[VOL. LXVI, No. 1699

tic clarity and abounds in happy and stimulatingallusions. It ean not fail to suggest to every seriousreader numerous subjects for further observationaland theoretical investigation. In future editions thevalue of the book would be enhanced for the reviewer,and possibly also for other observational astronomersof slender mathematical attainments, by the interpo-lation of a chapter between III and IV giving a briefand as nearly as possible self-contained mathematicaldevelopment of the whole theory. Reference couldthen be made forward from this chapter to the sub-sequent chapters, which would retain much theirpresent form, for a detailed discussion of the variousassumptions, for the quadrature of the differentialequations, for the discussion of the opacity law and soon. I do not believe that such an interpolation wouldseriously interfere with the logical treatment of thetheory, and it is conceivable that it might eliminatesome of the many cross references, forwards and back-wards, which are essential with the present arrange-ment.

H. H. PLASKETTDOMINION ASTROPHYSICAL OBSERVATORY,

VICTORIA, B. C.

ARTIFICIAL TRANSMUTATION OFTHE GENE

MOST modern geneticists will agree that gene muta-tions form the chief basis of organic evolution, andtherefore of most of the complexities of living things.Unfortunately for the geneticists, however, the studyof these mutations, and, through them, of the genesthemselves, has heretofore been very seriously ham-pered by the extreme infrequency of their occurrenceunder ordinary conditions, and by the general unsue-cessfulness of attempts to modify decidedly, and in asure and detectable way, this sluggish "natural" muta-tion rate. Modification of the innate nature of or-ganisms, for more directly utilitarian purposes, hasof course been subject to these same restrictions, andthe practical breeder has hence been compelled to re-main content with the mere making of recombinationsof the material already at hand, providentially supple-mented, on rare and isolated occasions, by an unex-pected mutational windfall. To these circumstancesare due the wide-spread desire on the part of biolo-gists to gain some measure of control over the heredi-tary changes within the genes.

It has been repeatedly reported that germinalchanges, presumably mutational, could be induced byX or radium rays, but, as in the case of the similaipublished claims involving other agents (alcohol, lead,antibodies, etc.), the work has been done in such away that the meaning of the data, as analyzed fromii

a modern genetic standpoint, has been highly dispu-tatious at best; moreover, what were apparently theclearest cases have given negative or contrary resultson repetition. Nevertheless, on theoretical grounds, ithas appeared to the present writer that radiatiois ofshort wave length should be especially promising forthe production of mutational chanlges, and for thisand other reasons a series of experimrients concernedwith this problem has been undertaken during thepast year on the fruit fly, Drosophila melanonaster,in an attempt to provide critical data. The v ell-known favorableness of this species for genetic study,and the special methods evolved during the writec'seight years' intensive work on its mutation rate (in1-eluding the work on temperature, to be referred tolater), have finally made possible the finding of soinedecisive effects, consequent upoin the applicationi otfX-rays. The effects here referred to are truly muta-tional, and not to be confused with the well-kinowineffects of X-rays upon the distribution of the ebLo-matin, expressed by non-disjunction, non-inheritedcrossover modifications, etc. In the present condenseddigest of the work, only the broad facts and conelu-sions therefrom, and some of the problems raised, calbe presented, without any details of the genetic iiietli-ods employed, or of the individual results obtained.

It has been found quite conclusively that treatmenitof the sperm with relatively heavy doses of X-rays in-duces the occurrence of true "gene mutations" in ahigh proportion of the treated germ cells. Severalhundred mutants have been obtained in this way in ashort time and considerably more than a hundred ofthe mutant genes have been followed through three,four or more generations. They are (nearly all ofthem, at any rate) stable in their inheritance, andmost of them behave in the manner typical of tleiMendelian chromosomal mutant genes found in or-ganisms generally. The nature of the crosses M axsuch as to be much more favorable for the detectioniof mutations in the X-chromosomes than in the otherchromosomes, so that most of the mutant genes dealtwith were sex-linked; there was, however, am-lple proofthat mutations were occurring similarly throughoutthe chromatin. When the heaviest treatment was givenito the sperm, about a seventh of the offspring thathatched from them and bred contained individuallydetectable mutations in their treated X-chromosonie.Since the X forms about one fourth of the haploidchromatin, then, if we assume an equal rate of muta-tion in all the chromosomes (per unit of their lengtlh),it follows that almost "every other one" of the spernicells capable of producing a fertile adult containedan "individually detectable" mutation in some eliroriio-some or other. Thousands of untreated parent flieswere bred as controls in the same way as the treated

