The American Naturalist MSS intended for p publication and books, ete., intended for review should be sent to the Editor of THE AMERICAN NATURALIST, Garrison-on-Huds on, New York. Articles containing research work bearing on the problems of organic evolu- tion are especially welcome, and will be given ee nd pu tion f authors free of charge. T ee One hund Further w hy a supplied at cost. Subscriptions and advertisements should be sent to the Fag ager! he ; Subscription aipa is four dollars a year. Foreign is Can twenty-five cents additional. The charge for single poii = thirty-five cents. The savertiiine rates are Four Dollars for a page. sce ees THE SCIENCE PRESS Lancaster, Pa. Garrison, N. Y. rete _ NEW YORK: Sub-Station 84 , Pa., under the Act of : PT Congress of Maron 3, 1879. ; m BULLETIN—For bargains i in Piinoak ii ical and Pre-historic Specimens. Books on Natural shee Comme || Fifty Years of Darwinism pate ; Comprising the eleven addresses in honor : — < ed - ? =a =|} of Charles Darwin delivered before the ee ong a American Association for the Advance- : “ment of Science. = > pp- $2.00, et “Henry | Holt & eee pany 34 West 33d St., New York 378 Safco: sehen veneer THE AMERICAN NATURALIST THE We AMERICAN NATURALIST A MONTHLY JOURNAL DEVOTED TO THE ADVANCEMENT OF THE BIOLOGICAL SCIENCES WITH SPECIAL REFERENCE TO THE FACTORS OF EVOLUTION VOLUME XLV Mo. Bot, Garden 1912 NEW YORK THE SCIENCE PRESS IQII THE AMERICAN NATURALIST VoL. XLV : January, 1911 No. 529 ORGANIC RESPONSE! D. T. MACDOUGAL, Px.D., LL.D. Desert LABORATORY, Tucson, ARIZ. Ar no time in the history of natural science has such a large share of thought and research energy been di- rected to the solution of evolutionary problemsas at pres- ent. Methods of work, plans for experimentation and modes of interpretation have recently undergone such rapid development and improvement that our potenti- ality for solving questions in heredity and origination is vastly greater than even at such recent date as the be- ginning of this new century. With increased facility in attack has also come wider vision and altered view- points with regard to almost all phases of biology. Biological thought once quickened and broadened by evolutionary ideas was by this same means led to be- come entangled in a maze of illusive assumptions as to purpose and plan in organisms from which it is being but slowly freed, to view functions as inevitable reactions, however complex they may be. The variables included in the equations of protoplasmic action are numerous and large, but they do not exceed the undefined prin- ciples of osmotic action, surface tension and unknown phases of association and dissociation that are concerned in the interplay of substances in the cell, and upon which 1 Presidential sigan Society of American Naturalists, Ithaca, New York, December 29, 5 6 THE AMERICAN NATURALIST [ Vot. XLV depend the chemico-physical relations of tissue compo- nents and structures of all kinds. If physiology escapes the soporific and deadening influence of the vitalistic con- ceptions, now appearing in some profusion, it may in turn furnish the secure means for a long and rapid advance in genetics, and it may be assumed with some certainty that the chief superstructures of evolutionary science will be those securely raised upon a foundation of physiolog- ically tested facts. In taking this direction, natural history is not alone; the briefest comprehensive view of the physical sciences will show that here also the chief advance lies along the way of the study of energetics, and that the fundamental problems are those lying about the mode and means of transformations of energy. Recent events in the field of evolution comprehend a number of movements and accomplishments of extraor- dinary interest. The rediscovery of the facts of alter- native inheritance, the formulation of the concepts of equivalent, balanced, paired or differential characters, the results of statistical studies of variability, the analyses of species of various constitution by pedigree cultures, in which the value of fertilization from various sources is carefully measured, the distinction of the biotype or geno- type as a hereditary entity, the possibilities in the ac- tion of pure lines within a specific group, the cytological contributions of fact and forecast upon the physical as- pects of heredity, and lastly the presentation of the facts and allowable generalizations identified with the muta- tion theory, comprise a series of advances, of accretions to knowledge, furnish a broadened foundation for biolog- ical science, and disclose additional possibilities in all lines of experimental research with living things, besides open- ing up new realms for speculative thought, and stimulat- ing the scientific imagination to renewed fruitfulness. Biological literature has also been recently enriched by a series of formal papers commemorative of the life and work of Charles Darwin, by more than fourseore workers No. 529] ORGANIC RESPONSE A representing the laboratories and national cultures of the world. This group of addresses and essays, fortu- nately written chiefly within four languages, taken col- lectively, constitutes a critical and evaluatory discussion of the mass of fact and galaxy of theory concerning or- ganic evolution, and furnishes the most complete and thorough appraisal ever made of any subject in modern biology. The moment, therefore, is one of consciousness of achievement, of realization of increased powers of pene- tration, and charged with desire for the exploitation of the unknown, and is vibrant with the inspiration coming from such a rapid march of events. With this quick- ening in activity, the outeries of acrid controversies no longer monopolize our attention, but it must not be sup- posed that differences of opinion have vanished from among us. The agreement as to the value of methods of experimentation and calibration is a most gratifying fact, but the harmonies of opinion as to interpretation of results have not yet come to a monotone. On the contrary, the pressure of new and undisciplined evidence has awakened a freshened chorus of voices cry- ing the virtues of special interests and extolling the suff- ciency of theories dignified by age and more or less weighty with authority. Those busy with vitalism of various patterns have spun a moiety of their favorite fabric to mend the breaks in the fragile web made by the impact of new facts. Isolation and the mechanism of geographical distribution have again been elaborated to account for all differentiation and what their exponents are pleased to term speciation. The anticipatory forma- tion of structures in a rudimentary condition with a long prefunctional progress, guided by the morphological pos- sibilities and actuated by internal impulses, has again been offered to us, fortified by paleontological fact and clever logic, in such manner as to avoid most of the serious objections to orthogenesis except those of physiological morphology. 8 THE AMERICAN NATURALIST [Vou. XLV Natural selection with diverse meanings and manifold implications has been made to explain development, dif- ferentiation and general evolutionary progress. The tu- mult'is greatest at the present time, however, about the idea of mutation. Standing to one side, the biologist hears a medley of assertions ‘‘that mutations have long been known,” ‘‘do not exist,’’ ‘‘were discovered by Dar- win,’’ ‘‘are always an evidence of hybridization,’’ ‘‘re- sult in the formation of nothing but elementary species,”’ ‘sive only weakened derivatives that are quickly swamped by parental forms,’’ ‘‘are encountered only among cultivated plants,’’ ‘‘the mutation theory is based upon the conception of unit characters,’’ ‘‘constitutes the only adequate means of accounting for the enormous number of living forms and myriad characters of living things,’’ ‘‘unit characters are unreal, have never been seen, do not exist and are incapable of demonstration.’’ ‘‘The difference between mutation and variation is one of amplitude only,’’ and lastly mutation signally ‘‘refutes Darwinism,’’ and ‘‘swings us back in harmony with the theologian’s arguments for special creation.’’ The absurdity of the many injudicial assertions by the partisans concerned need not blind us to the stubborn fact that saltatory changes do occur in hereditary pure lines in a large number of forms in both plants and animals. Observations and experiments have estab- lished beyond doubt that mutation is one way by which organisms bearing new combinations of qualities may arise, although it is probable that its importance as a general procedure varies in different groups of organisms and certain that many shades of opinion as to its exact part in the evolution of living things will always be held. Our appraisement of the value of all the protheses cited may also be amended from time to time with view- points altered by the advance of knowledge. The situa- tion with regard to one hypothesis is far more serious, however. This is the theory which predicates direct adaptational adjustment of the organism, quickly or No. 529] ORGANIC RESPONSE 9 slowly as the case may be, to environie factors, and the inheritance of the somatic alterations constituting such variations. The various corollaries of this theory have the force of a certain obviousness, its assumptions have been of ready service to the systematist and bio- geographer, and its conclusions have long been tolerated in the absence of decisive tests which are not to be easily made or readily carried out. The time has now arrived, however, when the claimants for Neo-Lamarckianism and all of its conclusions must show cause for its further con- sideration, or else allow it to drop from the position of being seriously taken as a method of evolutionary advance. It is unanimously agreed that organisms, plants as well as animals, change individually in aspect, in form and structure of organs, in functionation and habit as they encounter swamps, saline areas, gravelly uplands or slopes, climatic differences identifiable with latitude or elevation, and other physical and biological factors. It is assumed that these somatic alterations are accommodative and adaptive, making the organism more suitable for the conditions which produce the changes. Such an assump- tion is an over-reaching one. Any analysis of the changes which an organism undergoes after transportation to a new habitat will disclose one or a few alterations which might be of advantage in dealing with the newly encoun- tered conditions, but with these are many others, direct, necessitous, atrophic, or hypertrophic as to organs which have no relation whatever to usefulness or fitness. Fur- ther, a critical examination fails to disclose any theoret- ical considerations or any actual facts which would con- nect inevitably the somatic response with the nature of the excitation, outside of the specialized tropisms in which specific reactions are displayed. Even in these the adjustment is of such nature that a mechanism spe- cially perceptive to contact, for example, may react to changes in temperature, as illustrated by the action of tendrils, and many similar cases might be cited. It is evident that the soma of a plant or animal is not to be 10 THE AMERICAN NATURALIST [ Vou. XLV considered as capable of adaptive alterations to every new agency which may cause changes in its form, structure or functionation. Next we come to the very crux of the whole matter: do the unusual forms or activities of organs resulting from environic causes act in any manner upon the germ-plasm connected with such altered bodies? If we are to con- sider the activities of the organism or of the cell to de- pend mainly upon its chemical structure and constituency and such a generalization seems unavoidable, then we have means by which the soma might cause its proper- ties to be reflected from the germ-plasm in a succeeding generation, since the chemical mechanism of the soma and germ-plasm must be of the most intimate nature. That some such connection does actually exist is strongly suggested by the behavior of a great number of organ- isms which have been seen to carry marked environic effects to the second or even third generations; if the interrogation be made as to why the induced qualities are earried no further it may be said that the reply may be suggested by the results of long-continued action of the exciting agency, such as has been used by Woltereck with. Daphnia. If a general view be taken of the available information of interest in this connection, three classes of facts will be discerned. One group is comprised in the mass of in- formation obtained by the operations of the horticul- turist, the agriculturist and the breeder as to the behav- ior of crops, plants and domestic animals, when trans- ferred from one habitat to another. The greater part of such data is the result of observations which do not com- ply with the ordinary requirements in the avoidance of error so that strict comparisons as to the behavior of organisms under conditions of various habitats are im- possible. A consideration of the literature yields many suggestions for experimental research and the simple generalization that the direct effects of climatic com- plexes on the seasonal cycle, and upon color, or struc- No. 529] ORGANIC RESPONSE 11 tural features of the individual, may be repeated or car- ried over two or three generations, in a habitat where the specific causal combinations are lacking. This is the available total of knowledge furnished us by economic operations, and by the introduction operations of botan- ical gardens and plantations. In contrast with these the fortunate experience of Zed- erbauer with Capsella has yielded some conclusions of exceptional importance. A genotype of Capsella Bursa- pastoris resembling taraxicafolium was found on the | lower plains of Asia Minor, and displayed the well- known characters of this form, including broad leaves, whitish flowers, and stems 30-40 em. high. A highway leads from these regions to a plateau at an elevation of 2,000 to 2,400 meters. The conditions of distribution are such as to indicate that the plant has been carried up this thoroughfare by man, and in this elevated habitat it has taken on certain alpine characters, including elon- gated roots, xerophytic leaves, stems 2-5 em. high, red- dish flowers, with a noticeable increase of the hairiness of the entire plant. That the distributional history has been correctly apprehended seems entirely confirmed by the fact that when seeds are taken from the lowlands the alpine characters enumerated are displayed at once as a direct somatic response. When seeds are taken from plants on the elevated plateau where their ancestors may have been for many years or many centuries (perhaps as , long as 2,000 years) and sowed at Vienna and in other cul- tures carried through four generations the leaves lose their xerophytie form and structure, but the other charac- ters are retained within the limits of variability. The stems show an increase in average length of 1 or 2 em., the roots change as much, but the reproductive branches and floral organs retain their alpine characters. The slight modifications undergone by these features were seen to reach a maximum and to decrease in the latest generations cultivated. The structural changes and im- plied functional accommodations are indubitably direct 12 THE AMERICAN NATURALIST [Vou. XLV somatic responses, there is no escape from the conclu- sion that the impress of the alpine climate on the soma has been communicated to the germ-plasm directly or indirectly in such manner as to be transmissible, and the suggestion lies near that repeated and continued excitation by climatic factors may have been the essential factor in such fixation.” Among the most noteworthy investigations of the fea- tures of interest in connection with habitat changes are those being made by the anthropologist in which somatic calibrations of immigrating races and linguistic studies _ of peoples of known origin, geographical movement, and established relationship are being used to great advan- tage. No more fascinating chapters of scientific litera- ture are to be found than those which delineate the migra- tory movements, segregation and habital reactions of Polynesian islanders, of North American Indians, or of Asiatie peoples, yet their value as actual contributions to the phase of biology of interest to this society is hardly recognized. The investigator of problems in anthropol- ogy has the advantage of dealing with an animal whose psychology, history, traditions and records are readily intelligible to him, so that a much wider range of facts may be brought within the zone of reliability than when we deal with an organism whose actions, at best, are but imperfectly understood by us.? A second series of results of great interest and suggest- iveness are those which have been obtained in various laboratories as to the individual modifications in cyclical activity, functionation and structure of plants and ani- mals in response to unusual stimuli, or under the influ- ence of unusual intensities of the common environic com- ponents. The behavior of organisms in constant illumi- nation, equable and variable temperatures, salinity, alka- linity or acidity of the medium, unusual pressures of at- ***Versuche ueber Vererbung erworbener Eigenschaften bei Capsella — pastoris,’’ ait r. Bot. Zeitschr., Vol. 58: pp. 231-236, 285-288, 1908. e Boas, F., ‘‘Changes in Bodily Form of Descendants of Immi- ar The Tmimnigration Committee, Document No. 208, presented to the 61st Congress, 2d Session, Washington, D. C., U. B. A., 1910. No. 529] ORGANIC RESPONSE 13 mospheric constituents, to unusual compounds and unac- customed food-material, make up an important propor- tion of the sum total of information ordinarily classified as physiology. The morphogenic and accommodative adjustments presented afford by analysis the best con- ceptions available as to the nature of the physiologic activity of organisms. The experimental results of Stockard with fish eggs subjected to the action of various chemical substances are of unusual interest in the present connection. The eyclopean embryos of Fundulus formed in sea-water con- taining magnesium salts offer the first known example of the induction of an abnormality in the vertebrates occurring in nature, by specific reagents. Suggestion of a common cause is obvious as it is in the instances in which similar divergences have been secured in the labo- ratory with plants. As will be pointed out later, such analytical tests constitute a very important part of the procedure in the study of acclimatization results.‘ In very few cases, however, has the permanency or heritability of the deviations induced been tested, and in most of such tests the agencies employed might ha acted upon both soma and germ-plasm, as will be ap- parent upon an examination of the work of Standfuss, Fischer, Pictet and Houssey. The work of these older experimenters has been reviewed so many times that it will be unnecessary to discuss their results further in the © present paper. This was done at the Darwin memorial meeting in 1908, and quite recently by Bourne in his ad- dress before Section of Zoology of the British Associa- tion for the Advancement of Science, at the Sheffield meeting.” The present opportunity may well be used to make a presentation of the results of the last few years obtained “Stockard, ©. R., ‘‘The Development of Artificially Produced Fish.— The Cyclopean Embryo,’’ Jour. Exper. Morphology, Vol. 7, No. 2, p. 285, 1909 * Nature, Vol. 84, p. 378, 1910, September 22, 1910. 14 THE AMERICAN NATURALIST [ Vou. XLV by investigations, using a more perfected technique, and having the advantage of a keener insight into the real nature of the problems to be solved. That the general hypothesis with its corollaries is being subjected to the most critical examination and that the assumptions implied in the conception of inheritance of acquired characters are being put to exact and conclu- sive tests, is readily apparent when a review is made of recent and current researches in which living material from widely separated groups of animals and plants is being subjected to a variety of nutritive conditions and climatic agencies. Klebs, who has long been concerned with the morphogenic reactions of plants, has determined a series of conditions under which the stages of mycelial development, asexual zoospore and sexual or oospore formations in filamentous fungi may be inhibited or var- iously interchanged. Much more important reactions were obtained from Sempervivum, the live-forever of tae garden. In this plant, inflorescences were replaced by sin- gle flowers by experimental excitation while it was found the number and arrangement of the floral organs as well as of the stamens and carpels could be altered. Further- more, the deviations in question were found to be trans- missible to the second or third generation in guarded seed- reproductions.°® Microorganisms with a short cycle offer peculiarly ad- vantageous material by reason of their simple reproduc- tive processes, and also by the fact that it is possible to control environic factors with exactitude. The volu- minous literature of bacteriology shows that much at- tention has been devoted to the building up of characters by selection, and to the study of the behavior of morpho- logical divergences occurring in special cultures. The experiments of Buchanan with Streptoccus lacti- cus yields the conclusion that phases of fluctuating varia- tions in the bacteria induced by cultures may not be fixed, ° Alterations in the development and forms of plants as a result of environment, Proe. Roy. Soe. Lond., Vol. 82, No. B. 559, p. 547, 1910. No. 529] ORGANIC RESPONSE 15 and are not transmissible, which is in accord with the main body of evidence upon this point. There are, how- ever, a number of records of the appearance of definite qualities or morphological characters in the yeasts, which were transmissible and permanent. These de- partures were so striking as to be capable of being re- garded as mutational, and their origin has been ascribed to the influence of the environment by experimenters of notable skill, such as Beijerinck, Winogradsky, Lepesch- kin, Hansen and Barber. It may be recalled in this con- nection, that environic responses are generally sudden, and that the entire range of departure may be made in a single generation, at most in two or three.’ Pringsheim after a comprehensive review of his own work and of other available evidence obtained by a study of accommodations or adaptations of yeasts and bacteria to unusual temperatures, culture media, and poisons, concludes that some of these variations are fixed and transmissible both asexually and by spores, while others are not. It is not easy to analyze contributions upon this subject with reference to the differential action of the exciting agencies upon soma or germ-plasm, neither is it clear as to the action of the selection in the experimentation. It is important, however, to note that the alterations concerned are direct functional responses to the exciting agencies.’ The researches of Jennings with paramecium deals with conditions of morphology and physiology not widely dissimilar from those offered by the bacteria with regard to the present problems, and his work has been carried out with an extensiveness and thoroughness impossible to the worker with more massive and more slowly moving organisms. Cultures were carried through hundreds of generations with no progressive action in fluctuating "For a brief review of this subject see Buchanan, ‘‘Non-inheritance of impressed variations in Streptoccus lacticus,’’ Journal of Infectious Dis- eases, Vol. 7, p. 680, 1910. * Pringsheim, H., ‘‘Die Variablitit niederer Organismen,’’ Berlin, 1910. 16 THE AMERICAN NATURALIST [ Vou. XLV variability; while the organism as a whole was strongly resistant to all kinds of environic influences, and actual alterations were extremely rare. Most of the supposedly acquired characters disappeared in two or three gener- ations by fission, although one was followed for twenty- two generations. The new é¢haracter was borne by only one of the pair produced by a division, except in rare instances, and in only one case was there found such modification as to produce a race bearing the odd char- acter in which the feature in question was imperfectly transmitted in series of asexual generations.’ The results of Woltereck with Daphnia offer some- thing by way of contrast and also serve to illustrate the necessity for continuation of parallel cultures for the purpose of comparison of divergent forms and the nor- mal. The particular group of this crustacean furnish- ing the experimental material is taken to be very var- liable, and it was subjected to over-feeding with the im- mediate result that the variability of the form of the head appeared to be widened, the size of this structure being increased. This disappeared when lots from the culture ‘were restored to normal conditions in the earlier stage of the work. After three or four months of over-feeding, the form of the head came within narrower limits, and fewer aberrants were seen, while lots returned to normal conditions, showed a slower restoration of the original form of the head. Two years after the cultures were begun, it was found that the original head form was not displayed by young restored to normal nutrition condi- tions, the larger helmet being persistent. It seems fairly certain that a new genotype resulted from the long- continued action of the culture medium. * Jennings, H. $S., ‘‘ Heredity and Variation in the Simplest Organisms,’’ AMER. Nart., Vol. 43, No. 510, June, 1909; and other papers by the same author. ” 6 s DOUBLE A No. 529] ‘Ex ‘38a Sutsojoul ey Jo yok ay} sp; 33ə pasojourt ay} Jo SulA| SSEW YAVp oy, ‘palvededd A[Ava[D Ba oy JO s}uəzuo2 oy] pur uədo nə səurıquəow Joq SMOYS SULE 'E DIA ‘Ex 's}jsodəp yons woaz ƏƏ Áp uə st ‘favajquoo əy} Uo ‘auvaAqUIeU JeuU, Əy} “S39 ay} JO puə Je[;eUIS ay} uo A[[vPVedse ‘Swpy JO S}Isodep SMOYS sUBAQUIBU J9INO VY} AYM RY} VION ‘VUBIqMIBUT JOUUT OY} [BeAII 0} SB OS YOR pauANn} ə ey} JO Ulsivw aeddn əy} pug əugaquəw Ja}NO Ə} Up epvuUl Used PRY UOIS|PUT Uv 194ze UMOYS s; 339 əy} Yduasojoyd s) ul ‘Z ‘PIA ‘$x +830 oy} Jo ydvaSojoyd y `I IA qJ9, 284} 04 56 THE AMERICAN NATURALIST [ Vou. XLV peculiarities that have not appeared in any previous ac- count, seems to be worthy of record. It should be kept in mind, however, that the value of such descriptions does not come from the morphological features revealed, no matter how bizarre these may be, but from the fact that considerable light is often thereby shed on the complex physiology of the reproductive organs, and it is with this idea in mind that the following record is made. The egg in question, which was kindly handed to me by Mr. S. L. Pinckney, Austin, Texas, was laid March 28, 1910. It was stated that several double eggs had been received from the flock from which this egg came, but whether they were all laid by the same hen could not be ascertained. The egg was large, measuring 85 mm. in its long axis and 62 mm. in the short axis, and was slightly smaller at one end than at the other, so that we may speak of the blunt and pointed ends. It was prac- tically a soft-shelled egg, in that the amount of lime de- posited on the shell-membrane was very small, and for the most part was collected into little nodules scattered about over the surface (Figs. 1 and 2). A microscopical examination of the shell-membrane did not reveal any- thing unusual, for it consisted of the two characteristic layers, a thick outer and a thinner inner; but on cutting it open I was surprised to find another shell-membrane lying almost directly beneath it (Fig. 2). The two mem- branes were separated by the very thinnest layer of watery albumen. This second or inner membrane was in every way normal, and perfectly white, but was entirely void of lime deposits, reminding one very much in its gen- eral appearance of the membranes on eggs which have just reached the isthmus. The contents of the inner shell-membrane consisted of much albumen in which were imbedded a hard-shelled egg anda yolk (Fig. 3). Upon examination the inclosed egg was found to be perfectly normal in every respect, and its yolk contained a healthy blastoderm. The in- closed yolk, although normal in structure, was much dis- No. 529] A DOUBLE HEN’S EGG 57 torted, owing to the pressure exerted upon it by the ap- proximation of the hard-shelled egg. The albumen closely adhered both to the egg and to the yolk, but much of it was of a liquid nature, as was indicated by the ease = with which it flowed out of the cut first made in the inner membrane. The accompanying diagram will make clear the rela- tion of the various parts of this interesting monstrosity (Fig. 4). The inclosed egg lies toward the pointed end 4. Diagram of a median section of the egg. i.s.m., inner shell-mem- brane, the two lines representing its two layers; 0.s.m., outer shell-membrane; 8, Shell of the inclosed egg; y, yolk of the inclosed egg; y’, yolk of the inclosing egg. Natural size. of the inclosing egg, and its long axis meets the corre- sponding one of the double egg at an oblique angle. On account of this inclination of the inclosed egg its pointed end lies nearer to the blunt than to the pointed end of the inclosing egg. The inclosed yolk occupies the blunt end of the inclosing egg and is considerably distorted by pressure. The chalaze are but poorly developed, but the axis formed by a line passing through their points of attachment to the vitelline membrane approximately coincides with the long axis of the inclosing egg, showing that the yolk has maintained its original orientation. 58 THE AMERICAN NATURALIST [ Vor. XLV As I have pointed out above, the most interesting ques- tion regarding this egg pertains to the physiology of its formation. Parker states that two hypotheses have been advanced to explain how inclosed double eggs are formed. According to one of these, which was first advocated by Panum,* the inclosed egg remains in the distal part of the oviduct until overtaken by a second one, when both are then surrounded by a common envelope; according to the other a completely formed egg is carried by antiperi- stalsis back up the oviduct, where it meets a second one, and the two passing down become covered by a second shell and are laid. It seems quite evident from the de- scription of the egg just given that it is the product of antiperistalsis, but the especial interest lies in the fact that this process has taken place twice. The first antiperistalsis took place immediately after the hard-shelled egg was formed, and of course caused its migration to the upper or proximal end of the oviduct where it met the second egg. This meeting must have taken place very close to the infundibulum, for otherwise the yolk of the second egg would have possessed much larger chalaze. The second antiperistalsis occurred immediately after the inner of the two shell-membranes had been laid down, and must have succeeded in carrying the double egg up the oviduct to a point where albumen is secreted, that is, to a place slightly above the beginning point of the isth- mus; for it is only on this assumption that we are able to explain how a thin layer of albumen came to exist be- tween the two shell-membranes. The small amount of lime deposited on the outer of the two shell-membranes indicates that the egg did not remain long in the uterus, but must have been laid shortly after having entered that organ. In many respects this egg conforms to the facts already seen in the inclosed types of double eggs; thus the in- * Untersuchungen über die Entstehung der Missbildungen zunächst in den Eiern der Végel,’’ by P. L. Panum, Berlin, 1860. No. 529] A DOUBLE HEN’S EGG 59 closed egg lies near the pointed end of the inclosing one, and it was laid during the time of year when such eggs most frequently appear, that is, in the winter or spring; but it differed in one rather important respect. The pointed end of the inclosed egg does not lie in the same direction as that of the inclosing one. This unusual position of the inclosed egg doubtless has been brought about by crowding, and does not indicate necessarily that it was at first incorrectly oriented. Among the more important things so far revealed by a study of inclosed double eggs is the light thrown on the problem of the orientation of the egg in the oviduct, a problem in which the writer has been deeply interested. These eggs clearly demonstrate that when an egg has once entered the oviduct its original orientation in that organ is maintained during the formation of the enve- lopes, no matter to what extent it may have been moved up and down the reproductive passage. This fact strongly supports the conclusion reached by the writer‘ in a recent contribution, in which it was pointed out that the definite orientation of the egg in the reproductive passage is not a matter of chance, but is something that is handed on to the oviduct by the ovary, that is to say that the ova in the ovary have a definite polarity which is passed on to the oviduct through the mechanism of the infundibulum. *‘‘The Early Development of the Hen’s Egg, I., History of the Early Cleavage and of the Accessory Cleavage,’’ by J. Thomas Patterson, Journal of Morphology, Vol. 21, 1910. NOTES AND LITERATURE HEREDITY One of the most important papers relating to heredity that has appeared in recent months is that of Tower, dealing with hybridization investigations with species of the genus Leptino- tarsa. I shall not here attempt an extensive review of this paper, but mention it rather as a means of calling attention to its importance and suggesting that any one interested in theo- retical discussions of heredity should not fail to read it. Tower has done an immense amount of work with this genus. His re- sults lead him to accept the factorial hypothesis as an explana- tion of Mendelian phenomena but to discard wholly the de Vriesian interpretation of these factors. The most important contribution in this paper is the apparent fact that the environ- ment at the time when eggs are fertilized may change very ma- terially the nature of the hereditary factors. It is unfortunate that the author does not give more details in connection with this conclusion. The data he does give are mixed and contra- dictory. There has possibly been an error in printing Tower’s paper, but if not there was a serious error in its preparation, as will be seen from the following. In experiment 409, which was several times repeated, the results were exactly as if one of the parents had been a heterozygote between the two species. F, con- sisted of two types, one of which was identical with the female parent and the other intermediate between the two. The one like the female parent bred true to that type, while the other behaved in all respects as a heterozygote. In experiment 410- the same two species were utilized, but the temperature and humidity conditions at the time the eggs were fertilized were made quite different. This experiment, which was repeated eleven times, gave in every case ordinary Mendelian phenomena. In experiments 409/411, which was performed seven times, one set of eggs from the same cross as above was produced under conditions identical with those of experiment 409, and in * Tower, Wm. L., ‘‘ The Determination of Dominance and the Modifica- tion of Behavior in Alternative (Mendelian) Inheritance, by Conditions Surrounding or Incident upon the Germ Cells at Fertilization,’’ Biol. Bull., XVIII, No. 6, May, 1910 60 No. 529] NOTES AND LITERATURE 61 each case gave identical results with those of 409; that is, F, consisted of two types, one heterozygote, and the other homozy- gote of the maternal type. Using the same individuals which produced a set of eggs of this kind to secure another set of eggs produced under the conditions of experiment 410, the results in each of the seven experiments gave F, which behaved in all respects as a homozygote of the maternal type. This fact is set forth in considerable detail and Plate III illustrates it just as here described. This result occurred in all cases whether the set of eggs produced under the conditions of experiment 410 was produced before or after the set which gave the results of experiment 409. Now the remarkable thing about this experi- ment is this. While in experiment 410 the results in each of the eleven cases gave ordinary Mendelian heterozygotes in F,, in each of the seven cases of 409/411 the eggs produced under the same conditions as 410 gave homozygotes of the maternal type. Thus, the conditions of experiment 410 in eleven cases gave one result, in seven other cases they gave an entirely different re- sult, and the only difference in the conditions was that in 409/411 the female either had produced or was in the future to produce a set of eggs under the conditions of experiment 409 (p. 295). Thus, on page 294, in describing experiment 410, it is stated that experiment 410 gave F, all heterozygote; on page 330 it is stated that experiment 410 gave F, all homozygote of the ma- ternal type; on page 295-6, in describing experiment 409/411, it is stated that, under the conditions of experiment 410, this experiment gave F, all homozygotes of the maternal type; and on page 304 it is stated that experiment 409/411, when per- formed under the conditions of 410, gave results identical with those of 410. These statements are directly contradictory. We must withhold judgment on this point of influencing the hereditary factors at the time of fertilizations, until Mr. Tower informs us which of these statements are correct. In these exceptional cases, where the F, hybrid behaved as a homozygote of one of the paternal races, the author does not tell us whether the F, and later generations were each time pro- duced under the conditions which produced the aberrant F,. One would infer, however, that they were not, and that the change which occurred in the fertilization of the eggs which produced F, was permanent and not reversible. It is hoped 62 THE AMERICAN NATURALIST [Vou. XLV that he will give us fuller data on this point in future papers which are promised. In several cases Tower mixed three species which interbred freely and left them under natural conditions for several years. A careful study of the progeny in each case showed that a new type arose, consisting of a complex of the characters of the old types, and that this new type rather rapidly replaced every other type, although some of these other types were known to be quite capable of existing under the conditions of the experi- ment. This would indicate that in some way the new type had a distinct advantage over the other types with which it com- peted for food, or possibly the repeated crossing of the types was in some way inimical to all the types except the one. Ex- periments of this character show that hybridization may be an important factor in the development of new varieties or pos- sibly new species. From the fact that when the same species are mixed together in two places where the conditions are different, the resulting type which finally wins out and becomes practically the sole representative of the mixture, is different under different con- ditions. Tower draws the conclusion that the conditions sur- rounding the germ cells at the time of fertilization ‘‘ profoundly - modify the behavior and the relationships of the characters en- tering into the crosses.’ This conclusion seems hardly justified. Of course it may be correct, but the well known fact that from complex crosses of this kind a great many types may result from what we know of the behavior of Mendelian characters ‘and that these types would naturally bear different rela- tions to the environmental conditions offers apparently a much simpler explanation of the reason for the survival of the one type under one set of conditions and another type under another set of conditions. It seems hardly necessary to assume that the conditions existing at the time of the fertil- ization of the egg determine the characters which were to result from the fertilization to explain this particular phenomenon. Subsequent investigation of these new types of mixed origin showed that in all cases they occasionally produced individuals different from the general population but which in all cases ex- hibited characters which were present in the original parents of the complex mixture. Tower repeatedly compares this phe- nomenon to the phenomena which de Vries observed in (no- No. 529] NOTES AND LITERATURE 63 thera lamarckiana, and suggests that the mutations which de Vries observed are probably due to previous hybridization. This is a very interesting suggestion, but the writer is inclined to believe that the phenomena observed by de Vries were due to a different cause. It is definitely proved that in some of de Vries’s mutants the chromosome numbers are dif- ferent from those of the parent form. Cytological investiga- tions have also shown that in the reduction division in these (nothera mutants there are frequent irregularities in the dis- tribution of chromosomes. It seems probable that de Vries’s mutations are not the result of previous hybridizations but rather are due to irregular behavior of chromosomes in the re- duction division. If this is true then the phenomena observed by de Vries would be due to a different cause from that which presumably produced the results which Tower observed. In the case of Tower’s results we can explain the facts by the as- sumption of simple Mendelian segregation. In de Vries’s work there is evidence that the phenomena are due to a different cause. It is gratifying that Tower takes a very broad view of the fac- torial hypothesis of Mendelian phenomena. On page 323 he remarks: This factorial point of view is in no wise, of necessity, to be tied to or confounded with such speculations as the id-determinant-biophore fabrie of Weismann, nor with the pangene complex of de Vries, which have no foundation in fact. : This is the view which the writer has held for years and has fre- quently set forth in these pages. I have also frequently pointed out that we do not yet have sufficient knowledge of the phys- iological processes of living matter to permit us at the present time to formulate an adequate theory of the phenomena ob- served in hybrids. I think we can, however, point out the gen- eral nature of the causes underlying these phenomena, as I have attempted to do in my theory of Mendelian phenomena.*® In speaking of the difference in germ cells with respect to given characters, he has the following to say: What this difference in the gametes is we do not know, but observed behaviors are interpreted as being, most probably, due to the mechan- ical separation into different germ-cells of whatever it is that produces the contrasting attributes—segregation during gametogenesis. * American Breeders’ Magazine, Vol. I, No. 2. 64 THE AMERICAN NATURALIST [ Vou. XLV He further remarks on page 328: At present in biology we have no business with ultimate conceptions, and the two thus far attempted of germinal composition—the “ par- ticulate conception ” and the “crystalline entity” are both equally dismal failures and equally useless as working hypotheses. The statement on page 335, that characters which Mendelize are in the main unimportant attributes of the organism and only rarely are of importance in the struggle for existence, is a little bit strong. Apparently it would have been better to state that those characters which have been shown to Mendelize are of this nature. Unfortunately, most of the work of the Mendelians has been done with these superficial, easily observed characters. I see no reason why any character whatever might not, from the failure of some chromosome to perform a usual function, give a variation which would behave in Mendelian fashion if the resulting type were capable of propagating and crossing with the parent type. Tower’s paper will undoubtedly have an important influence on biological thought, as it deserves to have. W. J. SPILLMAN. ee et N The Anatomical Laboratory of Charles H. Ward 189 West Avenue, Rochester, N. Y. x OUR HUMAN SKELETONS are selected specimens scientifically ' and mounted. They are undoubtedly the finest and strongest skeletons obtainable, and are purchased by the leading Medical and Literary Colleges, Schools, Surgeons, ete. We make a number of special skeletons for demonstrating dislocations, muscular areas, anthropometric ks, muscles, ete. The mounting of the articulations permits movements as in life. Strength ge Higiity are — by — use of a special bronze wire of ) tl g oxidation. Portability and ease of d trati ttained by our nickeled steel clutch stand- ard, a is a great protection as well. ese skeletons are shipped entirely set up, carefully wrapped, and with a directions for unpacking and handling. Our Catalogue gives further details. : OUR SKELETONS OF TYPES OF VERTEBRATES are large specimens, principally of American species, mounted in characteristic Eoo with nickel-plated brass standards. | ae We We offer a a type collection‘for schools and colleges. 2 _ ANATOMICAL MOORS 8: sand T Universities. Ge ar Seapire head, brain, | f which ho tones. These EE pay 1 no duty, no trans- portation, no middlemen’s profits, but are sold direct, at very m moderate s, í nd can be supplied promptly when needed The American Naturalist A Monthly i established in 1867, Devoted to the Advancement of Factors of Organic Evolution and Her th Special Reference to the the ag Sciences CONTENTS OF THE JULY NUMBER A SS of the “Species Plantarum” of Linnaeus r the Starting Point of the oe of percha Professor W. G. Far Notes on Some Beaufort Fishes. E. W. vrei On the Effect of — Conditions = ae Reproduc- aphnia. Dr. J. F. McCLEN Are Fluctuations aay Dr. HARRY ne ‘Love. P Professor EDWARD Fr East. ~~. ap ~ and Se ap seacivayeighy The Age AS or oa ae a pee ae : Russ A bis “Determir eey on and neg Sfodifeaton of the Mendelian s, Pro- The Bubonic Peas "Proessor H E Was: os ai P Plants, WILLIAM L. Bray, CONTENTS OF THE AUGUST NUMBER Chromosomes and Heredity. Professor T. H. MORGAN. Spiegler’s “White Melanin” as Related to Dominant or Recessive White. Dr. Ross AIKEN GORTNER. Shorter Articles and Correspondence: A Pickwickian Contribution to Our Knowledge of Wasps: Professor KARL PEARSON. Notes and Literature: Heredity, Dr. W. J. SPILLMAN. CONTENTS OF THE SEPTEMBER NUMBER Nuclear swig of Sexu: eae in the Alge. Dr. BRADLE Y Moor: D TE Phenomena of Sexual ceai in Fungi, R R. A. HARPER. The Pose A the Sauropodous Dinosaurs. Der. W. D. ATTHEW. Shorter Articles and Discussion: Evolution without Iso- lation, Dr. oe T. GULICK. Retroactive Selection, Casper REDFIELD. The Logic of Chance in Problems of Genetics, ARTHUR Notes and Literature: Animal Structure and Habi Professor G. H. Parker. Plant Physiology, C. Ñ SHEAR. CONTENTS OF THE OCTOBER NUMBER hrant = Urosalpinx. Dr. HERBERT EUGENE AAE D ~ Sexual Reproduction in Gym- nosperms. ERLA ES J. CHAMBERLAIN. Nuclear Phenomena ‘a Sexual rar E in Angio- sperms. Profi D. M. MOTT Shorter Articles and Discussion : ‘mae Dr. Max MORSE. Notes and Lite Davin ae cian Professor T. D. A. COCKERELL. : Notes on Ichthyology, Presiden’ The Mammals of ere CONTENTS OF THE NOVEMBER NUMBER gassed of Skin Pigmentation in gee GERTRUDE C, VENPORT and CHARLES B, DAVENPORT. The yes Sense of the pany mesg: Bees distinguish Colors? JOHN Shorter Articles ane sa sya Anta oF the Product apea Method o oef cient of Correla’ ¥ Dr, J: j sa HARRIS. Notes and Literature: Schlosser on Fayûm Mammals, Dr. W. D. MATTHEW. Pei rere acgeen Genus Grayia: Professor T. D. A. Coc CONTENTS OF THE DECEMBER NUMBER Heredity of cosmid Pi “een in Man. GERTRUDE c. DAVENPORT RLES B. DAVENPORT. oe and Larva a Amie Jeffersonianum. Pro- essor W. H. PIERSO The RA ena of Sizes a Shapes in Plants. Professor Shorter icles and Discussion : The Modification of Mendelian Inheritance Pd a al Co r T. D. A. Cock Conditions. Notes ae Literature: Heredity, eh W. J. SPILLMAN. Index to Volume XLIV. Single Number 35 Cents Yearly Subscription, $4.00 The NATURALIST will be sent to new subscribers for four months for One Dollar THE SCIENCE PRESS Garrison, N. Y. Sub-Station 84: NEW Y P Lancaster, a VOL. XLV, NO. 530 ~ FEBRUARY, 1911 THE AMERICAN The American Naturalist MSS. intended for publication and books, etc., intended for review should be sent to the Editor of THE comedy NATURALIST, Garrison-on-Hud son, New York. Articles containing research work bearing on the problems of ae evolu- tion are os welcome, on will be given arelarenas | in publicatie e hundrea nts of pisao Seg are supplied to authors is of charge. Further reprints will be supplied at cos Subscri and advertisements should be sent to the Tatapi The subscription price is four dollars a year. Foreign postage is fifty cents and Canadian enty- er cents additional. The charge for maa copies is thirty-five cents. The advertising rates are Four Dollars for a oo : THE SCIENCE PRESS _ Lancaster, Pa. Garrison, N. Y. ee es _ NEW YORK: Sub-Station 84 _ Entered tter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act ot Congress of March 3, 1879. THE BULLETIN—For bargains in Ethnolograph- | MARINE wage HOLE Mas LABORATORY ae ETE E À SUPPLY DEPART ical and Pre-historic Specimens. Books on Natural l rahn. Preserved m aterial of all a animals for PE S i r ea Sa nas Fes oes TE | reamgiloel Mela soma dara nen a i : post ‘or 3 cen p. R a, ter 4 Duke St., Adelphi London—Englaad Liverworts and bk gn ina så 3 GEORGE M. GRAY, Curator, nia Hole, Mass. = ambriage University Press. | £ {E nugian | W. BATESON, M.A., F.R.S., Director of the John Innes Horticultural R.C. F JNNETT, MA. Professor of Biology in the Uni- ublication of sande original research in Heredity, Variation and allied f Deore time to time, contain articles summarizing the existing state of f seine but reviews and abstracts rigs yang published elsewhere f- equate illustra tion will be provided, and, ` where the subject t matter en T oE V Vol, J No. 1. November, 1910. iss White Flowered Varieties of Primu at fuse $ = Colour and o Characters | ex Figures. mee E “The Mode oee daa Ton of Doubleness in Flowers. I. Petunia. (are. 7 “The Bieta of one-sided Ovarotmy on the Sex + la si Tt in the the Potato: F THE AMERICAN NATURALIST VoL. XLV February, 1911 No. 530 THE APPLICATION OF THE CONCEPTION OF PURE LINES TO SEX-LIMITED INHERI- TANCE AND TO SEXUAL DIMORPHISM! PROFESSOR T. H. MORGAN COLUMBIA UNIVERSITY In the same sense in which our ideas concerning variation and heredity have been entirely revolutionized since 1891, so has a similar change taken place in regard to our theories of sex determination. Sex is now treated by the same methods that are used for Mendelian char- acters in general. From this point of view I propose to consider to-day three questions, intimately associated. First, the treatment of sex as a Mendelian character; second, the relation between sex and the inheritance of secondary sexual characters; third, the bearing of the recently discovered cases of ‘‘sex-limited-inheritance’’ on the problem of the transmission of characters in general. Most modern theorists are in agreement that the heredity of sex can be best understood when one sex is regarded as a pure line, or homozygous, and the other sex is treated as a phenotype, i. e., as heterozygous. The experimental evidence has made it plain that in some animals and plants it is the female that is hetero- zygous, and in other animals and plants it is the male that is heterozygous. Hence have arisen through the necessities of the situation the two following classes of formule : ‘From a symposium on ‘‘The Study of Pure Lines of Genotypes,’’ before the American Society of Naturalists, December 29, 19-4, 65 66 THE AMERICAN NATURALIST [Vou. XLV Gametes od Female F 8 gő Male OOS g female male 99 Female g ? Q¢ Male oOo S. 2 female male In certain groups of animals, as in Abraxas amongst insects, and in poultry amongst birds, the first scheme is essential to an interpretation of the facts obtained by experiment. In other groups, as in Drosophila amongst insects, and in man amongst the vertebrates, the second scheme accounts for the experimental results. These methods of formulation are open to two serious objections. As the tables show, the combination of 2¢ stands for the female in one case, and for the male in the other. In order to avoid this apparent contradiction it is assumed that in some groups femaleness dominates maleness, and in other groups maleness dominates fe- maleness, which seems to me paradoxical at least. It will be observed also that in the first of these schemes the male carries none of the sexual characters of the female, and in the second scheme the female car- ries none of those of the male; both of which assump- tions do not seem to me to be completely in accord with fact. Cytologists represent these same two schemes in a different way. They represent in the one case the female character by X; and the male by the absence of X. Thus: Gametes XO Female x OO Male o oO ao CO female male This representation covers the first class of cases where the female is heterozygous. For the second class, where the female is homozygous, the following scheme is No. 530] CONCEPTION OF PURE LINES 67 used, in which the female is represented by two X’s and the male by one X: Gametes XX Female XO Male x (0) AX 30 female male The XO—OO scheme applies, as before, to the case of Abraxas and to poultry, and the XX—XO scheme to the other class of cases. The latter expresses also exactly what takes place in the chromosomes of those groups where two classes of sperm exist (in relation to the X element), as has been demonstrated by Stevens and by Wilson. In both of these two latter schemes the production of the female is ascribed to the presence of the chr e X, but in the first formula one X makes the female and its absence stands for the male, while on the second formulation two X’s make the female, while one makes the male. In one case XO is female and in the other XO is male. Again we meet with the same paradox as in the first two formulations. The chief drawback to these formule is, in my opin- ion, the absence of any character to stand for maleness. Absence of femaleness does not appeal to me as a suffi- cient explanation of the development of a male; for the male is certainly not a female minus the female char- acters. Nevertheless, despite these objections I am inclined to think that these two methods of formulation indicate the direction in which we must look for an explanati of the experimental evidence, and that they may be still utilized provided we can so modify them that their in- consistences can be made to disappear. It seems to me that if we are to succeed in bringing Sex into line with Mendelian methods we must be pre- pared to grant that there are representative genes for the male condition and others for the female; and we must so shape our formule that the female carries the 68 THE AMERICAN NATURALIST [ Von. XLV genes for the male and the male carries those for the female. In fact, I am inclined to think that the evidence forces us to accept Darwin’s original view, that in each sex the elements of the other sex are present; a view that has been largely given up by modern theorists (except by Strasburger). I think that we must accept this in- terpretation for several reasons. Every zoologist is familiar with cases in which the same individual may at first function as a male and later as a female. More re- markable still is the case of the Nematodes in which in some species the female has come to produce both eggs and sperm as shown by Maupas and more recently by Potts, while in another closely related species it is prob- ably the male, according to Maupas and Zur Strassen, that has come to produce eggs as well as sperm. There is further the class of cases where the female develops the male secondary characters and the male those of the female. This class of cases I shall discuss later, for the value of this evidence will turn on whether these second- ary sexual characters are represented by independent genes, or are expressions of the presence of one or the other sexual condition; or due to a combination of these two possibilities. By means of the following formule we can meet the requirements that the situation seems to me to demand. If we admit that in the first class one of the genes has become larger than the other female genes, and if we admit that in the second class one of the female genes has become smaller than its sister genes we can account for the results as the following formule show: Gametes Fmfm Female fmfm Male fm fm Ffmm ffmm female \ male FmFm Female Fm X Fm Fmfm Male Fm \ fm PASSES, geome FFmm Ffmm No. 530] CONCEPTION OF PURE LINES 69 It should be carefully observed that in this scheme the female genes, F or f pair when they meet (allelomorphs) ; likewise the male genes‘ pair only with male genes. In fact, both genes are carried by all of the gametes. Sex- ual dimorphism may appear either because one female gene has become stronger than the others, or, because one has become weaker. On the first view we have the case where the female is heterozygous in its female genes; in the latter case it is the male that is heterozy- gous in its female genes. If in this latter case we as- sume that the weakened female gene is contained in the so-called Y-chromosome we can then understand how it is that we have a degraded series of this chromosome leading in some forms to its final extinction, for even its disappearance leaves the formule unaffected. On the same grounds we may anticipate that in those species in which the X elements are alike in the male, one X in the female may be found larger than its partner, al- though visible size differences in the chromosomes are not essential to the scheme, since these chromosomes undoubtedly contain many other factors than those of sex whose presence might obscure size relations even when such exist in the sex genes. These formule appear more complicated than those previously given, but in reality they are not so. It is the presence of m in all of the gametes that gives the appearance of complication. If this is omitted, as in the formula given below, the formule are no more complex than those given earlier. Gametes Ff Female ff Male f f Ff ff FF Female F F Ff Male F £ FF Fe The formule might be further simplified, if it seemed desirable to do so, by simply indicating the determining factor in each case as shown below; thus: 70 THE AMERICAN NATURALIST [ Vou. XLV Gametes FO Female OO Male O O FO (010) OO Femalə (0) O Of Male O f o0 Of But this last simplification is misleading, if the thesis that I shall here maintain in connection with sex-limited inheritance is correct; because the F’s and the f’s omitted in the last case are supposed to be carried in definite bodies, the chromosomes, which also carry other factors than sex factors, and it is essential to indicate their presence in some way in order that these other fac- tors may have some means of transportation. In a recent paper on sex determination in phylloxerans and aphids (1909) I discussed at some length different theories of sex determination, and adopted provisionally the view that the outcome is determined by a quantita- tive factor. The present hypothesis is little more than a further development of this same view,” but I hope in a form more in accord with the Mendelian treatment of the problem. Sex is still represented as the result of a quantitative factor F (or f), but its relation to the male factor is now expressed, for maleness is not assumed, as before, to be no femaleness or less femaleness. Here, as there, more of a particular factor turns the scale towards femaleness in the first class of cases, and less of the fe- male factor allows the scale to turn in the opposite direc- tion in the second class of cases.