Dendrocometes paradoxu (Stein)s . Part II.—Reproductio ...HISTORY or THE GENUS AND ANATOMY OF THE "ADULT" 341 IV. ... dish. Our method has been to accommodate some fourteen or fifteen

DKNDBOCOMETES PARADOXUS (STEIN). 337

Dendrocometes p a r a d o x u s (Stein).

Part II.—Reproduction (Bud-formation).

ByOeofTi-cy L.apage, M.Sc,

Late Assistant Lecturer in Zoology, Victoria Universityof Manchester,

Ana

J. T. Wadswortli,Research Assistant in Zoology, Victoria University of Manchester.

With Plates 25 and 26, and 16 Text-figures.

CONTENTS.PAGE

I. INTRODUCTION . . . . . 338

II. MATERIAL AND METHODS . . . . 3 3 8

III. HISTORY or THE GENUS AND ANATOMY OF THE " A D U L T " 341

IV. THE BUD . . . . . . 3 4 3(1) The Morphology and Habits of the Free Bud . 343(2) The Formation of the Bud . . . 3 4 7(3) Early Stages of Bud-Formation . . . 348(4) The Position of the Bud in the Parent . . 355(5) The Birth of the Bud . . . . 3 5 7(6) The Fate of the Bud . . . . 3(30(7) The Relation of Bud-Formation to the Size and

Maturity of the Parent . . . 363V. T H E N U C L E I . . . . . 3 6 4

(1) The Meganucleus . . . . 3 6 4(2) Division of the Meganucleus . . . 368(3) The Micronucleus . . . 3 7 3(4) Division of the Micronucleus . . . 375

VI . SUMMARY . . . . . . 3 7 8

VII. L I T E R A T U R E . . . . . 3 7 9

338 GKOFFREY LAPAGE AND J. T. WADSWOETH.

INTRODUCTION.

THE following account of reproduction by internal buddingin the Acinetarian species Dendrocometes p a r a d o x u s(Stein) embodies the results of observations made from timeto time in the zoological laboratories of the University ofManchester during the two years 1913-1914. In 1902Hickson and Wndsworth (5) published in this journal anaccount of the anatomy of the species, together with a fulldescription of the process of syugamy (conjugation). Thepresent paper is a continuation of that work. It was under-taken at the suggestion of Professor Hickson, who pointedout that a careful study of the nuclei of the Protozoa bymodern methods of technique cannot fail to produce usefulresults. Dendrocometes is a very favourable species forstudies of this kind, since the meganucleus and micronucleiare comparatively large, and the animal itself is easily obtainedand handled. Professor Hickson has very kindly placed atour disposal all the material that had been collected duringthe preparation of the paper referred to above. While theprocess of syngamy was being studied many observationswere also made upon bud-formation, and we have been ableto revise all the evidence so gained and to supplement it byfurther work upon fresh specimens. Throughout the workwe have been much indebted to Professor Hickson for hisconstant interest and help, and we are very glad to have thisopportunity of thanking him for criticisms which havematerially aided us in bringing tltis paper to a successfulconclusion.

MATERIAL AND METHODS.

Dendrocometes pa radoxus lives epizoically upon thegill plates of the Amphipod G a m m a r u s pu lex , and is ofvery common occurrence in that situation. Most of ourmaterial was taken from specimens of Grammar us pulexcollected from ditches and ponds around Northenden, nearManchester ; but we also used specimens from the ditches in

DT3NDR0C0METES PARABOXUS (sTEIN). 339

Delamere Forest, and from streams at Nantwich in Cheshire.One batch of G a m m a r u s , whose gills were particularly wellinfected, was very kindly collected for us by Mr. RaymondWilliamson, B.Sc, from the neighbourhood of Whaley Bridgein Derbyshire, and we have received material from othersources. G-ainmarus is not difficult to obtain, and a supply ofspecimens of D e n d r o c o m e t e s is, therefore, easily secured.The genus is widely distributed, having been recorded fromall parts of this country, from Europe, and from NorthAmerica (Hickson and Wadsworth, loc. c i t . ) . Owing to thiswide distribution, and to the ease with which it is obtainedand handled, since it is attached to such relatively largeobjects as the gill plates, it is a very favourable object forstudy,and maybe recommended as a suitableAcinetarian forclass work.

We have found that G-ammarus may be kept in a healthycondition for some months in ordinary earthenware pie dishes,provided that not too many of the animals are put into onedish. Our method has been to accommodate some fourteenor fifteen specimens in each dish, together with some of thesand and water from its native haunts, a'little fresh aquariumwater being also added. The dishes were then kept in a coolroom, ordinary laboratory temperature being usually unsuit-able, and periodically the water was oxygenated by theaddition of more fresh aquarium water. In this way we havebeen able to keep G a m m a r u s in a healthy state for longperiods.

In most of the specimens which were examined, at least afew D e n d r o c o m e t e s were found on each gill. Especiallyare they abundant on the two smallest gill plates. They mayeven occur here when all the other gills are free from infec-tion. Further, we have generally been able to observe somecases of bud-formation in each batch of gills examined, at anyrate during the summer months. In the winter months, budsare not so frequently seen. Whether this variation isseasonal, or related rather to the variations in the state of thefood-supply, we cannot yet say. We are inclined to believe

340 GEOFFREY LAPAGE AND J. T. WADSWORTH.

that the latter interpretation is the correct one. Often, whenthe G-arnmarus were freshly collected, no buds were foundin any of the Dendrocometes examined; but, after theGrammar us had lived in the dishes for a few days, budsbegan to appear. The reason for this is probably that theDendrocometes required a few days in which to accommo-date themselves to the new conditions of life. The multiplica-tion of the other organisms in the water, and the consequentincrease of food-supply probably also operated as an addi-tional stimulusto bud-formation.

The method we have employed for the preparation of livingindividuals for observation has been to place the hostGammarus in a watch-glass with a little pond-water, andto destroy the head with forceps. The legs are then detachedfrom the body, and the gill plates, which are attached to thethoracic legs, come away easily. They are generally set freeat the same time as the legs, and either float freely in thewater or sink to the bottom. They can then be taken up witha wide-mouthed pipette and transferred to a slide.

While low powers only are used it is wise to avoid the useof a cover-glass. If this can be avoided altogether so muchthe better. After the gill has been detached, Dendro-cometes is very sensitive to local conditions, and will quicklydie unless it is well supplied with oxygen audwi th waterfrom its native pond. The slide, when not in use, should,therefore, be kept in a moist chamber, and fresh pond-watershould be frequently added to it. In spite of elaborate pre-cautions and many experiments with apparatus designed tokeep up a constant, slow current through the water on theslide, we have never been able to keep the animnls in ahealthy state for longer than two days. Frequently they diedearlier than this. The reason is probably to be found in thefact that the gill itself very quickly shows signs of death anddegeneration of its tissues. Obviously it is impossible tokeep the gill in a natural healthy state when it has beendetached from the host. As we shall point out later(p. 363), Dendrocometes is peculiarly sensitive to any

DENDJROCOMETES PARADOXUS (STJEIJj). 341

change in the gill to which it is attached, and seems to reactdirectly and very quickly to any alteration in the health of its" host."

For observations of the formation of buds in the livingstate we usually covered the gill with a cover-glass, andfound the Zeiss Apoch. water immersion lens of great value,provided that efficient and critical illumination was employed.All the main features of the process can be seen, however,with the Zeiss D., the long working distance of which enablesone to use it without the intervention of a cover-slip.

All our observations on the living animal have been checkedand supplemented by the study of stained preparations. W ehave studied both whole mounts, prepared by immersing thewhole gill in corrosive-acetic or Schaudinu's fluid and subse-quently staining with brazilin, liEeinatoxylin, borax carmine,or alum carmine. The best results were obtained by the twofirst-named stains. In addition to these we have preparedserial sections by embedding the gill in the usual way. Thebest stain for such sections is undoubtedly iron liEematoxylin,but brazilin gives exceedingly good results. I t is anadvantage to stain the gills with borax-carmine or liEemii-toxylin before embedding.

Preparations of the free swimming bud are easily obtainedwith a little care. Being large enough to be seen under theordinary powers of a binocular microscope, they cau be easilyfollowed and caught in a fine pipette. They are then t rans-ferred toanalbumiuised slide, fixed, and v e r y c a r e f u l ly passedthrough, the stain and mounted. Extreme caution is neces-sni-y after fixation, both to avoid damaging these very delicaleobjects and to prevent their being carried away during thefrequent washings. W e found that they could be fixed wellby Hermann's fluid or by corrosive-acetic and stained withbrazilin or iron haeinatoxylin.

HISTORY OV THE GENUS AND ANATOMY OF THE " A D U L T . "

The history of the genus and the anatomy of the " a d u l t "

stage have beeu fully and accurately described by Hickson


and Wadsworfch (loc. cit.) and in the works of earlier authors.We have nothing to add to these observations; but in orderto facilitate a clear understanding of the process to be de-scribed, it is necessary to give here a brief outline of thegeneral structure and to consider the orientation of theanimal in relation to its surroundings.

The body of Dendrocotnetes is plano-convex in shape,the flat side being applied to the gill of the host. The wholebody is enclosed in a firm and well-defined cuticle, which iscontinued along the arms; but there is no theca or protectiveenvelope of any kind other than the cuticle. The absence ofa theca is correlated with the absence of a stalk or peduncle,such as exists, for example, in the genus Acineta and inmany other Acinetaria. The body of Dendrocometes is,therefore, directly applied to the surface of the gill. A basalplate or attaching disc of cement substance is secreted by theflat surface, and by means of this the animal is firmly fixedin position.

From the sides of the convexity of the body—that is tosay, from the sides of the dorsal surface (cf. p. 343)three or four, or even five arms project. Each arm is pro-vided with tentacles, and the whole apparatus serves tocapture, kill, and absorb the protoplasm of the prey, whichconsists of free-swimming organisms present in the surround-ing water. There is a single, large meganucleus and typicallythree micronuclei embedded in the finely granular cytoplasm.In addition, there are usually numerous large food-bodies andother granules of various kinds.