84 SC7IENCE

on

Aug

ust 7

, 200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

JuLY 22, 1927] 85

ones. Comparison of the mutation rates under thetwo sets of conditions showed that the heavy treatmenthad caused a rise of about fifteen thousand per cent.in the mutation rate over that in the untreated germcells.Regarding the types of mutations produced, it was

found that, as was to have been expected both ontheoretical grounds and on the basis of the previousmutation studies of Altenburg and the writer, thelethals (recessive for the lethal effect, though somewere dominant for visible effects) greatly outnum-bered the non-lethals producing a visible morpholog-ical abnormality. There were some "semi-lethals"also (defining these as mutants having a viability or-dinarily between about 0.5 per cent. and 10 per cent.of the normal), but, fortunately for the use of lethalsas an index of mutation rate, these were not nearly sonumerous as the lethals. The elusive class of "in-visible" mutations that caused an even lesser reduc-tion of viability, not readily confusable with lethals,appeared largr than that of the semi-lethals, but theywere not subjected to study. In addition, it was alsopossible to obtain evidence in these experiments forthe first time, of the occurrence of dominant lethalgenetic changes, both in the X and in the other chro-mosomes. Since the zygotes receiving these never de-veloped to maturity, such lethals could not be de-tected individually, but their number was so greatthat through egg counts and effects on the sex ratioevidence could be obtained of them en masse. It wasfound that their numbers are of the same order ofmagnitude as those of the recessive lethals. The "par-tial sterility" of treated males is, to an appreciable ex-tent at least, caused by these dominant lethals. An-other abundant class of mutations not previouslyrecognized was found to be those which, when hetero-zygons, cause sterility but produce no detectablechange in appearance; these too occur in numbersrather similar to those of the recessive lethals, andthey may hereafter afford one of the readiest indicesof the general mutation rate, when this is high. Thesterility thus caused, occurring as it does in the off-spring of the treated individuals, is of course a sepa-rate phenomenon from the "partial sterility" of thetreated individuals themselves, caused by the domin'antlethals.

In the statement that the proportion of "individu-ally detectable mutations" was about one seventh forthe X, and therefore nearly one half for all the chro-matin, only the recessive lethals and semi-lethals andthe "visible" mutants were referred to. If the domi-nant lethals, the dominant and recessive sterility genesand the "invisible" genes that merely reduce (orotherwise affect) viability or fertility had been takeninto account, the percentage of mutants given would

have been fax higher, and it is accordingly evidentthat in reality the great majority of the treated spermcells contained mutations of some kind or other. Itappears that the rate of gene mutation after X-raytreatment is high enough, in proportion to the totalnumber of genes, so that it will be practicable to studyit even in the case of individual loci, in an attack onproblems of allelomorphism, etc.Returning to a consideration of the induced muta-

tions that produced visible effects, it is to be noted thatthe conditions of the present experiment allowed thedetection of many which approached or overlappedthe normal type to such an extent that ordinarily theywould have escaped observation, and definite evidencewas thus obtained of the relatively high frequency ofsuch changes here, as compared with the more con-spicuous ones. The belief has several times been ex-pressed in the Drosophila literature that this holdstrue in the case of "natural" mutations in this or-ganism, but it has been founded only on "general im-pressions"; Baur, however, has demonstrated the trathof it in Antirrhinum. On the whole, the visible muta-tions caused by raying were found to be similar, intheir general characteristics, to those previously de-tected in non-rayed material in the extensive observa-tions on visible mutations in Drosophila carried outby Bridges and others. A considerable proportion ofthe induced visible mutations were, it is true, in lociin which mutation apparently had never been observedbefore, and some of these involved morphologicaleffects of a sort not exactly like any seen previoudy(e.g., "splotched wing," "sex-combless,"' etc.), but, onthe other hand, there were also nuimerous repetitonsof mutations previously known. In fact, the majorityof the well-known mutations in the X-chromosome ofDrosophila melanogaster, such as "white eye," "minia-ture wing," "forked bristles," etc., were reobtained,some of them several times. Among the visible muta-tions found, the great majority were recessive, yetthere was a considerable "sprinkling" of dominants,just as in other work. All in all, then, there can beno doubt that many, at least, of the changes producedby X-rays are of just the same kind as the "gene mu-tations" which are obtained, with so much greaterrarity, without such treatment, and which we believefurnish the building blocks of evolution.