* 2 In 1903 I suggested that in the case of the bee a quantitative factor determines sex, viz., the chromatin; two nuclei producing a female and one a male, Wilson (1905) has identified the quantitative factor with a special chromosome and this interpretation of the quantitative factor is here fol- lowed. On Wilson’s view the male condition is represented by the absence of the X-chromosome in some cases, and by the presence of only one X- chromosome in the others, (see ante); but on my view the determination of sex is regulated by this quantitative factor in relation to another factor, the male determining element. *It should be pointed out that these formule are in no way related to a suggestion that I made in 1907 in regard to dominance and recessiveness No. 530] CONCEPTION OF PURE LINES 71 These formule have certain advantages over those now in vogue, first, because the male gene is not ignored as a factor in sex determination; second, that its pres- ence, both in males and females, explains how under cer- tain conditions the male or the female may assume some of the characters of the opposite sex; third, that the paradox of the female being the heterozygous form in one class and the male in the other class is, in part at least, resolved; fourth, that the ease with which species pass from the hermaphrodite condition to that of sexual dimorphism and the reverse is understandable; fifth, that the production of males by parthenogenetic females can be accounted for by the loss of one of the female genes in the polar body; and lastly, we see how there may be two kinds of eggs, as in Dinophilus apatris, both of which can be fertilized; for, in such cases the sperma- tozoa should be all alike. I do not wish to urge this view too positively, for I am acutely aware that we are only at the beginning of our understanding of the problem of sex determination, but I believe that the difficulties of the current hypotheses must be clearly understood and met if possible.* THE INHERITANCE OF SECONDARY SEXUAL CHARACTERS From the point of view reached in the preceding dis- cussion let us now examine the problem of the inherit- ance of secondary sexual characters. Males are distinguished from females not only by the presence of sperm in place of eggs, but by the presence in general. That view I have entirely abandoned. In the present hypoth- esis the relation of the determining elements is stated in the same form as in other Mendelian formule, with the possible exception that here one gene is represented as larger or smaller than its allelomorphs, and the scale is turned by the mass relation between these female genes and those of the male, *I have not discussed here the possibility of selective fertilization, because if we can explain the facts without this problematical assumption we simplify the problem greatly. Moreover, the evidence brought forward by Payne, Brown and myself, while admittedly insufficient, stands definitely opposed to the view of selective fertilization. 72 THE AMERICAN NATURALIST [ Von. XLV of different kinds of ducts, glands, copulatory organs, or other accessory sexual apparatus; and also by structures not essential to reproduction. These last we call the secondary sexual characters. It has long been known that in the embryonic develop- ment of the vertebrates some of the accessory organs of the male appear in the female, and conversely some of the accessory organs of the female in the male. This evidence seems to me to point with no uncertain mean- ing to the conclusion that each sex carries the genes of the other. It is however the secondary sexual characters rather than these accessory organs of which I wish to speak now; for, these often appear to be present in one sex only. Are these characters represented in all eggs and sperm or are they by-products of the sexual condition of the animal? Fortunately there is a good deal of experi- mental evidence that bears on this question, but it is also true that the evidence teaches that the matter must be handled with care, and if I seem to speak dogmatically it is for lack of time rather than for want of caution. It has been shown by Meisenheimer that removal of the gonads of the caterpillar of Ocneria dispar fails to produce any effect, or very little, on the secondary sex- ual characters of the moth. It would seem, therefore, that these characters are represented in the germ cells in the same way as are other characters, and are not depend- ent for their development on the presence of the gonads. Some mechanism must exist by means of which the genes of these organs are distributed so that two kinds of in- dividuals are produced. It has been suggested by Castle that the secondary sexual characters may be carried by the Y-element in the formule XX — female, XY — male, but this hypothesis fails to explain the results when the Y-element is absent, as E. B. Wilson has pointed out. It also fails to explain how the male secondary sexual or- gans can appear in the female after castration. On the sex formule that I have suggested it is pos- sible to account for the results, if the genes for these No. 530] CONCEPTION OF PURE LINES 73 characters are carried by all cells alike; possibly they go along with the male-group, but this is not essential. Whether they develop, or not, will depend on the pres- ence of other genes in the cells. Thus when the Fmfm group is present they will be suppressed, or when, as on the other formule, the FmF'm group is present. We can understand on this view why in the insects the male sec- ondary sexual organs do not develop in the female after removal of the ovaries, because in this group it is not material derived from this source, but from materials produced in the cells themselves, that bring about the suppression. It has been demonstrated by Geoffroy Smith that when the young males of the spider crab, Inarchus mau- ‘retanicus, are infested by Sacculina the secondary sex- ual characters of the female develop. It appears that the parasite produces some substance that inhibits the activity of the male-producing group in each cell, or counteracts some materials produced there, so that the female characters now find the situation favorable for their development. When the young female crab is in- fected by Sacculina she does not develop the male sec- ondary characters, which is in harmony with the view just stated for the manner of action of the parasite. In birds and in mammals it has long been recognized that some substance is produced in the ovary that in- hibits the development in the female of the male second- ary sexual characters, for, after removal of the ovaries the male characters may to some extent develop. It seems fairly clear that here the female group in each cell fails to entirely suppress the male characters; the inhibiting effect from this source must be reinforced from something produced in the ovary. Whether after castration of the male the secondary sexual characters of the female develop is not so clear, since some at least of the characters that characterize the castrated male may be juvenile. But on my view the possibility exists for the castrated male to produce the secondary sexual 74 THE AMERICAN NATURALIST [ Vou. XLV characters of the female, if their development is in part suppressed by substances made in the testis. The view here presented also allows us to explain how the secondary sexual characters of the male are trans- mitted through the female, as they may be so transmitted. THE INHERITANCE OF SEX-LIMITED CHARACTERS In recent years a new class of facts has been discov- ered that promises to throw a flood of light not only on the sex-determination problem, but also on the problem of inheritance in general. I refer to the cases of sex- limited inheritance. We mean by sex-limited inheritance that in certain combinations a particular character appears in one sex only. An example will make this clear. In one of my cultures of the red-eyed fly, Drosophila, a white-eyed male appeared. Bred to red-eyed females, all of the off- spring, male and female alike, had red eyes. These inbred produced red-eyed males and females, and white- eyed males. In other words the white-eyed mutant transmitted his character to half of his grandsons, but to none of his granddaughters. Yet this white-eyed condition is not incompatible with femaleness; for, it can be artificially carried over to the female by making a suitable cross. If, for instance, a white-eyed male is crossed with a heterozygous red fe- male, there will be produced red-eyed males and females and white-eyed males and females. There are certain combinations of sex-limited char- acters that give results outwardly similar to sexual dimorphism. If a black langshan cock is crossed to a dominique hen, all of the sons are barred and all of the daughters are black. If a white-eyed Drosophila female is crossed with a red male all of the sons will have white eyes, and all of the daughters will have red eyes. I have another strain of these flies with small wings and still another strain with truncated wings. If a female of the former is crossed with a male of the latter strain all of No. 530] CONCEPTION OF PURE LINES 16 the daughters will have long wings and all of the sons will have small wings, like their mother. These cases conform to Mendel’s principle of segrega- tion. Were there time I could show by an analysis of the problem why these sex-limited characters behave in in- heritance in a different way from secondary sexual characters, although the results in both cases may be accounted for on the assumption that there are genes in the cells for both kinds of characters. In a word, this difference exists because one of the factors for the sex- limited characters in question is absent from one of the female determining chromosomes, while the genes for the secondary sexual characters of the male are contained in other chromosomes, possibly in those that contain the male determinants. This interpretation of the relation between the X- chromosomes and sex-limited characters makes it now possible to demonstrate a point of great theoretical im- portance. I invite your serious attention for a few moments longer to this question. Three other characters have appeared in my cultures that are sex-limited; one of these only I may now speak of. A male with wings half the normal length suddenly appeared. He trans- mitted his short wings to some of his grandsons, but to none of his granddaughters. I tried to see if the other sex-limited character, white eyes, could be combined in the same individual with short wings. As the next dia- gram shows a red-eyed short-winged male was bred to a white-eyed female with normal wings. All of the off- spring had long wings; the female had red eyes and the males white eyes. These were inbred and produced white and red-eyed males and females with long wings, red-eyed males with short wings, and white-eyed males with short wings. In the last case the transfer had been made. The reciprocal cross also given in the diagram is equally instructive. 76 THE AMERICAN NATURALIST [ Vou. XLV LWF — LWF Long-winged, white 9 SRF O Short-winged, red g LWFSRF — LWF LWF SRF SWF LEF = 9 Gametes F— oO g Gametes LWFLWF Long-winged Ọ white eyes SRFLWF Long-winged Ọ red eyes SWFLWF Long-winged ? white eyes LRFLWF n SRF Short-winged ¢ red eyes SWF Short-winged & white eyes LRF Long-winged ¢ red eyes LRF LRF Long-winged, red 9 SWF O Short-winged, white ¢ LRFSWF — LRFO LEF SWF LWF 8RF 9° Gametes LRF O g Gametes LRFLRF Long-winged ? red eyes LRF ong-winge red ey SWF Short-winged 4 white eyes LWF Long-winged ¢ white eyes SRF Short-winged g red eyes In both cases the combination is possible because in the female of the hybrid (F,) a shifting of the gene for long and that for short wing (both carried by the X- chromosome) takes place. This interchange is possible during the synezesis of the two X-chromosomes. On the other hand the male contains only one X-chromosome which has no mate, hence the gene for long wings in the hybrid (F,) can not leave that chromosome to pass into the male-producing group. If it could do so short-winged females would also appear, but as I have shown they are not present in the second generation. Interpreted in terms of chromosomes these results can have, in my opinion, but one meaning. During union of homologous chromosomes (during synezesis, perhaps) homologous genes pair and later separate to move to op- No. 530] CONCEPTION OF PURE LINES 77 posite sides (or enter the chr sometimes one way and sometimes the other). All the genes contained in the X-chromosomes can thus shift in the female be- cause in this group two X’s are present. Sex-limited inheritance is only possible where similar conditions exist (either in the male or in the female) and since in man color blindness follows the same scheme as does white eyes in my flies, we have an experimental proof that in the male of homo sapiens there is only one X-chromo- some, and this, in fact, Guyer has just shown to be the case from cytological evidence. But by parity of rea- soning it is the female in Gallus bankiva that should have _ only one X present, but Guyer is persuaded that here too (at least in the race of fowls he studied) the male has only one X-chr . There is then in this case a contradiction between the experimental evidence and that furnished by cytology and it remains to see which is correct. Bateson has shown that some of these cases of sex- - limited inheritance can be explained on the grounds that there is a repulsion between the female-determining fac- tor and that character that is sex-limited. The view that I maintain does not involve the idea of a repulsion be- tween unlike elements, not allelomorphic. Spillman’s hypothesis also involves this idea of repulsion between unlike elements. On my view, on the contrary, an at- tempt is made to show how the results may be due to a connection existing between certain material bodies in the egg; a connection that is consistently carried through successive generations, and subject only to the ordinary interchange of genes between homologous chromosomes (when a pair of chromosomes is present).° For several years it has seemed to me that the chromo- some hypothesis, so called, could not be utilized to ex- plain the Mendelian results in the form presented by ë The hypothesis advanced here to explain sex-limited inheritance applies also to Abraxas if the latter follows the Fmfm scheme and if in the egg there is no interchange between the F-bearing and the f-bearing chromo- somes, 78 THE AMERICAN NATURALIST [Vou. XLV Sutton, because, if it were true, there could be no more Mendelian pairs in a given species than the number of chromosomes present in that species. Even if this ob- jection could be avoided® the more serious objection still remained, namely, that with a small number of chromo- somes present many characters should Mendelize to- gether, but very few cases of this sort are known. De Vries was the first, I believe, to point out that this objec- tion could be met if the genes are contained in smaller bodies that can pass between homologous pairs of chro- mosomes; and Boveri has admitted this idea as compat- ible with his conception of the individuality of the chromosomes, In the case of the inheritance of two sex- limited characters in the same animal we have an experi- mental verification of this hypothesis. °Spillman’s suggestion that the difficulty exists only when it can be shown that more dominant characters can occur in the same individual than the number of chromosomes seems to me only to push back the difficulty. PURE LINES IN THE STUDY OF GENETICS IN LOWER ORGANISMS! PROFESSOR H. S. JENNINGS THe JoHns HOPKINS UNIVERSITY Art the meeting of this society a year ago I asked in a paper read,? whether the pure line idea did not deserve agitating a little before this society, and I tried to agitate it. This was because I saw that for practical purposes of future work it would be necessary to make up my mind as to the importance of this idea, and it seemed that other members of the society might be in the same situation and that we might help one another. My method of agitation was to give the apparent relations of the results of work along this line up to that time, to one of the burning problems in our field—the problem of selection. In the few minutes that each of us have here the purpose of agitation can be served and general results brought sharply into view only by naked and dogmatic statements, such as one would never use under other conditions. Such naked and dogmatic presentation has serious dis- advantages—felt most decidedly by the author when his critics hold the mirror up to nature. I have therefore at times regretted giving forth this paper. But if it has in any way acted as an irritant to arouse the discussion fore- shadowed in our present program, I shall feel that its good results outweigh its painful ones, and that it was worth while after all. We are apparently to have > brought before us a part of that ‘‘thorough try out’’ that I asked for, and from a study of our program I think I can see that it is not all to be a pean of praise for the pure line work. Such illumination and such interest as comes m a symposium on ‘‘The Study of Pure Lines or Genotypes,’’ be- * Fro; fore the Ameriean Society of Naturalists, December 29, 1910. * This JouRNAL, March, 1910. 79 80 THE AMERIGAN NATURALIST [ Vou. XLV from having both sides presented I believe that we have before us. What I wish to attempt is to give some concrete illus- trations of the answer to the question discussed by Dr. Webber— What are genotypes? I note that some of the titles on our program speak of the genotype hypothesis, the pure line theory. What I wish to emphasize is that these things, whatever we call them, are concrete realities —realities as solid as the diverse existence of dogs, cats and horses. I find in many biologists not working in genetics an incorrigible bent for seeking under such a term as genotype something deeply hypothetical or meta- physical, and for characterizing it therefore boldly as purely imaginative. This is merely because such workers have not the things themselves before them. The genotype is merely a race or strain differing heredi- tarily in some manner from other races. Neither the idea nor the fact is a new one, and we should perhaps do better to discuss merely the importance of distinguishing in our work the diverse existing strains—rather than to introduce an unfamiliar term for a familiar thing. But investigation has shown the existence of these strains to play a part of such hitherto unsuspected importance that it has seemed worth while to introduce a more precise term, which shall emphasize their importance for work in genetics. In work with a certain lower organism— Paramecium—t have found the existence of these diverse strains or genotypes to be the guiding fact, not only for work in genetics, but for all exact work in comparative physiology. I wish to show how this is true. We must then distinguish clearly these concrete realities called genotypes from any theories that have been built up in connection with them; from any generali- zations based on their study up to this time. The exist- ence and importance of genotypes are not bound up with any particular theory regarding selection or any other single point. In lower organisms, at least, genotypes OT pure lines are merely the name for certain actual exist- ences that you have before you; for facts that strike you No. 530] PURE LINES IN STUDY OF GENETICS 81 in the face. We have, side by side in the laboratory, a lot of diverse sets of our organisms, each set derived origin- ally from one individual, and each differing characteris- tically but minutely from the others—the differences per- sisting from generation to generation. The behavior and properties of these things are of course a matter for further study. Can selection change them? Can envi- ronmental action permanently modify them? These are matters quite distinct from the existence of the genotypes. To get a clear grasp of the matter, I believe that those not working with lower organisms will find it worth while to try to realize the condition which the investigator in this field has before him. A comparison may help. In lower organisms the genotype is actually isolated, each in a multitude of examples, which live along without admix- ture, visibly different from all others, for many genera- tions, before again plunging into the melting pot of cross- breeding. In higher organisms we should have the same thing if every rabbit, every dog, every human being, multiplied by repeated division into two like itself, till there were whole counties inhabited by persons that were replicas of our absent president; cities made up of copies of our secretary, and states composed of duplications of the janitor I saw outside. Every human being, as it now stands, represents a different genotype (save perhaps in the case of identical twins), and these genotypes become inextricably interwoven at every generation. It is there- fore easy to see how the genotype idea might appeal to workers among higher organisms as a mere hypothesis. What then are these visible, tangible, isolated geno- types (or races, or strains) of lower organisms, and how are they distinguished? Taking Paramecium as a type: 1. Some of them differ in size—the size of each remain- ing closely constant, under given conditions, for hundreds of generations; for years. This was the first difference observed, and I tried to demonstrate it by giving meas- urements of successive generations of the different races. But to the worker in the laboratory these differences are evident without refined measurements; the student is at 82 THE AMERICAN NATURALIST [ Vou. XLV once struck with the fact that one culture is formed of individuals that are throughout and constantly larger than those of another culture. And here, in view of that extraordinary cry ‘‘no here- dity without a correlation table’? (a ery that at once annihilates most Mendelian evidence of heredity), it may be well to define a little more precisely what is meant by saying that the diverse sizes are hereditary in the differ- ent races. It means that if you keep your different geno- types side by side under precisely the same conditions, you will find whenever you choose to examine and meas- ure them, that each has a characteristic size, differing from that of the others. If therefore you follow the diverse lines from generation to generation you will get a set of chains, each with links differing characteristically throughout from the links of the other chains. It means that it is possible to predict the diverse relative sizes that will be found in the different races, and that when you examine them a hundred generations later, you will find the prediction correct. These striking facts are what are meant by the statement that the diverse sizes are heredi- tary in the different lines—and the way to determine whether the statement is true or not is to examine the lines from generation to generation to see if the state- ment is verified. To neglect this obvious fact; to mix all your lines together and then, in order to find out if size 1s inherited, to laboriously work out coefficients of correla- tion by refined biometrical methods—is like cutting serial sections ten microns thick of an eel, in order to find out whether it has an alimentary canal. Persons have been known to so bedevil material with refined histological methods as to quite miss the alimentary canal of an eel. The way to see it is to open the animal up and take a look atit. The way to see diverse genotypes is to isolate them and look at them and measure them and compare them. If the use of correlation tables should succeed in obscur- ing these striking facts (as should not be the case with proper handling) this would merely show the worthless- * Compare Pearson, Biometrika, 1910, Vol. 7, p. 372. No. 530] PURE LINES IN STUDY OF GENETICS 83 - ness of this method of attempting to learn the important biological facts under consideration. 2. Some of the genotypes show slight but constant differences in structure, which I shall not dwell upon here.* 3. They show most varied differences in their physio- logical characters. These physiological differences may go with differences in form and structure, or apparently they may not—so that we find types that differ, so far as detectable, only in physiological peculiarities. This fact becomes of great practical importance for all physiological investigations, as a few examples from Paramecium will show: (a) The races or genotypes differ in the conditions, both external and internal, that induce conjugation. A worker, using a certain strain, works out the conditions inducing conjugation and gives precise directions for accomplish- ing this. His colleague, with another strain, finds this work all wrong, and the controversy on this ancient question continues. One of my strains can be absolutely depended on to conjugate monthly if certain definite con- ditions are furnished; another under the same condi- tions never conjugates; others show intermediate con- ditions. These differences require no biometric methods for their demonstration. (b) Again, the genotypes differ in rate of multiplica- tion; under the same conditions some divide once in twelve hours; others once in twenty-four or more hours; others have intermediate periods. (c) The genotypes differ as to the conditions required for their existence and increase. Several strains, out- wardly alike, living in the same medium, are cultivated side by side on slides, in the.usual hay infusion. One flourishes indefinitely. Another multiplies for ten gen- erations, then dies out completely, and this is repeated invariably, no matter how many times we start anew our “For a detailed, illustrated account ug “ characters, both structural and ge pene of these races , see Jennings and Hargitt, ‘‘ Character- isties of the Diverse Races of Peia, >” Journal of ‘Morpholo ogy, De- cember, 1910 f a 84 THE AMERICAN NATURALIST [Vou. XLV cultures of this genotype. A third lives along in a sickly way, barely maintaining its existence. Thus we get in our laboratory striking cases of nat ural selection between genotypes. To recall our com- parison with human beings, if we could mix an entire community composed homogeneously of, let us say, Roosevelts, with another of copies of your ash man— which would be likely to survive? If we place together in the same culture two genotypes of Paramecium, as I have many times done, almost invariably one flourishes while the other dies out. This ruins many a carefully planned experiment; it must take place on a tremendous scale in nature. What distinguishes the different genotypes then is, mainly, a different method of responding to the environ- ment. And this is a type of what heredity is; an organ- ism’s heredity is its method of responding to the envir- onmental conditions. Under a given environment the genotype A is large, while the genotype B is small. Under a given environment the strain C conjugates, while D does not. Under a given environment the strain E divides rapidly, F slowly or not at all. The various strains thus differ hereditarily in these respects, and we may say that the differences are matters of heredity. And yet we can get these same contrasts within any genotype (as our diagram illustrates), by varying the environment. The genotype A under one environment is large; under another it is small. Under one environ- ment the type C conjugates; under another it does not. Under one environment E divides rapidly ; under another, slowly. / Are then size, conjugation and rate of fission after all determined by heredity or by environment? Such a question, when thus put in general terms, is everywhere an idle and unanswerable one. All environ- mental effects are matters of heredity when we compare types differing in their reaction to the environment; all hereditary characters are matters of environmental ac- tion when we compare individuals of the same heredity under effectively different environmental conditions. No.530] PURE LINES IN STUDY OF GENETICS A A i B A | | | , i 3 | F E | E | E Different Race Same Race Same Environment Different Environment DIAGRAM TO ILLUSTRATE THE RELATION OF HEREDITY TO ENVIRONMENTAL ACTION IN DETERMINING CHARACTERS. xt.) Heredity has a meaning only when we (explicitly or im- plicitly) compare two concrete cases; when we say: To what is due the difference between these two cases? Otherwise we can demonstrate either that all character- 86 THE AMERICAN NATURALIST [Vou. XLV istics are hereditary (as we heard maintained at Woods Hole some summers ago); or, with Brooks, that there is no such thing as heredity. If we always compare two concrete cases, asking to what is due the difference be- tween them, and remembering that a difference in hered- ity means different response to the same environment, we shall avoid these confusions, and shall find the con- cept of heredity most useful. Do hereditary differentiations ever arise within our genotypes, so that from one genotype we get two? In other words, do we get from a single type strains that differ in their behavior under the same environment— the differences persisting from generation to generation? This is of course one of the fundamental questions. The genotypes of Paramecium, like those of most other or- ganisms that have been carefully studied, are singularly resistant, remaining quite constant in most respects, so far as has been determined. This is an example of what gives the genotype concept its practical and theoretical importance. This is what is meant by saying that selec- tion and environmental action are usually without in- herited effect within the genotype. To find differentia- tions within the genotypes of Paramecium, we must examine certain characteristics that are most delicately poised in their responses to all sorts of conditions; such is the rate of multiplication. Studying carefully this most sensitive character, we find that differences do arise within the genotype. Under given conditions, certain rare individuals are found that divide more slowly than usual, others more rapidly, and these differences are perpetuated from generation to generation indefinitely. How are these hereditary differentiations produced? The origin of these differentiations is in Paramecium as elusive as in most other cases where they have been discovered. Apparently they arise in our organism as a result of conjugation within the genotype. Certainly if after an epidemic of conjugation within the genotype we cultivate many isolated exconjugants, we find a certain small number of strains that differ in their rate of fission No.530] PURE LINES IN STUDY OF GENETICS 87 from that which is typical. But the experimental analy- sis of this matter is still in progress, and conclusions can not yet be drawn. It is only in rate of multiplication that I have thus far found hereditary differences arising within the pure line, and these but rarely. But this encourages one to hope that the same may be found for other characters when these are extensively studied with sufficient minuteness. The negative results thus far reached do not (as many critics have pointed out) exclude the possibility that rare cases of hereditary variation within the pure line will yet be found. What the negative results have demonstrated is that a very large share of the observed variations in organisms are not hereditary, and that se- lection based on these variations leads to no result—a conclusion of such great importance as to make the pure line work epoch-marking in character. Finally, what happens when diverse genotypes mix in conjugation? To my disappointment, I have found this much more difficult to determine for the infusorian than I expected. This is owing to the fact that the condi- tions for conjugation differ in the diverse genotypes, so that it is almost impossible to get them to conjugate at the same time. Further, in the rare cases where two are conjugating at once, the assortative mating discovered by Pearl results in the two sets remaining separate. Thus I have not yet been able to get crosses between two genotypes whose characteristics are known beforehand; and this will be necessary before a study of inhoritaned, exact in the modern physiological sense, can be made. On the other hand, it is possible to get conjugations in wild populations that include many genotypes, and to com- pare the results with conjugations where but a single genotype is involved. Certain most interesting results appear. In these conjugations of mixed populations, a great number of diverse combinations are produced ; the variability increases greatly, in size and in other re- spects. Numbers of the strains produced die, or multiply so slowly that they have no chance in competition with 88 THE AMERICAN NATURALIST [Vou. XLV those that are strong and multiply rapidly. Thus many of the combinations produced are canceled; only the strongest combinations survive. We have then on a most extensive scale an operation in natural selection and the survival of the fittest; the production of many combinations, some of which survive, while others fail. As already set forth, there is some indication of the same process in the case of conjugation within the genotype. At our last meeting I tried to summarize the facts as to the relation of genotypic investigation to selection; it turned out that much which had been deemed a progres- sive action of selection was not such; and up to that time the action of selection in modifying genotypes had not been demonstrated. Similarly, I had earlier summar- ized the facts regarding selection in behavior, showing that it there plays a large part. I have hence suffered the peculiar fate of being belabored as an anti-selectionist in genetics, while subjected in the field of behavior to rough treatment as the champion of selection. What I tried to do in both cases was, to determine how far we had actually seen the effectiveness of selection—holding this question quite apart from what we believe must occur, or believe will be found to occur when we have seen it. It appeared clear, and still appears clear, that a very large share of the apparent progressive action of selection has really consisted in the sorting over of preexisting types, so that it has by no means the theoretical signi- ficance that had been given to it. When operating on a single isolated type it appeared that the progressive action of selection had not been seen. These are facts of capital importance to the experimenter; besides their theoretical significance, they open to each of us the oppor- tunity to direct our efforts upon precisely this point, and so perhaps to be the first to see examples of this funda- mental process not yet seen. I hoped to accomplish this myself, but after strenuous, long-continued, and hopeful efforts, I have not yet succeeded in seeing selection effec- tive in producing a new genotype. This failure to dis- cover selection resulting in progress came to me as a No.530] PURE LINES IN STUDY OF GENETICS 89 painful surprise, for like Pearson I find it impossible to construct for myself a ‘‘philosophical scheme of evolu- tion’’ without the results of selection and I would like to see what I believe must occur. It is therefore with some pleasure that I am able to record for Paramecium this extensive operation of selection among the diverse exist- ing lines, and particularly in this extensive production of new combinations at conjugation, with cancellation of many of the combinations. It would seem that the diverse genotypes must have arisen from one, in some way, and when we find out how this happens, then such selection between genotypes will be all the selection that we require for our evolutionary progress. What I hope, therefore, is that some one on our program, more fortunate than myself, will be able to record seeing the actual production of two genotypes from one, or the transformation of one into another, by selection, or in any way whatever. Yet even if this is done, we shall make the greatest possible mistake if we therefore conclude that the exist- ence of genotypes is unimportant, and throw the matter aside ; for work with a mixture of unknown genotypes will always give confused and ambiguous results, whose signi- ficance no one can know. If on the other hand we work with single genotypes, or with known combinations of them, we shall understand what our results mean. And this applies to work in other fields of biology as well as to genetics, SOME EFFECTS OF TEMPERATURE UPON GROWING MICE, AND THE PERSISTENCE OF SUCH EFFECTS IN A SUBSE- QUENT GENERATION? DR. FRANCIS B. SUMNER Woops Hote, Mass. I must preface my remarks by an apology for coming before you with some results which have already been published pretty fully within the past year.? My appear- ance here may seem the more unwarranted in view of the limited amount of evidence which I am about to offer upon those subjects which form the focus of attention at this meeting, namely heredity and evolution. However, aside from the fact that I am acting at the instance of our president, I will say two things in my own defense. First, the results which I offer, meager as they doubtless are, appear to be the only ones of just this sort which are in evidence at present. And secondly, I am bold enough to believe that I have developed a promising method of attacking a few of the many knotty problems which are bound up together in the time-honored question: Are ‘Read before the American Society of Naturalists, December 30, 1910. ? (í Some Effects of External a upon the White Mouse,’’ Journal of Experimental Zoology, August, 1909. ‘‘The Reappearance in the Off- spring of Artificially Produced piena Modifieations,’? AMERICAN NAT- URALIST, January, 1910. ‘‘An Experimental Study of Somatie Modifica tions and their Reappearance in the Offspring,’’ Archiv fiir Entwicklungs- mechanik der Organismen, June, 1910. *Since writing this statement, I have received Semon’s highly interest- ing paper, entitled ‘Der Stand der Frage nach der Vererbung erworbener Kigenschaften’’ (Fortschritte der naturwissenschaftlichen Forschung, Bd. TI, 1910). From this I learn that some of the more important features of my results have been obtained by Przibram, in the course of experiments upon rats, conducted at about the same time as my own. I have not yet seen Przibram’s own report of his work. This confirmation will, I trust, dispel any doubts as to the statistical aM of my own figures, what- ever interpretation we may choose to give them. 90 No. 530 | TEMPERATURE ON GROWING MICE 91 acquired characters inherited? It is my hope to convince you that the method which I have employed conforms to certain a priori requirements, on the one hand, and, on the other hand, is workable in practise. That my results are not thus far more imposing is due, I think, to no defect in the method itself, but to the limitations which encompass a solitary investigator, deprived of some of the generally acknowledged desiderata for successful work in animal breeding, such, for example, as assistants, funds and ade- quate equipment. As to the logical requirements for such a test—to begin with, what is it that we are going to test? The ‘‘inherit- ance of acquired characters’’?—yes and no. First of all, that threadbare expression itself must be relegated to limbo where it belongs. For, not only does it fail to indi- cate with any precision the subject-matter of our inquiry, but historically the expression has been applied to a wide range of phenomena, real and alleged. Some of these we now know to be fictitious; others, on the contrary, are acknowledged facts; while others yet are more or less debatable. It is with the debatable group, of course, that we are here concerned. But, even among these, we en- counter not one problem but many. Suppose, then, that we drop all vague generalized expressions and consider one more or less restricted problem: Are specific struc- tural effects, resulting from the action of external condi- tions upon organisms of one generation ever repeated in the next generation under such circumstances that the immediate and parallel modification of the germ-cells may not be invoked as an explanation? Under ‘‘specifie structural effects,” I do not wish to include general con- ditions of health, metabolism, ete. What are some of the necessary conditions for a fair test of this question? To begin with, we must effect our modifications in the first generation. And since these modifications, if repeated at all, will probably reappear in a much-diminished degree, it would seem far prefer- able to select characters which lend themselves readily 92 THE AMERICAN NATURALIST [ Vou. XLV to accurate measurement. Qualitative differences, such as those of color or of physiological reactions, do not seem well adapted to such experiments, although they have commonly been the ones dealt with in studies of this sort. In the second place, we must choose such an organism and such a physical agency that the latter may act upon the former without immediately influencing the germ- cells. This would seem to rule out of consideration as really crucial tests of this problem all experiments, how- ever instructive otherwise, in which modification has been brought about through the influence of foods, unusual in amount or in character. For the effect of these upon the parent body is, of course, a chemical one, and the specific substances responsible for the modifications are presum- ably free to enter the germ-cells. The experiments of Arnold Pictet upon lepidoptera and of Houssay upon fowls are to be recalled in this connection. Similar con- siderations apply with equal force to any results from experiments in which invertebrate animals or ‘‘cold- blooded’’ vertebrates have been influenced by tempera- ture. The recent work of Kammerer upon lizards* and that of various investigators upon butterflies and moths occur to us at this point; likewise certain features of Tower’s work on Leptinotarsa. In such cases, by pretty general consent, we have to do with an ‘‘immediate effect upon the germ-plasm,’’ and not with transmission at all. Later, I shall inquire a little into the validity of this assumption. ~ In the meantime, I will point out that for certain classes of animals this objection cannot be raised, at least in its original form. I refer to the so-called ‘‘warm- blooded’’ ones. Iam not very well versed in that branch of physiology which deals with temperature regulation, but the published evidence which I have examined seems to show that mammals normally undergo but slight fluc- tuations of body temperature as a result of even very * Archiv fiir Entwicklungsmechanik der Organismen, September, 1910. No.530] © TEMPERATURE ON GROWING MICE 93 considerable changes in the penparainre of the surround- ing atmosphere.® Assuming, provisionally, the truth of this proposition, we may discount in advance the objection that the germ- cells of a mammal may be influenced by differences of temperature as such. If these differences affect the germ-cells at all, and it is reasonable to believe that they may do so, they must act upon them indirectly.* I shall revert to this point again shortly. _ As some of you may perhaps already know, I have succeeded for several years past in producing very decided quantitative differences in certain of the external parts of mice through the action of widely differing tem- peratures. ... (This part of the discussion has been omitted in the printed report, since the results in ques- tion have already been fully published.) In experiments such as those which I am describing, it, is obviously impossible to subject a single individual to both extremes of temperature during growth, and to com- pare the differing effects of these upon structure. We therefore, of necessity, resort to a comparison of aver- ages, based upon as many individuals as possible. If each of the contrasted groups is sufficiently large, and if its members have been taken at random, the presumption *Przibram found that the body temperature of his rats was somewhat C. This | raised when kept in a room at 30° to 35° ast was, however, con- siderably higher than the me n temperature of my own warm m, a e his experiments. (See Semon, op. cit., pp. 45, 46.) On the other hand; Pembrey (Journal of Physiology, 1895) found that the body temperature of mice did not rise appreciably above the ea when the animals were kept at a temperature of 29.5° or even 32.5° C. for an hour or more. effects of a more prolonged stay were not determined. I have myself Heese commenced experiments with mice, using a special clinical ther- meter made for the purpose. I have already (January 21) shown pretty Saian that mice may ws almost precisely the same rectal tempera- ture at —'6° C, as at + 30° e is true even Oy tunndbatty: ritak than temperature per se, is the factor ahia concerned in these modifications. As "o in my T paper (1909), the relative humidity of my heated room wa very m lower than that of the unheated room. Thus far I have not PPa the effects of these two factors. 94 THE AMERICAN NATURALIST [Von XLV is that the mean potential (that is to say, congenital) value of every character is about the same for the two lots. I fully realize that the study of genetic problems by the use of mass averages has recently received a decided set-back, largely through the labors of some of those who have contributed to our present program. But until some one is ingenious enough to produce a strain of par- thenogenetic or self-fertilizing mice, I fear that my only practical. method of procedure in these experiments is to deal with mass statistics based upon ‘‘heterozygous”’ stock. It must also be pointed out that the technique of the problem which I am discussing is inevitably different from that involved in the endeavor to find, or to produce, ‘‘mutations’’ or single abrupt deviations from the parent stock, which appear at once in full force, if they appear at all, and thereafter breed true. On the contrary, the distribution of the lengths for the tail, ear and foot, within each of the temperature groups in my experi- ments, appears to follow the normal probability curve, just as in the case of the so-called ‘‘fluctuating varia- tions,” whose heritability is nowadays so much in question. In the splendid paper of Professor Johannsen, to which we listened yesterday, occurs the following statement: ‘as yet no experiment with genotypically homogeneous cultures has given any evidence for the Lamarckian view, the most extreme ‘transmission’-conception ever issued.”’ Leaving aside for the time being the question whether results such as mine, even when every possible defect of technique has been eliminated, are to be regarded as ‘fevidence for the Lamarckian view,” let us consider for a moment whether the fact that I have not myself found it practicable to use ‘‘genotypically homogeneous cul- tures’’ does, in reality, invalidate the evidence which I offer. Ay y Professor Johannsen would hold this to be true. Bo far as I can see myself, the only difference between results from pure and from mixed lines in the No.530] © TEMPERATURE ON GROWING MICE 95 present case would be this. Individuals belonging to a single pure line would probably respond with much greater uniformity to the effects of an environmental change than would those belonging to a composite stock, consisting of a number of lines. It is quite conceivable that among these last some would respond in a much greater measure than others. Or indeed some might not be affected at all. But here the much-scorned ‘‘mass statistics’? would reveal the mean tendencies of the two lots, and the resulting data, though confessedly capable of further analysis, would be none the less valuable. If it be objected that the differences between the two aver- ages may be due to the presence in one or both of the con- trasted lots of a few ‘‘mutants,’’ while the remaining individuals may not have been affected at all, I will only point out, as above, that the frequency distributions are directly opposed to such an assumption. Having produced modifications of the sort mentioned, it remained to be seen whether these effects persisted beyond the generation immediately influenced. . . . (This part of the discussion, including an account of the method employed, the results, and certain of the possibilities of interpretation, I have thought it best to omit here, in view of the fact that I have covered practically the same ground in statements already published. I will merely note that the offspring of warm-room and cold-room mice, although themselves reared under identical temperature conditions, presented differences of the same sort as had been brought about in their parents through the direct effect of temperature, viz., differences in the mean length of tail, foot and ear.) There remain two principal alternative explanations, which are not wholly distinguishable from one another, and neither of which admits of being stated except in rather vague terms. One of these is the assumption that the changes under- gone by the parent body are in some way registered in the germ cells, so as to be repeated, in a certain measure, in. 96 THE AMERICAN NATURALIST — [Vou. XLV the body of the offspring. This conception has taken various forms, commencing with Darwin’s hypothesis of ‘‘nangenesis.’’ The same general view has recently been restated in chemical terms, and in a manner which is perhaps far less shocking to our common sense. The other alternative is that of a ‘‘ parallel induction”’ or ‘‘simultaneous modification of the germ-plasm,’’ through the direct action of the modifying agent. This explanation, as we all know, has been freely used by Weismann and others to account for a considerable range of phenomena, notably the persistence of temperature effects in a second generation of butterflies. The phrase has indeed become so familiar through long repetition that few of us stop to consider just what it implies. ‘Parallel modification of the germ-plasm?’’ How can the unformed material of the germ cells be modified in the same manner as certain groups of somatic cells—say in a butterfly’s wing—even by an all-pervading influence like temperature? This is obviously not what is intended. What we mean, concretely stated, is this: the germinal matter is so affected by the temperature that, after some hundreds or thousands of cell generations, certain of the resulting cells will show peculiarities in their pigment- producing powers of the same nature as those which arose directly in the somatic cells of the parent. And a most curious feature of this coincidence is that these modified cells are situated in precisely the same parts of the body in the one case as in the other.” Thus do these very simple explanations have a way of losing their simplicity when examined critically. In the present instance, the hypothesis stated may, for all we know, be the one that most nearly represents the truth. But it should be stated frankly, in all its complexity, and not palmed off upon us as a readily intelligible hypoth- esis, which relieves us of the necessity of adopting an ‘‘inconceivable’’ one such as that of pangenesis. * Weismann’s ‘‘determinant’’ hypothesis offers at least a formal solution of this difficulty, but I think that most biologists will agree with me that the solution thus offered is almost wholly a formal one. No.530] TEMPERATURE ON GROWING MICE 97 In the case of a warm-blooded animal, of course, such an explanation as the foregoing could not be offered with- out still further modification. It might be conceded that temperature, as such, could not affect the germ-cells to any appreciable extent. But it might, on the other hand, be contended that the effects of temperature, even upon the parent body itself, may not be direct, but may be due to the formation of specific chemical substances, which, through the medium of the blood, may be supposed to simultaneously influence the body and the germ-cells. Thus we should, after all, be invoking a ‘‘simultaneous modification of the germ-plasm,’’ as in the case of cold- blooded animals. Such a conception, vague as it is, has certain decided elements of strength. Let me point out, however, as I have already done more than once, that any such chemico- physiological mechanism as is here assumed would be of nearly or quite the same value for evolution as the ‘‘inheritance of acquired characters’’ in the old sense. An interpretation of this sort might ‘‘save the face’’ of certain speculative students of heredity, but the differ- ence between the two views would have little but academic interest. At the present time, I am continuing these experiments with mice, and am not only using much larger numbers than hitherto, but am resorting to several variations of the original theme, by which I hope to reduce the number of possible interpretations to a minimum. A friend wrote to me recently, wishing me no end of ‘‘good results, not Lamarckian.’’ This doubtless represents the attitude of » a large number of persons toward the whole subject. By many, anything with a taint of ‘‘Lamarckism’’ about it would seem to be, ipso facto, beyond the pale of legitimate scientific investigation, belonging rather to the same cate- gory as pre-natal influences, telepathy and the ‘‘border- land’? phenomena of psychical research. But the dawn of better times is already with us. In conclusion, let me state that my own attitude toward 98 THE AMERICAN NATURALIST [ Von. XLV this group of problems is one of indecision. If I confess to you, as I am bound to do, that positive results from my own experiments will give me far greater satisfaction than negative ones, this is chiefly because negative results commonly prove nothing. The question would be left very nearly as it was before. This, of course, constitutes a serious defect in my own vaunted method of attacking the problem, a defect which it shares, however, with any other which could be devised. But any results are better than no results, and these problems seem worth afar more thorough testing than they have yet received. The pres- ent experiments ought, as Professor MacDougal has pointed out, and the author keenly realizes, to be sub- jected to various checks and controls, and to be continued through a considerable series of generations. It is my own fervent hope to be able to carry out such a program. * Presidential address before the American Society of Naturalists, read at Ithaca, December 29, 1910. THE MENDELIAN RATIO AND BLENDED IN- HERITANCE! SHINKISHI HATAT THE Wistar INSTITUTE or ANATOMY Tue indefatigable efforts of neo-Mendelists have suc- ceeded in bringing numerous cases of inheritance, which had previously been considered incompatible with Men- del’s law, into their domain by widening the original limitations. We still have. many instances such as blended inheritance which can not apparently be harmon- ized with the law of Mendel. Recent experiments which demonstrate the existence of various degrees of domi- nance as well as the mutability of the determinants in their behavior, suggested to the writer that various forms of inheritance might be considered as degrees of modification of the law of Mendel. With this view in mind, I have attempted to obtain some general expres- sion for the underlying principle of the law of inherit- ance by which means Mendel’s original law may possibly be theoretically connected with the other cases. In fact, I was compelled to pursue this investigation in connec- tion with my own experiments on the inheritance of the weight of the central nervous system, though this is not yet ready to present at this time. In carrying out this investigation, I have assumed that the germ plasm is composed of many factors, the true nature of which is unknown, but which in one way or another determine the characters in the offspring. It is these hypothetical factors which are here provisionally called determinants. With this understanding, we ma now proceed to the argument. i Suppose a gamete of one parent after the reducing division contains n determinants, the whole group of t Read before the American Society of Naturalists, December 30, 1910. = 99 100 THE AMERICAN NATURALIST [ Vou. XLV determinants being designated p, and the gamete of another parent also contains after the reducing division n determinants, the whole group being designated q. Then in the first hybrid zygote (F,) there will be con- tained at the time of the union of the gametes 2n determ- inants. As we know, rearrangement takes place during the maturation of the germ cells and we assume this rearrangement to involve a random sampling by which n determinants are taken from the group of 2n. From the theory of probabilities we find that n, n— 1, n— 2 . determinants of either parent contained in the gametes of F, are proportional to the successive terms of the following series: p+ np"—'q 4 nO) pig a n(n - oa oo es 5 ee ie ee a ) The same phenomenon e in the gametes of the other hybrid parent (F,) and since the gametic consti- tution of the two hybrid parents is assumed to be identi- cal with respect to the distribution of determinants (1), the frequency of the various combinations of the de- terminants in the second hybrid offspring (F,) will be — given by the square of (1) or e 2 r n(n — 1)(n — 2) T ar ao 2 (r m - Eri .) (2) which may also be written as follows: (p? + 2pq +a)". This series, or the square of the binomial series, is then the most general expression for the gametic com- position of any hybrid arising from a combination of p and q determinants and may therefore be considered as the underlying principle of any law of inheritance where the idea of determinants is used. It is evident that since the somatic characters in ques- tion depend entirely on the behavior of the determinants, the relative frequency of various zygotes, as well as the character of the zygotes, depends on whether p or q de- No. 530] THE MENDELIAN RATIO 101 terminants are related as dominant and recessive, re- spectively, or whether they blend. Suppose p is recessive and q is dominant in the Men- delian sense, we at once obtain from (2) the general expression for the alternative inheritance or (RR + 2DR + DD)” where n refers to the number of allelomorphic pairs of characters, and the expansion gives a strict Mendelian ratio for any number of allelomorphic pairs of char- acters. On the other hand, if we consider that p and q determi- nants blend with an equal intensity the series (2) will give all grades of hybrid characters between the two pa- rental types, the frequency of which is proportional to the successive terms of a symmetrical point binomial curve, and the maximum frequency will be associated with the midparental types (case of equipotency). Castle’s (’09) experiments with the length of the ear of rabbits illustrates this case. Again let us suppose that p and q determinants blend, but with unequal intensity. According as p or q is pre- potent, the zygote will resemble more closely one or the other parent. The frequency of each type of zygote again will be represented by the symmetrical point bi- nomial curve. Thus the present series (2) represents both alternative and blended inheritance according to the behavior of the determinants. The fact just mentioned, that the expressions for both blended and alternative inheritance are obtained from the same series which represent the gametic composition, suggests that we may possibly obtain cases of blending in character which normally follow the law of the alter- native inheritance, and vice versa, and further we may even obtain both blended and alternative inheritance in the same offspring by subjecting the hybrid parents to different conditions, provided by such treatment we can modify the behaviors or functional activity of the de- 102 THE AMERICAN NATURALIST [Vou. XLV terminants, since as soon as the behavior is altered, we at once obtain from the series (2) another type of in- heritance. Although we have no clear direct evidence which demonstrates an occurrence of such extreme modification in the behavior of determinants, nevertheless the possi- bility of such an event is amply suggested by the recent experiments. For instance Tower (°’10) has shown not only a reversal of dominance and apparent failure of segregation by merely modifying the environment of the beetles, but also a case in which the same parents pro- duce offspring, some of which follow the law of Mendel while others show entirely different behavior with re- spect.to dominance and segregation. Tennent (’10) was able to obtain from a cross of Hipponoë esculenta with Taxopneustes variegatus, reversal of dominance by de- creasing the alkalinity of the sea water. Numerous samples of this sort can easily be found in the recent literature. Whatever be the real condition or conditions which control the behavior of the determinants, one point is clear from the above, that the determinants are not im- mutable in their behavior, but subject to modification. This fact naturally leads us to think that we may obtain various forms of inheritance which are more or less dif- ferent from the type form according to degree of func- tional modification. When a modification is maximum, we may even obtain a case of blended inheritance in a character which normally follows the law of alternative inheritance, or vice versa. The facts mentioned above then indicate that our de- duction from the properties of the formula is not at all improbable. Again the properties of the formula suggests that we can theoretically connect cases of blended inheritance with those of alternative inheritance by the mere as- sumption that p or q fails to dominate either completely or incompletely. Since as we have shown by the degree No. 530] THE MENDELIAN RATIO 103 of dominance, the formula reduces to either equipotent or prepotent blending inheritance. From this stand- point we may consider that blending inheritance is a limiting case of alternative inheritance where either — dominance is absent (equipotency) or is imperfect (heteropotency). If this hypothesis is accepted, then Mendel’s law of alternative inheritance may be taken as the standard, and all cases referred to it or blending in- heritance (though by this some more important features of inheritance are not suggested) may similarly be made the standard, the Mendelian ratios then becoming a special case. In this connection Professor Davenport’s (’07) view on the law of potency is of great interest. As his view of potency is so important, and especially as it clearly ex- plains the relation between Mendelism and cases con- sidered to be non-Mendelian, I shall quote his words at some length. After quoting various cases of inheritance, Professor Davenport says: _ Taking all cases into account, it is clear that Mendel’s law does not cover all; and if not, it must be a special case of a more inclusive law. Can we find a more general expression for the inheritance of charac- teristies which will cover all these cases? I think we can and that it may be called the law of potency. At the one extreme of the series we have equipotent unit characters, so that when they are crossed, the offspring show a blend, or a mosaic between them. At the other extreme is allelopotency. One of the two characteristics is completely recessive to the other. Between the two extremes of equipotency and al lelopo- tency lies the great mass of heritable characteristics which when opposed in heredity, exhibit varying degrees of potency. This sort of inherit- ance may be called heteropotency. Thus Professor Davenport shows also that Mendelian dominance is a particular case of potency, allelopotency, though he did not state that blending inheritance is a limiting case of Mendelism. Whether a new expression ‘‘the law of potency” should be introduced as Professor Davenport has sug- gested, or whether the various potencies may be consid- 104 THE AMERICAN NATURALIST [ Vou. XLV ered as a limiting case of Mendel’s law of alternative inheritance, thus saving the original name, is a matter for later decision, though the latter name seems to me ` more appropriate to retain owing to the fact that the phenomenon of segregation, most important of all, had been first stated by Mendel. Let us now consider a limiting case of our formula (2) when the values of. n (number of allelomorphic pairs of characters) increase. In the typical Mendelian ratio, the relative frequency of the various zygotes with re- spect to any given visible character is proportional to an expansion of (1+ 3)" which is the same as (1/4 + 3/4)" if we consider the relative values of the frequencies. Thus in all known cases of the inheritance, we have to deal with an expansion of (r -+ s)” where r+ s=1. A concise mathematical formula which represents a limit- ing case of the binomial series arising from an expansion » of (r + s)” will be very useful, especially when we are dealing with a quantitative measurement such as weight, length, area, volume, etc., since in these cases the values of the variates will be graded. Further, the theoretical frequency corresponding to each variate when the value of n becomes very large, can best be determined from such a mathematical expression which represents a limit- ing case. Without going into any detail of the mathematical treatment, it will be seen that we obtain two forms of expression according as r= s or rs. The former will be represented by the normal probability curve and the latter by a limiting case of a skew binomial curve. For representing a skew binomial curve we can best use DeForest’s formula (Professor Pearson’s curve of type 3). It may be useful to the reader to know that De- Forest’s formula degenerates into the normal probability curve as its simplest form, as will be seen below. DeForest’s formula (Hatai: 710) is usually written in the following form: No. 530] THE MENDELIAN RATIO ~ 105 1 X a?b—1 = ———_| ] aes eae A a ea +e) a where 1 k= 1 May aor For 12a%b ™ 288(a°b)? a =— quotient of twice the second moment divided by the third moment. b — second moment. Writing c for 1 ky 2mb we have log (2) = (e0 — 1) log (1 +3) æ i/Ææ\ that? 1oy" = (ab — Dia- la T 1 -a a |e a x? 1 x a eae | ey --35+(35- )a-(Ge-3) ab g l ey kab 8 [Nag Since for a vanishingly small value of the third moment, ab will be a very large number, consequently #/ab will be infinitesimal. Thus neglecting all terms in which x/ab is factor, we have Restoring the value of C and remembering that for large values of ab, k reduces to unity, we finally have which is the familiar formula for the normal probability curve. From the above it is clear that DeF orest’s formula and its limiting case represent the frequency distribution of the zygotes, whether we are dealing with alternative or 106 THE AMERICAN NATURALIST [ Vou. XLV blended inheritance. One, however, must not be misled to conclude that continuous variation necessarily means failure of segregation, since on the contrary apparent continuity may be a resultant of combinations of various segregating characters. Whether or not given data indi- cate a segregation, may be variously tested by some other means according to the nature of the experiment. From the above we draw the following conclusions: 1. The series obtained from the square of the binomial expresses the distribution of determinants for both alter- native and blended inheritance. 2. Blended inheritance may be considered to be a limit- ing case of alternative inheritance where dominance is imperfect. Thus Mendel’s law of alternative inheritance may be considered as the standard and all other cases referred to it. 3. DeForest’s formula with its limiting case ade- quately represents frequencies of all known cases of inheritance when the number of allelomorphic pairs of characters is large, especially when quantitative meas- urements are considered. LITERATURE CITED 1909. Castle, W. E. Studies of yong in Rabbits. Pub. of the ; Carnegie Inst. of Washington, No. 114, pp. 9-68. 1907. Davenport, C. B. Heredity and Mendel’s Law. Proc. of the Wash. Acad. of Sciences, Vol. 9, p. 179. 1910. Davenport, C. B. e Imperfection of pR and some of its Consequences. Am. Naturalist, Vol. 44, pp. 150-155. 1910. Hatai, §. DeForest’s Formula for ‘‘ An Unsymmetrical Probability Curve.’’ Anat. Rec., Vol. 4, No. 8, pp. 281-290. 1910. Tennent, D. H. The Dominance ps E or of Paternal Char- acters in Echinoderm RIET . f. Entwicklungsmechn. d. O 1910. Tower, W. L. The Determination of Dominance and Modification of Behavior in Alternative (Mendelian) Inheritance, by Conditions Surrounding or Incident upon the Germ Cells at Fertilization. Biol. Bull., Vol. 18, No. 6, pp. 285-352. DATA ON THE RELATIVE CONSPICUOUSNESS OF BARRED AND SELF-COLORED FOWLS! DR. RAYMOND PEARL, MAINE AGRICULTURAL EXPERIMENT STATION I. Puystcan Data Tue purpose of this note is to put on record a rather striking physical fact, and to discuss briefly its biological significance. Some two years ago Davenport? published a short note regarding the relative number of self-colored and of ‘‘penciled or striped’’ chicks killed by crows one afternoon, at Cold Spring Harbor. The rather striking result was that out of 24 birds killed, only one was other than self-colored. The communication closes with the following words: ‘‘This fragment, then, so far as it goes, indicates that the self-colors of poultry, which have arisen under domestication, tend to be eliminated by the natural enemies of these birds, and the pencilled birds are relatively immune from attack because relatively inconspicuous.’ | Some photographs taken on the poultry range of the Maine Experiment Station this past summer illustrate this point made by Davenport as to the relative conspic- uousness of self-colored birds in so striking and com- vlete a manner as to warrant their publication and a critical discussion of their significance. These photo- graphs were made without any thought whatever at the time that they were going to bring out the relative con- spicuousness of different plumage patterns. Indeed, it was not realized that they did so until the finished prints were given to me by the station photographer, Mr. Roy- 1 Papers from the Biological Laboratory of the Maine Experiment Sta- tion, No. 23. * Davenport, C. B., ‘‘ Elimination of Self-Coloured Birds,’’ Nature, Vol. 78, p. 101, 1908. 107 108 THE AMERICAN NATURALIST [ Vou. XLV i Photograph of a Golden Pencilled Hamburg g- Practically. a solid colored bird (red on body, black tail). The few barred feathers which the of this variety has are covered by solid colored feathers. In this picture one barred feather shows in the region of the saddle. The wind had displaced this feather. den Hammond, to whom I am indebted for developing and printing these pictures. As a matter of fact the four pictures which accompany this note were taken for the purpose of (a) testing a then new camera as to its usefulness in obtaining pictures to form part of a per- manent record system in poultry-breeding experiments, and (b) to get photographic records of certain particu- lar birds of interest from one standpoint or another. All the exposures were made by the same person (the writer) on the same afternoon and within an hour of each other. It was on a cloudless afternoon early in August, and the light conditions, shutter-opening, speed, and diaphragm opening were constant for all of the pic- tures. What differences appear in the pictures, then, are such as are referable to the different color patterns of the birds, when seen under the light conditions and against the kind of background which obtained in this case. No. 530] RELATIVE CONSPICUOUSNESS OF FOWLS 109 Fic very dark, practically solid black Fə g from the cross Cornish Indian pute Fı 9 from Cornish Indian g x Barred Rock Q. From these photographs the following poimia are to be noted : 1. As compared with self-colored birds the barred in- dividuals obviously are relatively much less conspicuous, when under the same light conditions, and when seen against the same kind of a background. The pictures of the barred birds (Figs. 3 and 4) are not, to be sure, like the ‘‘puzzle’’ pictures of supposedly protectively colored organisms, which one sometimes sees, where it is exceed- ingly difficult to distinguish the animal from the back- ground at all. In both Figs. 3 and 4 it is easy enough to see the bird, but at the same time these birds are obviously much less conspicuous than those shown in Figs. 1 and 2. 2. This inconspicuousness is equally marked whether the barred bird is in the bright sunlight (Fig. 3) or in a relatively deep shadow (Fig. 4). 3. These pictures furnish objective and unbiased phys- 110 THE AMERICAN NATURALIST [ Vou. XLV 2 ate cross-bred chick. .Sex 9. Produced by mating Fi barred ‘croes-breda inter s$ ical evidence regarding the relative conspicuousness of two types of plumage pattern. II. Dara on tHE Bronocican VALUE or THE [NCONSPIC- UOUSNESS OF THE BARRED PATTERN The physical fact set forth above is obvious: barred chickens are clearly less conspicuous than self-colored when seen against the background of grass on the range where they live. Has this physical fact any biological significance? Are the barred birds, by virtue of the possession of this color pattern, at any advantage in the struggle for existence? Is their relative inconspic- uousness any real protection against their natural ene- mies? It is the purpose of this section of the paper to present some numerical data regarding this matter. The only evidence which exists in the literature on this problem, so far as poultry is concerned, consists in the admittedly fragmentary statistics presented by Daven- port, which have been cited above. It should be pointed No. 530] RELATIVE CONSPICUOUSNESS OF FOWLS 111 F ure Barred Plymouth Rock Q, with barring of fine quality from the fyn i Spd It is to be noted that the bird in this figure is in the shadow of a T in contrast to that shown in Fig. 3, which is standing in the bright s out that Davenport’s data are fragmentary not alone in respect to the small number of deaths (eliminations in the technical sense) involved, but also because these deaths were due to but a single one of the natural ene- mies of poultry, namely the crow. There are, of course, many others. Under the conditions prevailing on or about the poultry plant of which the writer has charge the following animals are regular or occasional destroy- ers of young chicks: Rats, skunks, foxes, crows, hawks, cats.’ In different seasons the relative importance of these different enemies varies. Thus in the breeding season of 1908 many birds were killed by foxes. In 1909, the year “To this list one feels tempted to add that species of vermin which is in some respects the worst which attacks a poultry plant, namely the thief, but fortunately the range was free from his depredations in 1909. 112 THE AMERICAN NATURALIST [ Vou. XLV for which statistics are given below, not a single bird was killed by a fox so far as is known. Similarly in 1909 no birds were killed by skunks. In 1910 a skunk succeeded in getting into a house one night and killed a number of birds. On the Maine Station plant normally predaceous birds undoubtedly rank first in destructiveness. This is probably quite generally true of poultry plants, though because of the fact that the loss is distributed so evenly over the whole season the importance of this class of enemies is apt to be underrated. Next to predaceous birds stand rats, under our conditions. An important point to be noted is that on the plant under discussion here all killing of chickens by rats is done in the daytime. Rats burrow in the ground under the houses, and then when the chicks are playing about a rat will dash out, seize a chick and carry it back to the burrow. It is not an uncommon occurrence for a rat thus to kill as many as 12 chickens within the space of an hour. With rare ex- ceptions we never lose any chickens at night except those taken by thieves. The chicks are shut and locked in rat and (usually) vermin proof houses at night. Occasion- ally, as noted above, a skunk is able to effect entrance into a house. This, however, did not happen in 1909, the year which furnished the statistics given below. It should be clearly understood that in the statistics which follow all ‘‘eliminations’’ occurred in the daytime, when color and pattern might presumably be of some signifi- cance. It is my purpose to present some statisties, involving a relatively large number of individuals, regarding the relation of color pattern to the elimination of chickens by all of these natural enemies taken together. These statisties cover the hatching season of 1909 in which chickens were on the range, and subject to the attacks of enemies, from about April 1 to October 1. Birds of all colors and patterns ran together on the same open, turf-covered range, and, without regard to color or pat- tern, all were equally exposed to attack by all sorts of No. 530] RELATIVE CONSPICUOUSNESS OF FOWLS 113 natural enemies. The total number of chickens involved was 3,345. An account of the way in which the statistics were obtained is necessary. All of these 3,343 chicks were of known pedigree, and a numbered aluminum leg band was attached to each one when it was re- moved from the incubator in which it was hatched. A record was made of each chick’s number. This num- bered leg band was worn by the chick throughout its life. Whenever a chick died a record of this fact was made opposite its entry in the pedigree book. Dur- ing the season every living chick on the range was handled over twice and its leg band number checked back with the original entry, and at the end of the season all chicks remaining on the range were checked up. Now it is clear that dead chicks which come to autopsy will fall into two general classes: on the one hand, those that died from one or another of the many diseases which make the poultry raiser’s life a burden in the springtime, and on the other hand, those killed by some enemy but not carried away. In the latter class will fall the great ma- jority killed by rats, some killed by skunks, and a fair proportion of those killed by foxes. Usually a direct record can be obtained for practically none of the chicks killed by predaceous birds and cats. In 1909 we have reason to believe that substantially all unrecorded deaths were caused by predaceous birds. At the end of the season when the birds are checked up all will be accounted for as either (a) living, (b) dead from some disease, (c) killed by recorded enemies, or finally (d) missing. Of the missing birds there are two classes again. On the one hand are those killed by ene- mies which carried the carcases away, and on the other hand, are those that through accident lost their leg bands, and hence, while present in the flock, can not be entered upon the records. With the methods of work in use here the number of the latter class has always been small. Unfortunately I am not able to give exact figures for such birds for the season of 1909. It can be stated with cer- 114 THE AMERICAN NATURALIST [ Vou. XLV tainty, however, that they did not exceed 25. Of this number that lost their leg bands 8 were known to be self- colored birds. There were on the range in 1909 three classes of birds, in respect to color pattern. These were (a) barred birds, bearing either the pattern of the pure Barred Plymouth Rock, or a modification of it;* (b) solid (self-colored) black birds, resulting from the cross Cornish Indian Game 3 X Barred Rock 2; and (c) pure Cornish Indian Games of the dark variety which may for present pur- poses be classed as self-colored birds. With this somewhat lengthy explanation of the com- position of the flock and method of keeping records in hand we may proceed to examine the statistics of. elimi- nation. In compiling these statistics the blank birds which lost their bands (ca. 25) have been included with the eliminated. This does not affect the conclusions in any way because of the facts that (1) the number of such birds is so small relatively, and (2) the proportion of self-colored to barred birds among those which lost their bands is relatively higher than in the general population from which they came. The significance of this point will be apparent as we proceed. We have the following figures, it being understood that ‘eliminated’? means ‘‘killed by natural enemies’’ with the inclusion of the small number of birds which lost their bands as noted above. Total number of birds = 3,343. Number of barred birds = 3,007. Number of self-colored birds = 336. Total number of eliminated birds = 325. Number of barred birds eliminated = 290. Number of self-colored birds eliminated = 35. The above figures include all eliminated birds, those killed by recorded and unrecorded enemies together. If we take only those killed by recorded enemies, which *See Pearl, R., and Surface, F. M., ‘‘On the Inheritance of the Barred Color Pattern in Poultry,’’ Arch. f. Entwicklungsmech., Bd. XXX, Fest- Band fiir Roux), pp. 45-61, 1910. No. 530] RELATIVE CONSPICUOUSNESS OF FOWLS 115 under the conditions prevailing on the plant in 1909 means practically those killed by rats, we have: Number of barred birds eliminated by recorded ene- mies — 68. Number of self-colored birds eliminated by recorded enemies = 6. . From these figures the following proportions are de- rived: Of the total number of birds 10.05 per cent. were self-colored. Of all the eliminated birds 10.77 per cent. were self- colored. If we consider by themselves the birds eliminated by recorded enemies, we have: Of the birds eliminated by recorded enemies 8.11 per cent. were self-colored. Putting the figures in another way we have: Of the self-colored birds 10.42 per cent. were elimi- nated by all enemies. Of the barred birds 9.64 per cent. were eliminated by all enemies. Of the self-colored birds 1.79 per cent. were eliminated by recorded enemies (chiefly rats). Of the barred birds 2.26 per cent. were eliminated by recorded enemies. Of the self-colored birds 8.63 per cent. were eliminated by unrecorded enemies (chiefly predaceous birds). Of the barred birds 7.38 per cent. were eliminated by unrecorded enemies (chiefly predaceous birds). The conclusion to be drawn from these figures, which involve a large number of individuals, is obvious. It is that the relative inconspicuousness of the barred color pattern afforded its possessors no great or striking pro- tection against elimination by natural enemies, during the season (April 1 to October 1) of 1909 on the poultry range of the Maine Experimental Station. It might be objected that if the eliminations by predaceous birds alone could be separately recorded it would then be found that against this class of enemies the barred pattern had 116 THE AMERICAN NATURALIST [Vou. XLV great protective value, as suggested by Davenport’s fig- ures. This, however, can hardly be the case in the pres- ent statistics since if it be assumed that predaceous birds killed relatively few barred chicks and relatively many self-colored, then it must also be assumed that the other unrecorded enemies showed a preference for barred birds, since with all enemies taken together substantially equal proportions of both kinds of birds were eliminated. In other words, if we assume a selective elimination in the case of predaceous birds, we are obliged to assume an equal and opposite selective elimination on the part of other unrecorded enemies. There is no evidence on which such an assumption could be based. These figures, of course, cover only one year’s expe- rience, and are in no wise conclusive, but general obser- vation indicates strongly that essentially the same re- sult would be shown in other years if it were possible to tabulate the figures. Unfortunately neither the records of 1908 nor 1910 can be used for this purpose. In 1908 there were almost no self-colored birds on the range. In 1910, owing to the location of the houses on the range and other circumstances which can not be gone into in detail, thieves were active on the plant and the birds taken were not a random sample of the flock in respect to color. 1909 was a fortunate year for such a study as the present one. The thieves con- fined their attention to adult stock on a part of the plant | away from the chicks, and left the latter strictly alone. Definitely controlled observations regarding the elim- ination of animals by natural enemies, covering a consid- erable number of individuals and anything like a com- plete range of enemies, are exceedingly scarce. The whole question of the interplay of factors in the ‘‘strug- gle for existence” constantly going on in the organic world has been discussed very largely from the a priori standpoint, throughout the whole period since the ap- pearance of the ‘‘ Origin of Species.” The ‘‘rabbit with his legs a little longer,’’ the ‘‘fox with the little keener No.530] RELATIVE CONSPICUOUSNESS OF FOWLS ia sense of smell,’’ the ‘‘bird of dull colors which har- monized with the background,” et id genus omne, have been made to do valiant service. Ever since the first description, made by the Nurem- berg miniature painter Rösel in 1746,° of a case of pre- sumably protective coloration, we have been prone to argue that because an organism was colored or formed in such a way as to be inconspicuous it was, therefore, nec- essarily protected from attack by its enemies to a greater or less degree. The logic of such reasoning is flawless. It ought to be protected. But a conclusion may be per- fectly logical and still not true. In the study of pro- tective coloration, including mimicry, it is essential that a discovery that an organism is to human eyes inconspicu- ous or not readily distinguishable from some other or- ganism shall not be considered the final goal. Rather let such a discovery always be supplemented by an experi- mental or observational determination of whether this inconspicuousness really helps the organism, in actual practise, in avoiding elimination by natural enemies. It is worth noting that more than one recent critical stu- dent of these problems who has applied this method has brought to light results essentially similar in their gen- eral import to those set forth here.’ *Cf. Müller, H., ‘‘Schiitzende Aehnlichkeit einheimischer Insekten,’’ Senei, Jahrg. III, Heft 8, p. 114, 1879. . for mmnte the chapter on ‘‘Colouration of Organisms’’ in Dewar and Finn ’s ‘*The Making of Species’? (New York, seyi , and still more recently the thorough critical study by Punnett on ‘‘ Mimicry in Ceylon Butterflies, with a Suggestion as to the Nature of Polymorphism’? (Spolia Zeylonica, Vol. VII, Part XXV, September, 1910, pp. 1-24, 2 plates). SOME CONSIDERATIONS CONCERNING THE PHOTOGENIC FUNCTION IN MARINE ORGANISMS F. ALEX. McDERMOTT WASHINGTON, D. C. In two very interesting papers, Professor C. C. Nutting’ has brought forth evidence tending to show that in oceanic depths below the range of penetration of the sun’s rays, there exists a dim, phosphorescent light, quite general in its distribution, radiated from various photo- genic organisms of the abyssal regions, and having a defi- nite and valuable significance for the life of animal forms at these depths. That such a light actually exists is scarcely to be sanely doubted, in view of the evidence of the deep-sea explora- tions which have added so much to the knowledge of oceanic conditions. And that it has a purpose in the life of the forms inhabiting those portions of the ocean beds where it exists, seems to the writer equally undeniable, unless we accept Emerson’s poetic reasoning that “ Beauty is its own excuse for being.” Just what its purpose may be in hermaphroditic, simple forms not provided with definite organs of sight, and indeed also in many higher forms, may, of course, still be a legitimate subject for investigation and consideration. Professor Nutting’s remarks have been of special interest to the writer in connection with some recent studies made by the latter on the general subject of bio- photogenesis, with special reference to the: Lampyride.* *(a) ‘*The Utility of Phosphorescence in Deep-sea Animals,’’ page Nar., Vol. 3, 1899, pp. 792-799; (b) ‘‘The Theory of Abyssal Light,’’ Pros. VII Dano. Zool., advance reprint, 1910. 2 Amer. Journ. Physiol., 1910. Vol. 27, pp. 122-151; Canad. Entomol., 1910, Vol. 42, pp. 357-363; Popular Sci. Monthly, 1910, Vol. 77, pp. 114- 121. 118 No. 530] THE PHOTOGENIC FUNCTION 119 The coloring and photogenicity of the organisms found in the depths of the sea show some similarities to the corresponding features of life on land. Take the family Buprestide, of the genus Coleoptera, of the order of insects. The insects of this family are probably the most brilliantly colored of any of the beetles, and are colored quite as brilliantly as the insects of any other genus. The colors cover a quite wide range of metallic, polished, glistening greens, blues, reds, coppery and golden; many of the smaller species wear more somber dark blues, browns and blacks, but as a class they are brilliant and showy. Obviously, these colors would be invisible in the absence of light, and need a light of considerable intensity to bring out their full value. Now we find that almost without exception these Coleoptera are diurnal; they attain their maximum activity during the brightest daylight, and fly but little at night. But one species has been reported to be luminous, and unless this report is pretty definitely confirmed there is grave reason to doubt its authenticity. Now let us consider the Lampyride: The beetles of this family of almost eleven hundred species are in the great ° majority of instances, luminous; the non-luminous species form a decided minority of the true Lampyride. They are also, in the great majority of cases, mainly nocturnal in habit, hiding out of the sunlight during the day; those species which are markedly diurnal in habit are also those which are non-luminous, or in which the luminosity is relatively slight. In coloration, they show none of the bright metallic, showy colors of the Buprestide; black, gray, brown and yellow-brown predominate, with occa- sional red markings, yellow stripes and indistinct lines and spots. In them, the photogenic function possesses at least two definite significances: (1) it is an adjunct of the sexual organism of the insect, rendered of value to them by reason of their nocturnal habits, and (2) it has a pro- tective value. In the larve it might also be considered to have an aggressive value, in attracting the snails. etc., 120 THE AMERICAN NATURALIST [Vou. XLV on which they feed, but this argument would not hold for the imagos, which are much more active. Most of the above statements apply with equal force to the Pyrophorini, the luminous Elateride of the tropics; these insects are herbivorous, however, and the aggres- sive significance does not hold for them. It would seem, then, very probable that similar condi- tions obtain in the abyssal region, with its dim weird, phosphorescent light. The light produced by the Lam- pyride has recently been shown by Ives and Coblentz? to have the extremely high radiant efficiency of 96.5 per cent., against 4 per cent. for the best artificial illuminant. The spectrum of this light is a continuous band extending from the upper red to the lower blue with a maximum intensity in the yellow-green. This spectrum is of wider range than that of the sea-forms cited by Nutting,* but can hardly be of less efficiency. The light of the Lam- pyride is generally stated to be yellow, or greenish; there are some slight variations among different species, but in the main the lights are similar; it seems that a great many of the marine organisms also give a light of similar tone. Therefore colors whose wave-lengths are within the limits of those of the emitted lights of these forms, would be distinguishable in such a biophotogenic light. Although we do not yet know the full details of the process of the production of light by living forms, it is not too much to assume that Nature has developed it to a point very near to the maximum possible efficiency, and if such is the case, the luminous oceanic forms could emit a very penetrating illuminating radiation with very little expenditure of energy, and though this light might not be of any considerable intensity, as judged by our eyes, it could undoubtedly serve as quite a useful light to the large-eyed denizens of the deep. The photogenicity of Salpa, Noctiluca and other such simple forms, which are without definite organs of sight, * Bulletin of the U. S. Bureau of Standards, 1910, Vol. 6, pp. 321-336. t Supra b, page 10. No. 530] THE PHOTOGENIC FUNCTION 121 presents other difficulties. It is not, however, necessary to the faculty of perception of light that definite organs should exist. It is a quite well-known fact that certain worms, bacteria, and other low organisms are able to detect ultra-violet rays to which the human organism is wholly without sensible response, and yet these actino- tropic (if a coined word may be pardoned) forms show no definite organs such as might be adapted to the receiv- ing and recording of the very short wave-lengths of ultra- violet light. If, then, existing organisms are known to be affected by ultra-violet rays for which they have no special sense-organs, it is certainly logical to assume that they and other forms may also be susceptible to the longer and more easily discerned wave-lengths of visible light—especially when those wave-lengths comprise mainly the rays possessing the highest illuminating effect —and without the necessity for the existence of ‘‘eyes’’ or other definite light-receiving organs. As a matter of fact Noctiluca, and numerous other marine organisms have been shown to be susceptible to light, although they possess no specific organs for this function so far as we have been able to make out. Another consideration as to the purpose of the light presents itself here. We must consider the nature of the medium in which these creatures live. Water does not lend itself as readily as does air to the diffusion of the particles which produce the sensation of smell; and hence while odors, or speaking more properly, from the stand- point of marine organisms, flavors or tastes undoubtedly exist in the ocean water, they could not, on account of the water currents, lack of diffusion, etc., serve the purpose which the odors of land animals serve of giving indication of the presence and location of the creatures. It there- fore would not be unreasonable to assume that in the gre- garious simple luminous marine forms, the photogenic function takes the place to some extent of the animal odors of land forms. To sum up, then: 122 THE AMERICAN NATURALIST [ Vou. XLV From analogy to terrestrial forms, the photogenicity and coloration of marine organisms must play some essential part in their life histories; The absence of definite organs for the reception of the radiations of light may not necessarily indicate that the forms from which they are absent are insensible to these radiations; ' The photogenic function in certain simple marine forms may replace the olfactory function of terrestrial forms, to some extent. SHORTER ARTICLES AND DISCUSSION COMPUTING CORRELATION IN CASES WHERE SYM- METRICAL TABLES ARE COMMONLY USED In studying the assortative mating of Paramecium I have found occasion to compute the correlation in many cases for which double or symmetrical tables are commonly employed. I have found that in such cases the use of symmetrical tables is quite unnecessary and the computations can be performed with much less labor without them. It will, therefore, be worth while to show how the use of symmetrical tables can be avoided. When the two objects to be compared are alike, as when the two members, A and B, of conjugating pairs are examined, evidently either A or B might be entered in either the horizontal rows or the vertical columns of the correlation table. In such cases, the mean computed from the rows, and that computed from the columns are likely to differ, depending on which indi- viduals were entered in the rows, which in the columns. If, for example, the larger individual is always entered in the vertical columns, the smaller in the horizontal rows, as in Table II, then the means and standard deviations of the two sets will differ much. As a result the coefficient of correlation computed in the usual way will show varying values, depending on how the pairs are entered in the table. From the collection shown in Table II we can by varying the method of entering the pairs get coefficients of correlation varying from 0.132 to 0.523. Under such conditions Pearson (1901), Pearl (1907) and others enter each pair twice, once in the rows, once in the col- umns. This gives a ‘‘symmetrical’’ table, in which the sums of either the rows or the columns include all the individuals. This method is theoretically correct, since each individual func- tions both as ‘‘principal’’? and as ‘‘mate’’; the coefficient of correlation computed from such symmetrical tables is the cor- rect one. But such symmetrical tables are cumbersome and involve much labor. Pearl (1907) gives a formula by which the same coefficient can be obtained without making symmetrical tables, by computations involving the two means and standard 123 124 THE AMERICAN NATURALIST [ Vou. XLV deviations and the coefficient of correlation found in the usual way. But it is possible to find the correct coefficient of correlation from ordinary tables, and with much less labor than by either the use of symmetrical tables or by the method given by Pearl. To see how this can be done, it is well to examine a symmetrical table prepared for computation of the coefficient of correlation, such as is given in Table I. Here the large figures give the frequencies, while the subscripts in smaller type give the prod- ucts of the deviations from the approximate mean (37). There are two main points to be considered: (1) How the quantity 30311323 3.4/3 5/3 63 (7/3 8/3 940/41 2143/4445 30 ; k$ 2 31 421 224 4 32 o 1 33 sel y. 32 12 34 2b Bg 2a 13} lf |13 2e |l2 12 35 ld | la 24® 3344|54 |36 23 36 le 13| 32 7 |3| 54 63 30 3 Ht 2-47 ane - 38 la 54 3:|7 22| 5a 5a] 6s| le 18440 39 544 |22| 2 26| 28| Ap 19 4 34 | 36] 6a d | 53| 26| Bq 24 214 la| 2h JBL 41 224120 Sal 2e) 2226 14 42| 135 25 J |65| 29 As lab 17 43 id iLp iL 3 44 lak 22 3 4 ls l 2 |4 |1 1212838033/4019 1114/117/3 |3|1 R50 TABLE I. SEET CORRELATION TABLE FOR THE ag OF 125 PAIRS OF Paramecium au each individual orga ip twice, once in the vertical columns, once in hes fad aces rows. (Unit of AE 4 microns.) (xy) is to be correctly obtained; (2) how the mean and stan- dard deviation are to be correctly obtained. 1. With regard to the first point, it will be observed that such a table is divisible by a diagonal passing from the upper left- hand corner to the lower right-hand corner into two halves which are in all respects duplicates as regards both frequencies and deviation products. (The frequencies through which the diag- onal line passes are to be divided evenly between the two halves.) It is evident, therefore, that if we use only one of these halves No.530] SHORTER ARTICLES AND DISCUSSION 125 of the table in getting the sum S(wy) we shall get just one half the sum we should get by using the whole table; the sum for the whole table would therefore be obtained simply by doubling this half-sum. Now, if in place of making a symmetrical table we enter always the larger member of each pair in the vertical columns, the smaller in the horizontal rows, we shall get a table that is precisely one of these duplicate halves of the symmetrical table; this will be seen by comparing Tables I and II. The quantity (xy) from such a table will then be just half that from the symmetrical table; it may then be doubled, and the further computation will be identical with that for the symmet- 30, 31 32 33, 34 35, 86.37 38, 33, 40,41, 42,43, 44, 45, A B C 30, ] L 2 2 31 1, 1, ar ie 4 $2, wp 1 33, 2,6 22| ls 32 Epa 21/12 34, 1, | 26 }15] 4 | Is Qs} | ded [9] 3422 35, 2, |3.|4 || |3 | fiq 6 2s 36, z 2:114 |3: | 52 | 63 we 23| 7 |30 38, 2, |22| 53| 54| 6s| 16| |1s{22/18 |40 39, Lel 2e | 20} Re | 7}a2}19 40, ls | 22] 215] ne] 2a] | 8/23 |31 41, he 1]13 |14 42, i {| p e | n? 43, ; g 3| 3 i x ETE 3| 3 a ee Ske es NEE 45, zis 8 23 6 7 ele 12231916 3 3 1J. LE II. THE SAME TABLE SHOWN IN TABLE I, SAVE THAT EACH INDIVIDUAL ENTERED BUT ONCE—the larger member of the pair in the vertical column, the smaller in the horizontal row. rical tables. Or (as we shall see) this half sum, which forms the dividend in obtaining the coefficient of correlation, may be divided by a number half as great as in the symmetrical tables, giving the same result. It will further be seen that if in place of entering all pairs in the same way—the larger members in the columns, the smaller in the rows—we enter some or all of them differently, this will make no difference in the result. If in Table II, for example, the pair showing measurements 44 by 34 were entered in the reverse way, it would fall, no longer in the right upper quad- 126 THE AMERICAN NATURALIST [Vou. XLV rant, but in the left lower quadrant, at the point marked X. Here, as examination will show, it would receive the same sub- script that it has now, and would count as negative, exactly as it now does. Again, suppose the pair 36 by 31 were similarly transposed ; it would still fall in the left upper quadrant, at the point marked Y, where it would receive the same subscript as at present and count as positive, just as at present. And so of all other cases; the value of a pair is not altered in any way by changes in the way it is entered in the table. In making the table, therefore, the pairs may be entered only once and quite at random, or in any way that is convenient. 2. With regard to the mean and standard deviation, the ap- parent advantage of symmetrical tables is that they give us the actual mean of all the individuals; it is to this mean that our correlation must refer. But this actual mean ean readily be obtained from the tables in which each pair is entered but once, in any way that happens to be convenient. It is merely neces- sary to add together the sums of the rows and of the columns of the table. Thus in Table II the number of individuals having the length 35 is not 17 (sum from the row beginning.with 35), nor 6 (sum from the column headed 35), but 23 (sum from both the row and the column) and so for all other classes. It will be well to illustrate by an example certain of the steps in the com- putation. Table II shows a correlation table of single entry, as prepared for computation of the coefficients of correlation and other constants. After finding the sums of the rows (given in column A at the right) and of the columns (given in B, underneath), we place the latter sums (B) by the side of A, in the proper places (as at B’), then add the two sets, giving the sums shown in the column C at the right. These are the same sums that we should get from a symmetrical table; adding these we get the total number of individuals (250 in Table II). Now from this column C we find the approximate mean in the usual way; it lies in this case at the length 37 (with 38 individuals). Through the column and the row headed 37 we therefore draw the lines serving as axes of reference in finding the correlation- We now find the correlation in the usual way. In so doing (1) we make use always of the sums in the column C in finding mean, stan- dard deviation, ete. (2) We use for both horizontal and ver- tical axes of reference in computing the correlation in all cases No.530] SHORTER ARTICLES AND DISCUSSION 17 a row and column with the same heading (37 in this case). (3) We employ the ordinary frequencies in the body of the table in getting the sum of the deviations of (xy) for use in the formula for the coefficient of correlation, just as in ordinary cor- relation tables. The computation of the coefficient is of course (as in the case of symmetrical tables) considerably simpler than in the usual case, since we have but one standard deviation and one quantity d to deal with. Only one other point in the computation is peculiar, requiring careful observance. If we let n signify the number of pairs and N the number of individuals (so that N = 2n), then in find- ing the mean, standard deviation, and coefficient of variation, we use N (just as in symmetrical tables), so that the formula for the standard deviation is = (OD a 0nk But in getting the coefficient of correlation, the sum S(zy) : which we get from our unsymmetrical table is just half what we _ should get from a symmetrical table (as we have already seen). Therefore, to make the computations identical with those for symmetrical tables, we must either double this sum in the for- mula for the coefficient of correlation, or what is simpler, in place of doubling this sum we may halve the number by which we divide this sum, that is, we may use n in place of N. us the formula for the coefficient of correlation becomes by this method S 1 = (=e — «) x a This method lends itself readily to the valuable procedure recently described by Harris (1910) for finding the coefficient of correlation, the only point requiring careful attention being the fact that in finding the standard deviation we must use N (number of individuals), while in the formula for the coeffi- cient of correlation we must use n (number of pairs). The present plan is likewise well adapted for finding the coefficient of correlation by the ‘‘difference method’’ (see Harris, 1909). If the method we have described is used, the pairs are entered in the table but once, in any way that is convenient; the correla- tion computed will always be the same, and identical with that from symmetrical tables. It avoids the cumbersome and labo- 128 THE AMERICAN NATURALIST [ Von. XLV rious symmetrical table; at the same time it involves much less labor than the method given by Pearl. When there are many tables to be computed, the amount of drudgery it saves is great. PAPERS CITED 1909. Harris, J. A. A Short Method of Calculating the Coefficient of Correlation i in the Case of Integral Variates, Biometrika, 7, 21 1910. Harris, J. A. The Arithmetic of the Product Moment Method of Caleulating the Coefficient of Correlation. Amer. pate 44, iS) 693-699. 