The question of the orientation of the animal is a difficultone, since none of the organs are present by which such aquestion can be decided. Further, the animal is sedentary,and therefore no evidence can be obtained from its behaviourwhile iu motion. The matter cannot be properly discusseduntil the anatomy and orientation of the bud have been con-sidered ; but we may refer here to the views of Collin (3) onthe subject. Collin, after a comparative study of the wholeclass Acinetaria, comes to the conclusion that the whole of

DEKTDROCOMKTES PARADOXUS (STEIN). 343

the convex surface of the " a d u l t " D e n d r o c o m e t e s isdorsal, while the flat surface, which is applied to the gill, isventral. The grounds for this view are adequately discussedby him (loc. c i t . ) , and we do not propose to consider themhere. Suffice it to say that our observations on the parentand on the movements and anatomy of the bud, together witha study of its method of settling down on the gill, all tend toconfirm his view.

Adopting the view that the convex surface is dorsal andthe plane surface ventral, it is possible to distinguish twomain axes (v. Text-fig. 1). The first is dorso-ventral in direc-tion, passing vertically through the middle of the dorsalsurface and through the ventral basal plate of attachment.This is the true morphological axis (Text-fig. 1, A.B.). Thesecond is perpendicular to this, and, therefore, roughlyparallel to the ventral plane surface and to the surface of thegill. I t corresponds to the physiological axis of the bud(Test-fig. 1, C.D.) (of. below, p. 344).

Reproduction in D e n d r o c o m e t e s is effected solely bygemmation. The absence of simple binary fission is not re-markable, since it is well known that this process is extremelyrare in Acinetaria. Fission in Acinetaria takes the form ofgemmation in the vast majority of cases. While, however,gemmation may be either internal or external in other genera,it is always internal in D e n d r o c o m e t e s . Further, thenumber o£ buds formed is never multiple. Always a single,approximately plano-convex bud is formed inside the body ofeach individual parent. This subsequently escapes to theexterior, and swims about actively for a time by means of thecilia with which it is provided. Ultimately, however, itsettles down upon a gill and, developing arms, assumes theform and habits characteristic of the parent.

We propose first to describe the anatomy of this bud andsubsequently to consider the method by which it is formed.

THE MORPHOLOGY AND HABITS OP THE FKEE BUD.

When it is seen under low magnifications only, the general

344 GE0FFRI3Y LAPAGE AND ,T. T. WADSWORTH.

shape of the bud may be compared with that of an ovalplano-convex lens. In other words, it possesses a broad,relatively flat surface upon which four rows of cilia areinserted; and an opposite surface, which is markedly convex,even in contour, and entirely free from cilia.1 The dimen-sions of the bud are 0-06 X 0'04 mm.

When it is released from the parent the bud either swimsfreely in the surrounding water or " creeps " over the surfaceof a gill by means of its cilia, very much in the manner ofa hypotrichous ciliate. In the latter case a well-mai-kedbilateral symmetry is evident. One end is constantly anterior

in progression, the other posterior; the convex surface isalways uppermost;, the flat surface, on which, the cilia areinserted, being applied to the gill. On the analogy of anybilaterally symmetrical animal moving forward along a sub-stratum, the convex surface may therefore be described asbeing physiologically dorsal, the flat surface physiologicallyventral. Two physiological axes are thus distinguishable.First, an antero-posterior axis passing through the anteriorand posterior poles in a plane parallel to the flat ventralsurface (Text-fig. 2, C.D.), and, secondly, a dorso-ventral axisperpendicular to this (Text-fig. 2, A.B.).

1 The bud may be described as being hypotrichous in the sense thatits cilia are entirely confined to the ventral surface. The use of theterm in this connection can only be misleading and is better restrictedto descriptions of ciliate morphology.

DENDEOOOME'i'ES PAEADOXUS (STEIN). 345

Since the bud subsequently settles down with its flatventral surface applied to the gill, the convex surface of theadult is to be regarded as dorsal.

Although by their position in relation to forward move-ment the anterior and posterior end of the bud are readilydistinguishable, there is very little difference in structurebetween the two. In a bud which has just emerged theposterior pole may be recognised by the presence on itsdorsal convex surface of a small papilla (Text-fig. 15), whichis the remnant of the stalk by which the bud was attachedto the parent at the moment of birth (cf. p. 359). Plate (8)and Oollin (loc. cit.) have both recorded this papilla, andthe latter also mentions an interruption of the cilia at theposterior pole. We have never been able to see any suchgap in the continuity of the ciliated bands. The presence ofthe papilla at the posterior pole proves, however, that thepole which leaves the parent last is subsequently posterior inprogression. In other words the embryo is born, so to speak,head first.

The dorsal surface is always entirely free from cilia. Thetuft of longer cilia, which constitutes the so-called " adoralzone " of some other Acinetarian embryos (e.g. T o k o p h r y acyclopum) is never present here. The dorsal surface isalvvays uniformly even in contour, and may be described asextending almost as far as the ciliated bands (i.e. the line C.D..Text-fig. 2). At this point a more or less prominent ridgemarks the ventral limit of the dorsal surface (PI. 26, fig. 12).Turning round this we are, so to speak, on the ventral sideof the body.

When the ventral surface is carefully examined, it is atonce obvious that it is quite incorrect to speak of it as beinga plane surface. The most that can be said is that it is muchflatter than the dorsal surface.

After studying it from various points of view we havecome to the conclusion that it presents two surfaces ofdifferent contours, namely, an outer area on which the ciliatedbands are inserted and which is slightly convex ventralwnrds,

346 GEOPMEY LAPAGE AND J. T. WADSWOETH.

aud a central area which is concave ventralwards. The wholebud is therefore better described as being like a biconvexlens from which the greater part of the convexity on one sidehas been obliquely cut away and a concavity substituted.Reference to Text-fig. 3, where the dotted line represents thecentral concave area, will elucidate this.

Taking first the outer portion, namely, the area uponwhich the ciliated bands lie, we find that its convexity isnever so marked as the convexity of the dorsal surface. Itsoutline forms part of the circumference of a much largercircle, and it is partly due to this fact that the ventral surfaceas a whole looks flat in comparison with the dorsal. Thepresence of the ridge between them (referred to on p. 345)also accentuates this,.appearance.

In addition to this, the surface extent of the convex areais greater behind than in front. In sagittal section, there-fore, the ventral surface has the appearance of an inverteddome the summit of which has been cut off obliquely, alonga line which is not parallel to the antero-posterior axis of thebud, but at an acute angle to it (v. Text-fig. 3E.F.). Further-more, th& ciliated bands themselves lie parallel to this linealso (E.F.), the innermost one mai-king the internal limit ofthis convex area of the ventral surface. It follows, therefore,that they are not parallel to the antero-posterior axis either,but are set obliquely to it.

Turning now to the central part of the ventral surface,namely, that part which is surrounded by and enclosed insidethe ciliated ridges, it is, as we have already pointed out,slightly concave, the concavity being directedventralwards.It' a bud is viewed, therefore, while lying on its dorsalconvex surface so that the ventral surface is seen en face, byfocussing downwards one sees first the i n n e r m o s t ciliated

DENDEOOOMETES PAEADOXUS (sTKIN). 347

ridge, and subsequently the other ciliated ridges and thebottom of the concavity in the centre.

The ciliated ridges themselves are quite simple in structure.Each projects slightly from the sui'face of the body, pre-senting a double outline when carefully focussed. The ridgesare always four in number—we have never seen more orfewer than this—and the cilia which they bear are insertedalong the free edges. As we have explained above, theridges lie obliquely to the plane of the antero-posterior axis,owing to the conformation of the convex area of the ventralsurface on which they lie. The outermost ridge, which liesvery near to the dorsal limit of the ventral surface, is,however, more nearly parallel to this plane than are theothers.

THE FORMATION OF THE BUD.

The method by which the bud is formed will be betterunderstood, perhaps, if a brief summary is given before theprocess is considered in detail.

Broadly speaking, it may be said the bud is, so to speak,cut out of the cytoplasm of the parent. At one point, nearerthe dorsal side, the cytoplasm of the parent loses its normalconsistency. This change shows itself by the appearance of asharply outlined, refractive line, which extends across thebody and subsequently turns ventralwards at each end, so asto assume the outline of a semicircle. This narrow, refractiveline is the so-called brood-chamber of previous authors. Itacquires an extension towards the surface of the body, meetingthe cuticle of the parent at a point where the bud will subse-quently emerge. No open communication, however, everexists between the brood-chamber and the exterior, except atthe moment of birth (cf. Text-fig. 5).

The outline of the brood-chamber encloses a dome-shapedmass of protoplasm, which is later organised into the buditself. On this dome-shaped mass four ciliated ridges appear,and a contractile vacuole appears-in it. The micronuclei ofthe parent, which are situated, usually in its ventral region,

348 GEOFFREY LAPAGE AND .). T. WADSWOJftTH.

divide by a primitive kind of mitosis, three of the daughtertnicronuclei passing into the bud, three remaining in theparent. The meganucleus, also situated in the ventral regionof the parent, divides, however, by amitosis, a very markeddumb-bell stage being seen, during which one head of thedumb-bell lies in the ventral region of the parent next to thegili, the other head in the bud. The stalk connecting thesetwo heads gradually elongates, becoming narrower at thesame time, and finally is severed while the bud is still withinthe parent. When all the orgnnellas of the bud are com-pletely formed, it is expelled very rapidly to the exteriorthrough a pore which appears at the point where the brood-chamber meets the cuticle of the parent.

EARLY STAGES OP BOD-FORMATION.