In addition to the gene mutations, it was found thatthere is also caused by X-ray treatment a high propor-tion of rearrangements in the linear order of thegenes. This was evidenced in general by the frequentinherited disturbances in crossover frequency (at least3 per cent. were detected in the X-chromosome alone,many accompanied but some unaccompanied by lethaleffects), and evidenced specifically by various easesthat were proved in other ways to involve inversions,

SCIENCE

on

Aug

ust 7

, 200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

[VOL. LXVI, No. 1699

"deficiencies," fragmentations, translocations, etc., ofportions of a chromosome. These cases are makingpossible attacks on a number of genetic problemsotherwise difficult of approach.The transmuting action of X-rays on the genes is

not confined to the sperm cells, for treatment of theunfertilized females causes mutations about as readilvas treatment of the males. The effect is producedboth on o6eytes and early oogonia. It should benoted especially that, as in mammals, X-rays (in thedoses used) cause a period of extreme infertility,which commences soon after treatment and later ispartially recovered from. It can be stated positivelythat the return of fertility does not mean that the newcrop of eggs is unaffected, for these, like those matureeggs that managed to survive, were found in the pres-ent experiments to contain a high proportion of miu-tant genes (chiefly lethals, as usual). The practice,common in current X-ray therapy, of giving treat-ments that do not certainly result in permanent sterili-zation, has been defended chiefly on the ground of apurely theoretical conception that eggs produced afterthe return of fertility must necessarily represent "un-injured" tissue. As this presumption is hereby demon-strated to be faulty it would seem incumbent for medi-cal practice to be modified accordingly, at least untilgenetically sound experimentation upon mammals canbe shown to yield results of a decisively negative char-acter. Such work upon mammals would involve ahighly elaborate undertaking, as compared with theabove experiments on flies.From the standpoint of biological theory, the chief

interest of the present experiments lies in their bear-ing on the problems of the composition and behaviorof chromosomes and genes. Through special geneticmethods it has been possible to obtain some informa-tion concerning the manner of distribution of thetransinuted genes amongst the cells of the first andlater zygote generations following treatment. It isfound that the mutation does not usually involve apermanent alteration of all of the gene substance pres-ent at a given chromosome locus at the time of treat-ment, but either affects in this way only a portion ofthat substance, or else occurs subsequently, as anafter-effect, in only one of two or more descendantgenes derived from the treated gene. An extensiveseries of experiments, now in project, will be neces-sary for deciding conclusively between these two pos-sibilities, but such evidence as is already at handspeaks rather in favor of the former. This wouldimply a somewhat compound structure for the gene(or chromosome as a whole) in the sperm cell. Onthe other hand, the mutated tissue is distributed in amanner that seems inconsistent with a general appli-cability of the theory of "gene elements" first sug-

gested by Anderson in connection with variegatedpericarp in maize, then taken up by Eyster, and re-cently reenforeed by Demeree in Drosophila virilis.A precociously doubled (or further multiplied) con-

dition of the chromosomes (in "preparation" for latermitoses) is all that is necessary to account for theabove-mentioned fractional effect of X-rays on a giveiilocus; but the theory of a divided condition of eachgene, into a number of (originally identical) "ele-ments" that can become separated somewhat indeter-minately at mitosis, would lead to expectations differ-ent from the results that have been obtained in thepresent work. It should, on that theory, often havebeen found here, as in the variegated corn and theeversporting races of D. virilis, that mutated tissuegives rise to normal by frequent "reverse mutation";moreover, treated tissues not at first showing a muta-tion might frequently give rise to oae, through a "sort-ing out" of diverse elements, several generations aftertreatment. Neither of these effects was found. Ashas been mentioned, the mutants were found to bestable through several generations, in the great ma-jority of cases at least. Hundreds of non-mutateddescendants of treated germ cells, also, were carriedthrough several generations, without evidence appear-ing of the production of mutations in generations sub-sequent to the first. Larger numbers will be desirablehere, however, and further experiments of a differenttype have also been planned in the attack on thisproblem of gene structure, which probably can be an-swered definitely.

Certain of the above points which have already beendetermined, especially that of the fractional effect ofX-rays, taken in conjunction with that of the produc-tion of dominant lethals, seem to give a clue to theespecially destructive action of X-rays on tissues inwhich, as in cancer, embryonic and epidermal tissues,the cells undergo repeated divisions (though the oper-ation of additional factors, e.g., abnormal mitoses,tending towards the same result, is not thereby pre-cluded); moreover, the converse effect of X-rays, inoccasionally producing cancer, may also be associatedwith their action in producing mutations. It wouldbe premature, however, at this time to consider in de-tail the various X-ray effects previously considered as"physiological," which may now receive a possible in-terpretation in terms of the gene-transmuting prop-erty of X-rays; we may more appropriately confineourselves here to matters which can more strictly bedemonstrated to be genetic.