1907. Pearl, R. A Biometrical Study of Conjugation in Paramecium. Biometrika, 5, 213-297. 1901. Pearson, K. Mathematical Contributions to the Theory of Evolu- iors. On the Principle of Homotyposis and its Relation to Heredity, to the Variability of the Individual-and to that of the Race. Philos. Trans., A, 197, 285-379. THE JOHNS HOPKINS UNIVERSITY. H. S. JENNINGS. The Anatomical Laboratory of Charles H. Ward 189 West Avenue, Rochester, N. Y. = OUR HUMAN SKELETONS are selected specimens scientifically prepa: They are undoubtedly the finest and strongest skeletons obtainable, and are purchased by the leading Medical and Literary Colleges, Schools, Surgeons, ete. We make a number of special skeletons for demonstrating dislocations, muscular areas, anthropometric landmarks, muscles, etc. The mounting of the articulations permits movements as in life. Strength and rigidity are secured by the use of a special bronze wire of — tensile strengt and great eee 1o to oxidation. Portability ease of d y our nickeled steel clutch stand- ard which is a great protection as well. These are entirely set up, carefully wrapped, and with detailed en for omk and handling. Our Catalogue gives further detai OUR ste OF TYPES OF VERTEBRATES are large ‘specimens, principally of Cassini as mounted in characteristic poses on polished og nickel plated Ł es We offer a Pegs § ` r ee NON; lleges. _ ANATOMICAL MODELS _ anatomical models i in 1 great vi variety. These have been purchased b by many a ‘Schools and Universities. TI ead, brain, 2 a aoa as organs, ete., of most | oe of which we have large half tones. ‘These ERREF on no duty, no trans- p , no middlemen’s s profits, but are sold direct, at very moderate | ; when needed . i prices, ouu We: aso make BIOLOGICAL MODELS of the forms s commonly e laboratory. : The American Naturalist A Monthly Journal, established in 1867, Devoted to the A sca SEE etches. Aei a Ooa Ente aad of the Biological Sciences Heredity CONTENTS OF THE AUGUST NUMBER : Chromosomes and Heredity. Professor T. H. MORGAN., 3 Spiegler’s “White Melanin” ey oneness ae : Recessive White. Dr. Ross AIKEN GoRTNER. ‘Shorter Articles and Correspondence: A Š _ Contribution to Our Knowledge of Wasps: Tia ak Tela Heredity, Dr, W. J. Serran. . Pickwickian Professor CONTENTS OF THE SEPTEMBER NUMBER Nuclear Phenomena of Sexual mopar in the Dr, BRAD: Moore Dav: Nuclear Phenomena of Sexual aera in Fungi, PROFESSOR R. A. HARPER. The Pose of the Sauropodous Dinosaurs. Dr, W. D. MATTHEW. Shorter Articles and Discussion : Evolution without Iso- and Literature imal Structure and Habits, AA G. H. PARKER, Plant Physiology, ©. L. CONTENTS OF THE NOVEMBER NUMBER eee of Skin Pigmenta tation in Man. Geamkues S B. DAVENPORT. MARCH, 1911 The American Naturalist MSS. intended for publication and books, ete., intended for review should sent to the Editor of THE AMERICAN NATURALIST, Garrison-on-Hudso n, New York Articles containing research work bearing on ar e " iph evolu- tion are ph ea welcome, and will be given preference in publica e hundrea reprints Splat are sippii to authors m of charge. Further’ reprints will = supplied at asain should be sent to the publishers. The subscription price is y> dollars a year. — postage is fifty cents and postage -five cents additional. The narge for single copies is thirty-five cents. The advertising rates are Four Dollars for a pa o THE SCIENCE PRESS i Lancaster, Pa. _ Garrison, N. Y. ook EW YORK: Sub-Station 84 Entered as tter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of ; : of March 3, 1879. x THE BULLETIN—Porta in Ethnolograpn. | MARINE BIOLOGICAL LABORATORY ical and Pre-historie io Specimens, Books on Natural. | Ser oE ARTEN nt lr nimi “pores repre ape e e ney > BERA & Prva: ed of Algae, K Liverworts and ‘oases. 4 Duke Sty Adetphi—London—England | For price lists information, ad GEORGE S GRAY. Curator, Woods. Hole, Mass. Sep oR ee cree | z r ambridge > University Press. d TE ON, M.A cll. Dicter ot thie John tombs T aaier ea - PUNNETT 1 M-A., Professor of ot Biology in the Un t THE AMERICAN NATURALIST VoL. XLV March, 1911 No. 531 THE GENOTYPE CONCEPTION OF HEREDITY! PROFESSOR W. JOHANNSEN UNIVERSITY OF COPENHAGEN Brotoey has evidently borrowed the terms ‘‘heredity”’ and ‘‘inheritance’’ from every-day language, in which the meaning of these words is the ‘‘transmission’’ of money or things, rights or duties—or even ideas and knowledge —from one person to another or to some others: the ‘theirs’’ or ‘‘inheritors.’’ _ The transmission of properties—these may be things owned or peculiar qualities—from parents to their children, or from more or less remote ancestors to their descendants, has been regarded as the essential point in the discussion of heredity, in biology as in jurisprudence. Here we have nothing to do with the latter; as to biology, the students of this science have again and again tried to conceive or ‘‘explain’’ the presumed transmission of general or peculiar characters and qualities ‘‘inherited’’ from parents or more remote ancestors. The view of natural inheritance as realized by an act of transmission, viz., the transmission of the parent’s (or ancestor’s) personal qualities to the progeny, is the most naive and oldest conception of heredity. We find it clearly devel- oped by Hippocrates, who suggested that the different parts of the body may produce substances which join in the sexual organs, where reproductive matter is formed. * Address before the American Society of Naturalists, December, 1910. 129 130 THE AMERICAN NATURALIST [Vou. XLV Darwin’s hypothesis of ‘‘pangenesis’’ is in this point very consistent with the Hippocratic view, the personal qualities of the parent or the ancestor in question being the heritage. Also the Lamarckian view as to the heredity of “acquired characters’’ is in accordance with those old conceptions. The current popular definition of heredity as a certain degree of resemblance between parents and offspring, or, generally speaking, between ancestors and descendants, bears the stamp of the same conceptions, and so do the modern ‘‘biometrical’’ definitions of hered- ity, e. g., as ‘‘the degree of correlation between the abmodality of parent and offspring.’’ In all these cases we meet with the conception that the personal qualities of any individual organism are the true heritable elements or traits! This may be characterized as the ‘‘transmission-con- ception” of heredity or as the view of apparent heredity. Only superficial instruction can be gained by working on this basis. Certainly, medical and biological statisticians have in modern times been able to make elaborate state-- patti of great interest for insurance purposes, for the ‘‘eugenics-movement’’ and so on. But no profound insight into the biological problem of heredity can be gained on this basis, for the transmission-conception of heredity represents exactly the reverse of the real facts, just as the famous Stahlian theory of ‘‘phlogiston’’ was an expression diametrically opposite to the chemical reality. The personal qualities of any individual organ- ism do not at all cause the qualities of its offspring; but the qualities of both ancestor and descendant are in quite the same manner determined by the nature of the ‘‘sexual substances’’—i. e., the gametes—from which they have developed. Personal qualities are then the reactions of the gametes joining to form a zygote; but the nature of the gametes is not determined by the personal qualities of the parents or ancestors in question. This is ued modern view of heredity. No. 531] GENOTYPE CONCEPTION OF HEREDITY 131 The main result of all true analytical experiments in questions concerning genetics is the upsetting of the transmission-conception of heredity, and the two differ- ent ways of genetic research: pure line breeding as well as hybridization after Mendel’s model, have in that respect led to the same point of view, the ‘‘genotype- conception’? as we may call the conception of heredity just now sketched. Here we can not trace the historical evolution of the ideas concerning heredity, not even in the last ten years, but it must be stated as a fact that a very great number of the terms used by the modern biological writers have been created under the auspices of the transmission-concep- tion, and that perhaps the greater number of botanists and zoologists are not yet emancipated from that old con- ception. Even convinced Mendelians may occasionally be caught using such words as ‘‘transmission’’ and other now obsolete terms. The science of genetics is in a transition period, becom- ing an exact science just as the chemistry in the times of Lavoisier, who made the balance an indispensable imple- ment in chemical research. The ‘‘genotype-conception,’? as I have called the modern view of heredity, differs not only from the old ‘‘transmission-conception’’? as above mentioned, but it differs also from the related hypothetical views of Galton, Weismann and others, who with more or less effectiveness tried to expel the transmission-idea, having thus the great merit of breaking the ground for the setting in of more unprejudiced inquiries. Galton, in his admirable little paper of 1875, and Weismann, in his long series of fasci- nating but dialectic publications, have suggested that the elements responsible for inheritance (the elements of Galton’s ‘‘stirp’? or of Weismann’s ‘‘Keimplasma’’) involve the different organs or tissue-groups of the indi- Vidual developing from the zygote in question. And Weismann has furthermore built up an elaborate hypoth- esis of heredity, suggesting that discrete particles of 132 THE AMERICAN NATURALIST [Vor. XLV the chromosomes are ‘‘bearers’’ of special organizing functions in the mechanism of ontogenesis, a chromatin- particle in the nucleus of a gamete being in some way the representative of an organ or a group of tissues. These two ideas: that ‘‘elements’’ in the zygote corre- spond to special organs, and that discrete particles of the chromosomes are ‘‘bearers’’ of special parts of the whole inheritance in question are neither corollaries of, nor premises for, the stirp- or genotype-conception. Those special ideas may have some interest as expressions of the searching mind, but they have no support in experi- ence; the first of them is evidently erroneous, the second a purely speculative morphological view of heredity with- out any suggestive value. The genotype-conception of the present day, initiated by Galton and Weismann, but now revised as an expres- sion of the insight won by pure line breeding and Mendel- ism, is in the least possible degree a speculative concep- tion. Of all the Weismannian armory of notions and categories it may use nothing. It is a well-established fact that language is not only our servant, when we wish to express—or even to conceal—our thoughts, but that it may also be our master, overpowering us by means of the notions attached to the current words. This fact is the reason why it is desirable to create a new terminology in all cases where new or revised conceptions are being developed. Old terms are mostly compromised by their _ application in antiquated or erroneous theories and systems, from which they carry splinters of inadequate ideas, not always harmless to the developing insight. Therefore I have proposed the terms ‘‘gene’’ and ‘‘genotype’’ and some further terms, as ‘‘phenotype’’ and ‘‘biotype,’’ to be used in the science of genetics. The ‘‘gene’’ is nothing but a very applicable little word, easily combined with others, and hence it may be useful as an expression for the ‘‘unit-factors,’’ ‘‘elements”’ or ‘‘allelo- morphs” in the gametes, demonstrated by modern Mendelian researches. A ‘‘genotype’’ is the sum total of No. 531] GENOTYPE CONCEPTION OF HEREDITY 133 all the ‘‘genes’’ in a gamete or in a zygote. When a monohybrid is formed by cross fertilization, the ‘‘ geno- type’’ of this F,-organism is heterozygotic in one single point and the ‘‘genotypes’’ of the two ‘‘genodifferent”’ gametes in question differ in one single point from each other.’ As to the nature of the ‘‘genes’’ it is as yet of no value to propose any hypothesis; but that the notion ‘‘gene’’ covers a reality is evident from Mendelism. The Men- delian workers have the great merit of being prudent in their speculations. In full accordance with this restraint —a quite natural reaction against the morphologico- phantastical speculations of the Weismann school—it may be emphatically recommended to use the adjectival term ‘‘genotypical’’ instead of the noun ‘‘genotype.’’ We do not know a ‘‘genotype,’’ but we are able to demon- strate ‘‘genotypical’’ differences or accordances. Used in these derivated ways the term ‘‘gene’’ and ‘‘geno- type’’ will prejudice nothing. The very appropriate German term ‘‘Reaktionsnorm’’ used by Woltereck is, as may be seen, nearly synonymous with ‘‘genotype,’’ in so far as the ‘‘Reaktionsnorm’’ is the sum total of the potentialities of the zygotes in question. That these potentialities are partly separable (‘‘segregating’’ after hybridization) is adequately expressed by the ‘‘geno- type’’ as composed of ‘‘genes.’’? The ‘‘Reaktionsnorm”’ emphasizes the diversity and still the unity in the behavior of the individual organism; certainly, the partic- ular organism is a whole, and its multiple varying reac- tions are determined by its ‘‘genotype”’ interfering with the totality of all incident factors, may it be external or internal. Thence the notion ‘‘Reaktionsnorm’’ is fully compatible with the genotype-conception. The genotypes can be examined only by the qualities and reactions of the organisms in question. Supposing *They may therefore be characterized as ‘*mono-genodifferent’’; this term and the further terms ** di-genodifferent’’ and so on, may or may not be of any use. 134 THE AMERICAN NATURALIST [ Vou. XLV that some organisms of identical genotypical constitution are developing under different external conditions, then these differences will produce more or less differences as to the personal qualities of the individual organisms. By simple inspection of series of different individuals it will be quite impossible to decide if they have or have not the same genotypical constitution—even if we know them to be homozygotic.2 We may easily find out that the organ- isms in question resemble each other so much that they belong to the same ‘‘type’’ (in the current sense of this word), or we may in other cases state that they present a disparity so considerable that two or more different ‘‘types’’ may be discerned. All ‘‘types’’ of organisms, distinguishable by direct inspection or only by finer methods of measuring or description, may be characterized as ‘‘phenotypes.”’ Certainly phenotypes are real things; the appearing (not only apparent) ‘‘types’’ or ‘‘sorts’’ of organisms are again and again the objects for scientific research. All typical phenomena in the organic world are eo ipso phenotypical, and the description of the myriads of phenotypes as to forms, structures, sizes, colors and other characters of the living organisms has been the chief aim of natural history, which was ever a science of essentially morphological-deseriptive character. Morphology, supported by the huge collections of the museums, has of course operated with phenotypes in its speculations concerning phylogenetic questions. The idea of evolution by continuous transitions from one ‘*type’’ to another must have imposed itself upon zoolo- gists and botanists, because the varying external condi- tions of life are often‘ shifting the phenotypes in very fine gradations; but also—and that is an important point —because there may always be found considerable geno- typical differences hidden in apparently homogeneous populations, exhibiting only one single ‘‘type’’? around * Here we are not concerned with the question of variable dominance, ete. *Not always, as Bateson has the merit of having emphasized. No. 531] GENOTYPE CONCEPTION OF HEREDITY 135 which the individuals fluctuate. For the descriptive- morphological view the manifestations of the phenotypes in different generations are the main point, and here the transmission-conception immediately announces itself. Hence we may adequately define this conception as a “*phenotype-conception’’ in opposition to the genotype- conception. As already stated, the genotype-conception has been gained in two ways: pure line breeding and hybridization. The first way leads to an analysis of the existing stocks or populations, the second way may realize an analysis of the genotypical constitution of the individuals. The analysis of populations has its most obvious importance in all such cases, where the phenotypes are quantitatively characterized. Even where individuals with consider- able genotypical differences co-exist, the population may —by simple inspection or by statistical appreciation—seem to exhibit only one phenotype, this being usually charac- terized by the average measure of the individuals (dimen- sions, weight, intensity of any quality, number of organs and so on). This is due to the fluctuating variability swamping all limits between the different special pheno- types in question (see the diagram). Populations of self-fertilizing organisms (several cereals and beans, peas and others) have offered the Starting point for pure line breeding as a scientific method of research. A pure line may be defined as the descendants from one single homozygotie organism, ex- clusively propagating by self-fertilization. ‘‘Pure line’’ is a merely genealogical term, indicating nothing as to the qualities of the individuals in question. A ‘‘line’’ ceases to be ‘‘pure’’ when hybridization (or even inter- crossing) disturbs the continuity of self-fertilization. From a population of homozygotic self-fertilizers there can be started (isolated) as many pure lines as there are fertile individuals—of course very many of such pure lines will be quite identical in genotypical constitution and might in reality belong to one and the 136 THE AMERICAN NATURALIST [Vot. XLV REREN ARAS Nae E EEEN Ses SA Sd case the beans enclosed in glass- l are ostly impos- e ne to which it gs.—The fiu tions about the average length ( pure lines as well as in the mix the phenotype) within the ed population show no characteristic difference. No. 531] GENOTYPE CONCEPTION OF HEREDITY 137 same pure line if the genealogy was but sure. The guarantee of the descendence is thus a main point in the principle of pure lines. Identity of genotypical nature is not at all a proof for identical genealogy: the wide- spread confusion of ‘‘resemblance’’ with ‘‘ genealogical relation’’ is the root of much evil—of which the statis- tics of biometricians have given us some instances. The isolation of pure lines from plant-populations has been the instrument for gaining the conviction that se- lection is not able to shift the nature of genotypes. The well-known displacement of the ‘‘type’’ of a popu- lation by selection—this displacement proceeding from generation to generation in the direction indicated by the selection—is due to the existence a priori of geno- typical differences in such populations (see the dia- gram). By selection a relatively great number of those organisms, whose genotypical constitution is favorable for the realization of the desired degrees of any char- acter, will be saved for reproduction; hence the result of the selection! Within pure lines—if no mutation or other disturb- ances have been at work—or within a population in which there is no genotypical difference as to the char- acter in question, selection will have no hereditary influ- ence. This result has in recent years also been reached by several other experimenters in genetics. Here I also may recall the brilliant experiments of H. S. Jennings with Paramecium, experiments which have been carried out quite independently of my own researches and which have been of great importance for the propagation and support of the genotype-conception. The bearing It can not be detected by inspection that the five upper diagrams represent Phenotypes which are genotypically pgp te while the nethermost diagram —the sum of the others—indicates a phenotype masking five others. That ‘eae five phenotypes all are genotypically different is known a priori in this Special case, but it could not be discerned by simple inspection.—In the popula- tion genotypical pital are combined with merely individual fluctuations ; within the single pure line only such fluctuations are seen. Hence, while selec- tion within a pure line will have no hereditary op spate it is evident that any Selection in the populenion must shift or move the “type” of the progeny in the direction of selecti 138 THE AMERICAN NATURALIST [Vou. XLV of these experiments has been attacked on the ground that the Parameeciums multiply asexually; but this mat- ter seems to me of no importance in the present case. The experience that pure-line breeding of plants and pure-strain cultures of micro-organisms, in full agree- ment, demonstrate the non-adequacy of selection as a genotype-shifting factor, is a circumstance of the great- est interest. Also Woltereck’s experiments with Daphnias, the important researches of Wolff, and the highly interesting indications of C. O. Jensen as to bacteria may be mentioned here as further supports for this view. Quite recently Pearl has arrived at the same conclusion as to the egg-production by fowls. The famous Galtonian law of regression and its corol- laries elaborated by Pearson pretended to have estab- lished the laws of ‘‘ancestral influences’’ in mathemat- ical terms. Now, by the pure-line explanation of the well known action of selection in poly-genotypic popu- lations, these laws of correlation have been put in their right place; such interesting products of mathematical genius may be social statistics in optima forma, but they have nothing at all to do with genetics or general biol- ogy! Their premises are inadequate for insight into the nature of heredity. Ancestral influence! As to heredity, it is a mystical expression for a fiction. The ancestral influences are the ‘‘ghosts’’ in genetics, but generally the belief in ghosts is still powerful. In pure lines no influence of the special ancestry can be traced; all series of progeny keep the genotype unchanged through long generations. A. D. Darbishire’s laborious investigations as to the classical object of Mendel’s researches, green and yellow peas, may even convince-a biometrician that the ances- tral influence is zero in ‘‘alternative inheritance.’’ An- cestral influence in heredity is, plainly speaking, a term of the ‘‘transmission-conception’’ and nothing else. The characters of ancestors as well as of descendants are both in quite the same manner reactions of the geno- No. 531] GENOTYPE CONCEPTION OF HEREDITY 139 typical constitution of the gametes in question. Partic- ular resemblances between an ancestor and one or more of his descendants depend—so far as heredity is re- sponsible—on corresponding particular identities in the genotypical constitution, and, as we have urged here, perhaps to excess, the genotype is not a function of the personal character of any ancestor. The genotypic constitution of a gamete or a zygote may be parallelized with a complicated chemico-phys- ical structure. This reacts exclusively in consequence of its realized state, but not in consequence of the history of its creation. So it may be with the geno- typical constitution of gametes and zygotes: its history is without influence upon its reactions, which are de- termined exclusively by its actual nature. The genotype-conception is thus an ‘‘ahistoric’’ view of the reactions of living beings—of course only as far as true heredity is concerned. This view is an analog to the chemical view, as already pointed out; chemical com- pounds have no compromising ante-act, H,O is always H,O, and reacts always in the same manner, whatsoever may be the ‘‘history’’ of its formation or the earlier states of its elements. I suggest that it is useful to emphasize this ‘‘radical’’ ahistoric genotype-conception of heredity in its strict antagonism to the transmission- or phenotype-view. As to the evolution of human civilization we meet with true ancestral influences, viz., the tradition (comprising literature, monuments of art, etc., and all forms of teaching). Tradition is playing a very great réle, but tradition is quite different from heredity. Nevertheless there may often be danger of confusion, and here the use of false analogs is not harmless. So an obscure meta- phor is involved in archeologists’ reference to Greek temples as ‘‘ancestors’? of some types of Christian churches, or in their speaking of the descent of violins from more primitive ‘‘ancestors.’’ Certainly, evolution of types of tools, instruments and implements of all kinds is—at least partially—going on by means of select- 140 THE AMERICAN NATURALIST [Vou. XLV ive factors combined with tradition, the latter not only conserving the valuable types but actively stimulating their improvement. But all this has nothing at all to do with the biological notion of heredity. It is of course interesting to see that the idea of ‘‘evolution by selec- tion” has won credit in archeology, sociology, etc., but this involves nothing as to genetics, for which ‘‘tradi- tion’’ is irrelevant. The very ‘‘radical’’ form of the genotype-conception advocated here may be too ‘‘theoretical’’ to be carried through in all its consequences in cases of practical ex- periments in genetics. In nature and even in the chem- ical factories the chemical compounds are not always to be had in quite pure state. The history of a prepara- tion may sometimes be traced by accompanying impuri- ties. As to the analogy with the genotypes we touch here the question whether the genotypical constitution of a gamete may not be accompanied by some accessorial or accidental ‘‘impurities’’ from the individual organ- ism in which the gamete was developed. Here we meet with the cases of ‘‘spurious’’ heredity, e. g., the infections of the gametes or zygotes as may be seen in certain cases of tuberculosis, syphilis, ete. Such and other forms of spurious heredity may have the ap- pearance of ‘‘hereditary transmission’? or ‘‘ancestral influence’’; but theoretically they do not interfere at all with the genotype-conception of heredity. In such in- teresting cases as that detected by Correns, viz., the ‘‘heredity’’ of a special form of albinism by ‘‘trans- mission’’ through the plasm of the ovum—the sperm not transmitting this character—we may at the first glance be puzzled. Nevertheless, as Correns himself points out. here we have certainly to do with a pathological state 0 the plasm or the chromatophores in question, and that may perhaps be the reason for the lack of heredity through the sperm which carries no (?) plasm or only a small quantity. The etiology of such abnormalities being as yet quite unknown, it may often be very difficult to distinguish them clearly from ‘‘genotypically’’ de- No. 531] GENOTYPE CONCEPTION OF HEREDITY 141 termined abnormalities which show the normal form of heredity through both ovum and sperm. The case quoted demands further experience and seems not to be in ac- cord with results of Baur’s experiments. At any rate, there may be several difficulties to overcome in the full and consistent application of the genotype-conception, difficulties that may be characterized as perturbations by infection or contamination. And hereby it must be r bered that theoretically, as well as practically, there are no sharp limits between ‘‘normal’’ and ‘‘path- ological’’ manifestations of life. ‘‘Nature is beautiful, but not correct,’’ is a Danish saying. The principle of pure lines or, generally, pure culture, is of importance also for elucidating the celebrated ques- tion of the inheritance of ‘‘acquired characters.” Men- delism and pure-line researches are here in the most beautiful accordance, both emphasizing the stability of genotypical constitution; the former operating with the constituent unities, the latter with the behavior of the totality of the genotypes in question. The brilliant work of Tower with Leptinotarsa and the highly suggestive injection experiments of MacDougal indicate that changes of the genotypical constitution are produced by steps, discontinuously. And as yet no experiment with genotypically homogeneous cultures has given any evi- dence for the Lamarckian view, the most extreme ‘‘transmission’’-conception ever issued. As to bacteria, the important experiments recently made by C. O. Jen- sen for the purpose of changing their types through adaptation have given not only absolutely negative re- sults, but have demonstrated the fallacy of some posi- tive indications by previous authors. Lamarckism and selectionism are certainly at bottom the same thing: the belief in personal qualities being ‘‘transmitted’’ to the offspring. Observations in impure populations are now their places of resort; nevertheless, it is granted that their history in biology as suggestive ideas has been most glorious. : Apropos, some cases of apparent action of selection 142 THE AMERICAN NATURALIST [Vou. XLV may have direct touch with Lamarckian ideas, as, e. g., De Vries’s selection of buttercups, recently quoted by Jennings as ‘‘the only case that he has found’’ indi- eating hereditary action of selection: ‘‘Here, after selection the extreme was moved far beyond that before selection.” And Jennings says: ‘‘Possibly repetition with thorough analytical experimentation will show that something besides selection has brought about the great change. But at present the case stands sharply against the generalizations from the pure line work.”’ Certainly Jennings is in reason, when he, on the ground of his own masterly researches, looks out for ‘‘something besides selection.’’ There are three direc- tions for the inquiry here. First, the strong evidence that the buttercup-population was not at all homogene- ous. Secondly, the possibility of intercrossing. I only need to point out the beautiful researches of Shull as to the effect of intercrossing in maize. The heterozygotes were here larger and more productive than the pure strains. The surprises of heterozygotic ‘‘constructions’”’ or of new combinations in F, may perhaps be respon- sible for the case of De Vries’s buttercups; I shall not try to discuss it. But, thirdly, we have an instance pointed out several times by De Vries himself, viz., the combination of selection with nourishment: ‘‘la sélection c’est l’alimentation’’ as it has been said. I suppose that we have here the essential point. The buttercups in cul- ture have been better nourished than before the experi- ments. Hence, the ‘‘best’’ genotypes having been se- lected from the population and submitted to ‘‘better’’ nourishment, the result would easily be a moving of the extremes far beyond those before selection. The butter- cup-case seems to me to present no difficulties for the genotype-conception. The practical breeders are a somewhat difficult people to discuss with. Their methods of selection combined with special training and ‘‘nurture’’ in the widest sense of this word are mostly unable to throw any light upon questions of genetics, and yet they only too frequently No. 531] GENOTYPE CONCEPTION OF HEREDITY 143 make hypotheses as to the nature of heredity and varia- bility. Darwin has somewhat exaggerated the scientific value of breeders’ testimonies, as if a breeder eo ipso must be an expert in heredity. As to the principle of pure lines it has been occasionally vindicated by Ger- man authors, e. g., K. v. Riimker, that pure line breeding is a thing old and well known. This is quite true; nearly sixty years ago L. Vilmorin not only emphasized in a lucid manner the importance of pure breeding, but he even tried a little to use his experiences theoretically. But it can not be denied that the principle of pure lines, as a true scientific analytical implement, as an indispen- sable method of research in heredity—not merely as a questionable and, at any rate, unilateral and insufficient method of practical breeding—is a novelty from recent years. Had this analytical principle been used in the times of Darwin, or had it even been appreciated in due time by the biometric school, certainly the real bearing of selection might long since have been rightly under- stood also by the practical breeders of pure strains. The genotypes may then be characterized as some- thing fixed and may be, to a certain degree, parallelized with the most complicated molecules of organie chem- istry consisting of ‘‘nuclei’? with a multitude of ‘side- chains.’’ Continuing for a moment such a metaphor, we may even suggest that the genes may be looked upon as analogs of the ‘‘radicals’’ or ‘‘side-chains.’’? All such ideas may as yet be premature; but they are highly favored by the recent researches of Miss Wheldale. The fixity of a genotypical constitution in question is the conception arrived at by Mendelian and pure line work. Hence there is a discontinuity between different genotypes. This discontinuity has been energetically contested by several biologists, among whom Woltereck may be pointed out as an important representative. In his very interesting report on experiments with Daph- mas, Woltereck indicates, as said above, that selection was as yet ineffective; moreover he describes a case of discontinuous alteration of type (mutation), and his ex- 144 THE AMERICAN NATURALIST [Vou. XLV periments designed to confirm the Lamarckian view have given as yet negative results, even though these may be called ‘‘promising,’’ as he says. So all the evi- dence of his breeding experiments is in reality quite in favor of our genotype-conception ! But how much depends upon our mental eyesight, what we see. Woltereck confesses openly his belief in continuous evolution and remarks that for a convinced selectionist of the Weismann school the new genotype- conception is a ‘‘hard blow.’’ The aim of his paper in question is to parry off such blows. Of course this parry can not use his own statements just mentioned; as to their obvious but inconvenient accordance with our con- ception Woltereck might apply the famous words from Harvey’s times: ‘‘video sed non credo.’’ Hence the arguments must be taken from other observations, and some very instructive results of cultures under varying conditions have supplied the piéce de résistance for the discussion. Woltereck is within his right when assert- ing that we consider different genotypes as having con- stant differences (like different formulas in chemistry). This is an essential point; but Woltereck, admitting no constancy in the differences, tries to demonstrate that our view must be fallacious. In a very suggestive manner he presents ‘‘phenotype- curves’’ for several pure strains. These curves are graphical schemes expressing (for the strain in ques- tion) the average degree or intensity of any particular character as it manifests itself under different condi- tions, e. g., the relative length of heads by poor, inter- mediate and rich feeding, ete. Such ‘‘phenotype- curves’? may indeed be very useful as records of the behavior of the organisms in question, and they mark certainly a valuable progress in descriptive methods. The phenotype-curves of the Daphnias in question sometimes show rather constant differences between the pure strains compared; but mostly this is not the case. Especially under extreme conditions, e. g., with poor oT even with very rich feeding, some of the curves are con- No. 531] GENOTYPE CONCEPTION OF HEREDITY 145 fluent. So the differences between the phenotype-curves may vary considerably or may even vanish entirely. These experiences agree with numerous observations of Wesenberg Lund as to the Daphnias in the Danish lakes, and there is no doubt as to their correctness. But when Woltereck thinks that these facts are in- consistent with the existence of constant differences be- tween the genotypes, he shows himself to have totally misunderstood the question! Of course the phenotypes of the special characters, i. e., the reactions of the geno- typical constituents, may under different conditions ex- hibit all possible forms of transition or transgression —this has nothing at all to do with constancy or incon- stancy of genotypical differences. Every student of genetics ought to know this; some few examples may suffice to enforce it: Temperature has great influence upon the intensity of color in flowers; all shades of intensity from saturated reddish-blue to pure white may be observed with different temperatures in lilac flowers of the ‘‘colored’’ varieties. Such pure white flowering individuals are—as to color—pheno- typically not distinguishable from genotypically pure ‘white? varieties. Nobody will assume that there should be genotypical transitions here! Pure lines of beans may in one year be different in size, e. g., the average of the line A exceeding that of B. In another year B may exceed A, or their average sizes may be practically identical. Differences of soil may produce something similar, and it is well known to breeders that Some strains of wheat yield relatively much better than others on rich soil, while the reverse is realized on poorer soils. In four subsequent years two pure lines of barley, both characterized by a considerable degree of disposition to produce vacant spikelets (aborted grains) in the heads, presented the phenotypes here indicated in percentages of such vacancies. Pure line L: 30 33 27 29 Pure line G: 5 45 3 28 146 THE AMERICAN NATURALIST [Vou. XLV The genotype-differences are nevertheless constant; the ‘‘Reaktionsnorms’’ of the organisms in Woltereck’s cases, as well as in the examples just cited, are of course eo ipso ‘‘constantly different’’ just as well as the ‘‘ Reak- tionsnorms’’ of different chemical compounds. And as to chemical analogies it may perhaps be useful to state that different chemical compounds (the structural or constitutional differences of which surely are granted to be discontinuous and constant) may sometimes show ‘‘reaction-curves’’ highly resembling Woltereck’s ‘‘ phen- otype-curves.’’ It is, I suppose, quite sufficient to point out the temperature-curves of solubility for different salts of sodium and other metals. These curves inter- fere in different ways, cutting each other or partially confluent, in analogy with Woltereck’s phenotype-curves. The essential point in the whole matter is, of course, that a special genotypical constitution always reacts in the same manner under identical conditions—as all chemical or physical structures must do. Differences in genotypical constitution (as well as differences in chem- ical or physical nature) are not bound to manifest them- selves at all—and still less to do so in the same sense —under all conditions. Sometimes even quite special conditions may be required for the realization of possi- bilities (‘‘Potenzen,’’ as some German authors are say- ing), due to a special genotypical nature: This is a well- known fact in physiology as in the fine art of gardening. Baur has long since emphasized the importance of this point for the Mendelian researches. So the criticisms of Woltereck as to the genotypical discontinuity and constancy are only based upon a re- grettable misconception of the genotype-notion. Over and over we find in current literature this confusion of genotypes with phenotypes, and we even have met with the idea, that the Daphnias of a lake may in summer diverge in different races or varieties, but that in winter they converge into one single race! In this statement of Wesenberg Lund, the author regards of course only the phenotypes in a purely descriptive manner. It is evident No. 531] GENOTYPE CONCEPTION OF HEREDITY 147 that Woltereck’s view has been influenced by Wesenberg Lund in this matter; but what might be fairly excused in the latter is not allowable for an experimenter pretending to work with cardinal questions of genetics. Discontinuity and constant differences between the ‘‘oenes’’ are the quotidian bread of Mendelism, and here the harmony of Mendelism and pure line work is perfect. We have dealt with some recent criticism of the pure line results; now it is time to look at Mendelism. The aston- ishing evolution of this mode of research has given an almost interminable stock of special results, and cases that at first might seem incompatible with the Mendel- ian views have been analyzed more thoroughly on a large scale and have shown themselves quite in accord- ance with Mendelism. The magnificent book of Bateson gives a full account of this prosperous state of Mendel- ian research. And it may be evident that Mendelism gives the most striking verification of the essential point in Galton’s ‘‘stirp-hypothesis’’: the inadequacy of the personal quality in heredity. At the same time it overthrows totally the idea of ‘‘organs’’ as being repre- sented by the unities of the ‘‘stirp,’’ pointing out that the personal qualities of the organism in toto are the re- sults of the reactions of the genotypical constitution. The segregation of one sort of ‘‘gene’’ may have influ- ence upon the whole organization. Hence the talk of ‘“‘genes for any particular character’? ought to be omitted, even in cases where no danger of confusion Seems to exist. So, as to the classical cases of peas, it is not correct to speak of the gene—or genes—for ‘‘yellow’’ in the cotyledons or for their ‘‘wrinkles,’’—yellow color and wrinkled shape being only reactions of factors that may have many other effects in the pea-plants. It Should be a principle of Mendelian workers to minimize the number of different genes as much as possible. Here we meet with the questions of correlation and ‘coupling’? of genes. I can not here enter into a discus- Sion as to the notion of ‘‘correlation’’ with its several meanings; in my ‘‘Elemente der exakten Erblichkeits- 148 THE AMERICAN NATURALIST [Von XLV lehre’’ a rather full discussion is to be found. I may only point out here that many cases of presumed correlation may simply be cases of two or more characters (reac- _ tions) due to the presence—or even absence—of one single gene. The phenotypically distinct and even di- versely localized ‘‘characters’’ convey easily the impres- sion that they are reactions of different genes. The highly interesting experiences of Correns, Don- caster, Morgan, Spillman and others as to the sex-de- termining factors, are in some way connected with researches of correlation and ‘‘coupling’’ of genes. The discussion of the ingenious Bateson-Punnett scheme for Abraxas and Morgan’s suggestive schemes as to Droso- phila may favor the idea of what may be called ‘‘rami- fied’’ genes. Castle has in his splendid researches as to color-factors in rabbits, ete., initiated a systematic de- scription of the (partially) analyzed genotypes, some- what resembling the formulas of organic ‘‘structural chemistry.” If we suggest an analogy between the radicals of chemistry and the genes, the (partial) geno- type-formulas in Castle’s manner may be able to demon- strate ramifications of the genes inserted upon the main group of the genotype-constituents. Pausing a moment on this metaphor, it may be suggested that the ‘‘branch.”’ or ‘‘branches’’ of a ramified gene may be more difficult to separate from its ‘‘trunk’’ than the whole gene from the totality of the genotype. I shall here only ask if such views may be of any use as working hypotheses. Their bearing as to the realization of mutations is obvious,— but the purely speculative nature of these suggestions can not as yet warrant a longer discussion here. It should always be borne in mind that the Mendelian analysis is purely relative. Baur and Shull and even several others have emphasized this fact when discussing the segregations in their experiments, and Shull has clearly pointed out that it may be quite impossible to in- dicate whether a particular reaction (character) is due to something positive or to the lack of a factor in the genotypical constitution. All that can as yet be deter- No. 531] GENOTYPE CONCEPTION OF HEREDITY 149 mined in this regard by Mendelian analysis is the nwmber of differing points between the two gametes forming a heterozygote. Such differences may be termed ‘‘geno- differences.’’ The well-known facts, that a ‘‘character’’ may be dominant in some hybrids but recessive in others, and that segregation in different cases may be very dif- ferent, indicate that ‘‘characters’’ are complicated reac- tions. The famous case of Bateson’s fowl-hybrids as to the form of comb may here be quoted as an example: In Walnut comb X Rose comb the latter is recessive, in’ Single comb X Rose comb it is dominant, and in both casesthe segregation gives three dominants : one recessive. Now Bateson has shown that ‘‘Walnut’’ is a compound of Rose- and Pea-comb. Homozygotic Walnut differs from homozygotie Rose only in one point, as does Rose compared with Single. But Walnut-gametes differ from Single-gametes in two points; hence Walnut X Single, with Walnut as dominant, segregates in Walnut, Rose, Pea and Single in the proportions 9:3:3:1. Even with this analysis it is as yet not possible to decide whether Single or Walnut is the form of comb for the realization of which the greater number of positive factors are re- quired. Suggesting—what seems to be the most prob- able assumption—that Walnut is the most geno-compli- cated case, Single may even be an expression for a multitude of genes in the fowl-constitution. The rela- tivity of the analysis by segregation must in all such cases be remembered, and it is quite erroneous to think that dominance indicates the positivity of the ‘‘unit- factor’’ in question: So ‘‘Horns’’ are in Wood’s cases dominant in male sheep but recessive in female sheep. And as to analogs with chemical reactions it must be kept in mind that a characteristic reaction may be the consequence of lack of any substance as well as de- pendent upon the presence of any special compound in the solution in question. The elaborate work of Mendelians of recent years has shown very complicated segregations, and the most spe- cialized segregation is almost the most specialized analy- — 150 THE AMERICAN NATURALIST [Vou. XLV sis still known of any ‘‘character’’ in question. The ‘‘units’’ or ‘‘unit-factors’’ stated in Mendelian work are consequently quite provisory, depending essentially upon the number of genodifferences in the special crossing. Probably it may be discovered that several such ‘‘unit- factors’’ for one character may also be elements for the realization of quite other characters. If this be the truth, then the present state of Mendelism, characterized by the rapidly augmenting number of new ‘‘unit-factors’”’ demonstrated in the organization of different biotypes able to hybridize, may be replaced by a period in which ` many such unit-factors will be identified. At any rate there is no reason to believe that the further Mendelian analysis will augment the number of genes into absurd- ity. The enormously increasing possibilities of combina- tions by augmentation of the number of segregable genes are a source of interest also in this connection. As to cases of hybridization, in which segregation and combination do not suit the Mendelian ‘‘laws,’’ it must at first be stated that some apparent exceptions are prob- ably caused by non-homogeneity of the initial material for experiments. The experiments of Correns, Castle, Miss Saunders, Tschermak and others have shown to excess that phenotypes may seem totally ‘‘pure’’ and nevertheless be heterogeneous (e. g., white flowering stocks or albino mice). Thus constancy as to the pheno- type of the progeny is no sure proof for genotypical purity or unity. In discussing alternative inheritance we meet with difficulties of the same nature as in regard- ing fluctuating variability: the inadequacy of pheno- type-description as the starting-point for genetic in- quiries. Secondly, the more or less high vitality of the different combinations of genes in F, may perturb the Mendelian results, as Baur has illustrated; in other cases the dif- ferent degree of facility with which the union of special gametes is realized may influence the relative numbers of representatives in the F,-generation, as Correns has demonstrated. No. 531] GENOTYPE CONCEPTION OF HEREDITY 151 Here we can not discuss the difficulties in a complete carrying through of the Mendelian analysis; Bateson’s recent book contains a richness of instances concerning this matter. Only one instance of special importance may be mentioned here, viz., the so-called ‘‘ blended in- heritance’’ opposed to Mendelian segregation or ‘‘alter- native inheritance.’’ In cases of blended inheritance the genes in question might be supposed to ‘‘fuse together’’ by the act of hybridization, or, in accordance with the presence- and absence-view, the gene unilaterally carried to the zygote might here in some manner be ‘‘diluted.’’ In this way, which certainly is very badly compatible with the conception of genes as unit-factors, failing segrega- tion might be explained. Cases of failing segregation seemed to be abundant in the beginning of the modern Mendelian era; Mendel him- self pointed out some typical cases in the species-hybrids of Hieracium. And Correns’s indication as to the con- stant intermediate stature of maize stems seemed to be a crucial case. Now the insight won by breeding experi- ments as well as by cytological researches concerning the phenomena of apogamy has put the question in a new light. The discoveries of Murbeck, Raunkier, Ostenfeld, Rosenberg and others have led to quite other explana- tions as to the constancy of several intermediate hybrid forms. In such cases no segregation is realized, because no gametogenesis is going on—and in such cases there is no reason for supposing any ‘‘fusing’’ or ‘‘dilution’’ of genes. And as to Correns’s experiments, this careful author has himself withdrawn the suggestion in question. But still cases of ‘‘blending inheritance’’ remain. Among these Castle’s experiences as to the dimensions of rabbits, especially the length of ears, are the most impor- tant and most discussed instances. Castle has in a con- vincing and suggestive manner demonstrated that the complicated color-characters in rabbits agree with the Mendelian laws. Therefore much stress might be laid upon his indication of cases contrary to these laws. 152 THE AMERICAN NATURALIST [Vou XLV Crossing short-eared and long-eared races, he gained an F’,-generation with almost intermediate ears, and here no segregation was observed in F,. But even this case may agree with Mendelian laws. The idea for such interpreting is won—as Lang has clearly pointed out—by means of Nilsson-Ehle’s (and _ East’s) experiments, the former concerning the colors of wheat-grains, the latter dealing with the number of ‘‘rows’’ in the ears of maize. Nilsson-Ehle demonstrated that blending of red and white color in wheat is appa- rently a fiction: The red color is determined by several different genes, acting in the same sense and augmenting the effect of each other. Hence by segregation and new combinations of these approximately equipotent genes a whole series of gradations in red color will be realized. And these gradations must group themselves symmet- rically around the phenotype of the F, in question. If _we have to consider say three genes, A, B and C, we shall for F, use the formula AaBbCe, indicating the value 3 which is intermediate between aabbee as zero and AABBCC as 6. Even in case of no fluctuation such a series must present itself as an almost continuous grada- tion, and it is not difficult to find out that the progeny of every ‘‘class’’ here will breed true, i. e., the average of the progeny’s character will be like the ‘‘class’’ of the parent. Just so it is in the case of Hast’s experiments with maize, as East himself has clearly illustrated. Thus, well-analyzed instances of heredity in plants, concerning both color-factors and meristic factors may be compared with Castle’s case in question. Lang in his interesting criticisms points out that certain irregularities in Castle’s F,-material give strong evidence for the view that we have no blended inheritance but true segregation here as well as in the cases of Nilsson-Ehle (and, as we may add, in the cases of East). Further analysis may then prob- ably demonstrate in a more direct manner the true nature of the apparent blending in Castle’s case; as yet we can No. 531] GENOTYPE CONCEPTION OF HEREDITY 153 only say that this case does not seem incompatible with Mendelian views. It must also be borne in mind that certainly there have been very many genodifferences between the differing races intercrossed in Castle’s experiments. Hence these experiments are really operat- ing with highly poly-heterozygotic F,-generations. And how great influence upon dimensions (of ears and other parts of the body) those color-determining genes may have exercised can not be easily determined. As to beans, it is proved that genes, effective in color- reactions, may also have great influence upon the dimen- sions and forms. So in my crosses a special factor, which makes yellow color turn into brown and causes violet to be turned into black, has a very marked influence upon the size and form of the beans in question. Here exact data are not necessary; the instance exemplifies the two incident matters of fact, viz., that apparently simple ‘‘dimensional’’ or meristic characters may be determined by several different genes, and that one sort of gene may have influence upon several different reactions. Then it seems that Mendelian analysis is proceeding in a very prosperous way; but there may be even very narrow limits for this analysis: the entire organization may never be ‘‘segregated’’ into genes! But still there is much to do in carrying through the genotype-concep- tion as far as possible. As to cytological researches the genotype-conception is as yet rather indifferent. Certainly the process of segre- gation must be a cell-action intimately connected with division. But all the innumerably detailed results of the refined cytological methods of to-day do not elucidate anything as to segregation. It seems to the unprejudiced observer that the much-discussed cytological phenomena of karyokinesis, synapsis, reduction and so on may be regarded rather as consequences or manifestations of the divisions, repartitions and segregations of genotypical constituents (and all other things in the cell) than as their causes. This view is applicable even in those cases 154 THE AMERICAN NATURALIST [Vou. XLV where sex-determination can be diagnosticated cyto- logically. In the discussion as to the existence of true graft- hybrids the cytological configurations have of course a high importance as precisely defined characters of cells in such cases where the cytological elements of the two species in question are different. And, as it may be well known, cytological evidence is not at all favorable for the idea of graft-hybrids. But the use of cytological configu- rations for diagnosis is quite different from the idea that special cytological elements might have importance for the phenomena of heredity. The question of chromosomes as the presumed ‘‘ bearers of hereditary qualities’’ seems to be an idle one. I am not able to see any reason for localizing ‘‘the factors of heredity” (i. e., the genotypical constitution) in the nuclei. The organism is in its totality penetrated and stamped by its genotype-constitution. All living parts of the individual are potentially equivalent as to genotype- constitution. In botany there has been no doubt as to this conception, and as to animals, O. Hertwig has for a long time advocated the same view against the views of Weismann and others, who have suggested that ontogene- sis is partly determined or at any rate accompanied by a progressive simplification of the ‘‘anlagen’’ (as we say the ‘‘genotype-constitution’’) in the cells of the growing embryo. The agencies of normally varying ambient con- ditions and the interactions of specialized parts in the developing individual may exercise their strong influence upon the whole phenotypical state of the resulting partic- ular individual. But these factors will as a rule not change or shift the fundamental genotypical constitution of the biotype in question. Later on we shall touch the problem of such genotypical changes (the mutations) induced by external factors. Here we have to point out the fact that ‘‘living matter’’—or, with a more precise definition, those sub- stances or structures the reactions of which we call oe No. 531] GENOTYPE CONCEPTION OF HEREDITY 155 ‘‘manifestations of life,’’—is inter alia characterized by the property of autocatalysis. The autocatalysis of living beings must embrace the totality of their geno- typical constituents. It seems to me that this autoca- talysis as well as the compensative and complemental maintenance of genotypical equilibrium in the organisms, present some of the greatest enigmas of organic life. The discussion of cytological problems leads us to the question of pure or impure segregation. In the begin- ning of modern Mendelian researches several instances of presumed impure segregation of genes in gametogenesis were discussed, e. g., as to color factors in animals. But more thorough analytical experiments have in many such cases demonstrated ‘‘purity’’ in the gametes, the charac- ters in question having proved to be more complicated reactions than at first supposed. Recently Morgan has discussed the question in a quite new manner, suggesting —as a working hypothesis—that the segregation might be not of qualitative but of merely quantitative nature. Hence the gametes should as a rule not be pure. Never- theless, as the author illustrates by means of interesting diagrams, the F,-generation of a monohybrid with normal dominance might be composed of two classes of indi- viduals sharply defined. And the author suggests that this idea might be able to explain ‘‘the graded series of forms so often met with in experience and so often ignored or roughly classified by Mendelian workers.’’ Here we again touch the question of ‘‘blended in- heritance.” I suppose that the above-mentioned expla- nations by Lang and East are more consistent with the real nature of the graded series in question. Now the Mendelian work has not only been able to demonstrate that several cases of segregation apparently impure are pure segregations of complicated nature; but even the ““spotted conditions’? as to color in animals and plants, emphasized by Morgan as a puzzling case, does not seem to present any real difficulty for Mendelian explanation. Certainly such cases as Shull has pointed out, viz., hetero- 156 THE AMERICAN NATURALIST [Vou. XLV zygotic nature being necessary for ‘‘mottling’’ in some special bean-hybrids, may at first glance favor the idea of ‘‘snotted conditions’’ being due to irregular segregation or to different repartition of color-determining factors in the tissues in question. But a closer examination seems to vindicate the real existence of special ‘‘spotting factors.’’ The very interesting researches of Lock as to the ‘‘Inheritance of certain invisible characters. in peas’’ have clearly pointed out a ‘‘spotting’’ factor or a ‘‘pattern’’-determiner in peas, independent of any color- manifestation. It must be borne in mind that a multitude of characteristic epidermal ‘‘patterns’’ are found in animals and plants, these patterns concerning all epider- mal manifestations and often showing a widely fluctuat- ing variability. As to the realization of all such spots it might be suggested that in neighboring parts of the devel- oping epidermal tissue some little difference of ambient conditions may inhibit or even release reactions, the alter- nation of which produces the spots. The whole case seems to be somewhat analogous to the merely phenotypical phenomena of alternative variability first pointed out by De Vries, e. g., the alternation of decussated and contorted stems of Dipsacus. Here we touch the highly suggestive idea of ‘‘ sensible periods’’ in ontogenesis or histogenesis emphasized with so good experimental arguments by De Vries. Of course there must be a genotypical fundament for the existence of the alternating character in question, e. g., for the particular nature of the surface of the spots (or for the contortion in Dipsacus, ete.); strains without such genotypical fundament will not be spotted (nor produce contorted individuals at all)—These remarks are made only to point out that Morgan may have exaggerated a little his criticisms as to ‘‘spotting factors,” but I confess that this question is in need of closer analysis. : Then the problem of pure or impure segregation may still be open; but the tendency in modern genetics goes certainly in the direction of establishing pure segrega- No. 531] THE GENOTYPE HYPOTHESIS 157 tion as the normal case. If we accept the suggestion of autocatalysis as an essential factor for the propagation of living matter in general, and hence eo ipso, for the growth or multiplication of genotypical constituents, we might in case of impure segregation expect frequently to find ‘‘dominants’’ in the progeny of ‘‘recessives’’; and the numerical proportions of the dominants and recessives in consecutive generations must be rather irregular. But this is not the case. The recent experiments of Darbi- shire quoted above demonstrate in a beautiful manner the purity of segregation during subsequent generations in Mendel’s classical object, the pea. Francis Bacon says: ‘‘Human understanding easily supposes a greater degree of order and equality in things than it really finds.’? So we may in modern genetics be aware of the relativity and narrowness of our provisorial explanations, remembering Bacon’s warning that ‘‘many things in nature may be sui generis and irregular!’ Among the irregularities in heredity we may reckon the mutations, observed in nature as well as in more precisely defined conditions of artificial experiments. From the famous observations of De Vries and the indications of several earlier authors, to the modern experimental researches of MacDougal, Standfuss, Tower, Blaringhem and others, all evidences as to mutations point out the discontinuity of the changes in question. Here we need not enter the question; it is sufficient to state that the es- sential point is the alteration, loss or gain of constituents of the genotype. The splendid experiments of Tower as to Leptinotarsa have in the most evident manner shown that the factors which produce the mutations in this case, viz., the temperature and state of moisture, are able to act in a direct manner upon the genotypical constitution of the gametes; and Tower has noted the occurrence of Mendel- lan segregation in hybridizing his mutants with the original unaltered biotypes. There may in some cases be certain puzzling irregularities to be explained by future researches, but it is evident that in all such muta- 158 THE AMERICAN NATURALIST [Vou. XLV tions, discontinuity is the characteristic feature in the change of type. As to populations, the biotypes of which may practi- eally exhibit continuous transitions—like the case of my own populations of beans—the idea might be born that biotypes are evolved from each other by extremely small steps in genotypical change. Hence such mutations must be practically identical with ‘‘continuous’’ evolution. But there is no evidence for this view. Certainly in such populations the ‘‘static’’ transitions between the geno- typical differences manifesting themselves. in several characters may be called continuous—but such a ‘‘con- tinuity of museums,’’ as it might be called, is not at all identical with genetic continuity. Galton himself has emphasized the capital difference between the notions of continuity in collections and continuity in origin, and as yet the mutations really observed in nature have all shown themselves as considerable, discontinuous salta- tions. So in my own still unpublished experiments with pure lines. Natura facit saltus. The chemical analog to such mutations may be the formation of homologous alcohols, acids and so on. The greater mutations may be symbolized by more complicated molecular alterations. But such analogs are of very little value for the under- standing of genetic evolution. The genotype-conception supported by the great stock of experiments as to pure line work, Mendelism and muta- tions does not consider personal adaptation as a factor of any genetic importance. Phrases as ‘‘cl ters, won by adaptation and having successively been hereditarily fixed,” are without meaning from our point of view. Hence much talk of adaptive characters successively gained seems to us an idle matter. A closer study of desert-organisms and the like may elucidate such things; here the suggestive researches of Lloyd as to stomates in desert plants may be pointed out. And as to the old question of ‘‘mimiery,’’ this problem in the famous cases of butterflies has in a most convincing manner been put No. 531] GENOTYPE CONCEPTION OF HEREDITY 159 into Mendelian terms by the observations and experi- ments of Punnett, de Meijere and others. This strong- hold of the united Lamarckism and selectionism has now been conquered for Mendelism, i. e., for the genotype- conception. The genotype-conception here advocated does not pre- tend to give a true or full ‘‘explanation’’ of heredity, but may be regarded only as an implement for further critical research, an implement that in its turn may be proved to be insufficient, unilateral and even erroneous—as all working-hypotheses may some time show themselves to be. But as yet it seems to be the most prosperous leading idea in genetics. Heredity may then be defined as the presence of iden- tical genes in ancestors and descendants, or, as Morgan says in full accordance with this definition: ‘‘The word heredity stands for those properties of the germ-cells that find their expression in the developing and developed organism.”’’ And now it is time to end this communication, too long for its real contents, but too short for the importance and diversity of the great problem of heredity. In concluding this address I must highly emphasize the eminent merits of Hugo de Vries. His famous book ‘‘ Die Mutationstheorie,’’ rich as well in positive indications as in ingenious views, has been the mediator for the new and the old era in genetics. This monumental work is a land- mark in the progress of our science. Like the head of Janus it looks at once forward and backward, trying to reconcile—at least partly—the antagonistic ideas of con- tinuity and discontinuity in evolution and heredity; hence a great deal of the charm of De Vries’s work. But just these qualities have made the work of De Vries too eclectic for the stringent analytical tendencies of modern genetics—a tendency which has in recent years found a true home in American science. THE GENOTYPE HYPOTHESIS AND HYBRIDI- : ZATION?! PROFESSOR E. M. EAST HARVARD UNIVERSITY Ir sometimes seems as if the hypercritical attitude had become an obsession among biologists. A proper judi- cial spirit is of course essential to science, but do not biologists often require a large amount of affirmative data before assenting to a proposition which is in real- ity a simple corollary of one already accepted? For example, Darwin emphasized small quantitative variations as the method of evolution, although he rec- ognized the occurrence of larger changes both quantita- tive and qualitative. De Vries, on the other hand, emphasized large variations—especially qualitative variations—as the real basis of evolution, although he too admitted the existence of lesser changes. He dis- tinctly states that a mutation or new basis for fluctua- ting variation, may be so small that it is obscured by the fluctuations themselves. If relative frequency of occurrence is a criterion of the value of variations in organic evolution, which is not necessarily so, Darwin’s point of view is probably the nearer correct. If one could find a unit basis for describ- ing variations in terms of the physiological economy of the organism concerned, i. e., if one knew exactly what was a large change and what was a small change, he would probably find that a random sample of inherited varia- tions followed the normal curve of error. By this I *Read at the symposium on the ‘‘Genotype Hypothesis’’ at the meeting of the American Society of Naturalists, Ithaca, N. Y., December 28, 1910. Contribution from the Laboratory of Genetics, Bussey Institution of Harvard University. The experimental results are from cooperative work between the Con- necticut Agricultural Experiment Station and the Bussey Institution of Harvard University. 160 No. 531] THE GENOTYPE HYPOTHESIS 161 mean that small variations would center closely around a mode, and large variations would occur with a rela- tive frequency inversely proportional to their size. The point that I wish to emphasize, however, is that neither Darwin nor De Vries recognized the proper distinction between a mutation and a fluctuation. Darwin made no distinction. De Vries, however, considered fluctuations to be linear; that is, to be limited to increase and de- crease in characters already present. He thought that selection of such variations brought about changes in the selected population due to the inheritance of the fluctuations, but that the selected populations returned to the mean of the general population after selection ceased. Mutations, on the other hand, were gains or losses of entire characters—qualitative changes—which were transmitted completely, i. e., were constant, from the beginning. De Vries did indeed state that mutations could take place in any direction, which would involve the idea of linear change or quantitative mutations; yet it seems quite evident from his general attitude in ‘‘Die Mutationstheorie’’? that to his mind qualitative and quantitative variations were quite distinct. Many practical breeders had long known, however, that the selection of linear variations often produced new races which were as constant as any races, provided they were not exposed to crossing with individuals of the general population from which the selected race had come. Why this was true was unknown. It was felt that there was a real distinction between certain varia- tions, to which Darwin had not called attention; yet it was felt that the De Vriesian idea was not wholly cor- rect. It has been in making this distinction clear-cut and definite that Johannsen has rendered his great service. His elaborate extensions of the genotype conception of heredity have cleared up many debated points, and corroborative evidence has been received from so many lines that it can hardly be doubted that the main points of the hypothesis are correct. It may seem, therefore, 162 THE AMERICAN NATURALIST [Vou, XLV ‘as if the superstructure of this conception were too elaborate to rest upon a simple foundation; yet I can not see but that the basis of the entire hypothesis is the fact that a fluctuation is a non-inherited variation produced upon the soma by environmental conditions, while the inherited variation, the mutation if you will, is any variation qualitative or quantitative, that is germinal in character. This being so, it seems scarcely necessary for an elaborate proof of the proposition, for it is noth- ing but a corollary to that part of Weismannism which was already generally accepted. Of course it is recognized that pure Lamarckism still has followers to whom neither Weismannism in any form nor the genotype conception of heredity could ap- peal. But to thorough Weismannians and to those who believe in occasional germinal response to environmental conditions, it seems as if both propositions must be ac- ceptable and their interdependence apparent. Let us follow this line of reasoning to its logical con- clusion in regard to the physiology of heredity. The Mendelian notation has been generally accepted as a con- venient way of accounting for the facts of heredity in certain markedly discontinuous characters. It has been questioned by many, however, whether the Mendelian con- ception is not rather an apparent interpretation of a rela- tively small number of facts than a general law. De Vries has even suggested that there are definite physiological reasons why certain characters should Mendelize and others should not. His idea is that Mendelian segrega- tion occurs when a germinal determinant for a character (Anlage) meets an opposing determinant, and when no such opposition exists the character in the cross-bred organism breeds true. Now the universal tendency of the facts of breeding is towards an interpretation the oppo- site of this. When a determinant from one parent meets with no such determinant from the other parent (pres- ence and absence hypothesis), Mendelian segregation appears. When the same determinant is received from No. 531] THE GENOTYPE HYPOTHESIS 163 both parents, segregation can not be proved, for the char- acter breeds true. In fact the many results of experimental breeding dur- ing the past few years have convinced me that De Vries’s general conception of this matter is incorrect. There may be finally a considerable modification of our ideas regard- ing the ultimate nature of Mendelian unit characters and the exact meaning of segregation, yet the universal appli- eability of a strict Mendelian system to interpret the facts of heredity becomes more and more apparent every day. And the point that I wish to emphasize is that Mendelian inheritance is a genuine corollary of the geno- type hypothesis if the latter is applicable to a popula- tion in a state of natural hybridity. In my work with maize where free intercrossing does occur I am convinced of the existence of genotypes in a state of natural hybridization. Furthermore, these genotypes can be iso- lated by inbreeding. If it were true, then, that only certain markedly discontinuous characters such as color Mendelize, how could genotypes which differ from each other in size characters be isolated? It is not expected, however, that the statement that Mendelian inheritance and the genotype hypothesis are interdependent will be received without proof. Data that are believed to fur- nish such proof are submitted here. When Mendelism was a new idea it was natural that the behavior of many hybrids should be regarded as irreconcilable to such a system of interpretation. The earlier criticisms arose largely through the misconception that dominance instead of segregation was its essential feature. Later, when so many complex results from pedi- gree cultures were fitted into a strict and simple Men- delian notation, it was objected that the investigators could by expert juggling of a sufficient number of factors interpret according to their system any experimental results they might obtain. Perhaps a few biologists re- garded as a personal affront the gradual growth of the idea that the facts of heredity were complex, but it is 164 THE AMERICAN NATURALIST [Vou. XLV hardly likely that many could regard this complexity as an invention of Mendelians. The latter would only too gladly have the facts as simple as possible. There have remained, however, several instances in which hybrids apparently did not segregate in the F, generation. Mendel himself investigated one such case, the genus Hieracium. The investigation of Ostenfeld? made this case perfectly clear by showing that the hy- brids reproduced apogamously. Such asexual reproduc- tion may also explain the behavior of hybrids between species of brambles which are also said to breed true in all their characters. These cases, however, and others among animals of which human skin color is the example par excellence, may be left out of consideration because no exact data concerning them have been forthcoming. There remain the experiments of two careful investi- gators who observed no segregation in the F, generations of their hybrids, those of Lock? upon heights of maize plants and those of Castlet upon weights and ear lengths of rabbits. Lock expected that if segregation occurred it would be into two classes, i. e., simple mono-hybridism. For this reason he made no measurements which would show whether he obtained the kind of segregation which as is shown later in this paper, does occur in maize hy- brids. Castle® has recently admitted the possibility that his numbers were not large enough to prove definitely that segregation involving several small unit characters does not occur in the ear length of rabbits. The difficulty attending this earlier work was that there was no way of explaining different manifestations of the same character. Segregating characters could always be interpreted either as the presence and absence of a uni * Ostenfeld, C. H., 1904, ‘‘Zur Kenntnis der Apogamie in der Gattung Hieracium,’’ Ber. Detach, Bot. Ges., 22: 7 * Lock, R. H., 1906, ‘‘ Studies in Plant Breeding in the Tropics,’’ MI, Experiments with Maize, Ann. Roy. Bot. Gard. Peradeniya, 2: 95-184. * Castle, W. E., et al., 1909, ‘‘Studies of Inheritance in Rabbits, ”” ’ Car- negie Inst. Wash. Ta, 114: 5-70. *In lectures at the Tawi Institute, Boston, Mass., 1910. No. 531] THE GENOTYPE HYPOTHESIS 165 giving a 3:1 ratio, or as the complemental action of two different units each allelomorphic to its absence, giving :3:3:1 ratios or modifications of them. Nillson-Ehle® and the writer,’ however, have shown that several units each allelomorphic to its own absence may be the determi- nants of what appears to the eye as a single character. In the above paper the writer suggested that if such ratios as 15:1 and 63:1—di-hybrid and tri-hybrid ratios, respectively—were found in considerable numbers, then higher ratios of this kind might account for the apparent constancy of hybrids in characters that seemed to be con- tinuous. For, if—as is quite probable—the additional units increase the activity of the character in question, and if there is no dominance,’ it is quite evident that hybrids may be intermediate between the two parents. All the pure classes in a complex character of this kind would indeed be difficult to isolate, but segregation could be absolutely proved by a comparison of the variability of the F, and F, generations. Since writing the above paper I have obtained clear evidence of 15:1 ratios in two other cases. The first is a red pericarp color, the second is the condition of endo- sperm in maize which gives dented seeds as distinct from that which gives flinty seeds. There is even considerable probability that higher ratios oecur which affect the latter character. In another paper? I have shown photo- graphic evidence of size segregation in varieties of Nico. tiana rustica and stated that similar evidence of segre- gation of size character in maize had been obtained. The following figures and tables show sufficient of the evi- dence from the maize crosses to demonstrate conclusively ° Nillson-Ehle, H., 1909, ‘‘Kreuzungsuntersuchungen an Hafer und Weizen,’’ Lunds Universitets Årsskrift, N. F., Afd. 2., Bd. 5, Nr. 2, 1-122. ' East, E. M., 1910, ‘ʻA Mendelian RF OAN a Variation that is Apparently Canthmous: ?? AMER. NAT., 44: 65-8 ° One. dose, i. e., receiving the dante gene Sit: a single parent, would on the average increase the manifestation of the character half as much as two doses, . M., 1910, ‘‘The Rôle of Hybridization in Plant Breeding,’’ Pop. Sci. Pa Oct., 1910, pp. 342-354. 166 THE AMERICAN NATURALIST [Vou. XLV that size characters segregate. It is hoped that this evi- dence will make us more cautious about accepting uncor- roborated statements about characters which are definite exceptions to the Law of Mendel. It is by no means certain that no such exist, but no experimental proof of hybrids non-Mendelian in character has been made. A further proof of segregation of size characters has recently been made in a preliminary note by Emerson."° He states that definite segregation occurs in beans, gourds, squashes and maize. His full data are therefore awaited with great interest. Table I shows the frequency distribution of the heights of plants in a cross between no. 5 a medium-sized flint maize and no. 6 a tall dent maize. Sufficient seed was obtained in a previous season so that the entire series could be grown in rows side by side during one summer. This procedure eliminates any possibility that the varia- bility of the F, generation might have come from varying conditions of soil fertility. It will be noticed that the F, generation is nearly as tall as the taller parent. This increase in size is not due to dominance. It is the increased vigor that comes from crossing in maize, and while it obscures the hereditary differences in size, it is really a problem of development and not of heredity as was shown in a previous paper.” The distribution of heights in the F, generation is seen by simple inspection of the table to be more variable than the F, "generation in the case of each ear planted. Re- duced to simple terms by the calculation of the coefficient of variation in each case, however, the two generations can be compared more accurately. In the F, generation the C.V.—8.68 + .553 while in the various F, genera- tions from different ears the coefficients of variation run from 12.02 + .559 to 15.75 + .684. * Emerson, R. A., 1910, ‘Inheritance of Sizes and Shapes in Plants,’’ AMER, NAT., 44: 739-746, “ East, E. M., 1909, ‘‘ The Distinction lien Development and Hered- ity in | Inbreeding,” AMER. NAT., 43: 173-18 167 ‘sossu[o [vI}U0D eT} punose yueSs9Au0d A[Su0I4s ə19m [TY 'pə}unoə syuejd zo Iəqunu pue paINsBeUT IIM SIMAJXH p ‘uosgəs owes ur əpıs Aq OPIS UMOIS SoINSY Burrs Suornqeaqsic] er ‘S189 JUAIIPIP MIY WOIF UMOIY) ,, ‘S1BO JUJIIPIP OM WOIFJ UMOIH şr THE GENOTYPE HYPOTHESIS No. 531] ‘SIVI JUdIOIP JAY WOT UMOLY or ‘mosves oules UT ƏPIS Aq APIS UMOIH zr 60% | |@ | 80 |r | 4y | Fe | B42 | 09 | 89 | Ie | se | si) or] 9 | nid (to x 8-09) 833 Iiris io ié |% 1 | ke 9¢ 62 |ceierierio |¢ |T I | oxi ($9 X 8-09 6F9 e |e |ze| te | 69 |e | get! ziti 96 " pest 8 |1 $18 | | er ŠI (POX S09 og +e wi (#9 X 8-09 9% He | | H (eee eg prm ‘a (P9 X 9-09 00T née —> 09 “ON wr Itigi? ela eea | | i] rg ON r | s8 | ss | æ | ez |o |e | |z | ro | 1's] oo | ao | or | oF | er | oF | ze | ve | te |s | 9 | 2 [810.1 : “ON S}UB][J JO $14819 H 103 SaYOUT UT s1970 S880 (09 X P9) SSOUQ NI SEINVIG A0 SLHYITH AO NOILAAINLSIA. AONTAÒOTAA : «II @IAViL OGF’ F COZ | $98' F 8Z'OL | SIS’ + F4'I8 g If o |OTIITIG 91i Pz LS ITI Arg A 16 | Fl Si TI tanxe) 6S9 F ZOZI | 8E F G6 | 129° OF'6L I E rT 18 8 Or STi Stisti6 |L I> iF 3 „18 -(9%¢ GLO" F P8°Sl | L99°F SLIT | 208° + 888 | 3121/8 [OTIS 18 |e 19 IZLI8 lZli¢ 0 Iz iP Iz Z ag -(9x9) $89 F CLL | 98h F 9L'ZI | 989° F 00'T8 git t ig OTO [SLIT] 2 [OL/9T/ZZ/OT\OL/¢g g | 9 SIlTIs At Seq’ = 99° 619° F 128 | OL’ F S$C'K6 Lir t 9 19 16 16 eib ig a -(9x¢ 627 10°E | $86°F L09 | OOF’ E8TTOT| £ |8PT IPL Stig |S |T 9 IZ F 1¢° $68" F 6F'9 | 90F F 3389 $ 19 JOL er Li LT Sirig e g TTE | 80T | SOT | ZOT | 66 | 96 | £6 | 06 | 28 | #8 | T8 | 84 | GŁ | ZL | 69 | 99 | £9 | 09| Łe | +e | 1s | ‘AD ‘a's "Y ‘ON SIULA JO S]YSIOP] 10J soyouy UF S197U3Q SEVP | al PI4JVL ry x a ssouy 1 NI SINVId IZIV dO SLHÐITH AO NOILNJINLSIA AONTNOTAT 168 THE AMERICAN NATURALIST [Vou. XLV Table II shows a similar distribution of heights in cross between no. 60, a dwarf pop maize commonly known as Tom Thumb, and no. 54, a sugar corn known as Black Mexican. The distribution of heights of no. 54 was ob- tained in the same season as the F, generation. They were both grown upon the same plot of ground in which the soil appeared to be quite uniform. Unfortunately, the exact distribution of the heights of no. 60 and of the F, plants which were grown in previous seasons, is un- known. The range of the variates shown by the black lines, however, is correct. Furthermore, from notes re- corded at the time we know that the F, generation was comparatively uniform, the greater number of variates being distributed around classes 67, 70 and 73 inches. In this case, also, the effect of crossing is shown by the rela- tively high plants of this generation. The plants of the F, generation show a wide range of variation. ‘he highest individuals are practically the height of the highest individuals of the taller parent, no. 54. The lowest plants of F, do not reach the lower range of no. 60. I interpret this as due to continued heterozygosis in other characters and to physiological correlation. By the latter term I mean that since the plants of no. 60 are very small, F, segregates of the same size could only be expected where the ears and seeds also are very small. But since the ears and seeds of these plants also show segrega- tion in new combinations, normal growth correlation probably resulted in a somewhat larger average size. For example, little 40-inch plants were found with ears three times the length of normal ears of no. 60. It is likely that such plants might have been smaller if they had been recombined with the characters necessary for the production of smaller ears. Table III and Figs. 1-4, show the lengths of ears in the cross just described. In making this table the best ear from each plant that bore a well-filled ear was taken. The small ears, therefore, do not represent poor, unfilled or supernumerary ears. The coefficients of variability 169 THE GENOTYPE HYPOTHESIS No. 531] ‘SIVI 9014} WOIF UMOIY yz ‘SIVI GAI} WOIJ UMOIN yz ‘SIVO OM} WOIJ UMOID oz ‘SIBI JAY WOLF WAOIY) er “OLGT Ut ‘wed “g pue Fo ON “G0GT Up UMoIT ‘uo T pue 09 ON sr EI FOLGI | S80 F LTT | LPO F09 | €1%18 l6 |61|1z|2¢| 0F| Se lr | Fs|01\F | a Sa (FS ¥ 8-09) 66L F P8'6I | LEO FETI | L90 F79 T| 1/6 |8 |ZI|8i|8Z|egjLIj9T | Sts jT | ec (P9 X 8-09 chr F 9903 | B20 F GOT | seo Feo - I I 8 (ST 23 SF | 69 | $8 \STT| 86 | 6F| AL} 2 |T | ez al (p9 x 9-09 Z16 F 16ST | I0 FPS 6S0 F 9'F € |F ILI |SL/3t\¢ A (#9 X 2-09 E9 F FOF | PLO FIST | LOI FES #| GS /OLj/PI|L |9 |Z |3 bg 668° + PPPI | PZO F(E P80" FLT P | 86/33) 4 09 aa oe es|os gL OL 9'9 o'g | s'o oe es ri S'S) 08 | VS 0% | ‘AO ‘as V ‘ON Sp9eg CZ JO 814319 M JOJ SMBIDH UT S19} 00H SEVO (S X 09) Ssoup dO SATA9 JO SLHÐIAM JO NOLLAATALSIG] AONTNOAAT AI S@TaVi FPL 08'CS 180° 18% SEU FOSE | | S/T |ZE]TT|OL|ST |1823 jeg |se|se|zt|¢ |g 3 A (P9 X 8-09) SIF FPF LI ESO FSZ LO FGI I |6 |ST|98|68|89/89 |&4 | Lh 9G GT) OT) T oz e i963 FL9'SI 180° 66'T 890° F LCI I |9 | LT} 26] $9| 16 6ZI/SFT| 08/99/339 |F ot “A (FE X 9-09 LoL’ FSFI 880° 1¢'T (4 ties + |6 AT PL |SE/ S/T ‘a (PS X 9-09) ISS FETII 880° L8'L IZUFS'9L | Z| L | OT/SE| 9%) S| Sr| IL |g bo ON | ($82 F L331 10°F 18" 810° 9'9 |8 #6 |13| F 09 “ON | 1% | 06 | 6I | 8E | ZT {| OL er | +I | er ZI 1} or 6 | 8 |2 DFS O ‘a's Yy | ‘ON i | sivy JO SySue'y 10j ‘wD UT 81940900 ssBIO (PG X 09) SSOND NI Suv MO SHLONA'T JO NOILAOIINLSIA JAONTAOTAT s III WIdvViL 170 THE AMERICAN NATURALIST [Vou. XLV Fig. 1. No. 60, female parent, illustrating variation in length of ear (4). have again been calculated, but they hardly emphasize the real segregation as well as do the photographs which were made from representative ears of the different classes found in the actual crop. Table IV shows the segregation of weights of seeds in F, in this same cross. Fig. 5 shows the average size of Fic. 2. No. 54, male parent, illustrating variation in length of ear (p). No. 531] THE GENOTYPE HYPOTHESIS 171 Fic. 3. Variation in length of ear of F; generation of cross between No. 60 and No. 54 (4). the seeds of the two parents and the F, generation and the extremes of the F, generation. In making the weights for this table, it was necessary to use a scheme by which the sugary wrinkled seeds of the Black Mexican parent, no. 54 could be weighed as starchy seeds. This * Sabet tae Be ae Fic. 4. Variation in length of ear of F, generation of cross between No. 60 and No. 54 (}) 172 THE AMERICAN NATURALIST [Vou. XLV was done by planting the no. 54 between rows of the hybrid. Sufficient crossed seeds which had become starchy through Xenia were obtained to make the weights given. Not all of the ears, however, had 25 starchy seeds, which accounts for the small number of plants meas- ured. Furthermore, the seeds of no. 54 were a rather mixed lot, which of course resulted in a higher varia- bility than would probably have been found if only seeds Fic. 5. Average size of seeds of No. 60 and No. 54 and the F, generation of the cross between them. Extremes of the F generation. of the individual plant of no. 54 which was used as the male parent of the cross, could have been planted. Per- haps it should be noted here since the question might arise, that since the size of the seeds on an ear is gov- erned by the development of the pericarp, the sugar corn, no. 54, was unaffected in other ways than by having the pericarp filled out with starch by the hybridization which occurred attended by the resultant Xenia. In Tables III and IV the measurements and weights of the F, generation were recorded from only one cross, although three crosses between the two varieties were made. It might be said that one has the right to com- No. 531] THE GENOTYPE HYPOTHESIS 173 pare only the F, generation of cross of which the F, generation is given. If this were granted our conclu- sions in regard to segregation would be the same. It might be said, however, that sufficient records were made of the F, generations of the other crosses to know that they differed but little from the family of which the data were recorded. In addition, it is a fact that general Fic. 6. Avers a ears of No. 60 and No, 58 and the F; ge ation of the cross etween them, Extremes of the Fə generat populations of the two parents were studied, and their variation was undoubtedly greater than would have been that of the inbred progeny of the three parent plants of either variety. An additional cross between Tom Thumb pop maize and a small purple flint is illustrated in Fig. 6. The ears pictures show the average size of the two parents and the F, generation, and extremes of the F, generation. In conclusion there are two points I wish to notice. Unquestionable segregation in size characters has been shown by comparison of the F, and F, generations. It can scarcely be doubted that some of these segregates will breed as true as the parent forms, yet one can 174 THE AMERICAN NATURALIST [Von XLV scarcely do more than speculate in regard to the specific characters that are concerned in developing either organs or individuals of certain sizes. There are probably many characters that interact together in developing certain characters, although the actual determinants in the germ cells may be transmitted independently. These interde- pendent reactions during development obscure to us the real causes and what we regard as independent char- acters may be but indirect results of unknown causes. For example, the ability to evert their starch when heated has been the distinguishing character of the subspecies called Zea mays everta, the pop maizes. This character so called, however, is the resulting physical condition of the starch caused at least partially by the small size, the thickness and the toughness of the enveloping pericarp. . For these reasons it may not be possible—at least very soon—to point out even the number of characters con- cerned in size developments. From the number of ex- treme segregates obtained in each case I might venture to state that the size of ear in the cross shown in Fig. 5 is apparently due to not less than three characters, while the size of ear in the other cross pictured seems to be due to not less than four characters. NOTES ON GUNDLACHIA AND ANCYLUS DR. WILLIAM HEALEY DALL U. S. NATIONAL MUSEUM AsrourT seven years ago,! in the Nautilus I called at- tention to certain problems connected with the genera mentioned in the title of this paper, and urged investiga- tion of the subject from the hypothetical view of the two following propositions: : 1. That Gundlachia is merely an Ancylus which under favorable circumstances has been able to form a cal- careous epiphragm and survive the winter, which ordi- narily kills the great mass of individuals, and, while retaining the shell of the first season, to secrete an en- larged and somewhat discrepant continuation of it dur- ing the second summer. 2. That not all Ancyli necessarily have the ability to do this, but the practise may have developed in certain small species; and in tropical regions where the dry season takes the place of winter it is possible that sur- vival may become more or less habitual with some of these species. In this connection attention may be recalled to the estivation in dry mud behind a double epiphragm, in the Bahamas, of Segmentina dentata Gould,? and to the ob- servations of Erland Nordenskjold? on Ancylus mori- candi Orbigny, in Brazil. During the past four years I have received an inter- esting series of notes by Mr. John A. Allen, of Cleve- land, Ohio, connected with the Nungesser Electrice Works of that city, who has for some time been domesticating in small aquaria species of fresh-water shells, including * Nautilus, XVII, No. 9, pp. 97-98, January, 1904. *Smithsonian Miscell. Coll., Vol. 47, Pt. 4, No. 1566, pp. 446-448, April, 1905. * Zool. Anzeiger, XXVI, pp. 590-593, July, 1903. . 175 I6 THE AMERICAN NATURALIST [Vou. XLV Ancylus and Gundlachia. His observations extend over some six years and his notes contain so much of interest that it has seemed desirable to summarize and publish his data, thus placing on record facts which may stimu- late others to follow his example. Mr. Allen was kind enough to send to the museum a lot of Anacharis supposed to contain both Ancylus and Gundlachia in the living state, and numerous specimens of the former were observed in a jar to which the vege- tation was consigned, immediately after it was filled with water. We were not able to distinguish with cer- tainty any Gundlachia, though some may have been pres- ent, and the small aquarium was kept in good condition to await developments. This was in December, 1907. The Ancyli continued to exist in apparent health during the winter. In May, 1908, they seemed to go into hiding, but during the summer reappeared again in rather dimin- ished numbers, while a few young ones were observed. No particular change was noticed during the following winter and spring. While absent during the summer of 1909, it became necessary to transfer the collections to the new building of the National Museum and the aquaria were set aside. After the confusion of the transfer was measurably over, I examined the aquaria and, finding nothing visible, had the contents of the smaller one (about 8 X 4 X 10 inches in size) removed and submitted to the most careful scrutiny, the sand at the bottom being placed in a fine sieve for examination, but not a trace of Ancylus remained. I concluded that there had been sufficient carbonic acid in the water to completely dissolve these fragile shells after death, and that some unfavorable condition had exterminated the colony. In the other aquarium, which was about eight times the ca- pacity of the smaller one, the water had evaporated to about half its normal quantity and no mollusks except — a few small Lymnzas were visible, while the Anacharts had suffered considerably by the adverse conditions. This was towards the end of November, when it was No. 531] GUNDLACHIA AND ANCYLUS 177 difficult to get any fresh weed except by purchase. Being much occupied, I contented myself with having the aquarium filled with Potomac water from the tap. A short time afterward I was surprised to note a large number of young Ancylus with clean translucent shells, on the side of the tank. There had never been any Ancylus in the aquarium except such as might have been put in with Mr. Allen’s Anacharis. These had up to February 22, 1910, grown rapidly and continued to flour- ish, though the number then visible was only about half that which was noticed in November. In April the Ancylus completely disappeared again. I have not been able to discover where they went to, as the most careful scrutiny of the sparse amount of Anacharis remaining has not revealed any on the stems or leaflets. None of the specimens seemed to have formed any septum and nearly all of them were carrying a small colony of five or six minute hydroids on the posterior upper surface of the shell. The shells in February were still too fragile to admit of removal from the glass without crushing, and most of them kept on the side away from the window, on the sill of which the tank stood. They were about 3.0 mm. in length, and remarkably active, moving about on the glass with surprising speed. Subsequently Mr. Allen kindly furnished me specimens of all these stages in alcohol; and I also had the oppor- tunity of seeing some specimens in alcohol which had been sent to Mr. Bryant Walker and Dr. H. A. Pilsbry in 1908, and which were obviously identical with those sent as examples by Mr. Allen to me over a year later, and Dr. Pilsbry thought also with specimens collected at Rockford, Illinois, in the ancyloid stage. On account of its relations to the Gundlachia it will be referred to here as Ancylus meekiana, since, unless in the Gund- lachia stage, it seems not to have been described. Mr. Allen also sent a lot of the wild Ancylus collected in the Thornburg lagoon and which he was disposed to regard as something distinct from his aquarium ancy- 178 THE AMERICAN NATURALIST [Vov. XLV loids. After a careful examination under the microscope I have been unable to find any constant differences be- tween shells of the same age, except that the larger specimens of Ancylus seem to have grown continuously and evenly, while those ancyloids which attained a Gundlachia stage show the sharp contrast between the separate stage and that with the expanded third stage of the shell. As this is only what one might expect if the Ancylus attained its full growth without interruption, while the ancyloid becoming septate passed through a resting stage and then began to grow again, I consider this difference of no moment systematically. The young Ancylus and the ancyloid of the same length appeared generally quite identical, though I noticed that in both the obliquity of the apex varied to some extent, being more emphatically bent toward the posterior right side in some individuals than in others. Ancylus meekiana is, when young, for a time nearly parallel-sided, the growth toward maturity being more ex- panded than at first. The apex is behind the middle of the shell and slightly inclined toward the posterior right-hand side at maturity. The microscope reveals some very feeble radial striae from the apex, mostly vanishing be- fore they reach the base. The incremental lines are not strongly marked and the shell when clean is of a pale translucent yellowish color. At or near maturity the shell assumes a more oval form slightly more expanded in front than behind. The animal has short pointed © tentacles, well-marked black eyespots, and a bluish-white color, except about the mouth, where the yellow-brown jaws are laterally set and the buccal mass has a pinkish color. The shell is about 3.6 mm. long, 2.3 wide, and 1.0 high. In the dark-colored specimens of the wild Ancylus, on the inside, may often be seen a dark-brown line cor- responding to the margin of the young Ancylus and showing the more parallel-sided early outline. Miss Mary Breen, who has been studying the anatomy of the fresh-water gastropods of the District of Co- No. 531] GUNDLACHIA AND ANCYLUS ' 179 lumbia, was kind enough to undertake the removal and mounting of radule taken from specimens of the differ- ent stages, as well as from the wild Ancylus. This was a task of no little difficulty on account of the extremely minute size of the organ. The radule of ancyloids, sep- tates and Gundlachia were absolutely identical in ap- pearance and in number of teeth, the formula 5-10-1-10-5, holding good for all. The uncinal teeth are not gradu- ally modified from the laterals, but change abruptly and form a distinct band on each side of the radula. The lateral part of Stimpson’s figure of the dentition of his Gundlachia meekiana is imperfectly made out, and obvi- ously inaccurate; due doubtless to the fact that he had only a few specimens and a not very powerful microscope. Unfortunately his original material was destroyed in the great fire at Chicago of 1871. An examination of the radula of a septate form, col- lected in Nicaragua by Professor B. Shimek, showed a similar radula but with one more uncinal tooth on each side. In this case, unfortunately, while endeavoring to transfer the minute object to a slide for permanent pres- ervation, it mysteriously disappeared, and a trial with a second specimen was no more successful. The form of the laterals is fairly well given by Dr. Stimpson, and the rhachidian tooth is correct in his fig- ure; but the gradual modification and uncertain number of the outer teeth of the radula do not agree with our observations on the specimens from Ohio. Renewed correspondence with Mr. Allen led to the preparation of this paper, pending the continuation of his observa- tions. Since the different stages of Gundlachia need to be carefully discriminated, I have adopted the following nomenclature for them. In the first stage, when the young shell has a laterally compressed subconical shape without any trace of Septum, and is to all intents and purposes, concholog- 180 THE AMERICAN NATURALIST [Vou. XLV ically and anatomically, an Ancylus, I call the individ- uals ‘‘aneyloids.’’ In the second stage when the base of the conical shell is more or less closed by a flat horizontal septum con- tinuous with the margin around it, I call the individuals ‘“septates. ”’ Lastly, when the animal in its second season begins to form a marginal expansion external to the septum, and with its longitudinal axis sometimes at a considerable angle with the axis of the ancyloid shell, I reserve for this stage, up to and including maturity, the term **Gundlachia.’’ Mr. Allen kindly sent alcoholic specimens of ancyloids, septates and Gundlachias from his aquarium for ana- tomical examination. The posterior part of the foot en- tirely hides the septum when the living animal on the walls of the aquarium is examined through the glass. Nothing to distinguish it from ordinary Ancylus is visible in the soft parts. The creatures feed on the microscopic alge, etc., which grow on the walls of their domicile and when feeding the movement of the jaws and radula can be seen with ease by means of a magnifier. On the alcoholic specimens, on the exterior of the shell, were many minute lenticular capsules which, from anal- ogy with Neritina, Pompholyx, ete., were supposed to be the ovicapsules. The very young shells are very trans- parent and fragile. It is difficult to find them until they have reached a length of over a millimeter, and so far it has proved impracticable to detach them from their roost without crushing them, they are so extremely fragile. The smallest septate seen was slightly less than two millimeters in length and the animal had entirely with- drawn behind the septum, which covers more than two thirds of the aperture. The species in the Gundlachia stage agrees substan- tially with the form described from the District of Co- lumbia by Stimpson, under the name of Gundlachia meekiana. Asin many other fresh-water shells the newly No. 531] GUNDLACHIA AND ANCYLUS 181 formed shell is yellowish translucent, while the older part, especially when the pond or aquarium has a muddy bottom, often becomes darkened or even blackish, and more or less covered by a growth of conferva. Mr. Allen calls attention to the fact that the sharp line of demarkation which separates the dark encrusted shell of the septate from its translucent Gundlachia extension in the final stage, is evidence that the growth is not con- tinuous, but that a resting period of some duration sepa- rates the two stages. I have preferred for the most part to refrain from theorizing on the inferences to be drawn from the data, letting them speak for themselves. To me, however, the facts tend strongly to confirm the hypothesis suggested in the opening paragraphs of this paper. GENERAL NOTES The following notes are partly summarized from a rather voluminous correspondence with Mr. Allen, ex- tending over more than four years. The Thornburg lagoon is an abandoned channel’ of the Cuyahoga River. In 1903 the river was fairly well stocked with Unionidæ, but soon after that date the con- tamination of the river by drainage and sewage killed off the naiad population. This contamination is not believed by Mr. Allen to have seriously affected the water of the lagoon, though for some reason it does not seem to be a place favorable to vigorous growth of mollusks. It pro- duces a dwarf Planorbis parvus, a poorly developed Physa, a small form of Lymnea humilis modicella and a scanty supply of Amnicola. It is nearly filled with Nuphar on the leaves of which Ancylus is found; also Ceratophyllum, Potamogeton, ete., occur, especially where the water is shallow. At one place the bank bordering on the lagoon is steep and the water near it deep, so here even at low water mollusks would never be left dry. There is another por- tion of the lagoon where a wide zone, producing vegeta- 182 THE AMERICAN NATURALIST [Vor, XLV tion on which Ancylus occurs, is sometimes left uncov- ered when by dry weather the water becomes low. In this part of the lagoon three Gundlachia were found. In general the water of the lagoon is deep and constant, but owing to the presence of these shallows the hypoth- esis that the formation of a septum in Gundlachia may be due to alternation of wet and dry periods can not be wholly excluded. Ancylus occurs in one to three feet of water where Ceratophyllum is abundant. In the deeper water shore there is more Nuphar and less fine vegetation the Ancylus seems to be absent or rare. Mr. Allen attempted to domesticate the Thornburg Ancylus, placing many young ones in a 15 X 9-inch jar stocked with Anacharis from the lagoon. Apparently, all soon disappeared, although Lymnea and Amnicola, coincidentally transferred, lived a long time. NOTES ON THE SEVERAL JARS USED AS AQUARIA The 15 X 9-inch Jar—This originally contained a dwarf Nymphea which died. There was a mixture of peaty and ordinary soil about three inches deep in the bottom of the jar. This was stocked in 1906 with Anacharis and some specimens of Vivipara. The date of the first appearance in it of the ancyloid stage of Gundlachia was not determined. February, 1907, individ- uals were very numerous and, some being taken out to save in the dry state, the septate form was discovered. Mr. Allen had noticed the presence of the ancyloid form some time before. The first date at which Gundlachia had been obtained from the Thornburg lagoon was July 15, 1906, but Mr. Allen doubts if the copious swarm of ancyloid individuals of Gundlachia could have originated in the jar so quickly from individuals accidentally put in at that time. Some of the vegetation in the jar had been received from elsewhere in Ohio, and some from another state. The ancyloid stage of the Gundlacha can not be distinguished from the associated Ancylus by No. 531] GUNDLACHIA AND ANCYLUS 183 the external features as seen in the aquarium. In Feb- ruary, 1907, probably hundreds of the unseptate ancyloid form were present. There were several Vivipara in the jar that winter. Subsequently they were removed, Mr. Allen thinking that they might consume the food supply needed by the ancyloids. Having heard that the stunted growth of aquarium mollusks might be due to the pres- ence of their soluble excreta in the water, he thought the removal of the Vivipara might have had some influence in this way. However, the removal of the large snails did not stop septation. In the winter of 1906-07 the specimens of Planorbis parvus in the jar were large and healthy. In the winter of 1907-08 the individuals of this species appeared dwarfed. The water in the jar was then removed and replaced by distilled water. After that the Planorbis (and Mr. Allen thought also the Anacharis) took on a more healthy appearance. He thought that the concen- tration of saline matter due to refilling loss from evapo- ration with ordinary lake water might have been influ- ential injuriously, and the transfer to distilled water have lessened the tendency to septation. In the winter of 1907-08 septate individuals of which the exact number were not recorded were again found in the jar. In January and February, 1908, the ancyloid form was fairly plenty, though not so numerous as in the previous year. In spring they became fewer and in May, 1908, there were none visible (although in a smaller jar there were some). They reappeared in the first half of June, 1908. July 3, 1908, an immature septate individ- ual was taken, and another on July 20. On the theory that the septum is formed during a resting stage, these may have been forming during May, when nothing was in sight. August 3, 1906, another specimen was taken. January 11, 1909, a specimen was found which had be- gun to add the third or expanding stage of the shell ex- ternal to the septum. No mature Gundlachia were taken from this jar during the winter of 1908-09. 