We have shown in Text-fig. 4 the eai-liest stage which wehave been able to recognise in the living animals. Theidentification of a bud in its eai'liest stages is natm-ally verydifficult, since the diagnosis must depend, so far as livingobservations are concerned, upon the presence or absence ofeither active cilia, or of a brood-chamber, or of two contractilevacuoles. The presence of any one of these features is sufficientto arouse suspicion, and they are given in the order of theirrelative reliability (but not in the order in which they appear).In the case figured, neither active cilia, nor a duplication ofthe contractile vacuole were present, but the individualshowed a faiut, refractive line running transversely across thebody, perpendicular to the dorsoventral axis (Text-fig. 4,BB. CH.). This line was kept under observation for somehours, and eventually our patience was rewarded by itsfurther development. The line extended further across thebody, and subsequently turned downwards towards theventral side of the animal, ending about half way betweenthe dorsal and ventral surfaces. At this stage, therefoi-e, theanimal showed a refractive line, now quite well marked,which had become semicircular in outline, the convexity of

DEXDROOOMETES PARADOXUS (STEIN). 349

the curve being directed away from the g i l l—i .e . , towards

the dorsal surface of the animal. Within, it was marked off a

dome-shaped mass of cytoplasm, which was destined subse-

quently to give l-ise to the bud itself. A little later on a

TEXT-FIG. 4.

Br.Ch.

C. F.j Contractile vacuole. Br. Gh. Brood-chamber. M. Mega-nucleus.

further development of this refractive area occurred. Atone side au extension of it, which we shall call the birthpassage (Text-fig. 5, B. PASS.), passed out towards the dorsalsurface of the animal, meeting the cuticle of the parent at apoint where the bud was subsequently expelled. At this

C.R, Ciliated ridges. C.V. Contractile vacuole of bud. C.V^.Ditto of parent. B. pass. Birth passage. B.pap. Birth papilla.M. Meganucleus.

point a small papilla appears, which we shall call the birthpapilla (Text-fig. 5, B. PAP., and see also p. 351, et seq.).

The explanation of the nature of this refractive line isa matter to which we have given careful attention. By allprevious authors it has been regarded as the expression of adefinite cavity, the so-called brood-chamber (Bruthohle,Biitschli: cavi te embryonnai re , Colliu), inside which

VOL, 6 1 , PART 3 . NEW SERIES. 2 3

350 GEOFFREY LAPAGE AND J. T. WADS WORTH.

the bud develops, and the extension of it to the surface at theside has been regarded as a .definite canal, of the nature of avagina, opening to the exterior by a definite pore. For thesake of convenience we shall continue to refer to it as the" brood-chamber " ; but to regard it as a well-defined cavityis, we believe, somewhat misleading. Possibly such an inter-pretation has arisen from an erroneous idea as to the way inwhich itis formed. Biitschll (1) andPlate (loc. cit.) have bothstated that the brood-chamber begins in Dendrocometes ,as it apparently does in Podop l i rya q u a d r i p a r t i t a andsome other Acinetaria, as an invagination at the surface atthe point where the bud is ultimately expelled. The birth -passage, in this view,is formed first, and the "cavity" insidewhich the bud develops is regarded as a later extensionof this. In our opinion, however, this view is incorrect. Thebrood-chamber, in all the cases observed by us, began as atransverse refractive area across the body of the parent,which extended as described above, and only when fullyformed, reached the surface of the animal. We believe,therefore, that the birth-passage is formed last, not first.

In this connection the views of Stein (10) are interesting.Describing the meganucleus oE an individual in which bud-formation was in progress, this author states that it was" surrounded by a light area which resulted from a dissolutionof the granules . . . around it." This " light area" corre-sponds with the refractive line described above. AlthoughStein does not seem to have grasped the significance of his"light area/' his observation yet supports our contention thatthe brood-chamber begins in the parental cytoplasm near themeganucleus, and does not grow in from the surface. More-over, the words "dissolution of the granules" give a muchmore accurate idea of what is actually seen than does the-description of a definite cavity. We cannot agree that a.definite cavity exists. We have never been able to demon-strate it, either in the living animal or in serial sections, andcertainly if it were present, one would expect to find it inpermanent preparations. The so-called " brood-chamber " is-

DENDROCOMETES PARADOXUS (STEIN). 351

better described as an area over which the normal granularstructure and consistency of the cytoplasm are lost, and whichacquires an extension up to the cuticle of the parent. Thebud, therefore, does not lie in a space which possesses definitewalls, but is merely separated from the normal cytoplasmof the parent by a thin film, so to speak, of a more fluidnature, sufficiently fluid, in fact, to enable cilia to move init (cf. p. 352).

A further point which throws some doubt onBiitschli's andPlate's interpretation of the chamber as a definite cavity isthe structure of the birth papilla—namely, the point at whichthe brood-chamber meets the cuticle of the parent. BothBiifcschli and Plate describe an opening here. Indeed, if thechamber begins as an invagination, as they believed, anopening must be present at the commencement of theprocess. Plate stated that this opening closed up whilethe bud was being formed, and that a new pore was formedat the moment of birth to allow the bud to escape.Biitschli (2), however, threw some doubt upon this, and statedthat the original opening persisted throughout. Neither ofthese views is, in our opinion, correct, since they both dependupon the conception that the brood-chamber is a definitecavity, formed by invaginatiou. The true nature of the birthpapilla can only be properly made out in individuals whichare situated, not upon the edge of the gill, but on the surfaceof the gill internal to this. If such an individual is selected,it is possible to look directly down upon its dorsal surface,and to see the birth papilla from various points of view. Ifthis is done, it is seen that the papilla beai's not a definitepore, but a closed slit, very minute in size, situated betweentwo prominent lips. The papilla itself is formed, in fact,chiefly by the projection of these lips above the surface. Itis by the detection of this prominence that the position of theslit can be made out when it is seen obliquely (Text-figs. 7and 8), or from above (Text-fig. 6). We do not believe thatit is present at all at the very beginning of the process; butit is formed at an early stage, as soon, in fact, as the brood-


chamber itself reaches the cuticle of the parent. Moreover,we have never been able to show that it is ever actuallyopen—i.e., the brood-chamber does not ever openly com-municate with the exterior, except at the moment of birth.When birth occurs, as we shall describe later, the two lips arereadily seen, as they are forced apart by the passage of thebud between them.

Turning now to the dome-shaped mass of cytoplasmenclosed within the brood-chamber, indications may be seentliat it is beiug organised to form the bud itself. The first

--j-B.pap.

O.B. Ciliated ridges. G.VV Contractile vacuole of bud. C. Vs.Ditto of parent. B. pap. Birth papilla.

structures to appear are the ciliated ridges, one only appear-ing fii-st, the other three being added later. As a rule, theciliated ridges are completely marked off beEore the nucleihave begun to divide. On them the cilia are easily seen, andappear, when the bud is seen externally, as small, brush-liketufts actively beating about in the brood-chamber at the sides(Text-fig. 5). In some individuals the cilia may even stretchright across the chamber, and, coming into contact with thefirm resistance of the normal parental cytoplasm on theopposite side, they may beat against it and momentarily alterits shape by their movements.

The presence of active cilia is generally diagnostic of thepresence of a bud. However, it is well known that they arealso seen iu individuals which are about to leave the gill, and

DENDBOCOMETES PARADOXUS (STEIN). 353

which are not cases of bud-formation at all (cf. p. 363).By their activity or sluggishness -the state of health of thebud, and thus indirectly of the parent also, may be judged.When they stop altogether, it may be concluded that some-thing is wrong in the preparation, and a fresh example shouldbe sought (cf., however, p. 358).

With regard to the other organellas of the bud, we havebeen able in certain instances to distinguish the outline of themeganucleus in the broad stalk which connects the developingbud with the parent. It is comparatively difficult to see, how-ever, owiug to the density of the cytoplasm and the abundanceof food material. We have never been able to see the micro-nuclei in the living animals.

Coutractile vacuoles are always very easily seen. In fact,the presence of two such structures is generally an indicationthat bud-formation is in progress, especially when one of themis larger than the other. The contractile vacuole of the budappears earl}'—soon after the ciliated ridges are formed. Itmay be recognised by the fact that it is always smaller thanthat of the parent, and it pulsates more rapidly, systoleoccurring, on an average, every haK-mioute, or more oftenstill than that. Its position in the bud is more or less con-stant. It is usually found immediately dorsal to the ciliatedridges to one side of the middle line (Test-figs. 5 and 6, c.vi,and PI. 26, fig. 11). According to Biitschli (1) it is alwaysfound on that side of the bud which is furthest from theparental contractile vacuole. This author also describes adefinite outletting canal which opens into one of the groovesbetween the ciliated ridges on the ventral surface of the bud.We have not, however, been able to see this structure.

The position of the contractile vacuole of the parent is notso constant. Generally it lies towards the ventral side of theparent in relation to the dorsal region of the bud (Text-fig. 5,C.v2). This position is interesting in view of the fact thatin some other Aoinetaria (Collin, p. 169—Tokophryacyclopum, Choanophrya in fundibu l i fe ra , etc.) theparental contractile vacuole actually communicates with the

354 GEOFFREY LAPAGE AND J. T. WADSWOETH.

brood-chamber and expels its contents into it. In theseforms the brood-chamber has the form of a definite cavity,which dilates with systole when the vacuole discharges itscontents and contracts during diastole. We have never seenany such communication in Dendrocometes , whe "e, it willbe remembered, the brood-chamber is not a definite cavity.

In addition to the above well-defined organellse, there isone other constant feature of the cytoplasm of the developingbud which is worthy of note. In a bud that has almost com-pleted its development an area is observable in which thecytoplasm presents a different appearance than that of theremaining cell-substance (PI. 25, fig. 8, SP. CYT.). Thetexture, if one may so express it, of this area is denser andmore homogeneous in character than the rest of the cyto-plasm, and in addition it exhibits a very finely granularstructure. The absence of deeply staining bodies, such asparticles of food, etc., from this area is also characteristic.Its position is constant, on the antero-posterior axis of thebud, and adjacent to the nucleus. The micronuclei of thebud are genei^ally near this area, and in some cases they havebeen observed within it. Its staining reactions appear to besimilar to those of the remaining cytoplasm.

