitingany country or island, when first visited byman, would probably have been still more easi-ly tamed. Complete subjugation generally de-pends on an animal being social in its habits,and on receiving man as the chief of the herd orfamily. In order that an animal should be do-mesticated it must be fertile under changedconditions of life, and this is far from being

always the case. An animal would not havebeen worth the labour of domestication, at leastduring early times, unless of service to man.From these circumstances the number of do-mesticated animals has never been large. Withrespect to plants, I have shown in the ninthchapter how their varied uses were probablyfirst discovered, and the early steps in theircultivation. Man could not have known, whenhe first domesticated an animal or plant, whet-her it would flourish and multiply when trans-ported to other countries, therefore he couldnot have been thus influenced in his choice. Wesee that the close adaptation of the reindeer andcamel to extremely cold and hot countries hasnot prevented their domestication. Still lesscould man have foreseen whether his animalsand plants would vary in succeeding genera-tions and thus give birth to new races; and thesmall capacity of variability in the goose hasnot prevented its domestication from a remoteepoch.

With extremely few exceptions, all animals andplants which have been long domesticatedhave varied greatly. It matters not under whatclimate, or for what purpose they are kept,whether as food for man or beast, for draughtor hunting, for clothing or mere pleasure,—under all these circumstances races have beenproduced which differ more from one anotherthan do the forms which in a state of nature areranked as different species. Why certain ani-mals and plants have varied more under do-mestication than others we do not know, anymore than why some are rendered more sterilethan others under changed conditions of life.But we have to judge of the amount of variationwhich our domestic productions have under-gone, chiefly by the number and amount ofdifference between the races which have beenformed, and we can often clearly see why manyand distinct races have not been formed, name-ly, because slight successive variations have notbeen steadily accumulated; and such variations

will never be accumulated if an animal or plantbe not closely observed, much valued, and keptin large numbers.

The fluctuating, and, as far as we can judge,never-ending variability of our domesticatedproductions,—the plasticity of almost theirwhole organisation,- -is one of the most impor-tant lessons which we learn from the numerousdetails given in the earlier chapters of thiswork. Yet domesticated animals and plants canhardly have been exposed to greater changes intheir conditions of life than have many naturalspecies during the incessant geological, geo-graphical, and climatal changes to which theworld has been subject; but domesticated pro-ductions will often have been exposed to moresudden changes and to less continuously uni-form conditions. As man has domesticated somany animals and plants belonging to widelydifferent classes, and as he certainly did notchoose with prophetic instinct those species

which would vary most, we may infer that allnatural species, if exposed to analogous condi-tions, would, on an average, vary to the samedegree. Few men at the present day will main-tain that animals and plants were created witha tendency to vary, which long remained dor-mant, in order that fanciers in after ages mightrear, for instance, curious breeds of the fowl,pigeon, or canary-bird.

From several causes it is difficult to judge of theamount of modification which our domesticproductions have undergone. In some cases theprimitive parent-stock has become extinct; or itcannot be recognised with certainty, owing toits supposed descendants having been so muchmodified. In other cases two or more closely-allied forms, after being domesticated, havecrossed; and then it is difficult to estimate howmuch of the character of the present descen-dants ought to be attributed to variation, andhow much to the influence of the several pa-

rent-stocks. But the degree to which our domes-ticated breeds have been modified by the cros-sing of distinct species has probably been muchexaggerated by some authors. A few indivi-duals of one form would seldom permanentlyaffect another form existing in greater numbers;for, without careful selection, the stain of theforeign blood would soon be obliterated, andduring early and barbarous times, when ouranimals were first domesticated, such carewould seldom have been taken.

There is good reason to believe in the case ofthe dog, ox, pig, and of some other animals,that several of our races are descended fromdistinct wild prototypes; nevertheless the beliefin the multiple origin of our domesticated ani-mals has been extended by some few natura-lists and by many breeders to an unauthorisedextent. Breeders refuse to look at the wholesubject under a single point of view; I haveheard it said by a man, who maintained that

our fowls were descended from at least half-a-dozen aboriginal species, that the evidence ofthe common origin of pigeons, ducks and rab-bits, was of no avail with respect to fowls.Breeders overlook the improbability of manyspecies having been domesticated at an earlyand barbarous period. They do not consider theimprobability of species having existed in astate of nature which, if they resembled ourpresent domestic breeds, would have beenhighly abnormal in comparison with all theircongeners. They maintain that certain species,which formerly existed, have become extinct, orare now unknown, although formerly known.The assumption of so much recent extinction isno difficulty in their eyes; for they do not judgeof its probability by the facility or difficulty ofthe extinction of other closely-allied wild forms.Lastly, they often ignore the whole subject ofgeographical distribution as completely as if itwere the result of chance.

Although from the reasons just assigned it isoften difficult to judge accurately of the amountof change which our domesticated productionshave undergone, yet this can be ascertained inthe cases in which all the breeds are known tobe descended from a single species,—as withthe pigeon, duck, rabbit, and almost certainlywith the fowl; and by the aid of analogy thiscan be judged of to a certain extent with do-mesticated animals descended from severalwild stocks. It is impossible to read the detailsgiven in the earlier chapters and in many pu-blished works, or to visit our various exhibi-tions, without being deeply impressed with theextreme variability of our domesticated animalsand cultivated plants. No part of the organisa-tion escapes the tendency to vary. The varia-tions generally affect parts of small vital orphysiological importance, but so it is with thedifferences which exist between closely-alliedspecies. In these unimportant characters thereis often a greater difference between the breeds

of the same species than between the naturalspecies of the same genus, as Isidore Geoffroyhas shown to be the case with size, and as isoften the case with the colour, texture, form,etc., of the hair, feathers, horns, and other der-mal appendages.

It has often been asserted that important partsnever vary under domestication, but this is acomplete error. Look at the skull of the pig inany one of the highly improved breeds, withthe occipital condyles and other parts greatlymodified; or look at that of the niata ox. Or,again, in the several breeds of the rabbit, obser-ve the elongated skull, with the differently sha-ped occipital foramen, atlas, and other cervicalvertebrae. The whole shape of the brain, toget-her with the skull, has been modified in Polishfowls; in other breeds of the fowl the number ofthe vertebrae and the forms of the cervical ver-tebrae have been changed. In certain pigeonsthe shape of the lower jaw, the relative length

of the tongue, the size of the nostrils and eye-lids, the number and shape of the ribs, the formand size of the oesophagus, have all varied. Incertain quadrupeds the length of the intestineshas been much increased or diminished. Withplants we see wonderful differences in the sto-nes of various fruits. In the Cucurbitaceae seve-ral highly important characters have varied,such as the sessile position of the stigmas onthe ovarium, the position of the carpels, and theprojection of the ovarium out of the receptacle.But it would be useless to run through the ma-ny facts given in the earlier chapters.