184 THE AMERICAN NATURALIST [Vou. XLV August 19, 1909, a minute ancyloid specimen was taken, and another August 24. September 26 six ancyloids were visible at one time, but were not disturbed. It was noticed that the ancyloids came out in sight on the walls of the jar more freely on cloudy than sunny days. This jar, December, 1909, contains a dense and vigor- ous growth of Anacharis, also plenty of fresh-water algw. It stands in the factory room subject to the fall of factory dust, and to the changes of temperature in the room. When the room gets unusually cold the ancyloids mostly retire out of sight, temporarily. De- cember 9, 1909, two specimens with the third stage of the shell partly grown were taken near the top of the jar. A sudden spell of unusually cold weather having begun two nights previous may account for the ancyloids hav- ing gone, as they did, into hiding, but it was somewhat surprising that the more nearly mature form had not also hidden. The 8 X 6-inch Jar—This had sand on the bottom and was planted with Anacharis from the larger jar, carry- ing with it Ancylus, Gundlachia and Planorbis parvus in the summer of 1908. The following winter, having nothing but sand and water to live on, the vegetation had become rather attenuated and feeble looking. The ancy- loids were few and perhaps not more than half as large as those in the larger jar. January 19, 1909, two or three immature septate specimens were taken from this jar, and February 10 one about half grown. Very few ancy- loids were seen about this time in this jar. February 11 two immature septate specimens were taken, being all of either form which were at that time visible. Feb- ruary 24, 1909, for the first time since the eleventh, a small ancyloid was noticed. On the twenty-seventh one moderate-sized but fully septate individual was taken and one ancyloid seen. Another septate was taken March 8, and March 11-13 a solitary ancyloid was noticed. Fearing that there was not enough stock in the jar to . No. 581] |. GUNDLACHIA AND ANCYLUS 185 continue the race, March 15, Mr. Allen put in half a dozen ancyloids from the large jar. March 29 a mature septate was taken out, and it was noticed that the Plan- orbis looked frail as if insufficiently supplied with lime salts. October 11, 1909, two half-grown septates were taken from this jar. In the winter (1909-10) the Plan- orbis, for some unknown reason, completely disap- peared. From these data Mr. Allen concludes that about 80 per cent. of the stock in this jar had assumed the septate form, the conditions obviously being such as to stunt both Anacharis and ancyloids. In the 15 X 9-inch jar the vegetation is luxuriant and abundant, and the sep- tate individuals produced were only about two to five - per cent. of the ancyloids. From this Mr. Allen con- cludes that the formation of a septum is promoted by causes which tend to restrict or retard growth. The 9 X 7-inch Jar—This has a mixture of sand and soil at the bottom. There is plenty of algal growth, but the Anacharis is not as vigorous as in the 15 x 9-jar, from which it was stocked with ancyloids and Planorbis. In the winter of 1906-07 it yielded two septates. The winter of 1907-08 ancyloids were fairly numerous, more so than during the first winter, but no septates were de- tected. July 1, 1908, young fry, hatched that season, were visible. March 8, 1908, a fine large mature Gund- lachia was taken. The original ancyloid part was deep black and the flaring expansion beyond it was colorless and transparent. In the sand-bottomed jar the mature Gundlachia is uniformly yellowish translucent, but in the large jar with mud bottom the whole shell gets black- ish. December 13, 1909, a census of this jar was at- tempted. The day was dark and a count difficult, but the result was six septates and two ancyloids, all eight being small and immature. A Jar without Planorbis—Thinking it might be de- sirable to have a stock of the ancyloids not associated with Planorbis, Mr. Allen, about February, 1909, when 186 THE AMERICAN NATURALIST [Vou. XLV the Planorbis was not breeding, transferred some Ana- charis and a number of mature ancyloids to a new 15 X 9-inch jar, taking care not to introduce any Plan- orbis. May 3, 1909, the first ancyloid hatched in the jar was noticed; it was about half the size of the parents. Others appeared later. By December, 1909, the parent stock had disappeared and the stock hatched in the jar remains very small, indicating some unfavorable condi- tion. The bottom of the jar was covered with a mixture of ordinary and swamp soil, but the supply of swamp soil used in previous jars having been used up, that in the present jar was taken from another place, and may have contained some unfavorable matter. The Ana- charis in the jar is fairly flourishing, but there is no green algal growth. General Conclusions—The Gundlachia may repro- duce before assuming the completely mature form. The shell varies in apparent color in accordance with the muddy or sandy character of the bottom soil, but the dark coating in the former case is not incorporated with the shell structure. The ancyloid stage has a period of least activity im May. In July and August the septates appear. In au- tumn and early winter the third stage is developed, be- coming mature and complete in February or March. This course is, however, not invariable in the aquarium or domesticated specimens, since Mr. Allen has taken ancyloids in January or February, an irregularity prob- ably due to temperature and which might not have oc- cured in specimens under perfectly natural conditions. It is not certain that the ancyloids detected by Mr. Allen in July and August were the young of that season, since the minute creatures are very difficult to detect in the aquarium and can not be handled. They are so trans- lucent in the younger stages as to be practically invisible. However, it is probable that the eggs are laid during the winter and hatched in the very early spring. It seems likely that under average conditions only aà No. 531] GUNDLACHIA AND ANCYLUS 187 small proportion of the individuals advance beyond the septate stage; and also that, of the ancyloids, only part reach that stage. It is also probable, from Mr. Allen’s observations, that anything which tends to retard de- velopment may coincidently increase the tendency to form a septum. Since there is a period of least activity in May, a nat- ural observation year will be from one May to another. Mr. Allen summarizes the results obtained during the period, May, 1908, to May, 1909, as follows: None being taken before July nor after the following March, there were secured between July, 1908, and March, 1909, inclusive: 9 X 7-inch jar 1 septate 15 X 9-inch jar 4 septates total 15. 8 X 6-inch jar 10 septates -= From August 19, 1909, to December 13, 1909: 15 X 9-inch jar 8 septates fs 16. 9 X 7-inch jar 6 septates 8 X 6-inch jar 2 septates Further correspondence, during February, 1910, af- fords additional notes. A lot of the wild Thornburg Ancylus in alcohol was sent by Mr. Allen and, contrary to his expectation, on careful comparison with his series of ancyloids from his aquaria, no difference, beyond slight individual varia- tions, could be observed in the shells of the two series, while the radula and the soft parts, after repeated com- parisons, seemed to be identical in both. Mr. Allen especially notes that in the winter, 1909-10, the septates were the prevailing form in his aquaria, exactly the reverse of the case when the aquaria were freshly established. The generation, which appeared in ay and June, 1909, in the ‘‘Planorbis-free’’ jar, was dwarfed was not in sight during the latter part of the. winter, 1909-10, and may possibly have all died. Mr. Allen attributes the poor success of this jar to the use of 188 THE AMERICAN NATURALIST [Vou. XLV swamp soil from a different place from that previously used. February 15, 1910, being a dark day and therefore favorable for the septates to be out of sight, Mr. Allen counted those visible in the large aquarium. Six sep- tates and one ancyloid were noted. This illustrates the observation that (excepting the ‘‘Planorbis-free’’ jar) the septate is the prevailing form this season, and is promoted by causes which dwarf or retard growth. After noting the inexplicable way in which fresh-water mollusks sometimes appear and disappear from pools where they occur, Mr. Allen further suggests that the septate form may be a prelude to total disappearance of the species from a given place. Another count on February 17, 1910, gave three ancy- loids and three septates in sight, which Mr. Allen re- marks is the first time for a considerable period that the two forms have appeared in equal numbers. In the large jar every mature specimen seen this season has been conspicuously bicolored, the ancyloid or septate part being stained deep black, while the flaring extension is translucent and colorless, indicating that a resting period intervened between the completion of the septum and the formation of the mature shell. Three ancyloids seen February 17 were all translu- cent and about the same size. There can be little doubt that they date from the summer of 1909. Hence, Mr. Allen infers that the blackened original shells of the ma- ture Gundlachia date from the season previous. TABLE FOR JANUARY AND FEBRUARY, 1910 Specimens taken or observed Date Gundlachia Septates Ancyloids Janvary 12 (hig jar) 6.60685 00 83 0 Jannar 19 Gio MY gk, 1 0 January 31 (small jar) ............ 1 0 0 February 4 (Qui y 2 0 0 February 5 (big tar) -A se. 3 1 0 February 6 (medium jar) .......... 0 2 0 No. 531] GUNDLACHIA AND ANCYLUS 189 My last communication from Mr. Allen, dated De- cember 11, 1910, contains the following additional notes: As I have already written there was plenty of A-form (ancyloids) and no G-form (septates) visible in my original large jar last summer. But, since the latter part of November, besides ancyloids in various stages, young septates have been visible in fair abundance. I counted about a dozen in sight at one time. He concludes that ancyloids are present most of the year, but only young ones in May and mostly also in June. But septates appear to be a strictly winter form, that is, the immature septate stage appears in August or later, reaches maturity (Gundlachia) in February or March, and disappears about the end of April, after which and a shorter or longer interval the young ancyloids of the season begin to appear in the jars. If the hypothesis stated at the beginning of this paper be well founded, it would explain why mature Gund- lachias appear, if at all, usually as a few individuals in any given locality, and their presence can not be counted on, as in the usual case of fresh-water mollusks, and is distinctly a rarity in the temperate regions of the conti- nent, where there are no well-defined wet and dry seasons. NOTES AND LITERATURE MIMICRY IN some ways it would be a pity if the theory that mimicry has arisen through the operation of natural selection must be dis- carded since it is so ingenious in itself and was originated and fostered by such masters of theoretical biology. However, the old order seems to be surely giving place to new, here, as in other phases of the study of evolution. Since Wallace’s ‘‘Papilionide of the Malayan Region”’ the case of Papilio polytes has been a classic. The females of this butterfly are of three sorts: one like the male polytes, one like P. aristolochie and the third like P. hector. The two latter species are supposed to be distasteful to insectivorous animals while P. polytes is supposed to be edible. e two ‘‘models’’ are numerous in individuals and while ‘*P. hector and the hector form of P. polytes are confined to India and Ceylon, both P. aristolochiew and the aristolochie form of P. polytes have a wider range eastward.’’ The case is complete and has been convincing. : However, Punnett! found that in Ceylon The following statements may be taken as a fair presentation of the facts: 1. In the low-country the male form of polytes female is at least as numerous as either of the other forms, and may be the most abundant of the three. ; 2. In the northeast of the island, in the hector country, the aristo- lochiæ form polytes is nearly as abundant as the hector form, though its model is at any rate exceedingly scarce. 3. Higher up-country, where P. hector is rare or absent and P. aristolochie is common, the hector form of polytes is more abundant than the aristolochi# form. It is obvious that these statements are not in harmony with the ideas of those who look to the theory of mimicry for an explanation of the polymorphism that exists among the females of P. polytes. His observations concerning the enemies of butterflies con- firm those of other heterodox students, namely: that ‘‘as serious _enemies of butterflies in the imago state birds may be left out of *«¢Mimiery in Ceylon Butterflies, with a Suggestion as to the Nature of Polymorphism,’’ Spolia Zeylanica, Vol. VII, Part XXV, September, 1910. 190 No. 531] NOTES AND LITERATURE 191 account,’’ that lizards ‘‘certainly do not appear to exercise that nice discrimination with regard to butterflies which is necessary for the establishment of mimicking forms on the theory of nat- ural selection,’’ and that asilids are not averse to preying upon ** distasteful species.’ After pointing out that the resemblances on which the theory was based are far less striking in living, moving specimens than in their expanded museum state, he says Apart then from the questions whether the resemblances in many cases of mimicry are sufficiently close to be of effective service to the mimic, and whether the action of natural selection can be regarded as sufficiently stringent to have brought these resemblances into being, there are still the following difficulties in the way of the acceptance of the hypothesis of those who look to natural selection as an explanation of polymorphic forms in Lepidoptera: 1. The attribution of selection value to minute variation. 2. The absence of transitional forms. 3. The frequent absence of mimicry in the male sex. 4. The inability to offer an explanation of polymorphism, where the polymorphic forms ean not be regarded as mimics of a distasteful species. Moreover, the hypothesis assumes that minute variations of all sorts can be inherited, a position which at present is lacking in experimental proof. The gist of the constructive part of his paper is as follows: Natural selection plays no part in the formation of these polymorphie forms, but they are regarded as having arisen by sudden mutation, and series of transitional forms do not exist because such series are not biologically possible. Polymorphie forms may arise and may persist, provided that they are not harmful to the species, and it is possible to look upon their existence as due to the absence of natural selection rather than to the operation of this factor. . . . That polymorphism m a species should so frequently be confined to the female sex has long been remarked upon by those who study these matters, and the explana- tion most favored is that the female, burdened as she is with the next generation, is more exposed to the action of natural selection and in greater need of some protective adaptation. The weak point of such a view is that it does not explain why the male is not similarly protected. Tn Connection with this problem recent Mendelian research on sex- limited inheritance is highly suggestive. It has been shown that cer- tain types of inheritance receive their simplest explanation on the as- sumption that the female is heterozygous for a sex factor not contained qn the male and that this sex factor may, on segregation of the gametes, repel the factor for some other character for which the female is also 192 THE AMERICAN NATURALIST [Voi XLV heterozygous. From the beautiful experiments of Doncaster and Ray- nor it has been inferred that inheritance of this type occurs in the common currant moth (Abraxis grossulariata), where a distinct color variety, var. lacticolor, occurs. The factor for grossulariata pattern appears to segregate against the female sex factor, with the conse- quence that in only one type of mating, and that a rare one, is the lacticolor pattern transmitted to the male sex. Gametic formule are suggested and the conditions they im- pose are mentioned, but no breeding work was done. Whether the above explanation of the behavior of grossulariata is correct or not and also the correctness of the suggested formule for polytes are immaterial to the present discussion. It is now well known that ‘‘mutations’’ do occur in the females of insects and that the new characters can be transferred to the male by proper breeding. But, why do the mutants of P. polytes resemble greatly, even if they do not do so to such an extent as had been supposed, other species? On account of similar anatomical and physiological make up; or, in this case, did the proper gametic couplings once take place so that the then new female type was transferred to the males (as in grossulariata) and was there- after continued with such other modifications as were necessary to separate them taxonomically? In other words, the mimicking species came first and gave rise to the model! Mutation, in itself, is not the whole story. Granting it, we must be given a reason for the mutant resembling something else and while the amendment just made to Punnett’s paper may earry for this case, the chances are against it and we can not apply it to resemblances between species of different orders. In this connection, however, there seems to be an important thing which is often overlooked. It would be far more wonderful if, among the thousands of new forms which have arisen, there were no resemblances than it is that some of the forms are very much alike. As Punnett and others have pointed out, the same process which brought about such a close resemblance between, for ex- ample, earwigs (Orthoptera) and rove beetles (Coleoptera) that they are frequently mixed in entomological collections doubtless caused also the resemblances (here called mimicry because an advantage can be imagined) between certain flies and certam stinging Hymenoptera. If ‘‘chance’’ or ‘‘environment’’ is used in the former case it is not unlikely that it applies in the latter also. Frank E. LUTZ. The Anatomical Laboratory of Charles H. Ward 189 West Avenue, Rochester, N. Y. OUR HUMAN SKELETONS l ientifically prepared and mounted. They are undoubtedly the finest and strongest skeletons obtainable, and are purchased by the leading Medical and Literary Colleges, Schools, Surgeons, etc. We make a number of special skeletons for demonstrating dislocations, muscular ant landmarks, muscles, The mounting of i articulations permits movements as in tife. Strength sol Tigidity ar are = by the use a a special bronze wire of dia sdatimm normous oxidation. Portability 3 and ease of demonstration are attained by our sdokcsed! steel clutch — i ard, gee is a great protection as well. ji ese skeletons are shipped entirely "k up, carefully wrapped, ax with ay ies directions for papkig. a ae pe Danem qT gives further details. ey OUR SKELETONS OF TYPES OF VERTEBRATES are large i American species, mounted in b4 aa E T TORES of Sabra ae a E ms are leni era ema Schoolsan nd Universities. The series includes complete 3 ‘of which Se ialf Se ek Lite Masks « of Aborigi & si i Al a Ca t i of i The American Naturalist A Monthly Journal, established in 1867, Devoted to the Advancement of the Biological Sciences with Special Reference to the Factors of Organic Evolution and Heredi CONTENTS OF THE SEPTEMBER NUMBER Nuclear apga of Sexual AST in the Alge. Dr, LEY Moore DAVI ena! renna of spoe 1 Reproduction in Fungi, ESSOR Harp. Ss Pose fo the aaa Dinosaurs. Dr. W. D, Matt Shorter pre and Discussion: Evolution without Iso- lation, Dr. JOHN T. ETA saeth Seema, CasrER u on LD; The Logic of Chance in P , ARTHUR S. ex heag Notes and Literature: Animal Structure and Habits Professor G. H. PARKER. Plant Physiology, C. L. B. CONTENTS OF THE OCTOBER NUMBER LT Nuclear Phenomena of Sexual i in Gym- nosperms. CHARLES J. CHAMBE Nuclear Phenomena of Sexual 1 Reproduction in Angio- sperms. Professor D. M. M Shorter nega and Discussion : ‘ache. Dr. Max Mors Variations in Urosalpinx. Dr. HERBERT EUGENE WALTER. Notes and Literature: Notes on rato ae Davip serge JorpAN, The Mammals of Colorado, Prof D. A. COCKEREL LL, CONTENTS OF THE NOVEMBER NUMBER mp res He of Sieg tee sipe in Man. organ C. and CHARLES B, DAVEN The pes iii of the Hoey Bee—Can ach deta Colors? JOHN LOVE Shorter Articles and D : The Arithmetic of atic Product Moment Method "at ce atm cient of Correlation: Dr. J. ARTHUR E Notes and Literature: Schlosser on Fayû is, Dr. W. D. MEE The Ophidian mip Grayia: Professor T. D, A. COCKERELL. CONTENTS OF THE DECEMBER NUMBER seas of Skin Pi ee in Man. cient c. AVENPORT an B. DAVEN Spawn and Larva of mr cena Sii Pro- fessor W. H, PIERSO The Loge of Sizes a Shapesin Plants. Professor R. A. EMERSON. Shorter Articles and Discussion : The Modification of a erin bai „External Conditions. Profi D. A. COCKE Notes es Literature: Heredity, aes W. J. SPILLMAN. Index to Volume XLIV. —— CONTENTS OF THE JANUARY NUMBER Somatic Alteration: Its Origination and Inheritance. Dr. D. T. MacDoueaL, The Nature of Graft- hybrids. Professor Dovcias HOUGHTON CAMPBELL. A Double Hen’s Egg. Dr. J. THOMAS PATTERSON. Notes and Literature: Heredity, Dr. W. J. SPILLMAN. CONTENTS OF THE FEBRUARY NUMBER The Sp rey of Nave a tion Si Pure peas to Sex mited Inheri and to xual Dimorphism. Profe ssor T. H. prana ear Lines in cts age te’ ott Genetics in Lower Organisms. rofess Pad ‘outs of Tem sai upon Growing Mice, and the Persistence of Such Effects a a Subsequent Gen- ration. Dr, Francis B. SuMN The ee Ratio oa Biended regret SHINK- Data. - g=» Relative Conspicuousness n enga and Self- lored F Dr. RAYMOND PEA Some Considerations panime FA the Photogen Fane Shorter Articles and Discussion : Computing Correla” in Cases where Symmetrical — are com” manly used. Professor ame Single Number 35 Cents Yearly Subscription, $4.00 The NATURALIST will be sent to new subscribers for four months for One Dollar Garrison, N. Y. THE SCIENCE PRESS Lancaster, Pa. Sub-Station 84: NEW YORK VOL. XLV, NO, 532 APRIL, 1911 THE AMERICAN NATURALIST A MONTHLY JOURNAL Devoted to the Advancement of the Biological Sciences with Special Reference to the Factors of Evolution CONTENTS Page I. Genetical Studies on Oenothera. II. Dr. BRADLEY MOORE Davis - - 198 II. The Genotypes of Maize. Dr. GEORGE HARRISON SHULL - - - -234 II. Notes and Literature: Is the Female Frog Heterozygous in regard to Sex De- termination? Professor T. H. MORGAN. The Mutation Theory, Dr. R. R. GATES - ~ an ~ "a = - ~ - - - ~ - sete 263 THE SOIENCE PRESS LANCASTER, PA. GARRISON, N. Y. NEW YORK: SUB-STATION 84 The American Naturalist SS intended for publication and books, etc., intended for review should be sent tate Editor of THE AMERICAN NATURALIST, Garrison-on-Hucson, New York. Articles containing research work bearing on the problems of organic evolu- tion are especially welcome, and will be given preference in publication a. hundrea reprints of ere are supplied to authors free of charge. urther reprints will be supplied at co tions and adv srl _— roe sent et >a: a The posta rip subscription price is four dollars a year. twe yes -five cents additional. The charge for single copies is for a page. thirty-five cents, The advertising rates are Four Dollars —— THE SCIENCE PRESS ‘Lancaster, Pa. Garrison, N. Y. _ NEW YORK: Sub-Station 84 - Entered , tter, April 2, 1908, at th r, Pa., under the Act ot 3 Se i uaig g Se aaa | ARRE BIOLOGICAL LABORATORY = ical ———- Specimens. Books on Natural SUPPEN DEP ANTENE a of see = History, Seience, ‘ravel, Voyages, eto. See Tux | %2 Fobtyological Material of some invertebrates, fishes (in- cae 5 and mam- sae ae ee aaia al ales PAR = A Duke St, Adelphi—LondonEngland o PES bate lisis and M. GRAY, Tiai oh a r F Badge Dorani Press. | (E England) : JOURNAL OF GENETICS W. a. M.A., F. R-S., Director of the John Innes H i and R. C. PUNNET Variation ain nanas piares THE AMERICAN NATURALIST Vor. XLV | April, 1911 No. 532 GENETICAL STUDIES ON CENOTHERA. II Some Hysrips or Cnothera biennis ann O. grandiflora THAT RESEMBLE 0. Lamarckiana DR. BRADLEY MOORE DAVIS Tue status of Cnothera Lamarckiana is a matter which must be given serious consideration in any attempt to judge the value of De Vries’s mutation theory, for the reason that the behavior of this form in throwing off marked variants (mutants) from the type has been re- garded by De Vries as direct experimental proof of mu- tation. Indeed, the theory of De Vries may fairly be said to rest chiefly upon the behavior of this interesting plant, the account of which forms so large a part of his work, ‘‘Die Mutationstheorie,’’ 1901-1903. Aside from his claim of direct proof of the origin of mutations from nothera Lamarckiana, De Vries offers a considerable body of indirect evidence of the sort pre- sented in Darwin’s ‘‘Variation of Animals and Plants under Domestication,’’ and in that extensive and very carefully sifted account of Bateson, ‘‘Materials for the Study of Variations,” 1894. However, much of this in- direct evidence of De Vries deals with the origin of ‘‘sports’’ from domesticated forms or with the origin * Contribution from the Laboratory of Geneties, Bussey Institution of Harvard University No. 7. An investigation conducted with aid from the Elizabeth Thompson Science Fund for which the author desires to express his indebtedness, 193 194 THE AMERICAN NATURALIST (Vou. XLV of new forms under conditions that are not typical of those of nature in the wild. For these reasons such evi- dence could never appeal with so much force as would direct experimental proof that a wild species is in the habit of producing suddenly new types sufficiently dis- tinct from the parent form to rank as new species or even as strongly marked varieties. In ‘‘Die Mutationstheorie’’ of De Vries the behavior of Enothera Lamarckiana in giving rise to the so-called mutants is presented as evidence that new species have come into existence without intermediate steps from a form which is assumed to be typical of a species in na- ture. (Enothera Lamarckiana is made to bear the weight of an elaborate hypothesis, treating of funda- mental problems, very much as the apex might be made to bear the weight of an inverted pyramid. As the equi- librium. of the inverted pyramid depends upon the sta- bility of its apex, so the value for the mutation theory of the evidence from the behavior of Lamarckiana must rest with the status of this plant as a form truly repre- sentative of a typical species. De Vries from the beginning took it for granted that (Enothera Lamarckiana was a native American species introduced into Europe, an assumption that was perhaps not unnatural, although dangerous when the responsibil- ity of direct proof of the origin of species by mutation was laid upon its behavior. As far as the writer is aware, O. Lamarckiana, as a wild American species, is unknown. No American locality can be cited where it may be found as a clear component of the native flora. There are certain records of its presence under condi- tions that indicate the possibility of its being sometimes a garden escape, and there is some herbarium material, referred to Lamarckiana, which, however, has not been tested by culture and was collected at times when the im- portance of the most critical judgment in identification was not appreciated. It cannot be said that American botanists are not alive to the importance of the status of No.532] GENETICAL STUDIES ON (2ENOTHERA 195 Lamarckiana, for it is well known that a certain group would follow with persistence any clue that might give evidence of its being or having been an American native species. Crities of the evidence for De Vries’s mutation theory have been aware of the point of weakness that lay in the uncertain status of G@nothera Lamarckiana and the sug- gestions of Bateson and Saunders (’02, p. 153), Hast (707, p. 34), Boulenger (’07, p. 363), Leclere du Sablon (710, p. 266), Tower (710, p. 322), and others have prob- ably occurred to many, namely, that this plant is of hy- brid origin and that the appearance of its ‘‘mutations’’ is due to the continued splitting off of variants after the manner of hybrids. This view is held by a number of American botanists with whom the writer is acquainted and represents the attitude of those who are sceptical of the importance of mutation as a factor of organic evolu- tion in nature. If Lamarckiana is of hybrid origin it should be possible to obtain evidence of its probable parentage, and the present paper offers a hypothesis | with a considerable body of evidence in its favor. After the evidence has been presented the hypothesis will be discussed in the concluding section entitled ‘‘The Pos- sible Origin of @nothera Lamarckiana as a Hybrid of O. biennis and O. grandiflora.” None of the hybrids of biennis and grandiflora de- scribed in the following pages are identical with La- marckiana. There are important differences, chiefly of foliage and stem markings, which distinguish the hybrids at a glance, but on the other hand these characters in taxonomy would be considered of minor importance and the hybrids, if their origin were unknown, could not be placed elsewhere than next to Lamarckiana. Further- more, these differences are of a sort that are likely to be much less apparent when the results of crosses made this summer (1910) between certain types recently dif- ferentiated become known in succeeding cultures. In an investigation of this character the results, as every 196 THE AMERICAN NATURALIST: [Vou. XLV student of genetics knows, come slowly, and the writer feels no necessity of offering an apology in publishing preliminary data, since they are based on three seasons’ study, even though he hopes to present more conclusive evidence in the future. The cultures of the past season (1910) were grown partly at the Bussey Institution and partly in the Botanic Garden of Harvard University, where facilities were of- fered for which the writer is deeply indebted. METHODS To break the biennial habit and obtain flowering plants in one season it is only necessary to sow seeds in the hot house during the winter, where rosettes will develop, which may be set out in the open in the early spring. The cultures of 1910 were sown early in January and had developed large rosettes by May 5, when the plants were transferred to the gardens. It is best to sow the seeds thinly in large seed pans (with, of course, sterilized soil) from which each individual seedling may be potted. Cross pollination was always performed on flowers from which the unopened anthers had been removed. The best results are generally obtained when a castrated flower is left bagged for twelve to twenty-four hours before the pollen is applied, in order that the stigma may have time to mature fully which will be evident from the moist exudation on its surface. Strong manila bags tied firmly over the pollinated flowers have proved more satisfactory than special bags of parafin paper, which appear to hold the flowers in an atmosphere too moist for the best results. It is my practise to dip the forceps in a bottle of alcohol before each transfer of pollen and also to rinse the hands in alcohol. The pollen of nothera is so sticky that under ordinary conditions there is no danger from wind blown pollen, and by fol- lowing the practise outlined above there is little or no probability of impure pollination. In the future my plan will be to sow the seed capsule No. 532] GENETICAL STUDIES ON @NOTHERA 197 by capsule, which is the safest method to obtain quantita- tive results. In the past I have sown from bulk collec- tions of seed and my studies so far must be regarded as primarily qualitative in character. This practise was followed under the impression that probably only a small proportion of the seed from a cross, apparently as ex- treme as that between biennis and grandiflora, would germinate. The results, however, have shown that the seed of this cross is fertile to a very high degree. In conse- quence my cultures of this season gave three or four times more seedlings than it was practicable to bring to maturity. A process of selection became at once neces- sary, which was followed with the end in view of obtain- ing a variety of types, but it soon became evident, with the later development of the cultures, that it is impos- sible to select with accuracy among the young seedlings. Thus plants which showed certain tendencies as young rosettes or seedlings of five to nine leaves often changed very materially in later growth. For these reasons my cultures of this year even as qualitative studies are un- doubtedly not fully representative. (Enothera biennis L. My first crosses between biennis and grandiflora were made at Woods Hole, Mass., in 1908. This locality was carefully searched for rosettes of biennis with broad leaves and two plants of this character were transplanted from waste ground to the garden. The rosettes were similar and at maturity the plants proved to be the same form of biennis and were practically indistinguishable. These two plants were the starting points of two strains of biennis, designated A and B, which have been culti- vated through two and three generations, respectively, and have proved constant. It seems quite certain that under the name Œnothera biennis is included a number of races with well-marked peculiarities. These races are probably very pure, for the reason that close pollination is certainly usual, if 198 THE AMERICAN NATURALIST [ Vou. XLV Fic. 1. Mature rosette of @nothera biennis, B (10.21a). not universal, among the forms of the species. The lower portions of the stigma lobes in the bud lie below the tips of the anthers which discharge their pollen before the bud opens. As a result the stigma is not only well pollinated in the bud, but cytological studies on my strains A and B have shown that the pollen tubes reach the embryo sacs before there is any opportunity for cross pollination. Hybridization in nature could hardly occur in these forms unless their own pollen should be insuffi- cient for the number of ovules in the ovary or was much slower than foreign pollen in effecting fertilization ; al- ternatives that are very unlikely. These conditions, similar physiologically to those of cleistogamous flowers, are present in all of the forms of biennis known to the writer, and have been noted by De Vries for the Euro- pean types. As a result a strain once established is cer- tain to remain pure throughout at least the great mass of its seeds. The chief characteristics of the biennis strains A and No. 532] GENETICAL STUDIES ON ŒNOTHERA 199 Fic. 2. Mature plant of @nothera biennis, B (10.21a). B, employed in the hybrid studies of the present paper, when under good cultivation are as follows: 1. Rosettes.—The mature rosettes (Fig. 1), from 3 to 4 dm. broad, have about 40 closely clustered leaves, spatu- late, irregularly toothed at their base, and green except for occasional reddish spots. The rosettes are persistent and conspicuous during the early development of the main stem. 2. Mature Plants —The mature plants, 1-1.5 m. high, have a spreading habit (Fig. 2) with long side branches. Stems chiefly green (brownish below), the papillate glands at the base of long hairs also green. Basal leaves on the main stem narrowly elliptical, about 16 em. long (Fig. 3), leaves above lanceolate. 3. Inflorescence.—Bracts lanceolate, less than half the length of the buds (Fig. 4). 4. Buds.—About 5.5 em. long, the cone 4-angled (Fig. 4). Sepals clear green, their tips not markedly attenuate. 200 THE AMERICAN NATURALIST [ Vou. XLV 5. Flowers.—Small (Figs. 3 and 5). Petals about 1.3 em. long. Lower halves of stigma lobes (when closed) below the tips of the anthers. 6. Capsules.—Gradually narrowing from the base, 2- 2.5 em. long. 7. Seeds—Light brown. Those developed in the ovary of biennis after pol- lination by grandiflora are similar to the female parent. The most striking pecu- liarities of strains A and B in comparison with cer- tain other American types of biennis are:—the small flowers, green stems, and the absence of that red coloration in the papillate glands which is seen in - some forms of the species and is conspicuous in La- marckiana. Since differentiating the strains A and B in 1908, I have had an opportunity of observing somewhat extensively various forms of biennis in the vicin- ity of Boston, where l —! the prevailing type has eooo ace: k bogie rs xeon = larger flowers, frequently leaf from lower portion of main stem. a broader leaf, and usually stems punctate with red tinted glands. There is con- siderable variation in the characters above mentioned and I have this year selected certain plants growing wild on the grounds of the Bussey Institution that in some respects are likely to prove much more satisfac- tery for my purposes than the strains A and B. These No. 532] GENETICAL STUDIES ON ŒNOTHERA 201 l j biennis B srandiflora B biennis B | grandiflora B Fic. 4. Buds of biennis B and Fig. 5. Flowers of biennis B and grandifiora B. grandiflora B. plants (strains C and D) have been crossed this year with the best strains of grandiflora (B and D) and I expect to grow the hybrids in another season, when the strains will also be tested for their stability. The testing for purity is of course a necessary precaution, although, as explained above, the habit of self-pollination in biennis makes it very unlikely that any of these plants are tainted with foreign blood. These strains will not be further described until the prospective cultures have been grown. Sowings from the wild seed collected as (nothera grandiflora, described below, have given several plants of a southern form of biennis (strain S) which may prove of considerable interest since some of its characters (stem coloration, form of buds, size of flowers, etc.) are 202 THE AMERICAN NATURALIST [ Vou. XLV very favorable for combination with grandiflora with cer- tain ends in view. This southern strain was crossed this season with grandiflora, and the hybrids and the parent biennis will be studied through further cultures. (Enothera grandiflora. Ait. We are indebted to Dr. MacDougal (’05, p. 7) and to Miss Vail (’05, p. 9) for accounts of the rediscovery of this remarkable species of the southern United States and for a clear analysis of its probable history. Aiton’s original description (1789), from material grown at Kew, states that Gnothera grandiflora was introduced by John Fothergill, M.D., in 1778. An expedition of William Bartram in 1776, undertaken at the request of Dr. Foth- ergill for the purposes of botanical discovery, records the finding of a large-flowered @nothera near Tensaw (Taensa), Ala. Bartram’s picturesque and excellent de- scription of this new plant (see MacDougal, ’05, p. 7), together with a herbarium specimen in the British Museum from ‘‘Hort. Fothergill, 1778,’’ makes it evident that Bartram must have sent seed to Fothergill, through . whom the plant was introduced into England in 1778. Professor S. M. Tracy in 1904 visited the original lo- cality and found the species growing in considerable quantity along the east bank of the Alabama river in the vicinity of Dixie Landing, which is only a few miles from Tensaw. His material agrees with the descriptions of Bartram and Aiton and, according to Miss Vail, with the herbarium specimen of Fothergill, and there seems to be no doubt that nothera grandiflora, so widely culti- vated, has been rediscovered growing wild in its original locality. My seed of Œnothera grandiflora was collected by Professor Tracy at Dixie Landing in September, 1907. During the past three seasons I have sampled the col- lection to the extent of bringing to maturity from the wild seed thirty-four plants, and in addition some sev- enty rosettes and young plants have passed under my No. 532] GENETICAL STUDIES ON (2NOTHERA 203 inspection. Besides the above I have selected and cul- tivated from this material three strains (grandiflora A, B, and D), which have been carried through a second gen- eration represented by twenty-four mature plants. It became apparent, as my cultures progressed, that the Œnothera grandiflora growing near Tensaw is far from uniform in character. I have so far selected four distinct types of plants, only one of which, however, has been cultivated in later generations. The fact that the stigma of grandiflora is generally well above the tips of the anthers prevents pollination in the bud and in the opening flower, and offers a very much greater oppor- tunity for cross pollination than is possible in biennis. This condition is probably responsible for the hetero- geneity of the species. The type which I have under cultivation and which has proved stable is characterized by a somewhat broader leaf than is common to the species. From it have been differentiated the following three strains (A, B, and D), which have been used in the crosses with (nothera biennis. 1. Strain D came directly from a wild plant. 2. Strains A and B are the result of a cross between D and a similar plant, F, in the garden at Woods Hole in 1908. The cross was made to fix thoroughly the char- acters of a broad leaf and red coloration of sepals pres- ent in both parents. The two strains (A and B) are es- sentially similar, differing chiefly in the degree of red coloration present in the sepals, a character that is var- iable and probably cannot be depended upon as a fixed factor. As a matter of fact, strain D in a second genera- tion has proved very uniform and it is not likely that I Shall make further use of strains A and B. All three Strains are representative of the larger proportion of the plants of grandiflora that have passed under my ob- Servation, being, however, what a gardener would select as luxuriant forms with broader and larger leaves than the average. 204 THE AMERICAN NATURALIST [ Vou. XLV Fic. 6. Mature rosette of @nothera grandiflora, B (10.4a). In addition to the type represented by the strains A, B, and D (which is described in the paragraph below) there have appeared in the cultures from the wild seed the following three types markedly different from one another and from the general run of wild grandifloras. 1. A type represented by a large plant (2 m. high) peculiar for its light green broadly elliptical leaves, gen- erally green stem, green sepals, and a close rosette of erinkled leaves strongly resembling a half-grown rosette of Lamarckiana. This type, represented by a single plant (grandiflora I) appeared this year, 1910, and is likely to prove of great interest. It will not, however, be described until its behavior in later generations has been noted and its crosses with biennis have been grown. 2. A type represented by a relatively small plant (1.2 m. high) with stiff, broadly lanceolate, revolute leaves, and sepals deeply blotched with red. This peculiar form No. 532] GENETICAL STUDIES ON GENOTHERA 205 Fic. 7. Mature plant of @nothera grandiflora, B (10.4c). (grandiflora R) is too far from the general type of gran- diflora to be of value in the present study. 3. A type represented by a plant (grandiflora Z) with narrow lanceolate leaves, much too narrow to give re- sults, if crossed with biennis, that would approach La- marckiana. The chief characteristics of the grandiflora strains A, B, and D, employed in the hybrid studies of the present paper, when under good cultivation are as follows: 1. Rosettes—The mature rosettes (Fig. 6), 2-2.5 dm. broad, have about 20 loosely clustered leaves, broadly elliptical, irregularly and sometimes deeply cut at the base, slightly crinkled, and mottled with reddish brown blotches. The rosettes are transitory, the leaves with- ering during the development of the main stem. 206 THE AMERICAN NATURALIST [Vou. XLV Fic. 8. Flowering = branch of @nothera grandiflora, B (10.4a), with a leaf from the lower portion of the main ste . Mature Plants—The mature plants (Fig. 7), 1.5-2 m. pns have a more upright habit than biennis. Fre- quently the branching is profuse, the main stem and long side branches being covered with short shoots. Stems reddish, green only towards their tips, the papil- late glands following the color of the stem. Basal leaves on the main stem ovate or broadly elliptical, about 15 cm. long (Fig. 8); leaves above broadly lanceolate. 3. Inflorescence.—Bracts early in the season leaf-like and more than half the length of the buds (Fig. 4) ; later becoming very much smaller. No. 532] GENETICAL STUDIES ON (2NOTHERA 207 Fic. 9. Rosette of a hybrid (10.30 La), grandiflora B x biennis A, F, generation, 4. Buds.—F rom 9-10 em. long, the cone scarcely angled (Fig. 4). Sepals marked with reddish brown blotches, sometimes dull and faint; their tips attenuate. 5. Flowers.—Large (Figs. 5 and 8). Petals about 3.3 em. long. Stigma lobes generally 2-5 mm. above the tips of the anthers. 6. Capsules ——Tapering rather sharply from the large base, 2.5-3 em. long. T. Seeds—Dark brown. Those developed in the ovary of grandiflora, after pollination by biennis, have the Same color as the female parent. The above description is so worded as to emphasize the characters of the grandiflora strains A, B, and D in con- trast to the biennis strains A and B. It should be com- pared with the description of the latter forms to under- stand clearly the conditions that appear in the hybrids, which will now be described. 208 THE AMERICAN NATURALIST [ Vou. XLV > Np a Fic. 10. Mature plant of a hybrid (10. 30 eae grandiflora B x biennis A, F, generati Hysrivs 1N THE F, Generation The cultures of 1910 in the F, generation were hybrids of three different combinations of | parents, as follows: 1. grandiflora B X biennis A (10.30), grown at the Bussey Institution. 2. grandiflora B X biennis B (10.18) and the recip- rocal cross (10.19), grown at the Bussey Institution. 3. grandiflora A X biennis B (10.17) and its recip- rocal cross (10.20), grown at the Botanic Garden of Har- vard University. Bearing in mind that the biennis strains A and B are practically indistinguishable and that the grandiflora strains A and B are essentially similar, the cultures as a whole would not be expected to differ markedly, which was the case. The figures of hybrids published with this paper are from two plants in the first culture (grandiflora B X No.532] GENETICAL STUDIES ON ŒNOTHERA 209 1G. - Flowering side branch of a hybrid (10.80 La), grandiflora B x biennis a, F, generation. At the left is a leaf from the lower portion of the ste main biennis A). They were selected as likely to prove the most interesting for further cultures in the F, and later generations. These two plants 10.30 La and 10.30 Lb will be described in some detail, together with the general fea- tures of the cultures. l. grandiflora B X biennis A (10.30). This culture gave about 400 seedlings from which 57 were selected for the breadth of the cotyledons and the shortness of their petioles. These were brought to maturity. The char- acters of the parents were blended in the rosettes which had long, broadly elliptical leaves, toothed below, and col- 210 THE AMERICAN NATURALIST (Vou. XLV Fig. 12. Rosette of a hybrid (10.30 Lb), grandifora B x biennis A, F, generation. ored with large reddish spots and blotches. The mature plants likewise presented the characters of both parents blended in the habit, foliage, and flowers, all the charac- teristics of form and measurements being clearly inter- mediate. It was possible to distinguish certain rosettes as being more biennis-like or more grandiflora-like than the culture in general and the mature plants from these rosettes also exhibited similar differences. The plants of the culture therefore presented a certain range, the extremes being readily distinguished as more like one parent than the other although never approaching closely to either. Between the extremes were numerous transi- tions. Two rosettes of this culture were selected for their resemblance in certain particulars to Œnothera La- marckiana and the mature plants from these proved to be among the most interesting in the gardens. These hy- brids, 10.30 La and 10.30 Lb, were representative of a No. 532] GENETICAL STUDIES ON (ENOTHERA 211 Fic. 13. Mature plant of a hybrid (10.30 Lb), grandiflora B x biennis A, F, generation. type of hybrid of biennis and grandiflora that has proved not uncommon in my cultures. The description of these two plants will now follow, arranged to bring out the sal- ient features in comparison with the parent species and with Lamarckiana. Hybrid 10.30 La. 1. Rosette-—The mature rosette (Fig. 9), about 3.3 dm. broad, was persistent. Older leaves more elliptical and less spatulate than those of biennis (Fig. 1), deeply cut at the base as in grandiflora ig. 6), spotted with red. The younger leaves were markedly crinkled as in Lamarckiana (Fig. 15), but nar- rower and more pointed. 2. Mature Plant—The mature plant (Fig. 10), 1.8 m. high, had a straggling habit with long branches arising from the base, short clustered shoots above (grandiflora- like), a habit very different from the symmetry of La- 212 THE AMERICAN NATURALIST [ Vou. XLV Fic. 14. Flowering side branch of a hybrid (10.30 Lb), grandiflora B x biennis A, F, generation. At the left is a leaf from the lower portion of the main stem. marckiana (Fig. 16). Stem green above, mottled red below, occasional regions where red-tinted papillate glands lay on green portions of the stem as in Lamarck- iana. Basal leaves on the main stem (Fig. 11) elliptical, about 22 em. long, only slightly crinkled and not so long as the much-crinkled basal leaves of Lamarckiana (Fig. 17). Leaves on the upper portion of the plant broadly elliptical, slightly larger than those of Lamarckiana. 3. Inflorescence.—The inflorescence (Fig. 11) was strikingly similar to that of Lamarckiana (Fig. 17), the bracts being of about the same size and shape. No. 532] GENETICAL STUDIES ON G@NOTHERA 213 Fic. 15. Rosette of @nothera Lamarckiana (10.23c). 4. Buds.—F rom 6 to 6.5 em. long, about the same size as Lamarckiana and similar in form (compare Fig. 11 with Fig. 17), sepals green. Intermediate in size between those of parents. 5. Flowers.—Secarcely distinguishable from those of Lamarckiana (compare Fig. 11 with Fig. 17) and with the form and proportions of the parents blended. Petals about 2.2 em. long. Base of stigma lobes slightly below the tips of the anthers. 6. Capsules.—About 2.3 em. long, intermediate in size between those of the parents. T. Seeds—A shade of color clearly intermediate be- tween the light and dark brown of the parents. Hybrid 10.30 Lb. 1. Rosette——The mature rosette (Fig. 12), about 3 dm. broad, was persistent. Older leaves much broader than those of biennis (Fig. 1), cut at the base as in grandiflora (Fig. 6), a lighter green than the average of the culture, and with relatively few red spots as in Lamarckiana. The younger leaves were 214 THE AMERICAN NATURALIST [ Vou. XLV Fig. 16. Mature plant of @nothera Lamarckiana (10.23c). strongly crinkled, almost as broad as those of Lamarck- iana (Fig. 15), but more pointed. 2. Mature Plant—The mature plant (Fig. 13), 1.7 m. high, with long branches from the base, had a habit more spreading than that of Lamarckiana (Fig. 16), but was otherwise very similar. Stems green above, mottled red and brown below, the papillate glands of the same color as the portions of the stem upon which they lay. Basal leaves on the main stem (Fig. 14) broadly elliptical, about 20 em. long, without crinkles and not so long as the No.532] GENETICAL STUDIES ON ŒNOTHERA 215 Fic. 17. Flow ering seas branch of @nothera Lamarckiana cei with a leaf the lower portion of the main stem much crinkled basal leaves of Lamarckiana (Fig. 17). Leaves on the upper portion of the plant broadly ellip- tical, slightly crinkled, similar to those of Lamarckiana. 3. Inflorescence.—The inflorescence (Fig. 14) had bracts longer than those of Lamarckiana (Fig. 17), some- what crinkled and similar to the bracts in figures of (nothera scintillans (De Vries’s ‘‘mutant’’ from La- marckiana). 4. Buds.—F rom 6-6.5 em. long, about the same size as those of Lamarckiana, but with a more pointed cone and attenuated sepal tips (contrasted in Fig. 18), sepals green. Intermediate in size between those of the parents. 216 THE AMERICAN NATURALIST [ Vou. XLV Hybrid 10.30Lb Lamarckiana Fic. 18. Buds and flowers of a hybrid (10.30 Lb), grandiflora B x biennis A, F, generation, in comparison with those of @nothera Lamarckiana. 5. Flowers.—Scarcely distinguishable from those of Lamarckiana (contrasted in Fig. 18), and with the form and proportions of the parents blended. Petals about 2.2 em. long. Base of the stigma lobes slightly below the tips of the anthers. 6. Capsules.—About 2.5 em. long, intermediate in size between those of the parents. 7. Seeds—A shade of color clearly intermediate be- tween the light and dark brown of the parents. An examination of these two hybrid plants with respect to the contributions by their respective parents may be readily made by comparing the illustrations and descrip- tions of rosettes, mature plants, inflorescence, and flowers, bearing in mind that the parent biennis A is essentially indistinguishable from the strain biennis B which is here figured. It will be found that the hybrids present the characters of the parents in a blend. By a blend it must No. 532] GENETICAL STUDIES ON @NOTHERA 217 not be inferred that the characters of the hybrids are a perfect mean as to the measurement and proportions of their organs. This is certainly not the case for all of the characters of the hybrids and it would be a difficult matter to determine a perfect blend for a single character because of the fluctuating variations in the parent strains. All that I desire to demonstrate in the present account is the fact of blended conditions throughout these hybrids of the first generation, and the total absence of clear domi- nance of one parent over the other with respect to any character. It would be very difficult and probably impossible to support satisfactorily a claim that either of the two par- ent plants exhibited its influence to a measurably greater degree than the other. To illustrate this point let us examine hybrid 10.30 La. The rosette (Fig. 9) of this plant might be said to be more like that of biennis than of grandiflora, but the habit of the mature plant (Fig. 10) with respect to its short clustered shoots is more like the latter parent than the former, and thus two characteristic stages of the plant suggest opposite conclusions. This general balance of the influence of one parent over the other was manifest throughout the greater part of the culture, but, as previously noted, a small proportion of the hybrids was readily distinguishable as being more like biennis or grandiflora although never approaching closely to either parent form. The chief points of resemblance between the two hy- brid plants just described and @nothera Lamarckiana may be briefly summarized as follows: 1. The inflorescence was very similar to Lamarckiana, especially in the case of hybrid 10.30 La from which many branches might have been picked that as herbarium spe- cimens could not have been separated from a mixed and varied collection of Lamarckiana branches similarly pre- par 2i The only essential difference between the buds lay 218 THE AMERICAN NATURALIST [ Vou. XLV in the slightly greater attenuation of the sepal tips espe- cially in the case of the hybrid 10.30 Lb. 3. The flowers of the hybrids were scarcely distinguish- able from those of Lamarckiana, the small differences in the measurement of parts being no greater than might be expected in any reasonably large and varied culture of Lamarckiana. The chief difference with respect to the flower lay in the clear green color of the ovaries of the hybrids, the glands of which were not tinted red, as is characteristic of all material of Lamarckiana that the writer has seen. 4. Although the capsules were somewhat longer and more pointed than in the forms of Lamarckiana grown by the writer, they were not so long as in certain ‘‘long- fruited races’’ described by De Vries, who states that the fruits of Lamarckiana are highly variable and figures capsules as pointed as those of my hybrids (De Vries ’09, p. 528, et seq., Fig. 114). The points of difference concern chiefly the vegetative portions of the plants under discussion. 