It is first distinctly recognisable during the late dumb-bellstage of the division of the meganucleus and when thechromatic spindles of the micronuclei are undergoingabsorption. The same appearance is visible in stainedpreparations of those Dendrocometes which are preparingto leave the gills.

We are unable to state definitely what this appearancetruly represents; but we venture to put forward the sug-gestion that possibly a fluid substance, " a nuclear sap," isexuded from the nucleus at this stage, and under the in-fluence of fixatives and other reagents this substance be-comes coagulated, and gives rise to the appearance that wehave described above.

DENDliOOOMETES PARADOXUS (STEIN). 355

THE POSITION OP THE BUD IN THE PARENT.

The position of the bud inside the brood-chamber is shownin the diagrammatic Text-fig. 5, which represents a lateralview of a bud which is almost ready to emerge. All itsoi'ganellas are formed, and its orientation is for that reasonthe more easily understood. The first point to notice, andone which is all-important, is the position of the ciliatedridges. They lie on the ventral surface of the free bud, audtheir position therefore at once localises that surface. Fromthe figure it will be seen that while the bud is in the brood-cavity the ciliated ridges are always uppermost—that is tosay, on that surface of the bud which is turned away fromthe gill and towards the dorsal side of the parent. Theventral surface of the bud is therefore turned towards thedorsal surface of the parent, and it is attached to the floor ofthe brood cavity by its dorsal surface, which is connectedwith the parent by a broad stalk. In other words, theorientation of the bud while inside the brood-chamber is theexact opposite of that of the parent and of the position whichit will itself occupy when free. It is, so to speak, upsidedown while inside the parent. We have takeu special careto establish these facts with certainty, because Biitschli hasboth described and figured (1) the bud in exactly the reverseposition—namely, with its dorsal convex surface turned awayfrom the gill and its ciliated ridges (ventral surface) at thebottom of the brood-chamber. According to Butschli, there-fore, the bud lies in the brood-chamber on its ventral surface,the same way up, so to speak, as the parent. That such aview is incorrect we are convinced by the study of livingbuds and of the actual process of birth. The latter observa-tions are important in view of Biitschli's coufession that hesaw birth only once, and we shall have occasion to show thateven that instance was probably atypical. Had he seen onetypical birth, we are sure his views as to the orientation ofthe bud must have been different. Finally, the study ofserial sections places the matter quite beyond doubt. It is

356 GE01TREY r.APAGE AND J. T. WADSWORTH.

possible, in an individual cut at right angles to its dorso-ventral axis, to work down from the dorsal surface of theparent to the gill below. "When this is done, the ciliatedridges are always met with before the meganucleus. Sincein the free bud the meganucleus is dorsal to the ciliatedbands, it follows that the dorsal surface of the bud while itlies inside the parent is that part which lies nearest the gill.

Further study of the bud reveals, moreover, another point:namely, that the bud lies somewhat obliquely in its brood-chamber. The dorso-ventral axis which would pass throughthe centre of the parent body would not, therefore, passthrough the centre of the bud. Similarly the plane of itsventral surface is not exactly perpendicular to the dorso-ventral axis of the parent. The result of this is that asection which would divide the parent sagittally into twoequal halves would divide the bud very unequally, considerablymore of the ventral surface and ciliated ridges being cut offto one side than to the other.

When the bud is almost ready to emerge, the broad stalkby which it is still attached to the parent is to a certainextent reduced in width by an extension of the brood-chamberinto its base. This undermining, as one may call it, isalways more noticeable on the side on which the birth papillais situated. The bud thus has the appearance of "yearning"towards the slit through which it is destined to be born.

It must not be supposed, however, that the stalk everbecomes narrow ; not, at any rate, while the bud is still withthe parent. Further, we can confirm Biitschli's observationthat the bud is never completely cut off from the parentinside the brood-chamber. Whatever may be the case inthis respect in other Aciiietaria, the bud of Dendrocometesnever lies free in thebrood-chamber. Stein (loc. cit.) stated-that it moved back and forth in the chamber by means of itscilia. This is certainly wrong, as Biitschli has already pointedout. Stein was probably misled by the motion of the ciliathemselves, which very easily give the impression that thebud itself is moving.

DENDEOCOMETES PAJMDOXUS (STEIN). 357

THE BIRTH OP THE BUD.

It is characteristic of the actual process of birth that itoccurs with great rapidity and commences suddenly. Forthis reason we did not for some time observe the actualexpulsion, though we had no lack of individuals in whichbuds were being formed. Biitschli evidently experienced thesame difficulty, since he states (1) that he only saw the actual

TEXT-FIGS. 7-15.

12. 13. 14. 15.

birth of the bud in one solitary instance. To the best of ourknowledge the process has never been described in detail,and since for reasons adduced below, we are convinced thatthe case seen by Biitschli was not typical, we shall describefully what we have seen. All the steps in the process areillustrated in Text-figs. 7-15, which are diagrams made fromrapid sketches taken while birth was being watched. It isusually difficult to determine when a bud is about to emerge.In all its essentials, such a bud is in an advanced state ofmaturity. By keeping . such individuals under constant


observation, working turn by turn, we were able to see anumber of cases of birth.

In all these cases we noticed one very constant indicationthat the time for delivery was at hand. For a few secondsimmediately before birth the cilia of the bud become decidedlyless active and may cease to move altogether. This may bemistaken for death of the individual, and, indeed, was somistaken by us in several cases: bat it is nevertheless afairly reliable indication of what is to follow.

A second or two after the cessation of ciliary action thebud suddenly moves out towards the birth-pore. The processmay best be described as a heiniation of the bud, and is veryrapid. Whether it is due to a contraction of the mother orto some other agency we cannot say. But in some 3-5minutes from the first appearance of the bud at the birth-pore it is outside, free from the parent, and actively swimmingabout in the vicinity.

The actual herniation may be described thus. The bud ismoved rapidly to the birth papilla. In surface view it isseen that the two lips of the birth papilla are pushed apart,and between them appears a ciliated mass of bud protoplasm(Texb-fig. 10). The process continues until, viewed laterally,the parent shows a decided bulge at this point, formed by thebud itself. On this extruded portion of the bud a horizontalline of cilia, representing the ciliated ridges, is seen for amoment. Immediately afterwards this ciliated line hasapparently migrated lower down on the bud (Text-fig. 11),and later still takes up a position at its lower margin. Atthis moment cilia come into view along the upper marginalso, so that the bud bears a fringe of cilia all round itscircumference (Text-fig. 12). Immediately after this theciliated ridges again appear to be near the upper marginonly. They move to the middle again (Text-fig. 13), andfinally are seen at the lower border only (Text-fig. 14).

The shifting of the ciliated ridges is only apparent, ofcourse, and is due to a twisting of the whole bud around itsantero-posterior axis as it comes out. The lateral expulsion


is accompauied by a screw-like motion, so that, while at thevery commencement the ventral surface is uppermost and theciliated ridges are seen in profile along tlie upper border,the bud subsequently twists until, when half-way out, theventral surface and ciliated ridges are seen en face. Thetwist continues further, and finally, when the bud is com-pletely extruded, the ventral surface becomes lowermost,directed towards the gill, and the ciliated ridges are seen inprofile again, but now along the lower border.

When the bud is definitely outside it remains attached for atime to the parent by a rather broad pedicle (Text-fig. 14).This rapidly becomes thinner and longer owing to the efforts

TEXT-FIG. 16.

of the bud to escape. The fiual severance of the connectionis effected partly by a vigorous tugging movement executedby the bud itself, but also by another method. The bud mayvery rapidly rotate itself round the pedicle and twist itselfoff, so to speak, exactly as an apple might be twisted fromits stalk. This gyratory movement of the bud round itsstalk is quite remarkable.

Finally, when the stalk gives way, the bud swims awayfreely. A short peg-like remnant of the pedicle is always lefton the dorsal, convex side of the bud at the end which isposterior in forward movement (Text-fig. 15).

A similar stump may be seen on the parent also for a shorttime. It is often difficult to make it out here, however, sincethe wall of the parent remains for some time markedlycrumpled and shrunken (Text-figs. 15 and 16). Ultimately,


after the lapse of half an hour to an hour or so, the normal,even contour is regained.

One point, also discussed in the section dealing with tlienuclearphenomena,may be briefly referred to here. Butschli (1)has stated that the final division of the meganucleus does nottake place until tlie bud is outside the parent. It is divided,in other words, at the time when the stalk connecting budand parent is severed. We cannot agree with this view.Not only does the study of serial sections definitely disproveit, but also we were unable to detect any sign of meganuclearstructure iu the connecting pedicle during our observationson the living animals. This does not, however, necessarilymean that Biitschli's observations were inaccurate. The con-dition of affairs described by him is not impossible, for it iswell known that cases of what may be termed " pathologicalbir th" do occur in Acinetaria (cf. Oollin, loc, c i t . ) . Inthese cases the amitosis of the meganucleus may be verymuch delayed, and may, indeed, never occur at all, the budremaining attached to the parent and ultimately dying. Sucha case, we believe, was observed by Butschli, and mistakenby hi in for the normal and typical process described above.Since it was the only case he saw, he had no opportunityof comparing it with other instances. Iu active, healthyD e n d r o c o m e t e s , very little variation of the process ocqurs,and any such marked difference as that under discussion musthave been due to the unhealthy condition of the individual,or to the lack of proper precautions in preserving it in ahealthy state while under observation.

THE FATE OF THE BOD.

The aim and normal destiny of the bud is, of course, to seekout a favourable area on the gill and to settle down there,becoming firmly fixed by its ventral surface, and growingeventually into the ordinary " adul t" stage. The behaviourof the bud immediately-after liberation varies somewhat. In-the first place, it may remain in the neighbourhood of the

DENDROCOMETES PARADOXUS (STEJN). 361

parent for some time, swimming about freely; or, in othercases, it may swim away from the parent gill at once andseek out another gill altogether.