It is notorious how greatly the mental disposi-tion, tastes, habits, consensual movements, lo-quacity or silence, and tone of voice have va-ried and been inherited in our domesticatedanimals. The dog offers the most striking ins-tance of changed mental attributes, and thesedifferences cannot be accounted for by descentfrom distinct wild types.

New characters may appear and old ones di-sappear at any stage of development, beinginherited at a corresponding stage. We see thisin the difference between the eggs, the down onthe chickens and the first plumage of the va-rious breeds of the fowl; and still more plainlyin the differences between the caterpillars andcocoons of the various breeds of the silk-moth.These facts, simple as they appear, throw lighton the differences between the larval and adultstates of allied natural species, and on the who-le great subject of embryology. New charactersfirst appearing late in life are apt to becomeattached exclusively to that sex in which theyfirst arose, or they may be developed in a muchhigher degree in this than in the other sex; oragain, after having become attached to one sex,they may be transferred to the opposite sex.These facts, and more especially the circums-tance that new characters seem to be particular-ly liable, from some unknown cause, to becomeattached to the male sex, have an important

bearing on the acquirement of secondary sexualcharacters by animals in a state of nature.

It has sometimes been said that our domesticraces do not differ in constitutional peculiari-ties, but this cannot be maintained. In our im-proved cattle, pigs, etc., the period of maturity,including that of the second dentition, has beenmuch hastened. The period of gestation variesmuch, and has been modified in a fixed mannerin one or two cases. In some breeds of poultryand pigeons the period at which the down andthe first plumage are acquired, differs. Thenumber of moults through which the larvae ofsilk-moths pass, varies. The tendency to fatten,to yield much milk, to produce many young oreggs at a birth or during life, differs in differentbreeds. We find different degrees of adaptationto climate, and different tendencies to certaindiseases, to the attacks of parasites, and to theaction of certain vegetable poisons. With plants,adaptation to certain soils, the power of resis-

ting frost, the period of flowering and fruiting,the duration of life, the period of shedding theleaves or of retaining them throughout the win-ter, the proportion and nature of certain chemi-cal compounds in the tissues or seeds, all vary.

There is, however, one important constitutionaldifference between domestic races and species;I refer to the sterility which almost invariablyfollows, in a greater or less degree, when spe-cies are crossed, and to the perfect fertility ofthe most distinct domestic races, with the ex-ception of a very few plants, when similarlycrossed. It is certainly a most remarkable factthat many closely-allied species, which in ap-pearance differ extremely little, should yieldwhen crossed only a few more or less sterileoffspring, or none at all; whilst domestic raceswhich differ conspicuously from each other are,when united, remarkably fertile, and yield per-fectly fertile offspring. But this fact is not inreality so inexplicable as it at first appears. In

the first place, it was clearly shown in the nine-teenth chapter that the sterility of crossed spe-cies does not depend chiefly on differences intheir external structure or general constitution,but on differences in the reproductive system,analogous to those which cause the lessenedfertility of the illegitimate unions of dimorphicand trimorphic plants. In the second place, thePallasian doctrine, that species after havingbeen long domesticated lose their natural ten-dency to sterility when crossed, has beenshown to be highly probable or almost certain.We cannot avoid this conclusion when we re-flect on the parentage and present fertility ofthe several breeds of the dog, of the Indian orhumped and European cattle, and of the twochief kinds of pigs. Hence it would be unreaso-nable to expect that races formed under domes-tication should acquire sterility when crossed,whilst at the same time we admit that domesti-cation eliminates the normal sterility of crossedspecies. Why with closely-allied species their

reproductive systems should almost invariablyhave been modified in so peculiar a manner asto be mutually incapable of acting on each ot-her—though in unequal degrees in the twosexes, as shown by the difference in fertilitybetween reciprocal crosses of the same spe-cies—we do not know, but may with muchprobability infer the cause to be as follows.Most natural species have been habituated tonearly uniform conditions of life for an incom-parably longer time than have domestic races;and we positively know that changed condi-tions exert an especial and powerful influenceon the reproductive system. Hence this diffe-rence may well account for the difference in thepower of reproduction between domestic raceswhen crossed and species when crossed. It isprobably in chief part owing to the same causethat domestic races can be suddenly transpor-ted from one climate to another, or placed un-der widely different conditions, and yet retainin most cases their fertility unimpaired; whilst a

multitude of species subjected to lesser changesare rendered incapable of breeding.

The offspring of crossed domestic races and ofcrossed species resemble each other in mostrespects, with the one important exception offertility; they often partake in the same unequaldegree of the characters of their parents, one ofwhich is often prepotent over the other; andthey are liable to reversion of the same kind. Bysuccessive crosses one species may be made toabsorb completely another, and so it notorious-ly is with races. The latter resemble species inmany other ways. They sometimes inherit theirnewly- acquired characters almost or even qui-te as firmly as species. The conditions leadingto variability and the laws governing its natureappear to be the same in both. Varieties can beclassed in groups under groups, like speciesunder genera, and these under families andorders; and the classification may be either arti-ficial,—that is, founded on any arbitrary cha-

racter,—or natural. With varieties a naturalclassification is certainly founded, and withspecies is apparently founded, on communityof descent, together with the amount of modifi-cation which the forms have undergone. Thecharacters by which domestic varieties differfrom one another are more variable than thosedistinguishing species, though hardly more sothan with certain polymorphic species; but thisgreater degree of variability is not surprising,as varieties have generally been exposed withinrecent times to fluctuating conditions of life,and are much more liable to have been crossed;they are also in many cases still undergoing, orhave recently undergone, modification byman's methodical or unconscious selection.