1. The rosette of the hybrids consisted of mixed forms of leaves, only the younger resembling the markedly crinkled leaves of Lamarckiana. 2. The habit of the mature hybrid plants was more straggling, lacking the symmetry characteristic of La- marckiana. The basal leaves were not so large and were but slightly crinkled; the upper leaves, especially in the ease of hybrid 10.30 Lb, were similar to Lamarckiana. 3. The coloration of the stem was green above and mottled red and brown below, in contrast to the green stems of Lamarckiana punctate with red-tinted papillate glands. Similar glands were present in the hybrids, but their color (portions of 10.30 La excepted) was that of the regions of the stem upon which they lay. The hybrids, therefore, resembled Lamarckiana as to the inflorescence, floral parts, and fruits; they differed chiefly in certain vegetative characters and in the colora- tion of the stem. It remains to be seen through further No. 532] GENETICAL STUDIES ON ŒÆNOTHERA 219 cultures which of the two, the resemblances or the dif- ferences, are more stable in inheritance and variation. The type of Lamarckiana which has been compared in this paper with the hybrids of biennis and grandiflora is one with which I have been familiar for the past five years. It has been represented in my cultures by strains from seed that has come to me through three different sources, all of the seed, however, originally being de- rived from the cultures of De Vries. These strains have not differed materially from one another, and as grown in my small cultures have not exhibited marked varia- tion. The rosettes and mature plants have agreed in habit and foliage with the descriptions of Lamarckiana in ‘‘Die Mutationstheorie.’’ The flowers have, however, been uniformly smaller than the measurements and fig- ures of De Vries, the petals being about 2.5 em. long in- stead of measuring 3 em. or more. In the flower struc- ture the position of the stigma has proved more variable than one would be led to suppose by the figures and de- scriptions of Lamarckiana, the stigma generally being but slightly above the tips of the anthers or about at their level, and in some plants distinctly below. 2. grandiflora B X biennis B (10.18), and the recipro- eal cross (10.19). From about 200 seedlings of the first culture and about 250 seedlings of the second cul- ture, 66 and 70 plants, respectively, were brought to maturity, being selected for the breadth of the cotyle- don and the shortness of its petiole. As the rosettes formed there appeared much variation in the amount of anthocyan developed in the leaves, the larger propor- tion being marked with dull red spots and blotches, only a small number having few spots as in Lamarckiana. The form of the leaves likewise varied and certain of the rosettes were readily separated as being more biennis- like or grandiflora-like than the average. The plants that developed from these extreme forms of rosettes were also somewhat more like the respective parents than the mass of the culture which presented the characters of 220 THE AMERICAN NATURALIST [Vou. XLV the parents thoroughly blended in the form and propor- tions of habit, foliage, and flowers. Considering the cul- tures as a whole, there seemed to be no marked difference between the first cross and its reciprocal. The average types of hybrids in both crosses were es- sentially similar and a number of types were very close to the hybrids 10.30 La and 10.30 Lb of the previously described culture. Six plants in culture 10.19 were selected for special peculiarities, but these will not be de- scribed unless their behavior in the F, generation should justify a detailed account. 3. grandiflora A X biennis B (10.17), and the recipro- eal cross (10.20). From about 200 seedlings of the first culture and about 150 seedlings of the second culture, 49 and 60 plants, respectively, were brought to maturity, being selected for the breadth of the cotyledon and shortness of its petiole. These cultures were grown in a stiff clay at the Botanic Garden and presented an in- teresting contrast to the cultures previously described which were grown in a somewhat sandy well-fertilized soil. The plants were smaller and less vigorous vegeta- tively, although they flowered very freely. The rosettes and mature plants presented the characteristics of the parents well blended as in the other cultures. There were also a few extreme types that resembled one or the other of the parents more closely than the average. There appeared to be no significant differences between the first cross and its reciprocal. Two plants with marked peculiarities were selected from culture 10.17 and will be carried through an F, generation; they will be described if- their further cultivation proves of interest. Although the evidence, not being quantitative in char- acter, is incomplete, nevertheless the following points may be noted, at least provisionally, from these observa- tions on F, generations. (1) There was no indication from these cultures of a marked preponderance of either paternal or maternal influence upon the hybrids. No. 532] GENETICAL STUDIES ON G2NOTHERA 221 (2) No character of either parent was observed to be dominant. (3) It is doubtful whether there would be any material difference between a cross and its reciprocal if each were equally vigorous. (4) Although the extreme types in the culture, approaching somewhat the respec- tive parents, could be readily distinguished, they were connected by transitional forms and a sharp line could not be drawn between two sets of hybrids, such as have been described by De Vries (’07, ’08) as ‘‘ twin hybrids’’ and reported for crosses between the Onagra group and Lamarckiana. From observations on small cultures dur- ing the season of 1909 (Davis 710, p. 113) the writer was led to believe that ‘‘twin hybrids’’ might be present in this cross, but he no longer regards this as probable. Hyprips IN THE F, GENERATION In a recent paper (Davis ’10) I described two small cultures of hybrids between biennis B and grandiflora D that were grown at the Botanic Garden of Harvard Uni- versity in the season of 1909. Four of the plants of these cultures were of special interest as presenting flowers and inflorescences very similar to Lamarckiana, although differing markedly in foliage. I unfortunately was unable to observe the early development of these plants and for this reason they were not very good forms on which to base studies of their progeny in the F, gen- eration. Their seed was, however, sown this season and the cultures, described below, were of interest as indicat- ing the probable behavior of hybrid plants of biennis and grandiflora. These four lines will not be cultivated further, since I have in the hybrids 10.30 La and 10.30 Lb material better suited to the purposes of a quantitative study for the reason that the records of their life history have been kept in detail. The seeds of these hybrids proved fertile to a very high degree, but it was practicable to grow only a small proportion of the seedlings to maturity. 1. Progeny from hybrid 9ba, biennis B X grandiflora 222 THE AMERICAN NATURALIST [Vou. XLV D. This hybrid plant (Davis, ’10, pp. 112 and 113), an excellent blend of the parent forms, was similar to Lamarckiana in habit and floral structure, but differed in having smaller, uncrinkled leaves on the lower portions of the plant and larger bracts upon the inflorescence. From about 600 seedlings 73 plants were carried through the rosette stage and set in the ground, being selected to represent various types. The seedlings were strikingly diverse, some having long cotyledons similar to those of grandiflora, others having shorter and broader ones, and a large proportion with small light yellow, etio- lated cotyledons. Many of the latter seedlings died before the appearance of the second leaf, the others developed very slowly, forming rosettes one-fourth or one-third the size of the normal with more or less etiolated leaves. Twenty-three of the dwarf rosettes were set out in the garden and of these seven finally grew to be large plants similar to the average of the culture, but with a some- what etiolated foliage; of the remainder several died and the others developed into dwarf plants from 2-8 dm. high, small leaved, sparsely branched, and with flowers smaller than the average but larger than the biennis parent. The behavior of these etiolated dwarfs resembled De Vries’s account of the appearance of the form albida in his cultures of Lamarckiana. As the normal rosettes approached maturity it was possible to distinguish certain ones as somewhat more biennis-like or more grandiflora-like than the average, and the mature plants which developed from these showed similar points of resemblance to the respective parents of the cross. Nevertheless, the culture as a whole pre- sented these parental characters well blended, although exhibiting a much wider range of variation than the F, generation of this cross. This variation appeared to indicate a relative segregation of the parental characters deserving of detail studies upon larger cultures. There were a number of plants similar to the parent hybrid, but none markedly nearer to Lamarckiana. } No. 532] GENETICAL STUDIES ON GNOTHERA 223 2. Progeny from hybrids 9ba, 9bb, and 9be, grandiflora D X biennis B. These three hybrids (Davis, 710, p. 114) presented the parental characters well blended. They were essentially similar to Lamarckiana in flower struc- ture and inflorescence, but differed in foliage and habit, the leaves on the lower portion of the stem being but half the length of those similarly placed on Lamarckiana and with only slight traces of crinkles; the habit was biennis- like. From about 350 seedlings of hybrid 9ba, 76 plants were brought to maturity, being selected as representative types of the rosette stages. A small proportion of the rosettes was dwarfed and the nine selected representa- tives of this type developed small plants 1-2.5 dm. high, generally without side branches; these did not flower. The normal rosettes varied greatly in the forms of leaves and extent of the red coloration, certain ones being dis- tinctly more like the respective parents of the cross than the average; these differences were maintained in the mature plants, but to a less marked degree. The culture in general presented a habit more grandifloralike than biennis-like, but all of the characters remained blended, although there was a considerable range of variation in flower structure and foliage. While a number of the plants were similar to the hybrid parents, none proved to be appreciably nearer to Lamarckiana. There were about 550 seedlings of hybrid 9bb, from which 93 plants, selected as representative rosettes, were brought to maturity. Relatively few dwarf rosettes were present in this culture; eleven of these being selected grew into plants 1-4 dm. high, small leaved and without prominent side branches, the larger of the dwarfs develop- ing small biennis-like flowers. The culture in general was more uniform than the preceding, but certain rosettes and mature plants were noticeably more like one or the other of the original parents than the average, which presented these parental characters well blended. The foliage of the culture was distinctly crinkled so that the plants re- 224 THE AMERICAN NATURALIST [ Vou. XLV sembled Lamarckiana more closely than those of the other cultures in the F, generation, differing chiefly in the smaller size of the basal leaves and in the absence of red tinted papillate glands on a green stem, the stem being mottled with red. One plant of this culture (10.12 Lz) was selected for marked peculiarities, but will not be de- scribed unless its behavior in an F, generation proves of sufficient interest. About 800 seedlings of hybrid 9be appeared in the cul- ture, from which 95 plants were selected as representa- tive types of rosettes. A few dwarf rosettes were pres- ent, six of which set in the ground developed into un- branched plants about 1 dm. high, that failed to flower. The culture in general exhibited considerable variation, the most interesting types of plants being several with light green, smooth, obtusely pointed leaves, similar in shape to Lamarckiana, but without crinkles. There was shown the same previously described tendency on the part of a few rosettes and mature plants to depart from the average of the culture towards the characteristics of the respective parents of the cross, maintaining, how- ever, a blended structure of their parts. Considering these cultures of F, generations in com- parison with the F, generations that have been grown, the most striking feature is the greater range of varia- tion exhibited not only by the F, plants as a whole, but by their different parts. Since the studies were not quanti- tative in character, because such a large proportion of the seedlings were necessarily discarded, it has not seemed best to describe the variations in detail and such an in- vestigation is deferred for the present. However, 1 in this increased variation is clearly indicated at least a relative segregation of the parental characters in the F, genera- tion.” 2 Extensive cultures from the seed of the two hybrids 10.30 La and 10.30 Lb, described in this paper, are now (February, 1911) seedlings with 4-5 leaves which already show marked segregation in this F, genera- tion, with the extreme types closely resembling seedlings of the par the cross and between these a large range of intermediates. ents of No. 532] GENETICAL STUDIES ON ŒNOTHERA 225 THe Posstste Oricin or Œnothera Lamarckiana as a Hysrip or O. biennis anb O. grandiflora We have shown that hybrids between certain strains of Œnothera biennis and O. grandiflora may be synthe- sized, which approach somewhat closely to Œnothera La- marckiana, and there is good reason to believe that fur- ther experimentation will result in the production of forms with a more perfect resemblance. It is now im- portant to ascertain, as far as this is possible, whether there are any historical reasons why Lamarckiana may not have arisen either accidentally or intentionally from such a cross. Œnothera Lamarckiana appears to have been under cultivation in the gardens of the Muséum d’Histoire Nat- urelle at Paris in 1797, being described by Lamarck’ under the name grandiflora. Shortly afterwards Se- ringet renamed the form Lamarckiana, recognizing it to be distinct from the grandiflora of Aiton. As previously noted from the investigations of Mac- Dougal (’05) and Vail (’05), the evidence is very clear that grandiflora was introduced into England in 1778 and was at that time under cultivation at Kew. Forms of (Enothera biennis had of course been in European gar- dens for many years previous to 1778. There was there- fore a period of about eighteen years (1778-1797) during which hybrids between biennis and grandiflora might have arisen in Europe before the earliest known record | of the cultivation of @nothera Lamarckiana in Paris. So striking an American novelty as (nothera grandi- flora would almost certainly have been passed on from Kew to other botanical gardens and in the interval be- tween 1778 and 1797 is likely to have become widely dis- tributed and cultivated. On historical grounds then there seems to be no reason with respect to the date of the first recorded recognition of @nothera Lamarckiana why this form might not have arisen in Europe as a hybrid of biennis and grandiflora, *“*Eneyclopédie Méthodique Botanique,’’ Vol. IV, p. 554, 1797. * De Candolle’s ‘‘Prodomus,’’ Vol. III, p. 47, 1828. 226 THE AMERICAN NATURALIST [Von XLV Let us suppose that it should be shown that Œnothera Lamarckiana was in existence previous to the date 1778, what effect would such evidence have on the hypothesis that the form is a hybrid of biennis and grandiflora? It would not in the writer’s opinion have weight against experimental proof that Lamarckiana or forms closely resembling this plant may be synthesized as hybrids of these wild American species. It would not prove that Lamarck’s plant in Paris (1797) was not a hybrid. It would merely indicate that Lamarckiana, having arisen as a hybrid in America, was introduced as such into Eu- rope. As already pointed out, the position in grandiflora of the stigma well above the anthers gives ample oppor- tunity for chance hybridization in nature. Indeed. the diverse forms that have appeared in my cultures from seed of grandiflora collected in the field clearly show that the species is far from homogeneous in character, a con- dition that is probably due to a large amount of cross pollination. It may be expected that careful search, es- pecially in the southern United States, will bring to light occasional plants with characters intermediate between grandiflora and other species, such as, for example, southern types of biennis, but it is also probable that the behavior of such plants in culture will show them to be heterozygous in character, i. e., hybrids. There have been two attempts to establish the presence of Lamarckiana in Europe previous to 1778 when grandi- flora was introduced at Kew. MacDougal (’07, pp. 5, 6) refers to Lamarckiana a description and figure of an (Enothera by Miller, Plate 189, Fig. 2, for the ‘‘Gar- dener’s Dictionary,’’ 1760. This figure, published in 1757, is of a large-flowered Œnothera with petals 2 or 2.2 em. long and by its side (Fig. 1) is a smaller-flowered form. With respect to the point under discussion, the most important features of these figures, clearly shown by the drawing, is the position and form of the stigmas, well below the tips of the anthers and with the lobes unexpanded in open flowers. These are peculiarities of No.532] GENETICAL STUDIES ON GENOTHERA 227 biennis and, in the writer’s experience, are not charac- teristic of Lamarckiana where the stigma lobes are usu- ally expanded in the open flower and generally above or about on the level with the tips of the anthers. Further- more the size of the petals in the illustration of the large- flowered type (Fig. 2) is no greater and indeed not so great as in some forms of biennis. Both of the figures show the essential characteristics of the flower of biennis to which they have generally been referred in taxonomic accounts. For these reasons the view of MacDougal that the illustration of the large-flowered type (Fig. 2) is of Lamarckiana and establishes its presence in Europe previous to 1757, is to the writer not convincing. The second attempt to establish the presence of La- ~. marckiana in Europe previous to 1778 is the announce- ment of Gates (710) that certain marginal notes in a copy of Bauhin’s ‘‘Pinax,’’ 1623, give in Latin an accurate description of this plant although differing in one or two minor characters. Gates presents an outline of the points which indicate to him that the description refers to Lamarckiana, but the notes themselves are not pub- lished. A full account is promised, in which we may expect to see these Latin notes and judge of them for ourselves, and comments on this announcement will be reserved for the present. Finally we must return to the question of whether or not it appears probable that @nothera Lamarckiana is at present a component of the American native flora. De Vries (’05, p. 368) refers to Lamarckiana certain her- barium material at the New York Botanical Garden and Missouri Botanical Garden, both collected by A. W. Chap- man in Florida (1860 or earlier), and also material in the Philadelphia Academy of Science collected by C. W. Short at Lexington, Ky. A thorough search (MacDou- gal, ’05, p. 6) by several botanists in the vicinity of Lexington Ky., Nashville Tenn., Knoxville Tenn., and Courtney Mo., in the endeavor to find living plants that might be identified as Lamarckiana, was unsuccess- 228 THE AMERICAN NATURALIST [ Vou. XLV ful. Later, Miss Vail (MacDougal, ’07, p. 67) came to the conclusion that the plant from Lexington, Ky., is grandiflora, and a possible escape from cultivation. I have not seen the herbarium material mentioned above, but in the light of the fact that many dried specimens could be prepared from my hybrids which as such would be considered Lamarckiana, it is clearly necessary that evidence from herbarium material should be weighed with much caution. The average herbarium material of the G@notheras is generally not sufficient to show the pe- culiarities of the earlier phases of development (rosettes and basal foliage) which in the case of Lamarckiana fur- nish diagnostic characters that are necessary for a full identification. Unless the evidence of field collections is followed up by garden cultures, there is the possibility of numerous errors of interpretation. A specimen in the Gray Herbarium of Harvard Uni- versity is stated by MacDougal (’05, p. 5) to agree per- fectly with @nothera Lamarckiana, but in this view the writer can not accord. This plant was apparently grown in the Cambridge Botanical Garden, Massachusetts and bears the date 1862. The specimens are accompanied by the significant notes in the hand writing of Dr. Asa Gray ‘‘from seed of Thompson, Ipswich,” and ‘‘said by English horticulturists to come from Texas.’’ The flowers are large, with petals about 4.5 em. long and sepals about 5 cm. long, very attenuate, the tips projecting 1 em. beyond the folded petals in the manner characteristic of grandiflora. The stigma lobes are also grandiflora-like in their length, about 8 mm., and in their position, about 5 mm. above the tips of the anthers. A large detached leaf, about 18.5 em. long, with some evidence of former crinkles, suggests by its form (although rather small) the basal leaves of La- marckiana. The flowers and upper foliage of this speci- men, however, agree very closely with broad-leaved types in my cultures of grandiflora and do not resemble the Lamarckiana that I have grown from seeds of De Vries, or with his figures and descriptions in ‘‘Die Mutations- No. 532) GENETICAL STUDIES ON @NOTHERA 229 theorie.” If this plant could be established as derived from C£notheras introduced into England by Messrs. Carter and Co. at about 1860 from seeds said to come from Texas, it would be a point of great importance, as will appear in the following paragraphs. De Vries (’05, pp. 384-385) offers strong evidence that the strains of Lamarckiana at present cultivated in Eu- rope have a genetic relation to seed of Messrs. Carter and Co., of London about 1860. This seed is stated to have been received unnamed from Texas and plants grown from it were pronounced by Dr. Lindley to be Lamarckiana. A specimen from one of these plants is figured in ‘‘The Floral Magazine,’’ Vol. II, Plate 78, 1862, this plate being reproduced in ‘‘L’Ilustration Horticole,’’ Vol. IX, Plate 318, 1862. This plate shows an Œnothera with flowers about 10 cm. (4 inches) in diameter and with a large amount of red coloration on the sepals and ovaries; the stigma is figured both above and below the tips of the anthers. The flowers of this illustration are larger than those of Lamarckiana, as known to the writer, and would do for grandiflora except for the posi- tion of the stigma which is much closer to the anthers than is typical for this species. The red coloration of the sepals and ovaries is much too deep for typical La- marckiana and not unlike some forms of grandiflora, but the sepal tips, as drawn, are not so long or so pointed as in the latter form. Indeed the identification of this plate with any probable (nothera is very difficult and the reasons why it should be called Lamarckiana are to the writer far from convincing, although it would perhaps be as easy to argue for this identification as for any other. It is, however, possible that new light may be thrown on the composition of the cultures of Carter and Co. through the plant in the Gray Herbarium described above. The date of this specimen, 1862, together with the very suggestive notes of Dr. Gray 7 from seed of Thompson, Ipswich,” and ‘‘ said by English horticul- 230 THE AMERICAN NATURALIST [ Vou. XLV turists to come from Texas,’’ make it appear possible that this plant was derived from the cultures of Carter and Co. If this could be established it would indicate that forms very close to grandiflora were present in the cultures or seeds of this firm. It is not at all improbable that Texas with its immense area and very great range of climatic conditions may harbor grandiflora or related types especially since it is known to be rich in species of (nothera and to have a number of large flowered rep- resentatives. There may have been thus a second introduction into England of grandiflora-like types through Carter and Co. at about the year 1860. While there is of course no means of knowing whether their cultures were uniform, it is altogether probable that the results of their sowings gave a diverse progeny, since that has been my experi- ence with seed from Alabama. There seems to be no reason why chance hybrids may not have been present or why grandiflora-like strains might not have shortly hybridized with European forms of biennis. These pos- sibilities are mere matters of speculation to which little assistance is given by the puzzling plate in ‘‘The Floral Magazine” and in ‘‘L’Illustration Horticole’”’ referred to above. A search among the English herbaria might, however, result in the discovery of specimens which would materially assist in the solution of a very interest- ing question—the identity of the plants grown by Carter and Co. At present the specimen in the Gray Herba- rium appears to offer the most important evidence bear- ing upon the question. The contention that Lamarckiana was introduced in the form of a native American species at this date, 1860, seems to the writer to be without suffi- cient foundation. The American botanist will ask himself why, if La- marckiana was present in America as a native species mM 1860, no localities are known where it may be observed in the field. It will be hard for him to believe that so strong and vigorous a plant, if a wild species, has become No. 532] GENETICAL STUDIES ON ŒNOTHERA 231 so recently extinct when, as he well knows, the Hnotheras are established as remarkably successful forms in our flora. The fact that Lamarckiana is not known as a com- ponent of the native American flora stands as the most serious obstacle to the view that this plant is representa- tive of a wild species. The writer believes it very prob- able that plants more or less resembling Lamarckiana will occasionally, or perhaps rarely, be found in parts of America and under circumstances indicating that they are not garden escapes, but it seems to him equally prob- able that these forms when tested in culture will give evi- dence of a heterozygous, or hybrid nature. The mere records of such plants as handed down by the average type of herbarium specimen, unaccompanied by experi- mental cultures, will have little or no value for the pres- ent problem—the origin of @nothera Lamarckiana. SUMMARY This paper offers a body of evidence which shows that hybrids resembling Enothera Lamarckiana may be syn- thesized from certain strains of the American native species O. biennis and O. grandiflora. The resemblances of the hybrids to this plant are strongest with respect to the inflorescence, buds and flowers. The differences are chiefly manifest in the basal foliage of the mature plant, in the coloration of the stem, and in the more strag- gling habit of the hybrids. The rosettes of the hybrids present mixed forms of leaves, the younger with points of similarity to Lamarckiana. Bearing in mind that other strains of biennis have characteristics more La- markiana-like than those of strains A and B, herein de- scribed, it is more than probable that the hybrids from certain crosses made this season (1910), when grown in future cultures, will come closer to the desired end— the synthesis of a hybrid so similar to Lamarckiana as x be practically indistinguishable by the usual taxonomic ests. Exception is taken to the claim of MacDougal (’07, 232 THE AMERICAN NATURALIST [ Vou. XLV pp. 5, 6) that Miller’s Plate 189, Fig. 2, for the ‘‘Gar- dener’s Dictionary,’’ 1760, establishes the presence of La- marckiana in Europe previous to the date, 1778, when grandiflora is known to have been introduced into Eng- land. The view of De Vries that strains of Lamarckiana were introduced into England about 1860, through seed of Messrs. Carter and Co. said to come from Texas, is dis- cussed with reference to certain specimens in the Gray Herbarium of Harvard University and in the light of the author’s experience with seed from Alabama, indicating that Carter and Co. probably had grandiflora-like types in their cultures which were also likely to have been of a mixed character. The absence, so far as is known, of La- marckiana as a component of the native American flora is emphasized as a point of great importance against the claim that Lamarckiana was introduced into Europe as an American wild species. A working hypothesis is presented as a result of the writer’s experimental studies and in relation to such his- torical evidence as is available, to the effect that Gno- thera Lamarckiana arose as a hybrid between certain types of biennis and grandiflora, recognizing that under these names must, for the present at least, be included a number of races which can only be clearly defined by laborious genetical investigations. The precise time and place of such an origin for Lamarckiana is a matter of mere speculation, but there seems to the writer no good reason why hybridization between biennis and grandi- flora might not have taken place in Europe between 1778 and 1797 (when Lamarckiana was first recognized at Paris) and also at later dates, as for example about 1860. It is also possible that Lamarckiana may have been introduced as a chance hybrid from America, but the probability of such an origin is naturally rather remote. The bearing of the possible hybrid nature of @nothera Lamarckiana upon the claim of De Vries that the behav- ior of this plant demonstrates the origin of new species No.532] GENETICAL STUDIES ON GENOTHERA 233 by mutation from a form repr tative of a typical wild species will be sufficiently evident to require no com- ment at this time. A discussion of the matter will there- fore be reserved until the writer has proceeded further with his studies. CAMBRIDGE, MASS., November, 1910. LITERATURE CITED Aiton, Willi ’89. Hortus Kewensis: A Catalogue of the Plants Cu Hiva in the iii Pokak traii a Kew. London, 1789. Bateson, W., and Saunders, Miss E. R., ’02. Bepo to the Evolution Com- mittee of the Royal Society. Report I, London, 1902. Boulenger, G. A., ’07. On the Variations of the Evening Primrose (@no- thera biennis L.). Jour. of Bot., Vol. XLV, p. 353, 1907. Davis, B. M., 710. Notes on the Behavior of Certain Hybrids of Gnothera in the First Generation. AMER. NAT., Vol. XLIV, De Vries, Hugo, 705. Ueber die Dauer der TE bei (Bacthive Lamarckiana. Ber. deut. bot. Gesell., Vol. XXIII, p. 382, 1905 De Heed Hugo, 707. On Twin Hybrids. Bot. Gaz., Vol. XLIV, ; p. 401, 1907. De vag Hugo, ’08. Ueber die Zwillingsbastarde von Œnothera nanella. r. deut. bot. Gesell., Vol. XXVIa, p. 667, 1908. De hie ta Hugo, 709. The Mutation Theory. Chicago, Vol. I, 1909. East, E. E., 07. The Relation of Certain Biological Principles to Plant Breeding. Conn. Agri. Exper. Sta., Bulletin 158, 1907. Gates, R. R., *10. The ae a a of Gnothera Lamarckiana. Science, Vol. XXXI, p. 425, Leclere du Sablon, Mathieu ae ne la nature hybride de 1’Oenothére de Lamarck. 910. MacDougal, D. T., Vail, A. M., Shull, G. H., and Small, J. s 705. Mutants and apgr of the (Enotheras. Carnegie Inst., Pub. 2 1905. MacDougal, D. T., Vail, A. M., and Shull, G. H., ’07. wana Varia- ae and Relationships of the Œnotheras. Oarhegis Inst., Pub. 81 - 1907 Tower, W, L., 710. The Determination of Dominance and the Modification of Waievins in Alternative (Mendelian) Inheritance, by Conditions Surrounding or Incident = the Germ Cells at Fertilization. Biel. Bull., Vol. eae p. 285, Vail, hie M., Onagra ae (Ait.), a species to be included in the Noni ody flora. Torreya, Vol. V, p. 9, 1905. THE GENOTYPES OF MAIZE! DR. GEORGE HARRISON SHULL STATION FOR EXPERIMENTAL EVOLUTION, Cotp SPRING HARBOR, N. Y. Tar doctrine of evolution had to overthrow the con- ception of permanency of specific types, generally held when Darwin’s ‘‘Origin of Species’’ was published, be- ` cause that conception was then associated with the idea of a separate original supernatural creation of each such type. It was Darwin’s great triumph that he succeeded in marshaling such an array of facts pertaining to varia- bility, as to convince the scientific world—and through the scientific world, ultimately the whole world—that everything is in a state of flux, and that there is no such thing as permanency among living things. Owing to the work of De Vries and the other early students of modern genetics, permanency of: type again demands serious scientific consideration, for such per- manency is no longer incompatible with the doctrine of evolution, being now associated with some form of the mutation theory. The old idea of the immutability of specific types was based upon almost total ignorance of genetics, as was likewise the Darwinian conception of fluidity and gradual change, for although many appeals were made by Darwin to the experiences of plant and animal breeders, it is now known that these experiences were the result of no such careful control of conditions or analysis of results as has been found necessary for the discovery of genetic laws. The critical work of the past few years has wrought a great change and the new idea of permanency is gaining ground with the growth of experimental knowledge. Without granting that we have yet reached a position in which we can say definitely that types are absolutely *Read before the American Society of Naturalists, December, 1910. 234 THE GENOTYPES OF MAIZE 235 permanent and do not, at least in some cases, gradually the large accumulation of change into something new, ba ELA teen ae 2: RA Self fertilized during Five Generation Without Self Fertiliza. toh, Each ear in this exhibit represents a different pedigreed family. In e neh family the variation w 1 the ear chosen for the exhibit was The two self-fertilized ears under to two distinct strains, the left-hand one being 4 During the last two years this has been to its own modal number, while the right-hand ear has been selected te Welve rows throughout the course of the experimer 236 THE AMERICAN NATURALIST [ Vou. XLV experimental data now available makes it necessary to recognize a clear distinction between the evolutionary changes in types, on the one hand, and the fluctuations within each type, on the other hand. Quite naturally the first experimental evidence of the existence of permanent hereditary types involved only such characteristics as are clearly distinguishable upon inspection. Thus Jordan was able to demonstrate that within the systematic species Draba verna there are in- cluded as many as two-hundred hereditary forms, whose distinguishing characteristics appear unchanged from generation to generation, in such manner that his pedi- grees of these forms were clearly and permanently dis- tinguishable from each other by easily defined morpho- logical features. Such ‘‘petites espèces”? or ‘‘little species’’ (afterwards called by De Vries ‘‘elementary species,” and by Johannsen ‘‘biotypes’’ or ‘‘geno- types’’), have since been observed by Wittrock and his students, and by many others, in a great number of wild species, and they are now quite generally supposed to be of almost universal occurrence. About 1890 N. H. Nilsson made a similar discovery in connection with his breeding of wheat, oats, barley and other grains at Svaléf, Sweden, but his work remained practically unknown to the scientific world until it was brought to light several years ago by De Vries. Nilsson found in these grains elementary species, each with its own morphological characters and its own specific ca- ‘pacity to yield crops of given size or quality under given external conditions. More recently, sharp-eyed taxon- omists have been rapidly raising many of the elemen- tary species of wild plants to the rank of systematic species. It was natural that the earliest genotypes recognized, such as those of Jordan and Nilsson, should have pos- sessed visibly diserete characteristics, and that they should first have become familiar in normally self-fertil- ized plants, among which little confusion is occasioned No. 532] THE GENOTYPES OF MAIZE 237 by the rare crossing of unlike individuals. Great credit is due to Johannsen? for demonstrating that in such self-fertilized plants, types also exist which are not readily distinguishable by simple inspection, but whose occurrence may be completely demonstrated by the refined methods of the mathematician. Not only has Johannsen’s work been so extensive as to justify the conclusions arrived at by him, but various other investi- gators, working with different classes of research ma- _ terial, have shown that the conditions found by Johann- sen in beans and barley are duplicated in many other species and varieties. Perhaps the strongest support in this direction has come from the work of East? with potatoes and that of Jennings* with various microscopic organisms, especially with paramecium. The fact that Draba verna, and many other wild species in which the existence of numerous elementary species has been demonstrated, as well as wheat, oats, barley and beans, are all predominantly self-fertilizing, and that potatoes and paramecium have an asexual reproduction, has led some to the erroneous notion that the discrete- ness, uniformity and permanence of the types which have been discovered among these and other similar organisms, are in some way dependent upon the absence of crossing. It must be admitted that conclusions drawn from self- fertilized and asexual material do not necessarily apply to plants and animals whose successful existence is de- pendent upon repeated crossing. Nevertheless, the con- ception of pure and permanent genotypes in cross-bred material has become familiar simultaneously, owing to the work done in Mendelian heredity ; for homozygous * Johannsen, W., ‘‘Ueber Erblichkeit in Populationen und in reinen Linien,’’ 68 Pp., Jena, 1903. * East, E. M., ‘‘ The transmission of variations in the potato in asexual reproduction,’’ Conn. Exp. Sta. Report 1909-1910, pp. 119-160, 5 “Jennings, H. $., ‘‘ Heredity, variation and evolution in Protozoa— IT. Heredity and variation of size and form in Paramecium, with studies . growth, environmental action and selection,’’? Proc. Amer. Phil. Soc., 47: 393-546, 1908 238 THE AMERICAN NATURALIST [Vou. XLV combinations of the various characteristics of plants and animals ‘‘breed true’’ to those characteristics. Just as the first recognition of permanent differences in pure lines involved easily distinguishable characters, so also these first discoveries of permanent pure-breeding geno- types in cross-bred plants and animals involved easily definable morphological characteristics. The demonstra- tion that in normally pure-bred lines there are distinc- tions more minute than such easily visible features as characterize the elementary forms of Draba and many other species, was an important advance in our analysis of the populations which make up the species of plants and animals. A similar demonstration that populations of cross-breeding plants and animals are composed of fundamentally distinct types, intermingled but not changed by panmixia, and capable of being separated by appropriate means and of being shown to possess the dis- creteness, uniformity and permanence already demon- strated for the genotypes of self-fertilized and clonal races, will add greatly to the importance of the funda- mental conception of permanency of types involved in the work of De Vries and Johannsen. For the study of this problem there is probably no better plant than Indian corn. It is known to exist in a large number of obviously distinct strains or subspecies which cross together with the greatest ease. Many of its characteristics have been proved by different investiga- tors to be Mendelian unit-characters; such, for instance, as the color of the seed-coat, whether red, dark yellow, light yellow, variegated or colorless, the color of the aleurone layer, whether blue, red or white; the color of the endosperm, whether yellow or white; the chemical composition of the endosperm, whether starchy or sugary, the color of the silks and cobs whether red or white, ete. It has become known also, mainly through the excellent work done at the Illinois State Experiment Station, that oil-content and protein-content of the grains, the posl- tion of the ears, the number of ears on the stalk, and No. 532] THE GENOTYPES OF MAIZE 239 several other characters, are capable of accentuation by selection, so that different degrees of these qualities are capable of being made characteristics of particular strains of corn, without there being the least evidence as yet that these last-mentioned qualities bear any relation to the unit-characters with which the student of genetics generally deals. A further point in favor of maize as a subject for the study of genotypes among cross-breeding organisms lies in the fact that its flowers are so arranged that, while self-fertilization is possible, it is naturally almost completely excluded, thus ensuring the same re- lations as are presented by bi-sexual or dicecious plants and animals, while retaining the means of conveniently testing the genotypic nature of each individual by con- trolled self-fertilizations. I think I have demonstrated during the last five years that there are many genotypes of Indian corn which, although they can not always be distinguished by defin- able external characteristics, can be proved to be just as certainly and permanently discrete as the types whose distinguishing | ean be recognized as Mendelian unit-characters. I shall endeavor to show, in what fol- lows, a portion of the evidence which leads me to this conclusion. In 1905 I undertook a rather extensive series of com- parisons between cross-bred and self-fertilized strains of Indian corn for the purpose of discovering the effects of these methods of breeding upon variability, and these investigations have been continued each year since that time. Two phenomena immediately attracted my atten- tion: First, the well-known fact that the children of self- fertilized parents are inferior to those of cross-fertilized parents in height, yield and other characters dependent in any way upon physiological vigor. In every instance this phenomenon was plainly evident in the very first generation after self-fertilization. This decrease in physiological vigor due to self-fertilization has become an 240 THE AMERICAN NATURALIST [ Von. XLV extremely important relation in the study of the geno- types, as will be shown later. The second phenomenon which quickly made itself manifest, was first clearly appreciated in the second gen- eration after the beginning of the experiments; this was the fact that each self-fertilized family possessed mor- phological features which clearly differentiated it from all other families. In most cases the distinguishing characteristics of these families were of such elusive nature that it was impossible to recognize definite unit- characters, and indeed, morphological descriptions of the several pedigrees could often be made only in terms of greater or less intensity of the several qualities ex- hibited. However, the distinctions were real and applied to every member of the particular family. Thus one family might have a very slender, poorly developed male panicle, while another would have more thick and dense branches of the panicle. This difference might be quite small when given in actual measurement but inspection showed that every individual of the one family had the slender, illy developed panicles, while all of the offspring of the other family had the thicker, denser type. Simi- larly, one family might have a slightly broader and darker green leaf than another, and these characteristic differences were likewise uniformly present in all mem- bers of the single families contrasted. No such character as this is capable of being traced through the generations following a cross, in the manner usually pursued by the geneticist, and the matter must be approached by in- direct methods. The important point to be kept in mind here is simply that the self-fertilized families, derived originally from a common stock, do differ by morpho- logical characteristics, and that there comes to be great uniformity in regard to the presence of these character- istics in all the individuals of a given self-fertilized family. This relative uniformity, which is so obvious even to the casual observer, is not sufficient in itself, however, No. 532] THE GENOTYPES OF MAIZE 241 to positively demonstrate the existence of distinct geno- types in maize, because the slight variations which must always be present even in the most uniform progeny, can not be certainly distinguished as genotypic or fluctu- ating simply by inspection. Such demonstration must rest upon a combination of biometric and genetic evi- dence in order to prove acceptable. Most of the differ- entiating characters of my several strains of maize are such that they do not lend themselves readily to bio- metric methods, but the number of rows on the ear is well adapted for such study and several important re- sults have been derived from the consideration of this character. An important proof that the self-fertilized families derived from my common original stock of corn are genotypically distinct, and that they do not owe their different morphological and physiological qualities to fluctuations within a single genotype, was quickly found in the fact that two of these families selected respectively to 12 and 14 rows of grains on the ears, showed a regres- sion of row-number toward different centers instead of toward a common center. The mean of the original popu- lation was slightly above 14 rows. The selection to 14 rows was very near this mean and-the-selection-to12- rows was very near this mean and the selection to 12 rows considerably below it. According to Galton’s well-known aw of ‘‘regression toward mediocrity,” the mean of a family whose parents were selected to 12 rows should have lain somewhat above 12 rows, and that selected to 14 rows should have retained the mean approximately at 14 rows. The actual result in the case of selection to 12 rows was the production of a family having a mean row-number considerably below the number of rows selected, and the Subsequent generations have since shown a close ap- proach to an 8-rowed condition; while the family whose parents were in each generation selected to 14 rows has always had the mean very near to 14 rows. As these families were grown under as nearly uniform conditions as possible, the fact that the 14-rowed family continues 242 THE AMERICAN NATURALIST [Vou. XLV to have its mean row-number at 14 shows that the fall in row-number from 12 to 8 in the other family has been due to internal rather than to external causes. The change in variability in number of rows on the ears has also been studied from year to year. Continued self-fertilization has resulted in a gradual decrease of variability in the number of rows per ear in each of the self-fertilized lines. This is a fluctuating character, and so far as present evidence goes, the number of rows per ear in any strain can not be fixed at a definite num- ber. While it is probable that none of my self-fertilized families has yet reached an absolutely pure-bred con- dition, several of them have become so nearly pure-bred that their various relations can be used to demonstrate that they are approaching purity as a limit. In 1909 two of these nearly pure-bred families were compared with their reciprocal hybrids in the first and second generations, with reference to the variability in number of rows.’ It was found that the average varia- bility in these two self-fertilized families was 9.08 per cent. The variation in number of rows in their F, prog- eny was 9.06 per cent., and in the F, 12.63 per cent. A comparison of these coefficients of variability shows at once that the variation in number of rows in the F, is essentially identical with that in the self-fertilized lines used for the cross. Theoretically this should be so if the strains used were pure genotypes, because in that case all germ-cells in each pure strain were alike, and therefore, when individuals belonging to these two lines were crossed, equal sperms met equal eggs; consequently there should be no variability in their offspring due to germinal differences, but only those due to environment in the widest sense. As the pure-bred families and their F, and F, progenies were grown beside each other dur- ing the same season, they were subjected to as nearly identical environmental influences as can be attained. ë Shull, G. H., ‘‘ Hybridization methods in corn breeding,’’ Am. Breeders’ Magazine, 1: 98-107, 1910. No. 532] THE GENOTYPES OF MAIZE 243 Consequently, when the F, shows the same variability as the pure lines which entered into it we must conclude that there was at least approximate equality among the sperms which came from the one self-fertilized strain, and among the eggs which came from the other. In the F., on the other hand, genotypic differences appear, owing to the segregation of the different characteristics into the different germ-cells, and to this fact may be ascribed the increased variability in the F,. While other characters have not been studied by the same methods that have been used in the investigation of the number of rows on the ears, several features asso- ciated with the physiological vigor of the various pedi- grees have given evidence which appears to me to be strongly corroboratory of the uniformity of the germ- cells produced by plants which have become pure-bred through continued self-fertilization. The smaller size and less vigor of the offspring of self-fertilized plants as compared with those from a normally cross-bred plant were formerly taken to indicate that self-fertilization is injurious, and Darwin’s ‘‘ Effects of Cross and Self-fertil- ization in the Vegetable Kingdom” strongly impressed this point of view. I have been able to demonstrate, however, that this supposedly injurious effect of self- fertilization is only apparent and not real; or at least that if there is such injurious effect, it is relatively in- significant as compared with the increased vigor due to heterozygosis. The most important evidence of this is found in the fact that the continuation of self-fertiliza- tion in any pedigree does not produce a corresponding decrease in vitality and size. The decrease resulting from a second year of self-fertilization is not as great as that from the first year. The third year of self-fertiliza- tion produces still less deterioration, and as this process 18 continued a limit is approached in such manner as to Justify the inference that when complete purity is at- tained no further deterioration is to be expected, thus proving that self-fertilization is not in itself injurious. 244 THE AMERICAN NATURALIST [ Vou. XLV That this is also true of other plants is derivable from Darwin’s own work. This decrease in size and vigor is accompanied by the gradual lessening of variability, and when that state is finally reached in which there is no further decrease in size and vigor, it seems probable that there will be also no further noticeable change in variability. This does not mean, of course, that there will be no variability, for even the most uniform group of plants or animals will of necessity show slight. variations produced by differ- ent conditions of life, food supply and so forth. But present evidence does not warrant the belief that such fluctuations affect in the least the fundamental qualities of the genotype. In 1908 I suggested a hypothesis to explain the appar- ent deterioration attendant upon self-fertilization, by pointing out that in plants, such as maize, which show superiority as a result of cross-fertilization, this superi- ority is of the same nature as that so generally met with in F, hybrids. I assumed that the vigor in such cases is due to the presence of heterozygous elements in the hybrids, and that the degree of vigor is correlated with the number of characters in respect to which the hybrids are heterozygous. I do not believe that this correlation is perfect, of course, but approximate, as it is readily conceivable that even though the general principle should be correct, heterozygosis in some elements may be with- out effect upon vigor, or even depressing. The presence of unpaired genes, or the presence of unlike or unequal paired genes, was assumed to produce the greater func- tional activity upon which larger size and greater effi- ciency depend. This idea has been elaborated by Dr. East and shown to agree with his own extensive experi- ments in self-fertilizing and crossing maize. He sug- gests that this stimulation due to hybridity may be anal- ogous to that of ionization. Mr. A. B. Bruce proposes a slightly different hypothe- * East, E. M., ‘‘The distinction between development and heredity in in-breeding,’? AMER. NAT., 43: 173-181, 1909. No. 532] THE GENOTYPES OF MAIZE 245 sis in which the degree of vigor is assumed to depend upon the number of dominant elements present rather than the number of heterozygous elements. While all of my data thus far are in perfect accord with my own hypothesis, and I know of no instance in which self-fertil- ization of a corn-plant of maximum vigor has not re- sulted in a less vigorous progeny, it is quite possible that I have still insufficient data from which to distin- guish between the results expected under these two hy- potheses. However, for the purpose of the present dis- cussion, it is not necessary to decide which of these two hypotheses (if either) is correct. Both of them are based upon the view that the germ-cells produced by any plant whose vigor has been increased by crossing are not uniform, some possessing positive elements or genes not possessed by others. Several different characters which are more or less dependent upon physiological vigor have been taken into account in my work, each of which gives its own support to the conception upon which both of these hypotheses are based. The number of rows of grains on the ears which has been most extensively used as a measure of variability, and as a guide in selection, is found to be somewhat affected by the vigor of the individual, and it is due to this fact, no doubt, that the row-number is a fluctuating character, even in the pure genotype. An- other characteristic which has been used as a measure of vigor has been the yield of corn computed in bushels per acre.