These phenomena are the rule in moist chamber prepara-tions. In a state of nature it may well be the rule that a budshall settle down at once on the samu gill. We think itdoubtful whether the bud would be able, powerful swimmeras it is, to ovei'Come the strong currents of the swift-runningstreams in which G a m m a r u s typically occurs. In the pond-inhabiting G a m m a r u s no doubt these currents would beconsiderably weaker, but the risk which the bud would run ofnever finding a resting-place if it left the vicinity of theparent Grammarus would be considerable, because Gam-m a r u s is a very powerful and rapid swimmer. Moreover,the active swimming movements and the respiratory currentscreated by the host would be quite strong enough to sweepthe bud away, even in a still pond. It is very probable,therefore, that in a state of Nature the bud generally settleseither on the gill to which the parent is already attached, orto one of the other gills on the same G a m m a r u s .

During its free life the bud exhibits at least two typesof movement. The first type is characteristic of the timewhen the bud is swimming freely in the water in the neigh-bourhood of a gill. I ts progress is rapid, and generally in astraight line. It does not, however, orientate itself with thedorsal surface upwards and the ventral surface downwards, usmight be expected; but swims, as it were, on one side. Inother words, the dorsal surface is turned generally to the left,the ventral surface to the right, so that the right-handmargiu is directed upwards. Moreover, the forward move-ment is accompanied by a very marked swaying- round itsiintero-posterior axis, so that the observer sees now the dorsalsurface uppermost, now the right-hand margin and now theventral sui'face. An apt imitation of this phenomenon mightbe secured if it were possible to sling an oval plano-convexlens upon a string passing through its long axis. If it werenow propelled along the string, and at the same time gently

362 GEOFFREY LAPAGE AND J. T. WADSWOBTH.

tapped so as to make it sway slowly and rhythmically fromside to side about the string, the movement of the bud wouldbe very closely reproduced.

The second type of movement is characteristic of the timewheu a bud, having arrived at a gill, comes down upon it andbegins to move over its surface. It moves now very inuch inthe same way as a hypotrichous ciliate would move, the ciliaon the ventral surface being in contact with the gill andacting after the manner of legs. In this way the bud" creeps/' so to speak, carefully over the surface, picking itsway among the Bpistylids, Dendrocometes , and otherfauna already established there. Its behaviour during these" creeping" movements strongly recalls that of a hypo-trichous ciliate in search of food, and it may easily bemistaken for such by the careless observer. We have noevidence, however, to show that the bud ever feeds at allduring its active, free-swiinming stage. No mouth is present,and we must conclude that the "creepiug" movements aremainly exploratory, and have for their object the " selection "of a suitable place for attachment.

As regards the positions on the gill most often chosen by thebud, we have not found that auy absolute choice is ever exer-cised. Upon well-popnl;ited gills the Dendrocometes arescattered indiscriminately all over the surface. On sparsely-populated gills the animals are most often found towards, oractually ou, the edge, often being folded over the edgeso that they are attached to both surfaces of the gill plate.The rest of the gill surface is left quite free. From this it isreasonable to conclude that the edge is the position mostpreferred. Doubtless it is the situation most favourable forthe capture of food.-

We have also noticed that the animals also seem to preferthe proximal end of the gill when the edges are fully occupied,a preference also shown, curiously enough, by the Spirochonasand Epistylids, when these are present.

Whatever the position chosen, however, the bud loses notime in settling down. It becomes quite motionless at the

DENmiOCOMETJDS PARADOKUS (STJSIN). 363

chosen spot, with the ventral surface applied to the gill, andsecretes a basal plate of attachment, by which it is per-manently fixed in position. The ciliated ridges disappear,and the bud nowpresents the appearance of a small Dendro-cometes without arms. In preparations in which bud-formation is actively going on many such individuals may befound. Text-fig. 1 shows the appearance which they present.It will be seen that the ridge marking off the dorsal from theventral side, referred to on p. 345, is particularly well markedat this stage.

It only remains now for the bud to put out arms and itbecomes a typical "adul t" Dendrocometes . From thisstage onward it leads the normal life of that phase in thelife-history, actively feeding and growing in size. Ultimatelyit may conjugate inthe manner already described by Hicksonnnd Wadsworth (loc. cit.), or may itself form buds.

THE RELATION OP BUD-FORMATION TO THE SIZE AND MATURITY

OP THE. PARENT.

With regard to this interesting question very little can bedefinitely said. Apparently there is nothing in the appear-ance of any given individual which will indicate whether it iscapable of forming buds or not. Either well or sparsely fedindividuals may form buds, and we have observed the processin forms of all sizes. In every case that we have seen, how-ever, the parent possessed arms. It seems probable, there-fore, that a young individual, but recently settled on a gill,does not normally form buds until it has developed arms, andhas commenced to feed in the usual way.

In this connection we may refer to a fairly commonphenomenon which is associated with the migration of theDendrocometes from one gill to another. An individual,which has hitherto exhibited all the structural and physio-logical features characteristic of the species, will at certainperiods draw in its arms altogether, taking on the appearanceshown in Text-fig. 1. Very soon ciliated ridges appear, and


almost the whole substance of the animal parts from its baseand swims away, leaving only the basal plate of attachmentand a crumpled residuum of protoplasm, enclosed in the oldcuticle, attached to the gill. The active individual so liberatedsubsequently settles on another gill, secretes a new basalplate of attachment, and, putting out arms ngain, resumes itsnormal mode of life. The process was observed and has beengraphically described and figured by Biitschli (1), who re-garded it as being essentially different from true bud-forma-tion, and we entirely concur in this view. The mass which isleft behind on the gill consists of nothing but a residuum ofprotoplasm incapable of physiological activity, and destinedto disintegrate and die. The phenomenon is more of thenature of a migration, correlated with, aud doubtless directlycaused by some change in the gill itself. We have noticedthat individuals exhibiting this phenomenon are attached togills which are nob quite normal. Sand (9) has pointed outthat it always takes place upon gills which are about tomoult. If this is so, it is probable that, when the hostGrammar us is about to shed its skin, physiological changescommence in the gill which iu some way influence theDeudi 'ocometes living on it, so that they are stimulated tomigrate in the manner described above. Biitschli, who firstobserved the phenomenon, regarded it as an excretory process.We have no evidence to support either view; but thephenomenon is so common that we have no doubt that it isto be reckoned with as a constant feature of the physio-logical relationships of the animal.

THE MEGANUCLEUS.

The meganucleus of Dendrocometes is a conspicuousstructure of relatively large size, measuring from 0"02-O034 mm. In shape it varies a good deal, according to thestate of its activity. In the so-called "resting stage"—thatis to say, when it is not taking part in any reproductiveprocess or in syngamy—it is usually oval; but it may bemore nearly spherical or elliptical and very much elongated.

DENDROOOMETES PARADOXES (STEIN). 365

It usually lies in the ventral region, over the attached base ofthe creature, and when it is oval its long axis is parallel tothe plane of the ventral surface (Text-fig. 4). In the livinganimal it is usually seen without difficulty, and presents afinely-dotted appearance, due to the presence in it of fluechromatin granules.

The nucleus is a good example of the granular type ofnucleus characteristic of the meganuclei of Heterokaryota ingeneral. In permanent preparations it is often surroundedby a clear zone (PL 25, fig. 1), irregular in outline, whichseparates it from the cytoplasm. This zone is merely anartefact due to shrinkage daring fixation, and does not repre-sent any part of the nuclear apparatus. The nuclear matteritself is arranged in the form of a well-marked network withwide interspaces (PI. 25, fig. 2) ; the strands of this networkare composed of small granules aggregated together upon aliniu framework. From the fact that they stain by all theordinary nuclear stains, and from their subsequent behaviourduring divisiou of the nucleus, it is obvious that they aregranules of chromatin, and represent the chromatinic elemeutsof the nucleus. The interspaces of the network, on the otherhand, invariably refuse to take up the stain, and in a healthymeganucleus are perfectly homogeneous iu appearance.

We have never been able to demonstrate the presence ofa true nuclear membrane. Previous workers, includingButschli ,(1) and Plate (8), have definitely stated that thisstructure is well marked and easily seen. Collin also affirmsits presence, and has figured what he believes to be a nuclearmembrane in Dendrocometes (3) (PI. 25, fig. 19). In thatfigure a very definite membrane is shown, lining the shrink-age space in which the meganucleus lies, and beariugchromatin grains adhering to its inner surface.

This latter fact, coupled with the fact that the structurefigured stains deeply with iron hasinatoxylin, proves, in ouropinion, its true nature. A true nuclear membrane is alwaysachromatinic in nature, and only takes up nuclear stains withdifficulty. The "membrane" described by Collin is therefore

VOL. 6 1 , PART 3 . NEW S.EEIES. 2 4

366 GEOFFREY LA.PAGB AND J. T. WADSWOKTH.

mucli more likely to be merely a portion of the superficiallayer of the chroraatin grains, which, lying together, con-stitute what Minchin (6) has termed a "false" or "chroma-tinic" membrane. We have found portions of such a mem-brane in some of our preparations, especially in those whichhave been badly fixed and in which the meganucleus is dis-torted. We have never seen it in preparations which arewell fixed and stained. Moreover, it is significant that,although Collin gives five fignres of Dendrocometes (3)(PI. 25, figs. 18, 19, 20, 2], and 22), in only one of these(fig. 19) is the so-nalled nuclear membrane shown. A furtherpoint of some importance is the fact that no trace of a nuclearmembrane is ever seen during division of the megauucleus.It is not uncommon, of course, for the nuclear membrane todisappear during division—it does so frequently duringmitotic division at any rate—but it is reasonable to assumethat, were it present, some trace of it would be discovei'ableat some stage of the process.