Domestic varieties as a general rule certainlydiffer from one another in less important partsthan do species; and when important differen-ces occur, they are seldom firmly fixed; but thisfact is intelligible, if we consider man's method

of selection. In the living animal or plant hecannot observe internal modifications in themore important organs; nor does he regardthem as long as they are compatible with healthand life. What does the breeder care about anyslight change in the molar teeth of his pigs, orfor an additional molar tooth in the dog; or forany change in the intestinal canal or other in-ternal organ? The breeder cares for the flesh ofhis cattle being well marbled with fat, and foran accumulation of fat within the abdomen ofhis sheep, and this he has effected. What wouldthe floriculturist care for any change in thestructure of the ovarium or of the ovules? Asimportant internal organs are certainly liable tonumerous slight variations, and as these wouldprobably be transmitted, for many strangemonstrosities are inherited, man could undoub-tedly effect a certain amount of change in theseorgans. When he has produced any modifica-tion in an important part, he has generally doneso unintentionally, in correlation with some

other conspicuous part. For instance, he hasgiven ridges and protuberances to the skulls offowls, by attending to the form of the comb, orto the plume of feathers on the head. By atten-ding to the external form of the pouter-pigeon,he has enormously increased the size of theoesophagus, and has added to the number ofthe ribs, and given them greater breadth. Withthe carrier-pigeon, by increasing through stea-dy selection the wattles on the upper mandible,he has greatly modified the form of the lowermandible; and so in many other cases. Naturalspecies, on the other hand, have been modifiedexclusively for their own good, to fit them forinfinitely diversified conditions of life, to avoidenemies of all kinds, and to struggle against ahost of competitors. Hence, under such com-plex conditions, it would often happen thatmodifications of the most varied kinds, in im-portant as well as in unimportant parts, wouldbe advantageous or even necessary; and theywould slowly but surely be acquired through

the survival of the fittest. Still more importantis the fact that various indirect modificationswould likewise arise through the law of corre-lated variation.

Domestic breeds often have an abnormal orsemi-monstrous character, as amongst dogs,the Italian greyhound, bulldog, Blenheim spa-niel, and bloodhound,—some breeds of cattleand pigs,—several breeds of the fowl,—and thechief breeds of the pigeon. In such abnormalbreeds, parts which differ but slightly or not atall in the allied natural species, have been grea-tly modified. This may be accounted for byman's often selecting, especially at first, conspi-cuous and semi-monstrous deviations of struc-ture. We should, however, be cautious in deci-ding what deviations ought to be called mons-trous: there can hardly be a doubt that, if thebrush of horse-like hair on the breast of theturkey-cock had first appeared in the domesti-cated bird, it would have been considered as a

monstrosity; the great plume of feathers on thehead of the Polish cock has been thus designa-ted, though plumes are common on the headsof many kinds of birds; we might call the wattleor corrugated skin round the base of the beakof the English carrier-pigeon a monstrosity, butwe do not thus speak of the globular fleshyexcrescence at the base of the beak of the Car-pophaga oceanica.

Some authors have drawn a wide distinctionbetween artificial and natural breeds; althoughin extreme cases the distinction is plain, in ma-ny other cases it is arbitrary; the difference de-pending chiefly on the kind of selection whichhas been applied. Artificial breeds are thosewhich have been intentionally improved byman; they frequently have an unnatural appea-rance, and are especially liable to lose their cha-racters through reversion and continued varia-bility. The so-called natural breeds, on the otherhand, are those which are found in semi-

civilised countries, and which formerly inhabi-ted separate districts in nearly all the Europeankingdoms. They have been rarely acted on byman's intentional selection; more frequently byunconscious selection, and partly by naturalselection, for animals kept in semi-civilisedcountries have to provide largely for their ownwants. Such natural breeds will also have beendirectly acted on by the differences, thoughslight, in the surrounding conditions.

There is a much more important distinctionbetween our several breeds, namely, in somehaving originated from a strongly-marked orsemi-monstrous deviation of structure, which,however, may subsequently have been aug-mented by selection; whilst others have beenformed in so slow and insensible a manner, thatif we could see their early progenitors weshould hardly be able to say when or how thebreed first arose. From the history of the race-horse, greyhound, gamecock, etc., and from

their general appearance, we may feel nearlyconfident that they were formed by a slow pro-cess of improvement; and we know that thishas been the case with the carrier-pigeon, aswell as with some other pigeons. On the otherhand, it is certain that the ancon and mau-champ breeds of sheep, and almost certain thatthe niata cattle, turnspit, and pug-dogs, jumperand frizzled fowls, short-faced tumbler pi-geons, hook- billed ducks, etc., suddenly ap-peared in nearly the same state as we now seethem. So it has been with many cultivatedplants. The frequency of these cases is likely tolead to the false belief that natural species haveoften originated in the same abrupt manner.But we have no evidence of the appearance, orat least of the continued procreation, undernature, of abrupt modifications of structure;and various general reasons could be assignedagainst such a belief.

On the other hand, we have abundant evidenceof the constant occurrence under nature ofslight individual differences of the most diver-sified kinds; and we are thus led to concludethat species have generally originated by thenatural selection of extremely slight differences.This process may be strictly compared with theslow and gradual improvement of the racehor-se, greyhound, and gamecock. As every detailof structure in each species has to be closelyadapted to its habits of life, it will rarely hap-pen that one part alone will be modified; but, aswas formerly shown, the co-adapted modifica-tions need not be absolutely simultaneous. Ma-ny variations, however, are from the first con-nected by the law of correlation. Hence it fo-llows that even closely-allied species rarely ornever differ from one another by one characteralone; and the same remark is to a certain ex-tent applicable to domestic races; for these, ifthey differ much, generally differ in many res-pects.

Some naturalists boldly insist (28/1. Godron'De l'Espece' 1859 tome 2 page 44 etc.) that spe-cies are absolutely distinct productions, neverpassing by intermediate links into one another;whilst they maintain that domestic varieties canalways be connected either with one another orwith their parent-forms. But if we could alwaysfind the links between the several breeds of thedog, horse, cattle, sheep, pigs, etc., there wouldnot have been such incessant doubts whetherthey were descended from one or several spe-cies. The greyhound genus, if such a term maybe used, cannot be closely connected with anyother breed, unless, perhaps, we go back to theancient Egyptian monuments. Our Englishbulldog also forms a very distinct breed. In allthese cases crossed breeds must of course beexcluded, for distinct natural species can thusbe likewise connected. By what links can theCochin fowl be closely united with others? Bysearching for breeds still preserved in distantlands, and by going back to historical records,

tumbler-pigeons, carriers, and barbs can beclosely connected with the parent rock-pigeon;but we cannot thus connect the turbit or thepouter. The degree of distinctness between thevarious domestic breeds depends on theamount of modification which they have un-dergone, and more especially on the neglectand final extinction of intermediate and less-valued forms.