Further facts concerning the nature of the genemay emerge from a study of the comparative effectsof varied dosages of X-rays, and of X-rays adminis-tered at different points in the life eycle and undervaried conditions. In the experiments herein re-

86 SCIENCE

on

Aug

ust 7

, 200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

JULY 22, 1927]

ported, several different dosages were made use of,and while the figures are not yet quite conclusive theymake it probable that, within the limits used, the num-ber of recessive lethals does not vary directly with theX-ray energy absorbed, but more nearly with thesquare root of the latter. Should this lack of exactproportionality be confirmed, then, as Dr. IrvingLangmuir has pointed out to me, we should have toconclude that these mutations are not caused directlyby single quanta of X-ray energy that happen to beabsorbed at some critical spot. If the transmuting5effect were thus relatively indirect there would be agreater likelihood of its being influenceable by otherphysico-chemical agencies as well, but our problemswould tend to become more complicated. There is,however, some danger in using the total of lethal mu-tations produced by X-rays as an index of gene muta-tions occurring in single loci, for some lethals, involv-ing ehanges in crossover frequency, are probably asso-ciated with rea.rrangements of chromosome regions,and such changes would be much less likely than"point mutations" to depend on single quanta. A re-examination of the effect of different dosages musttherefore be carried out, in which the different typesof mutations are clearly distinguished from one an-other. When this question is settled, for a wide rangeof dosages and developmental stages, we shall also bein a position to decide whether or not the minuteamounts of gamma radiation present in nature causethe ordinary mutations which occur in wild and in cul-tivated organisms in the absence of artificially admnin-istered X-ray treatment.As a beginning in the study of the effect of varying

other conditions, upon the frequency of the mutationsproduced by X-rays, a comparison has been made be-tween the mutation frequencies following the rayingof sperm in the male and in the female receptacles,and from germ cells that were in different portions ofthe male genital system at the time of raying. No de-cisive differences have been observed. It is found, inaddition, that aging the sperm after treatment, beforefertilization, causes no noticeable alteration in the fre-quency of detectable mutations. Therefore the deathrate of the mutant sperm is no higher than that ofthe unaffected ones; moreover, the mutations can notbe regarded as secondary effects of any semi-lethalphysiological changes which might be supposed tohave occurred more intensely in some ("more highlysusceptible") spermatozoa than in others.

Despite the "negative results" just mentioned, how-ever, it is already certain that differences in X-ray in-fluences, by themselves, are not sufficient to accountfor all variations m mutation frequency, for the pres-ent X-ray work comes on the heels of the determina-tion of mutation rate being dependent upon tempera-

ture (work as yet unpublished). This relation hadfirst been made probable by work of Altenburg andthe writer in 1918, but was not finally established untilthe completion of some experiments in 1926. Thesegave the first definite evidence that gene mutation maybe to any extent controllable, but the magnitude of theheat effect, being similar to that found for chemicalreactions in general, is too small, in connection withthe almost imperceptible "natural" mutation rate, forit, by itself, to provide a powerful tool in the muta-tion study. The result, however, is enough to indicatethat various factors besides X-rays probably do affectthe composition of the gene, and that the measurementof their effects, at least when in combination withX-rays, will be practicable. Thus we may hope thatproblems of the composition and behavior of the genecan shortly be approached from various new angles,and new handles found for their investigation, so thatit will be legitimate to speak of the subject of "genephysiology," at least, if not of gene physics and chem-istry.In conclusion, the attention of those working along

classical genetic lines may be drawn to the opportu-nity, afforded them by the use of X-rays, of creatingin their chosen organisms a series of artificial racesfor use in the study of genetic and "phaenogenetic"phenomena. If, as seems likely on general considera-tions, the effect is common to most organisms, it shouldbe possible to produce, "to order," enough mutationsto furnish respectable genetic maps, in their selectedspecies, and, by the use of the mapped genes, to ana-

lyze the aberrant chromosome phenomena simultane-ously obtained. Similarly, for the practical breeder,it is hoped that the method will ultimately prove use-

ful. The time is not ripe to discuss here such possi-bilities with reference to the human species.The writer takes pleasure in acknowledging his sin-

cere appreciation of the cooperation of Dr. DaltonRichardson, Roentgenologist, of Austin, Texas, in thework of administering the X-ray treatments.

H. J. MULLERUNIVERSITY or TEXAS

SCIENTIFIC APPARATUS ANDLABORATORY METHODS

AN INSTRUMENT FOR REPEATED DETER-MINATIONS OF BLOOD VISCOSITY

IN AN ANIMAL'IN experiments where it is desirable to make re-

peated determinations of the viscosity of the bloodof an animal, the withdrawal of the amount of blood

1 From the Physiological Laboratories of the Universityof Chicago and the University of Western Ontario, Lon-don, Canada.

SCIENCE 87

on

Aug

ust 7

, 200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from