Taking all these facts into consideration, we must conclude,therefore, that a true nuclear membrane does not exist in themeganucleus of Dendrocometes . A "false " or " chroma-tinic" membrane may, however, appear in the mannerdescribed above.

Kinetic elements, such as centrosomes, centrioles, etc., arealso entirely absent from the meganucleus, and division isstrictly amitotic. In this respect the meganucleus ofDendrocometes agrees with that of other Acinetaria.

Further, we have not found any nucleoli—i.e., pure lumpsof plastin, although they are present in the meganuclei ofother Acinetaria. Karyosomes, also, are not present asisolated lumps of chromatin apart from the general chromatinnetwork.

The interpretation of the features mentioned above hasgiven rise to conflicting views, and discussion has beenmainly centred round the question as to whether the chro-matin grains are supported upon a liuin framework or not.Collin (3), and with him Biitschli (1) and others, denies abso-

DENDUOOOMETES PARADOX US (STEIN). 367

lutely that any liniu framework is present. He regards theuieganucleus as a vesicle, enclosed in a definite membraneand filled with nuclear sap in which float quite freely thegrains of chromatin (microsomes) and larger granules of adifferent nature (raacrosomes). The conception is extendedalso to the structure of the cytoplasm, and is invoked furtherto explain the phenomena of the nuclear division and bud-formation. In support of it the author brings forward anextensive series of observations on the other Acinetaria, andseeks to supplement these with collateral evidence derivedfrom other groups of Protozoa (Ciliata). In our opinion,however, this conception does not fit the facts as we haveobserved them in the case o£ D e n d r o c o m e t e s . The wholestudy of the meganucleus in the "rest ing state," as wellas during division, goes to support the alternative conceptionof a number of chromatin grains borne upon and supportedby a framework of linin threads.

We are not able definitely to affirm that we have seen thisframework, at any rate while the meganucleus is in the" r e s t i n g " phase. Even in the thinnest and most carefullystained sections no framework is visible there. But we donot feel justified in concluding from this fact alone that noframework is present. Minchin (loc. ci t . ) has clearlypointed out that the fact that no linin threads can be seen isno proof that they do not exist. Linin, he argues, is achro-matinic in nature, and therefore difficult to render visible bythe ordinary chromatinic stains. Further, he points out thatwere no framework present, it would be difficult to explain thearrangement of the chromatin in a very definite manner,in this case a network. We may add to this the fact thatduring division a very obvious framework appears (cf. pp.369 et seq. , PI. 25, figs. 7 and 8). If there is no lininframework during the " r e s t i n g " stage, from what is itformed during division ? Furthermore, it is pretty generallyagreed that a linin framework supports the general cytoplasm,and is present there as a network similar to that which isfound in many nuclei. Now, in D e n d r o c o m e t e s it is no


more easy to demonstrate the presence of a framework in thecytoplasm than it is in the meganucleus. Are we to con-clude that, therefore, it is not present in either situation ?We, at any rate, are not prepared to accept such a view.

The whole matter is, perhaps, only part of the much widerquestion of the structure of protoplasm in general. Collin'sconceptions are no doubt the outcome of an acceptance ofButschli's theories on this subject. In our view the presenceof a linin framework is an essential point in a true conceptionof the structure of this rneganucleus. It is composed, webelieve, of a network of linin threads, bearing fine grainsof chromatin which are approximately equal in size. There nrenever any centrosomes, centrioles, or other kinetic centres, andthe presence of nucleoli is exceptional. The whole structureis bathed in a nuclear sap, which, when coagulated in perma-nent preparations, appears as a perfectly homogeneoussubstance filling up the interspaces of the network.

A word is necessary here in explanation of PI. 25, fig. 1,because we believe that atypical appearances, such as thisillustrates, have led to erroneous interpretations of the truestructure. The figure shows a state of the. meganucleuswhich we have met with fairly often. The interspaces of thenetwork appear to be filled with solid bodies which stsiinrather more darkly (especially with brazilin) than thecoagulated nuclear sap does as a rule.

We do not think that these apparently solid and separatebodies lying in the interspaces of the network are normalnuclear elements. We have come to the conclusion that theyare local coagulations of nuclear sap which have stained in anirregular and abnormal manner. They are not present in themajority of our preparations.

DIVISION OP THE MEGANUCLEUS.

It has already been pointed out that the meganucleus ofDendrocometes does not possess a centriole, nor are thereany controsomes or other visible kinetic centres in relation to.it. Division is always strictly amitotic.


The process of division consists simply of an elongation ofthe nucleus into a dumb-bell shape, and a subsequent separa-tion of the two heads of the dumb-bell to form the twodaughter-meganuclei. I t will be remembered that the mega-nucleus lies in the ventral region of the parent. In division itelongates towards the dorsal surface into the region wherethe bud is being formed. At the dumb-bell stage one headof the dumb-bell lies in the dorsal region of the bud, theother in the ventral region of the parent. Subsequently theconnecting strand divides be fo re the bud is ready to beexpelled, one half forming the meganucleus of the bud,the other half remaining as the meganucleus of the parent.

The structure of the megauucleus in the " rest ing" con-dition has already been described, and need not be gone intofurther. Soon after the commencement of the formationof the bud, the parent meganucleus begins to undergo change,and the first alteration in its structure is the disappearanceof the regular mesh work arrangement of the chromatiugrains. The nucleus now presents an even granular appear-ance, the chromatin being distributed more or less uniformlyall over the nuclear area, no definite interspaces being visible(PJ. 25, fig. 3, contrast fig. 2). Close examination of thechromatin elements now reveals the fact that the}' are nolonger grauules, equal in size, but are beginning to be elongatedinto elliptical, or even shorb rod-shaped particles, some ofwhich are noticeably larger than the others. We propose tocall this phase the " granular stage " (PI. 25, fig. 3).

In the nest stage (PI. 25, figs. 4, 5, and 6), the loss of thegranular condition of the chromatin elements is complete,and the nucleus is now composed of rod-shaped lumps ofchromatin, arranged along certain well-marked and obviouslines. All these figures show that at this stage the wholenucleus has the appearance of being involved in a kind ofvortex movement which we shall refer to as the "whirlstage." Usually this phase is very well marked and is quiteremarkable. The whole nuclear area seems to have beenfixed in the act of undergoing a general "churning up," so

370 GEOFFREY LAPAGE AND J. T. WADSWOBTH.

to speak, the lines of movement being always directed towardsa definite focus, the position of which in the nucleus mayvary. Pig. 4 shows this well. We have repeatedly lookedfor this movement m the living creature, but have neverbeen able to see it. Collin (loc. cit.), however, has actuallyseen it in progress, and describes it as follows :

" Le prelude de la division observee in vivo consiste enmouvements de brassage d'abord extr&mements lents, in-certains, comme hesitants, mais temoignant avec netteted'une rupture d'equilibre; . . ."

Coll in, in accordance with his conception of the mega-nucleus as a vesicle filled with fluid in which chromatingrains float freely, compares the phenomenon with that of anAniceba moving forward. Just as one can explain themovements of an Amoeba as a simple effect produced byinequalities of surface tension on its different faces, so, hesays, one can explain these movements in the meganucleus ofDen drocometes as being purely passive adaptations to adisturbance of the equilibrium of the cell. This disturbance,he believes, is directly caused by the development in the cellof the embryonic space (brood-chamber). This explanationis tempting and ingenious, but it depends entirely upon hisown conception of the structure of the meganucleus. Wehave already stated that we cannot accept that conception,since it denies the existence of a linin network. In ourview the whirl stage, above described, is much move trulycompared to the condition which is seen during the spiremestage of mitosis in metazoan nuclei. It differs essentially fromthat, however, in that it is not governed and directed by anyvisible kinetic centre. It is very probable, however, that thetwo processes are analogous and have a similar significance.

The "whir l" stage coincides in time roughly with thestage of micronuclear division shown in PI. 25,fig. 6. It is notconfined to the meganucleus of Dendroconietes alone, butoccurs also in the meganuclei of other Acinetaria. Probablyit is to be regarded as a regular and constant feature of theamitosis of these nuclei.

DENDE0C0MBTJ5S PABADOXUS (STEIN). 371

Following on this] phase, the meganucleua passes on atonce to the actual process of division. I t now has the verybeautiful and characteristic shape shown in PI. 25, figs. 7, 8.By this time the bud is well advanced, in some cases almostcompletely formed. PL 26, fig. 9, reconstructed from a seriesof sections, illustrates this point. The bud is complete, exceptfor the final division of the meganucleus. Six micronucleiare present, three in the bud and three in the parent, theciliated ridges are fully formed, and the connecting stem ofthe meganuclear dumb-bell is already thin and long-drawnout. A bud at this stage would, iu the living state, presentsomewhat the appearance shown in Text-fig. 5.

The structure of the dumb-bell itself differs in differentparts (v. PI. 25, fig. 7). Iu the broad band which connectsthe two heads the linin framework is clearly seen. I tconsists of a system of fine threads connecting the twoheads. On tliein lumps of chromatin, still rod-shaped in thisregion, are arranged in quite an irregular manner. PI. 25,fig. 8, shows a dumb-bell at a rather later stage, when theconnecting band bears little or no chromatin at all. In thisstage all the chromatin has been accumulated in the two"heads . " PI. 26, fig. 9, shows a still more advanced con-dition in which the connecting band has become attenuatedand will shortly be broken at its middle.

With regard to the structure of the "heads" of thedumb-bell, PI. 26, fig. 9, is very instructive. Here thedaughter meganucleus in the parent shows the chromatiuelements still in the form of rods arranged along certainlines, as if streaming into that end of the dumb-bell. Thedaughter meganucleus in the bud, however, shows thechromatin elements once more in the form of grains,arranged in the manner characteristic of the " granularphase" with which the process of division began. The samecondition is much more clearly shown in PI. 26, fig. 10. Inthis case the chromatm in the two " heads " is again organisedinto approximately spherical grains, more or less equal insize and uniformly distributed over this part of the nucleus.