It has often been argued that no light is thrownon the changes which natural species are belie-ved to undergo from the admitted changes ofdomestic races, as the latter are said to be meretemporary productions, always reverting, assoon as they become feral, to their pristineform. This argument has been well combatedby Mr. Wallace (28/2. 'Journal Proc. Linn. Soc.'1858 volume 3 page 60.) and full details weregiven in the thirteenth chapter, showing thatthe tendency to reversion in feral animals andplants has been greatly exaggerated, though no

doubt it exists to a certain extent. It would beopposed to all the principles inculcated in thiswork, if domestic animals, when exposed tonew conditions and compelled to struggle fortheir own wants against a host of foreign com-petitors, were not modified in the course oftime. It should also be remembered that manycharacters lie latent in all organic beings, readyto be evolved under fitting conditions; and inbreeds modified within recent times, the ten-dency to reversion is particularly strong. Butthe antiquity of some of our breeds clearly pro-ves that they remain nearly constant as long astheir conditions of life remain the same.

It has been boldly maintained by some authorsthat the amount of variation to which our do-mestic productions are liable is strictly limited;but this is an assertion resting on little eviden-ce. Whether or not the amount of change in anyparticular direction is limited, the tendency togeneral variability is, as far as we can judge,

unlimited. Cattle, sheep, and pigs have variedunder domestication from the remotest period,as shown by the researches of Rutimeyer andothers; yet these animals have been improvedto an unparalleled degree, within quite recenttimes, and this implies continued variability ofstructure. Wheat, as we know from the remainsfound in the Swiss lake- dwellings, is one of themost anciently cultivated plants, yet at the pre-sent day new and better varieties frequentlyarise. It may be that an ox will never be produ-ced of larger size and finer proportions, or aracehorse fleeter, than our present animals, or agooseberry larger than the London variety; buthe would be a bold man who would assert thatthe extreme limit in these respects has beenfinally attained. With flowers and fruit it hasrepeatedly been asserted that perfection hasbeen reached, but the standard has soon beenexcelled. A breed of pigeons may never be pro-duced with a beak shorter than that of the pre-sent short-faced tumbler, or with one longer

than that of the English carrier, for these birdshave weak constitutions and are bad breeders;but shortness and length of beak are the pointswhich have been steadily improved during thelast 150 years, and some of the best judges denythat the goal has yet been reached. From rea-sons which could be assigned, it is probablethat parts which have now reached theirmaximum development, might, after remainingconstant during a long period, vary again in thedirection of increase under new conditions oflife. But there must be, as Mr. Wallace has re-marked with much truth (28/3. 'The QuarterlyJournal of Science' October 1867 page 486.), alimit to change in certain directions both withnatural and domestic productions; for instance,there must be a limit to the fleetness of any te-rrestrial animal, as this will be determined bythe friction to be overcome, the weight to becarried, and the power of contraction in themuscular fibres. The English racehorse mayhave reached this limit; but it already surpasses

in fleetness its own wild progenitor and all ot-her equine species. The short-faced tumbler-pigeon has a beak shorter, and the carrier abeak longer, relatively to the size of their bo-dies, than that of any natural species of the fa-mily. Our apples, pears and gooseberries bearlarger fruit than those of any natural species ofthe same genera; and so in many other cases.

It is not surprising, seeing the great differencebetween many domestic breeds, that some fewnaturalists have concluded that each is descen-ded from a distinct aboriginal stock, more espe-cially as the principle of selection has been ig-nored, and the high antiquity of man, as a bree-der of animals, has only recently becomeknown. Most naturalists, however, freely admitthat our various breeds, however dissimilar,are descended from a single stock, althoughthey do not know much about the art of bree-ding, cannot show the connecting links, nor saywhere and when the breeds arose. Yet these

same naturalists declare, with an air of philo-sophical caution, that they will never admitthat one natural species has given birth toanother until they behold all the transitionalsteps. Fanciers use exactly the same languagewith respect to domestic breeds; thus, an aut-hor of an excellent treatise on pigeons says hewill never allow that the carrier and fantail arethe descendants of the wild rock-pigeon, untilthe transitions have "actually been observed,and can be repeated whenever man chooses toset about the task." No doubt it is difficult torealise that slight changes added up duringlong centuries can produce such great results;but he who wishes to understand the origin ofdomestic breeds or of natural species mustovercome this difficulty.

The causes which excite and the laws whichgovern variability have been discussed so late-ly, that I need here only enumerate the leadingpoints. As domesticated organisms are much

more liable to slight deviations of structure andto monstrosities than species living under theirnatural conditions, and as widely-ranging spe-cies generally vary more than those which in-habit restricted areas, we may infer that varia-bility mainly depends on changed conditions oflife. We must not overlook the effects of theunequal combination of the characters derivedfrom both parents, or reversion to former pro-genitors. Changed conditions have an especialtendency to render the reproductive organsmore or less impotent, as shown in the chapterdevoted to this subject; and these organs con-sequently often fail to transmit faithfully theparental characters. Changed conditions alsoact directly and definitely on the organisation,so that all or nearly all the individuals of thesame species thus exposed become modified inthe same manner; but why this or that part isespecially affected we can seldom or ever say.In most cases, however, a change in the condi-tions seems to act indefinitely, causing diversi-

fied variations in nearly the same manner asexposure to cold or the absorption of the samepoison affects different individuals in differentways. We have reason to suspect that an habi-tual excess of highly-nutritious food, or an ex-cess relatively to the wear and tear of the orga-nisation from exercise, is a powerful excitingcause of variability. When we see the symme-trical and complex outgrowths, caused by aminute drop of the poison of a gall-insect, wemay believe that slight changes in the chemicalnature of the sap or blood would lead to ex-traordinary modifications of structure.

The increased use of a muscle with its variousattached parts, and the increased activity of agland or other organ, lead to their increaseddevelopment. Disuse has a contrary effect. Withdomesticated productions, although their or-gans sometimes become rudimentary throughabortion, we have no reason to suppose thatthis has ever followed solely from disuse. With

natural species, on the contrary, many organsappear to have been rendered rudimentarythrough disuse, aided by the principle of theeconomy of growth together with intercrossing.Complete abortion can be accounted for onlyby the hypothesis given in the last chapter, na-mely, the final destruction of the germs orgemmules of useless parts. This difference bet-ween species and domestic varieties may bepartly accounted for by disuse having acted onthe latter for an insufficient length of time, andpartly from their exemption from any severestruggle for existence entailing rigid economyin the development of each part, to which allspecies under nature are subjected. Nevert-heless the law of compensation or balancement,which likewise depends on the economy ofgrowth, apparently has affected to a certainextent our domesticated productions.

As almost every part of the organisation beco-mes highly variable under domestication, and

as variations are easily selected both conscious-ly and unconsciously, it is very difficult to dis-tinguish between the effects of the selection ofindefinite variations and the direct action of theconditions of life. For instance, it is possiblethat the feet of our water-dogs and of the Ame-rican dogs which have to travel much over thesnow, may have become partially webbed fromthe stimulus of widely extending their toes; butit is more probable that the webbing, like themembrane between the toes of certain pigeons,spontaneously appeared and was afterwardsincreased by the best swimmers and the bestsnow-travellers being preserved during manygenerations. A fancier who wished to decreasethe size of his bantams or tumbler-pigeonswould never think of starving them, but wouldselect the smallest individuals which sponta-neously appeared. Quadrupeds are sometimesborn destitute of hair and hairless breeds havebeen formed, but there is no reason to believethat this is caused by a hot climate. Within the

tropics heat often causes sheep to lose theirfleeces; on the other hand, wet and cold act as adirect stimulus to the growth of hair; but whowill pretend to decide how far the thick fur ofarctic animals, or their white colour, is due tothe direct action of a severe climate, and howfar to the preservation of the best-protectedindividuals during a long succession of genera-tions?

Of all the laws governing variability, that ofcorrelation is one of the most important. In ma-ny cases of slight deviations of structure as wellas of grave monstrosities, we cannot even con-jecture what is the nature of the bond of con-nexion. But between homologous parts—between the fore and hind limbs— between thehair, hoofs, horns, and teeth—which are closelysimilar during their early development andwhich are exposed to similar conditions, wecan see that they would be eminently liable tobe modified in the same manner. Homologous

parts, from having the same nature, are apt toblend together, and, when many exist, to varyin number.

Although every variation is either directly orindirectly caused by some change in the su-rrounding conditions, we must never forgetthat the nature of the organisation which isacted on, is by far the more important factor inthe result. We see this in different organisms,which when placed under similar conditionsvary in a different manner, whilst closely-alliedorganisms under dissimilar conditions oftenvary in nearly the same manner. We see this, inthe same modification frequently reappearingin the same variety at long intervals of time,and likewise in the several striking cases givenof analogous or parallel variations. Althoughsome of these latter cases are due to reversion,others cannot thus be accounted for.

From the indirect action of changed conditionson the organisation, owing to the reproductive

organs being thus affected—from the directaction of such conditions, and these will causethe individuals of the same species either tovary in the same manner, or differently in ac-cordance with slight differences in their consti-tution—from the effects of the increased or de-creased use of parts—and from correlation,—the variability of our domesticated productionsis complicated to an extreme degree. The wholeorganisation becomes slightly plastic. Althougheach modification must have its own excitingcause, and though each is subjected to law, yetwe can so rarely trace the precise relation bet-ween cause and effect, that we are tempted tospeak of variations as if they arose sponta-neously. We may even call them accidental, butthis must be only in the sense in which we saythat a fragment of rock dropped from a heightowes its shape to accident.

It may be worth while briefly to consider theresult of the exposure to unnatural conditions

of a large number of animals of the same spe-cies and allowed to cross freely with no selec-tion of any kind, and afterwards to consider theresult when selection is brought into play. Letus suppose that 500 wild rock-pigeons wereconfined in their native land in an aviary andfed in the same manner as pigeons usually are;and that they were not allowed to increase innumber. As pigeons propagate so rapidly, Isuppose that a thousand or fifteen hundredbirds would have to be annually killed. Afterseveral generations had been thus reared, wemay feel sure that some of the young birdswould vary, and the variations would tend tobe inherited; for at the present day slight devia-tions of structure often occur and are inherited.It would be tedious even to enumerate the mul-titude of points which still go on varying orhave recently varied. Many variations wouldoccur in correlation with one another, as thelength of the wing and tail feathers—the num-ber of the primary wing-feathers, as well as the

number and breadth of the ribs, in correlationwith the size and form of the body—the num-ber of the scutellae with the size of the feet—thelength of the tongue with the length of thebeak—the size of the nostrils and eyelids andthe form of lower jaw in correlation with thedevelopment of wattle—the nakedness of theyoung with the future colour of the plumage—the size of the feet with that of the beak, andother such points. Lastly, as our birds are sup-posed to be confined in an aviary, they woulduse their wings and legs but little, and certainparts of the skeleton, such as the sternum, sca-pulae and feet, would in consequence becomeslightly reduced in size.

As in our assumed case many birds have to beindiscriminately killed every year, the chancesare against any new variety surviving longenough to breed. And as the variations whicharise are of an extremely diversified nature, thechances are very great against two birds pai-

ring which have varied in the same manner;nevertheless, a varying bird even when notthus paired would occasionally transmit itscharacter to its young; and these would notonly be exposed to the same conditions whichfirst caused the variation in question to appear,but would in addition inherit from their modi-fied parent a tendency again to vary in the sa-me manner. So that, if the conditions decidedlytended to induce some particular variation, allthe birds might in the course of time becomesimilarly modified. But a far commoner resultwould be, that one bird would vary in one wayand another bird in another way; one would beborn with a beak a little longer, and anotherwith a shorter beak; one would gain some blackfeathers, another some white or red feathers.And as these birds would be continually inter-crossing, the final result would be a body ofindividuals differing from each other in manyways, but only slightly; yet more than did theoriginal rock-pigeons. But there would not be

the least tendency towards the formation ofseveral distinct breeds.

If two separate lots of pigeons were treated inthe manner just described, one in England andthe other in a tropical country, the two lotsbeing supplied with different kinds of food,would they after many generations differ?When we reflect on the cases given in the twen-ty-third chapter, and on such facts as the diffe-rence in former times between the breeds ofcattle, sheep, etc., in almost every district ofEurope, we are strongly inclined to admit thatthe two lots would be differently modifiedthrough the influence of climate and food. Butthe evidence on the definite action of changedconditions is in most cases insufficient; and,with respect to pigeons, I have had the oppor-tunity of examining a large collection of domes-ticated kinds, sent to me by Sir W. Elliot fromIndia, and they varied in a remarkably similarmanner with our European birds.