372 GEOFFBEY LAPAGE AND J. T. WADSWORTH.

The impression given by the study of all these dumb-bellsis, therefore, that the dumb-bell stage is to be regarded asthe climax, so to speak, of a series of changes, commencingwith the "granular phase " and passing through the " rod "and "whir l" stages subsequently. The aim of these prepara-tions is to render the chromatin divisible between bud andparent. This is accomplished by the dumb-bell stage. Whenthe act is completed, the daughter nuclei commence a seriesof similar changes, but in the reverse order, which finallylead back to the assumption by the daughter meganucleiof the meshwork structure characteristic of the "res t ing"meganucleus.

In considering this process as a mechanism for the divisionof chromatin between parent and bud, we must conclude thatsuch division is only approximately equal. No doubt theassumption by the chromatin of the form of rods, theirintimate mixing in the "whir l" stage, and their subsequentseparation along certain definite lines in the dumb-bell stage,do all indicate an attempt, as it were, at equal division.But there is never any evidence of the existence of visiblekinetic elements or of the elaborate mechanism which ischaracteristic of mitosis. The process is essentially amitotic;and the feature which decides, we think, its amitotic characteris not so much the absence of kinetic elements, but the factthat at no stage are any"structures found which could beinterpreted as chromosomes. We would point out, however,that recent researches have tended to show that the dividingline between amitosis and the primitive forms of mitosis isby no means a sharp one. It appears necessary to recognisethat amitosis, no less than mitosis, presents grades of differingcomplexity. We are dealing here, without any doubt, witha form of amitosis which is certainly elaborate and highlyorganised. It is not surprising, therefore, that it shouldapproximate in unimportant features, to those forms ofgenuine mitosis which are very simple and primitive.


THE MICEONUCLEUS.

The first proof of the existence of a microuucleus in anAcinetarian was given by Biitschli (2) in 1876, when hedescribed it in Sphserophrya. Since then its presence hasbeen proved in all the Acinetaria which have beea described,and it is now generally agreed that the presence of oneor more micronuclei is characteristic of the group.

In the case of Den d roco m e t e s itself, the first certain proofof the presence of micronuclei was definitely given by Hicksonand Wads worth (loc. ci t .) when they described in detailtheir structure :md their method of division during syngamy.

There are typically three micronuclei in each individual;the bud, therefore, normally possesses three also. Thisnumber may, however, vary. We have found in some casesas many as four or five, or even six in one instance, but nevera greater number than this. In the vast majority of casesthere are three, and this number must be regarded astypical. I t is worthy of note that two of them usually lieclose together, side by side, as it were, while the third issome distance away. We can offer no explanation of thisfact, but have indicated it in the figures (cf. PI. 26, figs. 9and 10, and PI. 25, fig. 7).

In the living animal it is rarely possible to demonstrate thepresence of micronuclei. In stained preparations, however,they are very easily found. Under low magnifications(Zeiss D.) each micronucleus sliows itself as an apparentlysolid granule of chromatinic matter lying freely in a clearzone, which separates it from the surrounding cytoplasm. I thas the appearance, then, of the vesicular type of nucleuscommonly seen in Flagellata, Amoeba? of the Umax group,and some other Protista. But it is evident, when its structureis considered in detail, that it differs from these nuclei inimportant respects.

The structure of the resting micronucleus is shown inPI. 26, fig. 13. There is a central chromatinic body sur-


rounded by a clear zone in which no structure of any kindcan be detected. This clear zone is a very characteristicfeature of the micronucleus, and serves as a valuable guide indistinguishing it from the other granules which are usuallypresent in the cytoplasm. It is not an infallible guide, sincevery frequently, and particularly in well-fed individuals, foodmasses may occur in great abundance; and since theyrepresent ingested organisms in various stages of digestion,these masses usually contain chromatinic material. More-over, they are frequently surrounded by a clear zone similarin appearance to that which forms part of the micronucleus.They may, therefore, very closely resemble micronuclei, andit is often very difficult to establish their true nature. Theonly infallible test is their subsequent behaviour during bud-formation or syngamy.

With regard to the nature of the clear zone, which is socharacteristic of the micronucleus, at least two alternativesare possible. Either ib represents a purely artificial spaceproduced by the action of reagents, and similar to that oftenfound around the meganucleus in stained preparations(v. p. 365), or it is a part of the micronucleus itself.After careful consideration we have come to the conclusionthat the latter interpretation is the correct one. In the firstplace, the circumference of the space is always regular in out-line, fin almost perfect sphere. It is never irregular, as, forexample, is the space which often surrounds the meganucleus.Moreover, its constant presence points to some real relation-ship with the rest of the nucleus. Were it produced byshrinkage, we should naturally expect to find it around allother solid bodies of a shrinkable nature in the cytoplasm.We have already pointed out that such a space due toshrinkage is often present round the food masses, butwhen it is present, it is never so broad or so regular inoutline.

Remembering all these points, we think it right to concludethat this clear zone does represent a part of the micronucleus.Against this view is the fact that we could never in any case

DENDROCOMETES PABADOXUK (STEIN). 375

detect a membrane, or anything like one, separating the clearzone from the surrounding cytoplasm. This, however, doesnot necessarily invalidate our view. In many Protozoa, forexample, Amoebae of the limax group, simple, vesicularnuclei occur, which are composed of no more than a centralgrain of chromatin (karyosome) floating in a zone of clear,nuclear sap, and not enclosed in a membrane. Such anucleus has been called a " protokavyon," and we believe tha tthe micronuclei of D e n d r o c o m e t e s are of this type. Theonly differences between them and nuclei of the Amoeba typeare, first, the absence from the former of a ceutriole or anykind of visible kinetic cen t r e ; and, secondly, the fact tha tthe chromatiu of the central body is not present iu the formof a compact, single lump, but is arranged in a networkof strands, the whole having a structure which we may callspongy.

The structure of this central chromatinic body is shown iuPI. 26, fig. 13. In shape it is spherical, and it is composedof s trands of chromatin arranged in a more or less regularnetwork. Never in any " r e s t i n g " micronucleus have we seensmall grains of chromatin comparable to those chai'acteristic ofthe " r e s t i n g " meganucleus. The chromatin of the micro-nncieus is always laid down as a continuous strand on anachromatinic network, and this s t rand is never broken, noroverlaid with granules at any point. W e shall see, however,that granules do appear during certain phases of division.

DIVISION OF THE MICKONUCLEUS.

The division of the micronucleus during conjugation hasalready been described by Hickson and Wadsworth (loc. c i t . ) ,and we have found tha t their description applies in broadoutline to the process of division during bud-formation also.There are, however, one or two points which must bereferred to.

The micronuclei of D e n d r o c o m e t e s divide by what mustbe described as a primitive form of mitosis, resembling, but


not identical with, the mitosis of the micronuclei of Para-mcecium described by Hertwig (4). The process also showssome general resemblance to the so-called " promibosis"described by Nagler (7) in the primitive Amcebse. Butthere is one very important difference—namely, the completeabsence of centrioles, polar plates, or any such kinetic centresfrom the micronuclei of Dendrocometes . We have re-peatedly and very carefully searched for traces of suchstructures, since a p r io r i one would expect them to bepresent; but we have never, ,in any case, been able todemonstrate them, and we do not believe that they arepresent.

The first feature which indicates that a micronucleus iscommencing to divide is a noticeable increase in size of thecentral chromatinic body. This body swells up, occupyingmore of the nuclear area, and at the same time it loses itsmeshwork structure, and the chromatinic threads becomearranged in the form of an irregular coarse skein. Subse-quently the skein becomes finer, and may fill up almost thewhole of the clear zone.

Following on this, the chromatinic body elongates, as isshown in PL 26, fig. 14. Now, for the first time, smallgranules of chromatin are seen irregularly scattered throughthe elliptical figure, Tying on the longitudinal linin thrends,which are themselves easily seen at this stage. We havecome to the conclusion that these grains of chromatin repre-sent the stage in more typical mitoses in which chromosomesare formed. We have never been able to find any moredefinite evidence of the formation of chromosomes than this,and we do not believe that definite rods are ever formed,such as have been described, for example, in the divisionsduring conjugation (Hickson and Wadsworth, loc. cit.).The figures of Hickson and Wadsworth in their paper aredrawn from preparations made from the same material andby the same hand, so lhat our failure to find them in thedivision during bud-formation cannot be ascribed to badtechnique.

DBNDROCOMETES PABADOXDS (STEIN). 377

Further, afc a later stage (PI. 25, fig. 3), the granules justmentioned are again seen, but situated now at the two polesof the figure. At this stage it will be seen that the linitiframework is very well marked, and bears on it somechromatin in process of separation to the poles, in additionto the granules already in that situation.

Subsequent stages simply include the further elongation ofthe spindle. In PI. 26, fig. 15, the chromatin is collecting atthe poles, and this figure shows well the distinction betweena lighter staining chromatin, forming caps at each end of thespindle, and the granules at the extreme poles, which representthe chromosomes. In PL 25, figs. 5 and 6, the elongation of thespindle is illustrated. Often this may proceed to such anextent that the figure may reach almost across the wholewidth of the body of the animal. The chromatinic elementsgradully fuse, the granules disappearing again, until the pear-shaped masses shown in PI. 25, fig. 6, are formed. Subse-quently the spindle breaks down, and the " daughter nuclei "regain the meshwork structure characteristic of the restingstage. In PI. 25, fig. 7, three micronuclei are shown in thebud, three in the parent, and between them traces of the dis-appearing spindle can be made out.