If two distinct breeds were mingled together inequal numbers, there is reason to suspect thatthey would to a certain extent prefer pairingwith their own kind; but they would often in-tercross. From the greater vigour and fertility ofthe crossed offspring, the whole body would bythis means become interblended sooner thanwould otherwise have occurred. From certainbreeds being prepotent over others, it does notfollow that the interblended progeny would bestrictly intermediate in character. I have, also,proved that the act of crossing in itself gives astrong tendency to reversion, so that the cros-sed offspring would tend to revert to the stateof the aboriginal rock- pigeon; and in the cour-se of time they would probably be not muchmore heterogeneous in character than in ourfirst case, when birds of the same breed wereconfined together.

I have just said that the crossed offspringwould gain in vigour and fertility. From the

facts given in the seventeenth chapter there canbe no doubt of this fact; and there can be littledoubt, though the evidence on this head is notso easily acquired, that long-continued closeinterbreeding leads to evil results. With her-maphrodites of all kinds, if the sexual elementsof the same individual habitually acted on eachother, the closest possible interbreeding wouldbe perpetual. But we should bear in mind thatthe structure of all hermaphrodite animals, asfar as I can learn, permits and frequently neces-sitates a cross with a distinct individual. Withhermaphrodite plants we incessantly meet withelaborate and perfect contrivances for this sameend. It is no exaggeration to assert that, if theuse of the talons and tusks of a carnivorousanimal, or of the plumes and hooks on a seed,may be safely inferred from their structure, wemay with equal safety infer that many flowersare constructed for the express purpose of en-suring a cross with a distinct plant. From thesevarious considerations, not to mention the re-

sult of a long series of experiments which Ihave tried, the conclusion arrived at in thechapter just referred to—namely, that greatgood of some kind is derived from the sexualconcourse of distinct individuals—must beadmitted.

To return to our illustration: we have hithertoassumed that the birds were kept down to thesame number by indiscriminate slaughter; butif the least choice be permitted in their preser-vation, the whole result will be changed.Should the owner observe any slight variationin one of his birds, and wish to obtain a breedthus characterised, he would succeed in a sur-prisingly short time by careful selection. As anypart which has once varied generally goes onvarying in the same direction, it is easy, by con-tinually preserving the most strongly markedindividuals, to increase the amount of differen-ce up to a high, predetermined standard of ex-cellence. This is methodical selection.

If the owner of the aviary, without any thoughtof making a new breed, simply admired, forinstance, short-beaked more than long-beakedbirds, he would, when he had to reduce thenumber, generally kill the latter; and there canbe no doubt that he would thus in the course oftime sensibly modify his stock. It is improbable,if two men were to keep pigeons and act in thismanner, that they would prefer exactly the sa-me characters; they would, as we know, oftenprefer directly opposite characters, and the twolots would ultimately come to differ. This hasactually occurred with strains or families ofcattle, sheep, and pigeons, which have beenlong kept and carefully attended to by differentbreeders, without any wish on their part toform new and distinct sub-breeds. This uncons-cious kind of selection will more especially co-me into action with animals which are highlyserviceable to man; for every one tries to get thebest dogs, horses, cows, or sheep, without thin-king about their future progeny, yet these ani-

mals would transmit more or less surely theirgood qualities to their offspring. Nor is any oneso careless as to breed from his worst animals.Even savages, when compelled from extremewant to kill some of their animals, would des-troy the worst and preserve the best. With ani-mals kept for use and not for mere amusement,different fashions prevail in different districts,leading to the preservation, and consequentlyto the transmission, of all sorts of trifling pecu-liarities of character. The same process willhave been pursued with our fruit-trees andvegetables, for the best will always have beenthe most largely cultivated, and will occasiona-lly have yielded seedlings better than their pa-rents.

The different strains, just alluded to, whichhave been actually produced by breeders wit-hout any wish on their part to obtain such aresult, afford excellent evidence of the power ofunconscious selection. This form of selection

has probably led to far more important resultsthan methodical selection, and is likewise moreimportant under a theoretical point of viewfrom closely resembling natural selection. Forduring this process the best or most valuedindividuals are not separated and preventedfrom crossing with others of the same breed,but are simply preferred and preserved; yet thisinevitably leads to their gradual modificationand improvement; so that finally they prevail,to the exclusion of the old parent-form.

With our domesticated animals natural selec-tion checks the production of races with anyinjurious deviation of structure. In the case ofanimals which, from being kept by savages orsemi-civilised people, have to provide largelyfor their own wants under different circums-tances, natural selection will have played a mo-re important part. Hence it probably is thatthey often closely resemble natural species.

As there is no limit to man's desire to possessanimals and plants more and more useful inany respect, and as the fancier always wishes,owing to fashions running into extremes, toproduce each character more and more stron-gly pronounced, there is, through the prolon-ged action of methodical and unconscious se-lection, a constant tendency in every breed tobecome more and more different from its pa-rent-stock; and when several breeds have beenproduced and are valued for different qualities,to differ more and more from each other. Thisleads to Divergence of Character. As improvedsub-varieties and races are slowly formed, theolder and less improved breeds are neglectedand decrease in number. When few individualsof any breed exist within the same locality, clo-se interbreeding, by lessening their vigour andfertility, aids in their final extinction. Thus theintermediate links are lost, and the remainingbreeds gain in Distinctness of Character.

In the chapters on the Pigeon, it was proved byhistorical evidence and by the existence of con-necting sub-varieties in distant lands that seve-ral breeds have steadily diverged in character,and that many old and intermediate sub-breeds have been lost. Other cases could beadduced of the extinction of domestic breeds,as of the Irish wolf-dog, the old English hound,and of two breeds in France, one of which wasformerly highly valued. (28/4. M. Rufz de La-vison in 'Bull. Soc. Imp. d'Acclimat.' December1862 page 1009.) Mr. Pickering remarks (28/5.'Races of Man' 1850 page 315.) that "the sheepfigured on the most ancient Egyptian monu-ments is unknown at the present day; and atleast one variety of the bullock, formerlyknown in Egypt, has in like manner becomeextinct." So it has been with some animals andwith several plants cultivated by the ancientinhabitants of Europe during the neolithic pe-riod. In Peru, Von Tschudi (28/6. 'Travels inPeru' English translation page 177.) found in

certain tombs, apparently prior to the dynastyof the Incas, two kinds of maize not nowknown in the country. With our flowers andculinary vegetables, the production of new va-rieties and their extinction has incessantly recu-rred. At the present time improved breeds so-metimes displace older breeds at an extraordi-narily rapid rate; as has recently occurredthroughout England with pigs. The Longhorncattle in their native home were "suddenlyswept away as if by some murderous pestilen-ce," by the introduction of Shorthorns. (28/7.Youatt on 'Cattle' 1834 page 200. On Pigs see'Gardener's Chronicle' 1854 page 410.)

What grand results have followed from thelong-continued action of methodical and un-conscious selection, regulated to a certain ex-tent by natural selection, we see on every sideof us. Compare the many animals and plantswhich are displayed at our exhibitions withtheir parent-forms when these are known, or

consult old historical records with respect totheir former state. Most of our domesticatedanimals have given rise to numerous and dis-tinct races, but those which cannot be easilysubjected to selection must be excepted—suchas cats, the cochineal insect, and the hive-bee.In accordance with what we know of the pro-cess of selection, the formation of our manyraces has been slow and gradual. The man whofirst observed and preserved a pigeon with itsoesophagus a little enlarged, its beak a littlelonger, or its tail a little more expanded thanusual, never dreamed that he had made thefirst step in the creation of a pouter, carrier, andfantail-pigeon. Man can create not only anoma-lous breeds, but others having their wholestructure admirably co- ordinated for certainpurposes, such as the racehorse and dray-horse, or the greyhound and bulldog. It is by nomeans necessary that each small change ofstructure throughout the body, leading towardsexcellence, should simultaneously arise and be

selected. Although man seldom attends to dif-ferences in organs which are important under aphysiological point of view, yet he has so pro-foundly modified some breeds, that assuredly,if found wild, they would be ranked as distinctgenera.

The best proof of what selection has effected isperhaps afforded by the fact that whatever partor quality in any animal, and more especially inany plant, is most valued by man, that part orquality differs most in the several races. Thisresult is well seen by comparing the amount ofdifference between the fruits produced by theseveral varieties of fruit-trees, between the flo-wers of our flower-garden plants, between theseeds, roots, or leaves of our culinary and agri-cultural plants, in comparison with the otherand not valued parts of the same varieties. Stri-king evidence of a different kind is afforded bythe fact ascertained by Oswald Heer (28/8. 'DiePflanzen der Pfahlbauten' 1865.) namely, that

the seeds of a large number of plants,—wheat,barley, oats, peas, beans, lentils, poppies,—cultivated for their seed by the ancient Lake-inhabitants of Switzerland, were all smallerthan the seeds of our existing varieties. Ruti-meyer has shown that the sheep and cattlewhich were kept by the earlier Lake-inhabitantswere likewise smaller than our present breeds.In the middens of Denmark, the earliest dog ofwhich the remains have been found was theweakest; this was succeeded during the Bronzeage by a stronger kind, and this again duringthe Iron age by one still stronger. The sheep ofDenmark during the Bronze period had ex-traordinarily slender limbs, and the horse wassmaller than our present animal. (28/9. Morlot'Soc. Vaud. des Scien. Nat.' Mars 1860 page298.) No doubt in most of these cases the newand larger breeds were introduced from foreignlands by the immigration of new hordes ofmen. But it is not probable that each largerbreed, which in the course of time has supplan-

ted a previous and smaller breed, was the des-cendant of a distinct and larger species; it is farmore probable that the domestic races of ourvarious animals were gradually improved indifferent parts of the great Europaeo-Asiaticcontinent, and thence spread to other countries.This fact of the gradual increase in size of ourdomestic animals is all the more striking ascertain wild or half-wild animals, such as red-deer, aurochs, park-cattle, and boars (28/10.Rutimeyer 'Die Fauna der Pfahlbauten' 1861 s.30.) have within nearly the same period decrea-sed in size.

The conditions favourable to selection by manare,—the closest attention to every character,—long-continued perseverance,—facility in mat-ching or separating animals,—and especially alarge number being kept, so that the inferiorindividuals may be freely rejected or destroyed,and the better ones preserved. When many arekept there will also be a greater chance of the

occurrence of well-marked deviations of struc-ture. Length of time is all- important; for aseach character, in order to become stronglypronounced, has to be augmented by the selec-tion of successive variations of the same kind,this can be effected only during a long series ofgenerations. Length of time will, also, allowany new feature to become fixed by the conti-nued rejection of those individuals which re-vert or vary, and by the preservation of thosewhich still inherit the new character. Hence,although some few animals have varied rapidlyin certain respects under new conditions of life,as dogs in India and sheep in the West Indies,yet all the animals and plants which have pro-duced strongly marked races were domestica-ted at an extremely remote epoch, often beforethe dawn of history. As a consequence of this,no record has been preserved of the origin ofour chief domestic breeds. Even at the presentday new strains or sub-breeds are formed soslowly that their first appearance passes unno-

ticed. A man attends to some particular charac-ter, or merely matches his animals with unu-sual care, and after a time a slight difference isperceived by his neighbours;—the differencegoes on being augmented by unconscious andmethodical selection, until at last a new sub-breed is formed, receives a local name, andspreads; but by this time its history is almostforgotten. When the new breed has spread wi-dely, it gives rise to new strains and sub-breeds, and the best of these succeed andspread, supplanting other and older breeds;and so always onwards in the march of impro-vement.

When a well-marked breed has once been esta-blished, if not supplanted by still further im-proved sub-breeds, and if not exposed to grea-tly changed conditions of life inducing furthervariability or reversion to long-lost characters,it may apparently last for an enormous period.We may infer that this is the case from the high

antiquity of certain races; but some caution isnecessary on this head, for the same variationmay appear independently after long intervalsof time, or in distant places. We may safely as-sume that this has occurred with the turnspit-dog, of which one is figured on the ancientEgyptian monuments—with the solid-hoofedswine (28/11. Godron 'De l'Espece' tome 1 1859page 368.) mentioned by Aristotle—with five-toed fowls described by Columella—and cer-tainly with the nectarine. The dogs representedon the Egyptian monuments, about 2000 B.C.,show us that some of the chief breeds then exis-ted, but it is extremely doubtful whether anyare identically the same with our presentbreeds. A great mastiff sculptured on an Assy-rian tomb, 640 B.C., is said to be the same withthe dog still imported from Thibet into the sa-me region. The true greyhound existed duringthe Roman classical period. Coming down to alater period, we have seen that, though most ofthe chief breeds of the pigeon existed between

two and three centuries ago, they have not allretained exactly the same character to the pre-sent day; but this has occurred in certain casesin which no improvement was desired, for ins-tance, in the case of the Spot and Indianground-tumbler.