Considei'ing this process in relation to mitosis in general,the most striking features are: (1) The complete absence ofcentrioleSj centrosomes, or any visible kinetic centres; and(2) the absence of chromosomes in any defined form. Withregard to the latter point, Hickson and Wadsworth, in theirdescription of the mitosis of syngamy, figured and describeddefinite chromosomes arranged in the form of an equatorialplate. In none of our preparations have we been able tofind any definite rods such as those figured there. Thechromatin is never more highly organised than as the grainsshown in our figures. We must conclude, therefore, that themitosis of bud-formation differs from that of conjugation inthis respect.

The absence of visible kinetic centres is not in itself a veryremarkable feature, since numerous instances are now on


record of mitosis which normally occur without such struc-tures. Moreover, Hickson and Wadsworth were unable tofind them in the mitoses of syugamy. The process musttherefore be regarded as a mitosis of a primitive kind,resembling those primitive mitoses which occur generally insimple nuclei of the protokaryon type.

SUMMARY.

(1) This paper describes the reproductive process in theAcinetarian Dendrocometes pa radoxus (Stein), epizoicon the gills of Gammarus pulex.

(2) Reproduction is effected by internal budding. Bachindividual produces a single bud at each reproductive act.

(3) The bud is roughly oval and plano-convex in shape,measuring 0"06 x 0'04 mm. The convex surface is dorsaland is devoid of cilia. The opposite ventral surface bearsnumerous long cilia, inserted along four ciliated ridges.This surface, hitherto regarded as flat, is in reality convex iuthe outer ciliated area, concave in the centre. The budpossesses, like the parent, three micronuclei.

(4) Bud-formation begins by a linear dissolution of thecytoplasm of the parent, which proceeds until a dome-shapedmass of cytoplasm is marked off. This mass is subsequentlyorganised into the bud, acquiring a contractile vacuole andfour ciliated ridges. The orientation of the bud while stillwithin the parent is discussed.

(5) The area over which the above-mentioned dissolutionof the parental cytoplasm occurs constitutes the so-called" brood-chamber." Unlike the similar structure in someother Acinetaria, it is not to be regarded as a definite space.It extends to thesurface of the parent at one point at whichthe bud is subsequently born. There is, however, no definiteopening to the exterior, except at the moment of birth.

(6) The development of the bud and the actual process ofbirth are described in detail.

(7) Division of the meganucleus of the parent is amitotic.

DENDROCOMETES PARADOXES (STEIN). 379

The meganucleus does not possess a true achromatinic nuclearmembrane, but the surface layer of chromatin. is so arrangedas to form a false chromatinic membrane. Division of thetneganuoleus is always complete before the bud leaves thebody of the parent.

(8) Division of the micronuclei is, on the other hand, by aprimitive kind of mitosis. No chromosomes or visible kineticcentres are present.

(9) Reference is also made to the migration of the adultD e n d r o c o m e t e s from one gill to another. This processmay simulate bud-formation, and is either of the nature ofan excretory process or is associated with ecdysis in the hostGram m a r us.

(10) Attention is called to an area of special cytoplasmwhich is found in the bud during its formation. The signifi-cance of this structure is at present obscure.

LITERATURE.

1.—Biitschli, 0. (1877).—"Ueber den Dendrocomefces par adoxus,Stein" etc., ' Zeits. f. wiss. Zool.,' Bd. xxviii, p. 49.

2. (1889).—"Infusoria," in Bronn's 'Tliierreich.'3. Collin, B. (1912).—" J t̂ude inonograpliique sur les Acinetiens,"

'Arch, de Zool. exper.,' tome li.4. Hertwig, B. (1889).—"Ueber die Conjugation der Infusorien,"

' Abhandl. der bayer. Akad. der Wiss.,' cl. II, Bd. xvii.5. Hickson, S. J., and "Wadsworfck, J. T. (1902).—"Dendrocometes

paradoxuB," 'Quart. Journ. Micr. Sci.,' vol. 45, Part 3, p. 325.6. Minchin, E. A. (1912).—' An Introduction to the Study of the

Protozoa,' p. 75. Edwin Arnold, London.7. Nagler, K. (1909).—" Entwicklungsgesckiclitliclie Studien iiber

Aruoben," ' Arch, f. Protiatenk,' Bd. xv.8. Plate, L. (1886).—"Untevsuckungen einiger an den Kienienblattem

des Gammarus pulex lebenden Ektoparasiten," ' Zeits. f.wiss. Zool.,' Bd. xliii, p. 175.

9. Sand, R. (1901).—' fitude monograpliique sur le groupe des In-fusoires tentaculiferes.' Bruxelles : Alfred Castaigne.

10. Stein, F. (1851).—"Neue Beitriige sur Kenntniss der Entwick-lungsgeschichte mid des feineren Baues der Infusionsthiere,"' Zeits. f. wiss. Zool.,' Bd. iii, p. 475.

380 GEOFFREY LA PAGE AND J. T. WADSWOETH.

OTHER BEFERENCES TO D E N D R O C O M E T E S PARADOX US WILL

BE FOUND IN THE FOLLOWING PAPERS.

Doflein, F. (1911).—' Lehrbuch der Protozoenkunde.' Jena: GustavFischer.

Fraipont, J. (1878).—' Beclierches sur les Aeinetinieng de la coted'Ostend,' Bruxelles.

Hickson, S. J. (1900),—" The Nuclei of Dendvocometes." Eeportsof the British Association.(1903).—"The Infusoria or Corticata Heterokaryota" in Lan-

kester'a ' Treatise on Zoology.' London: A. and C. Black.Kent, W. S. (1880-1882).—'A Manual of the Infusoria.1 London:

David Bogue.Keppen, N. (1888).—" Remarks on the Infusoria Tentaculifera " (in the

Russian language), ' Mem. Soc. nat. Odessa,' vol. 13.Martin, O. H. (1909).—" Some Observations on Acinetaria," ' Quart.

Journ. Micr. Sci.,' vol. 53, Pt. 2, p. 351.Maupas, E. (1881).—"Contribution a l'etude des Acinetiens," 'Arch.

de Zool. exper.,' tome ix, p. 299.Schneider, A. (1886):—"Fragments sur les Infusoires," ' Tablettes

Zool.,' tome i. Poitiers.Wrzesniowski, A. (1877).—" Beitrage zur Naturgeschichte der Infu-

sorien," ' Zeita. f. wiss. Zool.,' Bd. xxix, p. 270.

EXPLANATION OF PLATES 25 AND 26,

Illustrating Mr. Geoffrey Lapage's and Mr. J. T. Wadsworth'spaper on " D e n d r o c o m e t e s p a r a d o x u s . " Part I I .—Keproduction (Bud-formation).

All the figures except figs. 7, 13, and 14 are drawn with the cameralucida, Comp. Oc. 6, Zeiss apo. 2 mm. Tube length 146, magnificationX 900 linear.

Fig. 7 is drawn with Ocular 2, and the same objective. MagnificationX 630 linear.

Figs. 13 and 14 are drawn with the Comp. Oc. 18, and the sameobjective. Magnification approximately 2, 400 linear.


PLATE 25.

Fig. 1.—Section of individual with atypical resting meganucleus(v. p. 368).

Fig. 2.—Section of individual witli typical resting meganucleuB (M)and two micronuclei (m).

Fig. 3.—Figure constructed from two consecutive sections. Mega-nucleus in " granular " phase (-fcT). Two micronuclei in early stage ofdivision (m.m.).

Fig. 4.—Section of individual with meganucleus (M) in " whirl"phase of division. (/.»<•.) Food masses, (gr.) Gill.

Fig. 5.—Section of individual with meganucleus (M) in " whirl"phase Two micronuclei (m) dividing with spindles (sp.). (OR.)Ciliated ridges of bud. The cytoplasm is heavily laden with foodmasses (f.m.). Contrast with fig. 6, where the cytoplasm is relativelyclear.

Fig. 6.—Section showing a rather later stage. The meganucleus(M) is still in the " whirl" phase, but the micronuclei (in) are almostcompleted divided. (C.B.) Ciliated ridges of bud.

Fig. 7.—Whole mount showing the micronuclei (m) completelydivided: three in the bud (m,) and three in the pareut (»ia). Theremains of one spindle are indicated. (G.B.) Ciliated ridges ofbud. Meganucleus in dumb-bell stage.

Fig. 8.—Section showing another meganucleus (M) in the dumb-bellstage. (C.B.) Ciliated ridges of bud. (Sp. Cyt.) Area of specialcytoplasm (v. p. 345). Only one micronucleus (in) is present in thesection, (f.m,.) Food masses.

PLATE 26.

Fig. 9.—Figure constructed from a series of sections to show thebud almost completely formed. The meganucleus (M) is still in thedumb-bell stage, but the connecting strand is almost severed. Threemicronuclei are present in the bud (m^) and three in the parent (m2),one being seen through the bud itself. (C.B.) Ciliated ridges of bud.

Fig. 10.—Figure constructed from a series of sections to show themeganucleus completely divided into a portion in the bud (m,) and aparental portion (m?). Three micronuclei are present in the bud (m,)and three in the parent (TO2). (C.B.) Ciliated ridges of bud.

Fig. 11.—Diagrammatic drawing of free bud fixed with osmic acidand stained with brazilin. (M) Meganucleus. (in) The three micro-nuclei, (c.v.) Contractile vacuole. (C.B.) Ciliated ridges.

VOL. 61 , PART 3. NEW SERIES. 25


Fig. 12.—Lateral aspect of the bud, diagrammatic.Fig. 13.—Resting inicronucleus.Fig. 14.—Two micronuclei in early stage of division. See description

in text.Fig. 15.—Three micronuclei in later stage of division. See descrip-

tion in text.

G.L.d.l.

f.rw

M.

LAPAGE &, WADS WORTH— DENDROCOMETES.

cv.

i s . • •C L d e l - Huth,LitUr London

LAPAGE & WADSWOBTH — DENDROCOMETES.

Documents

Dendrocometes paradoxu (Stein)s . Part II.—Reproductio ...HISTORY or THE GENUS AND ANATOMY OF THE "ADULT" 341 IV. ... dish. Our method has been to accommodate some fourteen or fifteen