De Candolle (28/12. 'Geographie Botan.' 1855page 989.) has fully discussed the antiquity ofvarious races of plants; he states that the blackseeded poppy was known in the time ofHomer, the white-seeded sesamum by the an-cient Egyptians, and almonds with sweet andbitter kernels by the Hebrews; but it does notseem improbable that some of these varietiesmay have been lost and reappeared. One varie-ty of barley and apparently one of wheat, bothof which were cultivated at an immensely re-mote period by the Lake-inhabitants of Switzer-land, still exist. It is said (28/13. Pickering 'Ra-ces of Man' 1850 page 318.) that "specimens of asmall variety of gourd which is still common in

the market of Lima were exhumed from anancient cemetery in Peru." De Candolle re-marks that, in the books and drawings of thesixteenth century, the principal races of thecabbage, turnip, and gourd can be recognised:this might have been expected at so late a pe-riod, but whether any of these plants are abso-lutely identical with our present sub-varieties isnot certain. It is, however, said that the Brusselssprout, a variety which in some places is liableto degeneration, has remained genuine for mo-re than four centuries in the district where it isbelieved to have originated. (28/14. 'Journal ofa Horticultural Tour' by a Deputation of theCaledonian Hist. Soc. 1823 page 293.)

In accordance with the views maintained by mein this work and elsewhere, not only the vari-ous domestic races, but the most distinct generaand orders within the same great class—forinstance, mammals, birds, reptiles, and fishes—are all the descendants of one common pro-

genitor, and we must admit that the whole vastamount of difference between these forms hasprimarily arisen from simple variability. Toconsider the subject under this point of view isenough to strike one dumb with amazement.But our amazement ought to be lessened whenwe reflect that beings almost infinite in num-ber, during an almost infinite lapse of time,have often had their whole organisation ren-dered in some degree plastic, and that eachslight modification of structure which was inany way beneficial under excessively complexconditions of life has been preserved, whilsteach which was in any way injurious has beenrigorously destroyed. And the long-continuedaccumulation of beneficial variations will infal-libly have led to structures as diversified, asbeautifully adapted for various purposes andas excellently co-ordinated, as we see in theanimals and plants around us. Hence I havespoken of selection as the paramount power,whether applied by man to the formation of

domestic breeds, or by nature to the productionof species. I may recur to the metaphor given ina former chapter: if an architect were to rear anoble and commodious edifice, without the useof cut stone, by selecting from the fragments atthe base of a precipice wedge-formed stones forhis arches, elongated stones for his lintels, andflat stones for his roof, we should admire hisskill and regard him as the paramount power.Now, the fragments of stone, though indispen-sable to the architect, bear to the edifice built byhim the same relation which the fluctuatingvariations of organic beings bear to the variedand admirable structures ultimately acquiredby their modified descendants.

Some authors have declared that natural selec-tion explains nothing, unless the precise causeof each slight individual difference be madeclear. If it were explained to a savage utterlyignorant of the art of building, how the edificehad been raised stone upon stone, and why

wedge-formed fragments were used for thearches, flat stones for the roof, etc.; and if theuse of each part and of the whole building werepointed out, it would be unreasonable if hedeclared that nothing had been made clear tohim, because the precise cause of the shape ofeach fragment could not be told. But this is anearly parallel case with the objection that se-lection explains nothing, because we know notthe cause of each individual difference in thestructure of each being.

The shape of the fragments of stone at the baseof our precipice may be called accidental, butthis is not strictly correct; for the shape of eachdepends on a long sequence of events, all obey-ing natural laws; on the nature of the rock, onthe lines of deposition or cleavage, on the formof the mountain, which depends on its up-heaval and subsequent denudation, and lastlyon the storm or earthquake which throws downthe fragments. But in regard to the use to which

the fragments may be put, their shape may bestrictly said to be accidental. And here we areled to face a great difficulty, in alluding towhich I am aware that I am travelling beyondmy proper province. An omniscient Creatormust have foreseen every consequence whichresults from the laws imposed by Him. But canit be reasonably maintained that the Creatorintentionally ordered, if we use the words inany ordinary sense, that certain fragments ofrock should assume certain shapes so that thebuilder might erect his edifice? If the variouslaws which have determined the shape of eachfragment were not predetermined for thebuilder's sake, can it be maintained with anygreater probability that He specially ordainedfor the sake of the breeder each of the innumer-able variations in our domestic animals andplants;—many of these variations being of noservice to man, and not beneficial, far moreoften injurious, to the creatures themselves?Did He ordain that the crop and tail-feathers of

the pigeon should vary in order that the fanciermight make his grotesque pouter and fantailbreeds? Did He cause the frame and mentalqualities of the dog to vary in order that a breedmight be formed of indomitable ferocity, withjaws fitted to pin down the bull for man's bru-tal sport? But if we give up the principle in onecase,—if we do not admit that the variations ofthe primeval dog were intentionally guided inorder that the greyhound, for instance, thatperfect image of symmetry and vigour, mightbe formed,—no shadow of reason can be as-signed for the belief that variations, alike innature and the result of the same general laws,which have been the groundwork throughnatural selection of the formation of the mostperfectly adapted animals in the world, manincluded, were intentionally and specially gui-ded. However much we may wish it, we canhardly follow Professor Asa Gray in his belief"that variation has been led along certain bene-ficial lines," like a stream "along definite and

useful lines of irrigation." If we assume thateach particular variation was from the begin-ning of all time preordained, then that plastic-ity of organisation, which leads to many injuri-ous deviations of structure, as well as the re-dundant power of reproduction which inevita-bly leads to a struggle for existence, and, as aconsequence, to the natural selection or sur-vival of the fittest, must appear to us superflu-ous laws of nature. On the other hand, an om-nipotent and omniscient Creator ordains every-thing and foresees everything. Thus we arebrought face to face with a difficulty as insolu-ble as is that of free will and predestination.