“ A third characteristic, which was not taken into account at the beginning of the experiments but which aig given confirmatory data in the later years, is the height of the stalks, a character which was much used by Darwin as a measure of vigor in his study of the effects of cross- and self-fertilization in plants. a = understood that this method of stating yields is seriously thnk Ge ae a st ig peas of a 2o poe noe a ; e each of my igrees has usually oceupie only about one one-hundredth of an acre. However, I believe that this defect is more than offset by the advantage of using a unit of yield with which all readers are familiar. 3 246 THE AMERICAN NATURALIST [ Vou. XLV We may now consider the behavior of these several measures of physiological vigor in relation to the theory that distinct genotypes of maize are gradually segre- gated from their hybrid combinations, by self-fertiliza- tion, and that the degree of vigor is correlated with the degree of heterozygosis. I have kept families selected to given numbers of rows on the ears—one series of families repeatedly self-fertil- ized and another series repeatedly crossed with mixed pollen in such a manner that self-fertilization is pre- cluded by artificial means. It is not practicable to do this crossing with mixed pollen in such a manner as to duplicate the conditions found in an ordinary corn-field for the simple reason that the number of individuals which contribute the pollen must be more greatly re- stricted than is true in the open field. While self-fertil- ization has been entirely prevented, there has been a degree of in-breeding somewhat greater therefore than will occur under non-experimental conditions. This degree of in-breeding is sufficient to slowly eliminate some of the hybrid elements which were originally in my strain of corn and should consequently lead to a gradual deterioration in case my theory of the relation between vigor and hybridity is correct. As a matter of fact, such deterioration has become apparent in the ‘‘cross-bred’’* families, when measured either by height of stalk or yield per acre, though both of these measures show that the deterioration has been slight. It is so slight, indeed, that it is very much exceeded by the fluctuations from season to season, and may only be demonstrated by the application of a correction which approximately elimi- nates this seasonal fluctuation. When we compare this continual slight fall in physiological vigor of the cross- *It should be noted that here and in wlat follows I use the expression ‘‘eross-bred’’ in a special sense, to denote the fact that all self-fertilization has been avoided. The more usual use of the term ‘‘ecross-bred’’ to denote a cross between individuals belonging to distinct strains, I replace in this paper by the expression ‘‘F,,’’ as I ean see no tangible distinction between such a cross, and hybridization in the older, more restricted, and more arbitrary sense. No. 532] THE GENOTYPES OF MAIZE 247 bred families with the changes produced in the self- fertilized families during the same period, there is a strik- ing contrast, for in the latter case there was great de- crease in height and yield in the first year, a consider- ably less decrease in the second year of self-fertilization, still less in the third year, and so on, and while I have evidence that none of my self-fertilized families has yet reached a state of perfect stability, they are at the present time decreasing in regard to both of these measures of vigor somewhat less rapidly under continued self-ferfil- ization than are the families in which self-fertilization has been absolutely precluded. Necessary corollaries of the view that the degree of vigor is dependent on the degree of hybridity, or, in other words, that it is dependent roughly upon the number of heterozygous elements present and not upon any injuri- ous effect of in-breeding per se, are (a) that when two plants in the same self-fertilized family, or within the same genotype, however distantly the chosen individuals may be related, are bred together, there shall be no in- crease of vigor over that shown by self-fertilized plants in the same genotype, since no new hereditary element is introduced by such a cross; (b) that first generation hybrids produced by crossing individuals belonging to two self-fertilized lines, or pure genotypes, will show the highest degree of vigor possible in progenies represent- ing combinations of those two genotypes, because in the first generation every individual will be heterozygous with respect to all of the characters which differentiate the two genotypes to which the chosen parents belong, while in subsequent generations, recombination of these characters will decrease the average number of hetero- zygous genes present in each individual; (c) that crosses between sibs among the first-generation hybrids between two genotypes will yield progenies having the same char- acteristics, the same vigor, and the same degree of hetero- geneity, as will be shown by the progenies of self-fertil- ized plants belonging to the same first-generation family. 248 THE AMERICAN NATURALIST [Vou. XLV All of these propositions have now been tested in a limited way. In 1910 nine different self-fertilized fami- lies were compared with nine crosses between sibs within the same self-fertilized family; ten crosses between sibs in F, families were compared with ten self-fertilizations in the same F, families; seven families were raised as first generation hybrids between individuals belonging to different self-fertilized families; and ten families were grown, in which self-fertilization had been entirely precluded during the past five years. The average height of plants in decimeters, the average number of rows per ear, and the average yield in bushels per acre, in these fifty-five families are given in the MS table: Selfed | ‘Safed F; = Fa |r, x Self |F, X Sibs | cox Av. Height | 19.28 | 20.00 | 25.00 | 23.42 | 23.55 | 23.30 | 22.95 Av. Rows 12.28 | 13.26 14.41 | 13.67 | 13.615 | 13.73 15,18 68.07 44.62 61.42 Av. Yield | 29.04 | 30.17 An examination of this table indicates to me that on the whole iny self-fertilized families are not yet quite pure-bred; for the sib crosses give on the average a slightly greater height, number of rows per ear, and yield per acre than the corresponding self-fertilized fami- lies, as shown by a comparison of the first two columns of the table. The same fact is apparent from a com- parison of the ‘‘F, X self” and ‘‘F, X Sibs’’ columns, except that in this case the heights and number of rows per ear are essentially equal while the yield per acre is significantly higher in the sib-crosses than in the self- fertilized families. An alternative explanation of these slight differences between the results of self-fertilization and of sib-crosses may attribute them to an injurious effect of self-fertilization, but in any event such injurious effect must be exceedingly slight as compared with the stimulating effect of heterozygosis. My practise of choosing for seed the best available ears tends to delay the attainment of complete genotypic purity, and this fact favors the view that whatever advantages the sib- No. 532] THE GENOTYPES OF MAIZE 249 erosses show, are attributable to this lack of purity, rather than to any advantage gained by crossing per se. The columns of the table representing the F, and F, show very plainly the superiority of the former over the latter in regard to both height and yield per acre. The fall in average height from F, to F, from 25 decimeters to 23.4 decimeters and the corresponding fall in yield per acre from 68.07 bushels in the F, to 44.62 bushels per acre in the F, show in a most striking way the economic advantage of using first-generation hybrids for produc- ing the corn crop. A comparison of the F, hybrids with the ‘‘cross-breds’’ shows the average yield of the former to be 6.55 bushels per acre greater than that in the fami- lies in which self-fertilization had been avoided. The relation of these results to the experiences of eco- nomic breeders of corn may now be considered. Perhaps in no other class of plants has the evidence been so strong for the possibility of gradual improvement through con- tinued selection as in corn, and this method has been generally followed. The selections of particular physical and chemical qualities which have been carried on at various experiment stations have produced noteworthy results. Most important instances of this kind are in- volved in the breeding experiences of Hopkins, Smith and other breeders at the Illinois State Experiment Station, which have been already mentioned. Here selections for high oil content, low_ oil content, high protein and low protein, high ears and low ears, and the angle which the ears make with the axis of the plant, as well as selection for increased yields, have all led to the production of strains which possessed the desired qualities to a much higher degree than that in which they existed in the foun- dation stock when the selection began. All of these re- sults may be readily explained on the ground that some hybrid combinations of genotypes have greater capacity for the production of the desired qualities than other com- binations, and that the selection has gradually brought about the segregation of those genotype-combinations 250 THE AMERICAN NATURALIST [ Vou. XLV which had the highest capacity for the production of the desired qualities. At least in regard to yield and not improbably also in regard to the other qualities for which selections were made, the results were dependent, not upon the isolation of pure types possessing the de- sired quality, but upon the securing and maintaining the proper combination of types. I have shown above that segregation takes place in a manner at least similar to, if not identical with, the well-known behavior of Men- delian characters. As a consequence of this, no strain of corn can be maintained at a high value with respect to any quality whose development is correlated with heter- ozygosis, except by continued selection for the particular qualities desired. If in any such specialized strain selec- tions should be made for a few years on the basis of some character independent of the one used in establish- ing the strain, the superior qualities for which it was originally selected would quickly disappear, owing to the breaking up of the efficient combinations which had been segregated ‘and maintained by selection. The principles here presented have very great poten- tial consequence for the practical grower of corn, and possibly for the breeder of many other cross-breeding plants and of animals. Their importance seems not to have been fully appreciated by any one however, until recently, though several breeders appear to have glimpsed the possibilities at one time or another. Thus G. N. Collins,’ of the United States Department of Agriculture, has recently shown that several breeders at different times began experiments to test the value of hybridiza- tion in the production of high-yielding strains of corn. The first attempt of this kind which he has found was that of W. J. Beal’ at the Michigan Agricultural College in 1876. At Professor Beal’s instance several other ex- periment stations undertook to work in co-operation with the Michigan Station in testing the value of hybrids 1m ? Collins, G. N., ‘‘The value of first ~ hybrids in corn,’’ Bull. 191, U. S. Bureau ‘of Plant Industry, 45 pp., ” Beal; W. J., Reports, Michigan Board i p 1876-1881. No. 532] THE GENOTYPES OF MAIZE 251 corn breeding, but only Professor Ingersoll,’ of Purdue University, reported results. Professor Sanborn!” ap- parently performed similar experiments in the late eight- ies at the Maine Agricultural Experiment Station. In 1892 G. W. McCluer'*® reported on a number of crosses made during the preceding two years at the Illinois Agri- cultural Experiment Station, and during the next two years Morrow and Gardner'* published bulletins from the same station, describing the results of a number of crosses. Apparently none of this work led to the subse- quent utilization of hybridization methods in corn breed- ing, as no work along this line appears to have been done between the time when Morrow and Gardner issued their second bulletin in 1893 and the publication of the first report of my work with corn at the Station for Experi- mental Evolution in 1908. The work of Beal, Ingersoll, Sanborn, McCluer, and Morrow and Gardner showed that increased yields from the hybrids, as compared with the strains used for the crosses, are the almost invariable result, though both MeCluer, and Morrow and Gardner found isolated instances in which the hybrids were in- ferior to the parent. strains. Hartley'® has since reported that among a number of crosses made by the United States Department of Agriculture also, some gave poorer yields than the parent strains used for the cross, while others gave superior yields, and reached the conclusion, which I think is justified by my own results, that pro- miscuous crossing is not necessarily advantageous but that certain combinations lead to increased yields while others may prove disadvantageous. Collins'® has “Seventh Annual Report of Purdue University, 1881, p. 87. “Sanborn, J. W., ‘‘Indian corn,’’ Agriculture of Maise 33d Annual Report, Maine Boa ‘a of Agriculture, 1889—90, p T8. McCluer, G. W., ‘‘Corn erossing,’’ Bull. 21, Minois Agr. Exp. Sta., 1892, p. 85. “Morrow, G. E., and Gardner, F. D., Bulletin 25, ah 179-180, and Bulletin 31, pp. 359-360, Ilinois Agr. Exp. Sta., 1893 and 1894. * Hartley, C. P., ‘*Progress in methods of Bei a higher yielding ered of corn,’’ Yearbook, U. S. Dept. Agr., 1909, pp. 309-320, 4 pls. p. cit. 252 THE AMERICAN NATURALIST [ Vou. XLV also reported on sixteen hybrid combinations all but two of which gave increased yields in the F,. From the work of all these men, especially from my own compari- sons between F, and F, hybrids, it has become obvious that the secret of the highest success in corn breeding from an economic point of view lies in finding those strains which will produce the largest yield and then utilizing the first-generation hybrids each year. The point which most interests us on the present occa- sion is not, however, the economic importance of using first generation crosses, but the evidence which appears to me clearly indicate that a normally cross-bred plant like Indian corn harmonizes in its fundamental nature with such normally self-fertilized material as beans, wheat and other grains, and such clonal varieties as pota- toes, paramecium, etc., that the egg-cells and sperm-cells of even the most complex hybrids present a limited num- ber of different types which can be assorted into homo- zygous combinations, and that, therefore, the progressive change resulting from continued selection may be simply explained as the gradual segregation of homozygous types or of the most efficient heterozygous combinations. The fact that yield and perhaps many other qualities attain their highest development in the case of complex hybrids naturally leads to the unconscious selection of heterozygous plants for the next year’s cultures, and the continual breaking up of these complex hybrids in sub- sequent generations gives a result which closely resem- bles fluctuating variation, but which is fundamentally different from it. The genuineness of the gains made by selection in corn might naturally lead to the conclusion that fluctuations are inherited were it not for the abun- dant evidence now available showing that a considerable portion of the variation presented is not fluctuational, but is due to the presence of a mixture of different types which any selection partially segregates. NOTES AND LITERATURE IS THE FEMALE FROG HETEROZYGOUS IN REGARD TO SEX-DETERMINATION ? THE evidence that sex is determined by an internal mechanism in unisexual animals has accumulated rapidly in the last few years. The one outstanding case is that of the frog. That extreme variations in the sex ratio occur in this amphibian has been evident from the early experiments of Born 1881, Pfliiger 1882, and Yung 1883-85. The effects were generally ascribed by the earlier workers to differences in the food of the tadpole. Most recent and more carefully controlled experiments, notably those of Cuénot and of King, have shown beyond doubt that food is not a factor that determines the sex of the tadpole. On the other hand, Richard Hertwig has effected astonishing changes in the sex ratio of the frog by delaying fertilization of the eggs. Over-ripe eggs produce a high percentage of males. This con- clusion has been recently confirmed and extended by a student of Hertwig’s, Sergius Kuschakewitsch.! By delaying fertiliza- tion of the eggs for 89 hours after the first eggs had been laid (which gave 53 per cent. of males) there was produced 100 per cent. of males. The death rate of the larve was so low (from 4 to 6 per cent.) that it could not have seriously affected the results. The following table gives the outcome of — s observations and those of Kuschakewitsch. Author ey Hour | Hours on f Wenn | Hours ‘eave Bonm | Boeri our Se 0 ‘4. Boy ea a | Mh | | 88 R. Hertwig, 1907| 58% | 54% | — | 55% — | — |87% | — | — 49 =i | — | 58% | — |59 Berl re : 48.5 37 a fee SOS a 18 = Kuschakewitsch | 53 da E s E This evidence shows beyond question that the sex ratio is affected by delay in fertilization, and may seem to show even that sex itself is determined by this factor. The evidence will, how- ever, bear closer scrutiny. The frogs, Rana esculenta, were captured while pairing, and were allowed to lay a few eggs in * Hertwig’s Festschrift, 1910. 253 254 THE AMERICAN NATURALIST [ Vou. XLV confinement, when they were separated. After 89 hours the female was killed, the remainder of her eggs placed on glass slides, and fertilized with a decoction of the testes of other (one or more?) males. If many of the eggs soon rotated within their membranes this was taken as a sign of successful fertilization. It will be noted that a different male from that employed for the normal fertilization was necessarily employed, because the original male had presumably lost his power to further fertilize. The employment of different males introduces a possible error into the results, for, if the male is heterozygous for sex determina- tion, it is conceivable, as I have previously pointed out in review- ing Hertwig’s results, that in different individuals the sperm may be differently affected in regard to its fertilization power. At present we have no evidence to show that in male frogs such differences exist, and it seems unlikely that such consistent results as these of Hertwig and of Kuschakewitsch can be explained in this way. An alternative view is, however, possible. If the female is heterozygous for sex production, and in consequence two kinds of eggs are produced, it may be that the female deter- mining eggs are more injured by delay than are those of the other class, the male-determining eggs. It becomes, therefore, impera- tive to know what proportion of eggs were fertilized in these experiments, Unfortunately this critical evidence is omitted from Kuschakewitsch’s paper. He states that the death rate of the tadpoles that emerge is low, but one looks in vain for informa- tion relating to the number of eggs that were fertilized. There- fore until this datum is forthcoming it is not possible to draw any certain conclusions in regard to sex determination from the evidence published by the author. T. H. MORGAN. COLUMBIA UNIVERSITY. THE MUTATION THEORY The publication of the first volume of DeVries’s ‘‘ Mutations- theorie’ in 1901, together with the rediscovery of Mendel’s principles, served to bring about a period of unprecedented activity in the study of the problems connected with variation, heredity and evolution. While the results of this decade of work have probably raised as many questions as they have answered, yet the period has undoubtedly been marked by advances of the first importance, both in methods of investigation and in No. 532] NOTES AND LITERATURE 255 results and the point of view achieved. This stimulus we owe in no small measure to the author of ‘‘Die Mutationstheorie.’’ During this period DeVries himself has continued his activities uninterrupted except by his two visits to America, in which he did much through his letters and the publication of his ‘‘ Species and Varieties’’ and ‘‘Plant Breeding,’’ to familiarize his views to American biologists. However, the actual detailed data upon which his theory was based, remained largely a sealed book except to readers of German. Even those engaged in active work on these subjects frequently failed to acquaint themselves suffi- ciently with ‘‘Die Mutationstheorie’’ before breaking into the field of controversy. Particularly is this true of the second volume, the contents of which have been in large part neglected. Professor Farmer and Mr. Darbishire have therefore per- formed an important service in translating this work into Eng- lish. The first volume of their translation? is the subject of this review. The second volume is promised for April. The work will undoubtedly receive a wide reading by English-speaking biologists, and by others as well. The translation is an excel- lent one, faithful to the German meaning but rendered into idiomatic English. Whatever the degree of one’s familiarity with the German edition, a perusal of the work in English will be found profitable and stimulating. A few remarks regarding the contents of the book itself may not be out of place. In a re-perusal of the work, one is struck with the optimism of its author and with the brilliancy and breadth of his exposition of the views set forth. It is not neces- sary to agree with these views in their entirety in order to appre- ciate these qualities of the book. The analysis of the data amassed by Darwin, in which it is shown that Darwin’s single variations are the same as De Vries’s mutations, seems to the reviewer particularly effective. The conception of elementary Species seems also one which will be of lasting value, having already shed a flood of light on many problems. : Probably the time will soon come when nearly all biologists will be ready to admit that mutation, or the sudden appearance of new forms, has been an important factor at least, in species formation in plants and animals. Admitting this, it remains to be discovered what relation these sudden appearances bear to the gen- 1 DeVries, Hugo, 1909, ‘‘'The Mutation Theory.’’ Translated by Pro- fessor J. B. Farmer and A. D. Darbishire. Volume I. Six colored plates, figs. 119, pp. 582, Chicago, The Open Court Publishing Co. 256 THE AMERICAN NATURALIST [ Vou. XLV eral trends of evolution, which are apparent in so many phyloge- nies. This larger problem, which may not be amenable to direct experimental attack, will probably occupy evolutionists for many years to come. For, granting the facts of mutation, we have only accounted for a micro-evolution, and it has still to be shown that the larger tendencies .can be sufficiently accounted for by the same means, without the intervention of other factors. While the supreme importance of DeVries’s investigations on mutation in Cnothera is fully recognized, his premutation theory has always seemed to the reviewer unsatisfactory as a hypothesis to explain the material basis of these phenomena. The cytological investigations of myself and others on these forms have determined the events of germ cell formation, some of which provide a possible basis for the sudden appearance of new types. They have, moreover, shown that different cyto- logical processes are involved in the origin of different mutants, and in this way have thrown much light on the relationships of some of the mutants to their parent form. It is probable that the whole question of the relation of the mutants to their parent will be found to be much more complex than at present supposed. R. R. GATES. p ee E re Ea SECOND EDITION, NOVEMBER, 1910 AMERICAN MEN OF SCIENCE A BIOGRAPHICAL DIRECTORY EDITED BY J. McKEEN CATTELL this aeaa directory r revision if it isto maintain ea ies earl third ofthe names in the ; apenas nan; and thes he which appeared in the first division have in nearly every case been revised. The — been as There has nO work. Greater strictness has been observed in confining its scope to the natural and — exact sciences, and for this ee a few names included in ea sted see have been omitted. have been exerted to — make the book as shee panko ages eaa are of course omissions, if only i even t t is for the information needed. The t ee a is The his group i laces on t editor’ s object in selecting this ph enc the conditions on which scientifie research dependa and 30 so far ie Seta vicar taneiny sea an appendiz the two statistical studies that have been to the Second Edition. — ` ? This directory should be in the hands of all those who are directly or indirectly interested in mento (1) Men of Science will find it indisponible: es not only the names, adress, asienta rosaria and the like of t their fellow workers, but also an A someens of the reach ook of he ona, in (2) — in science, ev terested in en th Will find much of interest and value to tom (3) Executives i in institutions of learning {4) Editors of Newspapers and periodicals will find it to be one of the works of reference that they 9 as meet oad frequently. | ties poweesicpenyeed and rag paper and bound in buckram with i increased in size by more more than 50 percent. it ie sold at the same pic a the at e Price: Five Dollars, net, THE SCIENCE The American Naturalist A Monthly Journal, established in 1867, Devoted to the Advancement of the Biological Sciences Reference to the Factors with Special ‘actors of Organic Evolution and Heredity CONTENTS OF THE OCTOBER NUMBER —- in Urosalpinx. Dr. HERBERT EUGENE ENRERE ee a te nosperms. CHARLES J. CHAMBERLAIN Nuclear Phenomena of Sexual 1 Reproduction in Angio- sperms, Professor D. M. Shorter Ar Articles and ET ths Sterility, Dr. Max Nets and Literatur: Notes on Tebtheysloey; President ‘Davip STARR JORDAN. The Mammals of Colorado, _ Fotis TeDe A Cocia - CONTENTS OF THE NOVEMBER NUMBER Heredity of Skin py cceena dt in Bey GERTRUDE C. DAVENPORT and DAVENPORT. paia te Honey Boba Bees distinguish Colors? JOHN LOVEL Arithmetic of the "ONTENTS OF THE DEOEMBER NUMBER Heredity of Skin P ion in Man. GERTRUDE C. DAVENPORT an sto amnengearadaceas an CONTENTS OF THE JANUARY NUMBER “A Double Heis Eeg. Se esate peace a _ Notes and Literature : ‘Heredity, Dr. W, J. nik f VOL. XLV, NO. 533 Hp e s THE AMERICAN NATURALIST A MONTHLY JOURNAL Devoted to the Advancement of the Biological Sciences with Special Reference to the Factors of Evolution CONTENTS - The Inheritance = — and Sex in Colias Philodice. ee JOHN H. GER Nucleus and bpi in Heredity. Professor MICHAEL F. GUYER A — Study of the Structure of the Photogenic Organs of Certain Ameri ampyride. F, ALEX. MCDERMOTT and CHAS. G. CRANE -~ Shorter NER and Discussion: A Neglected a on Natural SNES in the English Sparrow. Dr. J. ARTHUR HARRI - Notes and Literature: volando Yule’s isis to the Theory of Statis- PEARL ties. Dr. RAYMOND THE SCIENCE PRESS LANCASTER, PA. GARRISON, N. T. NEW YORK: SUB-STATION 84 MAY, 1911 Page 284 306 319 The American Naturalist MSS. intended for publication and books, etc., Intended for review should be sent to the Editor of THE AMERICAN NATURALIST, Garrison-on-Hudso n, New York. Articles Se ga rch work bearing on the problems of organic evolu- tion are especialiy welcome, and will be given preference in publicatie One hundred el by wg of rs 43+ fran are supplied to authors was of charge. Further reprints will be supplied at cost 3 a sina = gir hong sho m be sent to niia ar ngese hier subscription s four rs a year. Fore e T RO T eaa > THE SCIENCE PRESS Lancaster, Pa. Garrison, N. Y. B NEW YORK : Sub-Station 84 Office at Lancaster, Pa., under the Act of Congress of March 3, 1879. ees P ria a a TENTH EDITION. -ical and Pre-historic Specimens. Books on Natural THE MICROSCOPE, History, Soienoe; rio Voyages, eto. See Tur |% introduction to Microscopic Methods and lio Hister, z Marine Biological Laborato 7 ~ WOODS HOLE, MASS. 3 INVESTIEATION | Facilities for research in Zoology; Reperat aber ANA: THE AMERICAN NATURALIST VoL. XLV May, 1911 No. 533 THE INHERITANCE OF POLYMORPHISM AND SEX IN COLIAS PHILODICE! PROFESSOR JOHN H. GEROULD DARTMOUTH COLLEGE THERE is perhaps no phenomenon of greater general interest to students of organic evolution than polymor- phism, yet, although it is of frequent occurrence in insects, in few cases has it been investigated with long- continued and thorough experiments in breeding. Ento- mologists have usually been content to prove that dif- ferent forms arise from the eggs of a single female, or of similar females of the same species, without reference to the male parent or to the immediate ancestors of the female. The time has come when these interesting phe- noniena, lying at the very doors of those at least who live in the country, demand more serious attention than they have yet received. Colias philodice, the common yellow butterfly of the clover, called sometimes the clouded sulphur or roadside butterfly, is distinctly dimorphic in the female sex, in that the ground color of the wings is either yellow or White, the yellow female in most localities being much the more abundant. As this common species can readily be bred in large numbers, it affords excellent material for Studying the inheritance of dimorphism limited to one sex, "Read before the American Society of Naturalists, December 30, 1910. 257 258 THE AMERICAN NATURALIST [Vor XLV Moreover the color pattern, which is the same in both the typical yellow and the albinic variety of the female, differs in the two sexes to such an extent that they may be distinguished even in flight. The wings of the Fic. 1. Colias philodice. Male. Fic. 2. Yellow female. male (Fig. 1) are marked with a solid black band of nearly uniform width extending along their outer mar- gins, whereas in the female (Fig. 2) the marginal band is wider on the fore wings and usually invaded by spots Fic. 3. White female. of the ground color, but narrower on the hind wings and dusted with scales of yellow or, in the albinic variety, of white. The marginal band on the hind wings of the female in some individuals is absent altogether, the width of the bands in general in both sexes being very variable. Besides these pronounced secondary sexual differences No.533] INHERITANCE IN COLIAS PHILODICE 259 in color and the common occurrence of the albinie female, a few specimens. have been taken of an extremely rare mutation, the melanie male, in which the yellow is re- placed by smoky black, the margin being distinctly paler than the ground color. In geographical distribution this aberration does not seem to follow the general rule laid down by Scudder that melanie forms occur in the southern part of the range of a species, for two of the specimens were from near Montreal, three seen and one captured at Palmyra, N. Y., and one now in the collection of Mr. H. P. Richardson, of Concord, Mass., was taken at Plain- field, Mass. Partial melanism, or a melanistic tendency, often occurs in the female, though complete melanism has been found so far only in the male. This tendency reappears in suc- cessive generations independently of the environment, in certain strains that I have bred, though I think it pos- sible that the action of the surroundings on certain indi- viduals in a plastic condition may turn the germ cells in this direction. I have not yet had an opportunity to test this supposition,.and my observations have been almost entirely directed to the inheritance of the albinic variety of the female. Albinism in the genus Colias is due to the replacement of yellow pigment with another which is white (Fig. 3), all other pigments (black, red, ete.) remaining the same. The white is sometimes tinged with yellow, but there is a Sharp difference between the color of a typical yellow female and that of the albinic form. Albinism is not entirely confined to the females, though among the 900 descendants of white females that I have raised there has been not one white male. White males may be expected in regions where the white female is especially abundant. At Hanover, N. H., the proportion of white females to yellow is, perhaps, roughly five per cent. At Ithaca, N. Y., Professor Macgillivray informs me, the proportion of white females is considerably larger, being perhaps 10 or 15 per cent., and at Milton, Mass., Mr. W. L. W. Field estimates them at 20-25 per cent. In two localities where the white male has been 260 THE AMERICAN NATURALIST [Vot XLV taken, about 50 per cent. of the females are white. These regions are Lava, Sullivan Co., N. Y., reported by Mr. Geo. Franck, and Alstead, N. H., on the authority of Mr. W. L. W. Field, who has seen only one white male, how- ever, during several seasons of field work in that region. Scudder makes the statement that ‘‘In the north this rarely, almost never, occurs in the first brood of the sea- son, and is found much more abundantly in the latest than in the middle brood, the numbers increasing as the season advanees.’’ If this statement is true, it has an important bearing on the inheritance of the white char- acter,? for he means, of course, that the proportions of the white females to vellow in the spring broods is less than in the later broods. My observations at Hanover in 1909, and those of my collaborator Mr. P. W. Whiting in Cambridge, Mass., in 1910, do not bear this out, for we found the white females in both places quite as common in the spring brood as in those of summer or autumn. Edwards likewise states that in the south the white form is not infrequent in the spring brood. The probable rea- son for Sendder’s observation is that the population of the spring brood in the long run may be relatively small, because many of the hibernating caterpillars perish. The chances of finding white females in the field in the spring after a severe winter may therefore be less than during the flight of the more abundant summer broods that have not been affected by disastrous winter weather. One of the most interesting observations that I have made during the past two seasons was the discovery of a wild female Colias philodice of the spring brood closely resembling Colias nastes of Labrador (Fig. 4), with a greenish-yellow field overspread with brownish scales, giving a grayish effect. In the margin brown replaces ‘black. This form of female is not common, and I have seen it nowhere described. It was captured at Hanover, N. H., on June 10, 1909, and produced a brood of 34 butterflies (Fig. 5) of which 19 are males, all of which are yellow, 10 are yellow females, 5 are white females. The yellow and white colors of these offspring are of an * See foot-note, pp. 266, 267. No.533] INHERITANCE IN COLIAS PHILODICE 261 unusually clear hue, and quite unlike the color of the mother. The progeny of this female show that, supposing her to have mated with a pure yellow male not carrying white, as was probably the case, she is a heterozygote for The upper figure represents a spring form of female of Colias tele ri rom Hanover, N. H., resembling C. nastes of Labrador, shown in the lo fig color, potentially white, though modified probably by the effect of cold upon the chrysalis in early spring into a form strikingly like that of the Arctic species, Colias nastes. I hope to ascertain from caterpillars now hiber- nating whether this spring form may be produced at will from larve from a white mother by the action of cold upon the chrysalis. My attention was attracted to the problem of inherit- ance of the white color in this species by certain state- ments in Edwards’s great work on the ‘Butterflies of North America.” He says that the progeny of an albino female are partly albino and partly yellow, or it may be all yellow. ‘‘In one instance,’’ he says, ‘‘I had five butterflies from eges laid by an albino, ii there re- sulted one male and four yellow females, no albino. 262 THE AMERICAN NATURALIST [ VoL. XLV In another case of four females one was an albino. Mr. Mead has met with similar results, and neither of us have known an albino to be produced from the eggs of a yellow female.’’ These brief notes pointed Fig. 5. The spring form of female of Colias philodice (at the top) and her offspring. The two rows at the bottom are yellow females; the third row from >» their nastes-like the bottom are white females. .None of the offspring resemble mother. so clearly to Mendelian inheritance that I resolved to investigate the matter, and I have thus far bred over 900 individuals from white females and from the daughters and sons of white females. In general my results differ from Edwards’s observations at two points: (1) The white females always produce both white and yellow females, provided the family is large enough to represent all the possibilities. The family mentioned by Edwards, four yellow females, no white, was too small to indicate No. 533] INHERITANCE IN COLIAS PHILODICE 263 that the mother would produce no white offspring. (2) Yellow females under certain conditions do produce both yellow and white offspring. Thus the mothers of fami- lies ‘‘e,’’ “f,” “ʻi,” and ‘‘k,’’ 1910, were yellow, produc- ing, respectively: 13 white and 14 yellow daughters; 7 white and 5 yellow; 30 white and 14 yellow; 19 white and 19 yellow. The conditions under which these four yel- low females produced white offspring were alike. The female and the brother with which she mated were in each case the immediate offspring of a white female, the male mate in each case being presumably heterozygous for color, y(w), the yellow female homozygous for that color, yy. 1. [NHERITANCE OF THE WHITE FEMALE Summary of Results My observations, begun in the fall of 1908 and extend- ing through the two following seasons, some of the same stock now being in hibernation, may be summarized as follows: 1. The white female, of which I have tested 13 indi- viduals, is in all cases heterozygous for color, producing when crossed with a pure yellow male (wild, or her own brother), either equal numbers of white and of yellow female offspring, in accordance with Mendelian expecta- tion (stock from Cambridge, Mass., families a, b, ¢, d, 1910), or twice as many yellow females as white (stock from Hanover, N. H., families a and c, 1909). The male offspring of a heterozygous white female are all yellow, though presumably one half are heterozygous, y(w), and one half homozygous dominants, yy. 2. It is evident from these observations that white is dominant in the female, yellow in the male, these being the colors of the respective heterozygotes. This case is comparable to the results obtained by Wood in crossing horned Dorset with hornless Suffolk sheep, the male heterozygote of F, having horns, the ewes being hornless. The horned condition is therefore dominant in the male, while in the female hornlessness is dominant. 264 THE AMERICAN NATURALIST [ Vou. XLV TABLE I WHITE §* [HETEROZYGOUS, W(Y)] X YELLOW ¢ [HOMOzyYGoUS, YY] N f N b f Whi Yell Monier ra owed alee. Madrim Fenaies Wild, white Wild, yellow 1808 3 2 3 y ga 1909 a 39 7 16 Nostee-like formë - 3 "op 19 5 10 Wild, white ef " ee 13 3 7 Total in 1908 and 1909 743 8 178 8 369 9 Mother Father | Brood Males i nl Ae | Wild, whiteA | Wild, yellow | 1910 a 18 6 9 B ab T b 79 22 28 4 t C tt 6“ | c 24 (3 t4 D | 6c 64 | d 30 15 13 Total in 1910 | irga 5188) 5899 3. If a heterozygous white female is crossed with cer- tain other yellow males, her brothers and the sons of the same white female, that are heterozygous for color (as indicated by the fact that when mated with homozygous yellow females both white and yellow offspring result), a larger number of white females than of yellow are produced, though not three white to one yellow according to Mendelian expectation. The oe observed is approximately two to one (viz., 38:22, 13:8, g and h, 1910, and probably the 8:4 of family 2w, 1909). This propor- tion may be explained, as Dr. Castle first suggested to me, by assuming that in these families no homozygous whites occur, through infertility or abortion of the ‘‘white’’ germ cells that would naturally combine with white, so that the offspring consist of: dod 25 per cent. homozygous yellow, yy, and 50 per cent. heterozygous yellow, y(w). 92 25 per cent. homozygous yellow, yy, and 50 per cent. heterozygous white, w(y), no pure homozygous whites of either sex occurring in my es, ‘The sign 8 stands in Tables I-VI for the white female, ? for the yellow female. š This individual is possibly a i apes (yy) 9, modified by cold and mated with a heterozygous y(w) male. It is pee that this brood should be included in Table IV. No. 533] INHERITANCE IN COLIAS PHILODICE TABLE II 265 WHITE : [HETEROZYGOUS, W(Y)] X YELLOW & [HETEROZYGOUS, Y(W)] Mother Fath White Yell ( father ( yellow! = Main Females F Kang 1q3 *1d8 3 l 4 3 3 1q | + 1d38 g 63 38 22 1d | « h 16 13 8 Total 838 3 548 8 339 9 TABLE III WHITE § [HETEROZYGOUS, W(Y)] X YELLOW g [UNTESTED, YY OR TH Mothe ka Mal Fonda | Feon KA Wats | dain | mea | pme, | pomas | ennes vee | | aS 3 | asg 1909, 2w 20 8 | 4 w t 1910, m 9 5 | 3 | o 9 : 3 Pea Total 383 3 1688 1499 Thus here is a case comparable to that of yellow mice, which, as shown by Cuénot, Castle and others, are always al roergons. homozygous yellow mice not being pro- duced; so that the proportions obtained by mating yel- low mice together is 66.6 per cent. yellow, 33.3 per cent. non-yellow. 4. The wild white males which occur rarely are pre- sumably recessive homozygous whites, but none have yet appeared in my crosses of heterozygous yellow males with heterozygous white females (g and h, 1910), and I have not yet had an opportunity to test their possibili- ties in breeding. TABLE IV YELLOW 9 [HoMozyeous, YY] X YELLOW g OT, y(w)] Setomaneaes ———— n rell Tiwa Father Brood Males bac aR i i I ON F ag 1910, e 58 13 14 al? al? f 10 7 5 4d% tlds k 46 19 19 lą% t1d'8 i 27 30 uH oo 14133 6988! 5299 * Tested by crossing with 1d°°9 (yellow). See Table IV. t Tested Z Eo with 1d Ọ (yellow). See Table IV. t See Tabl 266 THE AMERICAN NATURALIST [ Vou. XLV TABLE V YELLOW ? [HOMOZYGOUS, YY] X YELLOW ¢ [HOMOZYGOUS, vr] oth | | Whit Yell ( prioni | ne | pees | pare | Fem ales Fem pel a at, am a | 1909, 27 | 20 Lge eee OUR) Umber OF MEON iusi i sas wes Ve oe hee ee 507 TO -CUMOr OL OMEA ce isis nae 5 bas oa se oe ee eae 412 TABLE VI THE RESULTS OF BREEDING EIGHT DIFFERENT PURE LINES OF STOCK DURING 1908-1910. Brackets enclose designations of individuals used in subsequent breeding Mother Father Brood és Females 88 1908 Wild, white Wild, yellow |_...... 3 2 1909 Ag ye H 1909 a |39[a®g,etal.J|} 7[a%}8] a® 8 ab gf 2w 20 8 atọ Either a®, a”, 2y 20 0 a’6 or af? 1909 B Nastes-like form | Wild, yellow b 19 5 1909 C Wild, white a e 13 3 1910 A at Wild, white teas ee 1910a | 18 [2o] 6 alto a'g e 58 13 al 9 alg f 10 7 1910 B 2 : 4b 3 4b2 8 Wild, white | Wild, yellow | b |79 [tes Z| \22 ib 4b?" 3 4b 3 m 9 1b” 8 is o 9 1910 © Wild, white Wild, yellow c 24 8 1910 D 1 la” Wild, white a ü d “Li de] 15] 14% 1s 148 1d'8 38 1d® 3 ut h 16 13 1d* 9 “ i 27 30 1da% ọ 1a? ż k 46 19 1036 g é l 4 3 Yellow Females 7? 16 [a% 9} 4 15 If the proportion yellow females to white is really greater in = spring brood than in ade would a evident that some of the yellow females of a states, then it spring brood are heterozygous for color. See p. 260. summer and especi ially autumn, as Seu No. 533] INHERITANCE IN COLIAS PHILODICE 267 5. Yellow daughters of a white female are probably all homozygous for yellow.*? When crossed with certain of their brothers, presumably heterozygous for color, these yellow females produce both yellow and white female offspring, sometimes in equal numbers, in accordance with Mendelian expectation (broods e, f, k), but some- times twice as many white as yellow (brood i). 6. Yellow homozygous females, daughters of a white female, when mated with other brothers presumably homozygous produce only yellow offspring (brood 2y, 909) 2. INHERITANCE IN Colias Edusa The numbers of typical orange and of white females of the European Colias edusa obtained by Frohawk (1901) from the eggs of four wild white females (var. helice), viz., 110 white 92 (helice) and 125 orange 92 (edusa) with 302 $3, are in approximation to equality, and point to the conclusion that, in this species also, the white female is heterozygous for color. Harrison and Main (1905) raised from the eggs of a white female (helice) of this species 79 3g, 52 99 helice (white), and 19 99 edusa (orange). The numbers indicate that in this case both the parents were probably heterozy- gous for color. All the male offspring were of the typical orange hue, so it may be assumed that the 25 per cent. of homozygous white males that would be expected from mating two heterozygotes together were aborted, though the numbers indicate that 25 per cent. of the females were homozygous in whiteness. The expectation in the distri- bution of the observed number (71) of females would be * It is of course not impossible that yellow females that are heterozygous for color may exist, and that this may account for the excess of yellow females over white in broods a, b and c, 1909. In order to test this matter and to determine whether, when a pair of yellows throw white, it is the male or the female that carries the white, the male crossed with any yellow Should first be mated with a female known to be of a pure yellow strain. 268 THE AMERICAN NATURALIST [Vo. XLV 173 homozygous white, ww, + 354 heterozygous white, w(o), +17} homozygous orange, 00, — 534} white [ww + w(o)] +172 orange, which accords closely with the actual count, viz., 52 white (helice), 19 orange (edusa). 3. GENERAL OBSERVATIONS ON THE Genus Colias Since the female color pattern is the one that prevails in both sexes when there is no differentiation (e. g., Colias nastes, C. hyale, ete.) I am inclined to the view that in this genus of butterflies at least, as probably in birds, the secondary sexual characters of the male represent a more highly modified, those of the female a more primitive, condition. We may recognize in this country, as in the eastern continent, a natural series of species of the genus Colias, at the beginning of which stands the undifferen- tiated. Arctic Colias nastes of Labrador, Greenland, northern British America and Alaska, with the female color pattern, and a dull greenish yellow ground color suffused with brown, common to both sexes. This ground color, as my brood of Colias philodice, 1909, b, shows, is closely related to white and probably interchangeable with it. Next in the series are the subarctic C. pelidne and C. scudderi, in the males of which the yellow color and black color pattern typical of many species of Colias attain their full development, while all the females are clear white, with faint marginal dark bands. The yellow ground color and the solid black marginal band probably arose by mutation in an undifferentiated nastes-like or white stock, and at once became dominant in the male, while the original colors and color pattern remained dominant in the female. Southward from the range of C. pelidne, in the Cana- dian faunal region, is the closely related C. interior, m which yellow females (var. laurentina) occur, though white females are ‘‘on the whole commoner’’ according to Seudder, and from this region southward extends C. philodice, in which the yellow females generally are far more adundant than the white. Finally, the orange color of C. eurytheme of the central and western states, 1m No. 533] INHERITANCE IN COLIAS PHILODICE 269 which species a most complicated polymorphism occurs, probably represents a stage in evolution beyond the yel- low, as does also the black of the melanic male mutant of Colias philodice. The view that the color and color pattern of the male butterfly diverge more widely from the typical colora- tion of the group to which the species belongs, than those of the female, though advocated by Darwin, 1871, was strenuously opposed by Scudder (’89, Vol. 1, p. 531), who cites the white female of Colias philodice as evi- dence to support his position. The case of Argynnis diana, in which the dark blue female differs much more widely from the usual tawny color of the fritillaries than does the male, certainly points strongly to Seudder’s view, but it may well be that no one rule applies to all genera of butterflies, though there are in butterflies and in birds few if any exceptions to the law that the plumage of the male is more brilliantly colored and more highly differentiated than that of the female. 4. INHERITANCE IN Papilio memnon Jacobson’s observations on the Javan butterfly Papilio memnon, in which there are three varieties of female, and the discussion of them by de Meijere, 1910, show that, as in Colias, the dominant form among the females is the one most unlike the male, viz., the brownish, tailed Achates; the form that is recessive in the female, as in Colias also, is the one most like the male, viz., the dark tailless Laocoon. The intermediate variety, Agenor, is heterozygous, epistatic to Laocoon but hypostatic to Achates. In the male the dark color, recessive in the female, is completely dominant. Inspection of Jacobson’s results leads one to believe that two, or probably three, pairs of unit characters are involved, and that not all of the individuals recognized as Achates or as Agenor are of the same gametic consti- tution. The remarkable fact brought out by Jacobson 1s that, in the various combinations made, only two of the three varieties of female were obtained in any one brood. As a working hypothesis, I regard the dominant female 270 THE AMERICAN NATURALIST [Vou. XLV form (the brownish tailed Achates) as the original type, from which the tailless dark-colored male and the some- what similar Laocoon have been derived by mutation, in the same way that the white color, dominant in the female but recessive in the male of Colias, may be pos- tulated as the ancestral color in that genus. 5. INHERITANCE oF SEX Discussion of the inheritance of sex in Colias philodice at present must deal in part with unverified hypotheses, because I have not yet secured and tested white males nor, if they exist, homozygous white females. Since, however, all other possible combinations have been rea- lized, these may now be reviewed, and tentative predic- tions made as to what progeny may be expected in the future from homozygous white stock in its various com- binations. Let us suppose that the male color pattern and all primary and secondary sexual characters of the male are dependent upon the presence of a ‘‘determiner’’ for which the male individual is a homozygous dominant (xx), while the female individual is heterozygous, one half of the gametes which it produces containing the determiner (x) and one half lacking it (o). Thus the gametic constitution of the female may be represented as xo, that of the male as xx. Taking color into consideration, the nature of the pure yellow male may be represented by the symbol: yyXX, that of the pure yellow female as yyox. Furthermore, if yellow is dominant in the male, and white in the female the male heterozygote would be y(w) xx, while the white female would have the symbol w(y) ox. Such a white female, being heterozygous in both color and sex, may be further assumed to produce in equal numbers gametes of four kinds. This hypothesis will appear perhaps more firmly grounded if we imagine that both of the mitoses which give rise to the polar bodies are dif- ferential divisions, instead of one being an equation division and one a differential division, as is usually as- No. 533] INHERITANCE IN COLIAS PHILODICE 271 sumed or demonstrated to be the case. The eggs of the white female are, accordingly, to be represented as fol- lows: yx, yo, wx, wo; those of the yellow female: yx and yo. It is not necessary to assume in gametogenesis of the heterozygous white female of Colias any repulsion between one determiner and another resulting in a coup- ling such as is believed to occur in Abraxas. The deter- miner for yellow and that for white have equal chances of passing into a gamete with the male determiner or into one without it. There are, of course, nine imaginable sets of combina- tions that would take place in the fertilization of the eges of a species with three sorts of females: yyox, w(y)ox and wwox by the sperms of the males: yyxx, y(w)xx and wwxx. We will consider first the combina- tions that up to the present time actually have been made in my cultures. 1. THE PURE YELLOW FEMALE X THE PURE YELLOW MALE ITOR X A IYI R yo = gametes of the female yx, yx= gametes of the male XX, yyox= 50 per cent. pure yellow gg, 50 per cent. pure yellow 99. (Brood 2y, 1909.) 2. THE PURE YELLOW FEMALE X HETEROZYGOUS YELLOW MALE yyox X y(w)xx YX, yo= gametes of the female yx, wx = gametes of the male YYXX, yyox, y(w)xx, w(y)ox 25 per cent. pure yellow gg, 25 per cent. pure yellow 99, 25 per cent. heterozygous yellow dd, 25 per cent. heterozygous white 99, all the males being yellow, and the females yellow and white in equal num- bers. (Broods e, f, k, i 3. WHITE HETEROZYGOUS FEMALE X PURE YELLOW MALE w(y)ox x yyxx a S wx, yo hee of the female yx, cae of the m y (w)xx, w(y)ox, yyxx, yyox dS 50 per cent. pure yellow, 50 per cent. heterozygous yellow, $F 50 per cent. pure yellow, 50 per cent. heterozygous white, all the male being yellow, the females ee and white in equal numbers, (Broods a, e, 1909; a-d, 191 272 THE AMERICAN NATURALIST [ Vou. XLV 4. WHITE HETEROZYGOUS FEMALE X HETEROZYGOUS YELLOW MALE w(y)ox X y(w) xx Wx, Wo, yo= gametes of the female yx, wx= gametes of the male y(w)xx, w(y)ox, yyxx, yyox y (w) xx hig wox. Assuming that the last two combinations (homozygous wiles} are cancelled, we should hav gg 25 "= cent. pure yellow, 50 per Sink heterozygous yellow, 9 25 per cent. pure carat 50 per cent. heterozygous white. This aki tion has also been accomplished in my cultures, e. g., broods g and h, 1910. The five possible remaining combinations may never be completely realized owing to partial or complete infer- tility of the homozygous white stock. However, white males do occur, and assuming that homozygous white zygotes might be successfully produced, the resulting combinations would be as follows: 5. PURE YELLOW FEMALE X Homozygous WHITE MALE yyox X WWXX yx, yo= gametes of the female wx, wx= gametes of the male y(w)xx, w(y)ox, that is, both males and females would be hetero- zygous for color, all the males being yellow, all the females white. 6. HETEROZYGOUS WHITE FEMALE X Homozycous WHITE MALE w(y)ox X wwxx WX, WO, yx, yo= gametes of the female wx, wx = gametes of the male WWXX, WWOX, y(W)xx, w(y)ox, giving dd 50 per cent. heterozygous yellow, 50 per cent. homozygous white, 99 50 per cent. heterozygous white, 50 per cent. homozygous white, thus all the females would be white, but the males yellow and white in equal numbers. 7. Homozycous WHITE FEMALE X PURE YELLOW MALE WWOX X yyxx wx, wo= gametes of the loran yx, yx= gametes of the Aion w(y)ox, or the males all heterozygous yellow, the females all erozygous white [the same result as in (5) No. 533] INHERITANCE IN COLIAS PHILODICE 273 8. HOMOZYGOUS WHITE FEMALE X HETEROZYGOUS YELLOW MALE wwox X y(w)xx wx, wo= gametes of the female yx, wx= gametes of the male y(w)xx, w(y)ox, Wwxx, wwox, or g 50 per cent. yellow, heterozygous, 50 per cent. pure white, 99 50 per cent. white, heterozygous, 50 per cent. pure white. 9. Homozycous WHITE FEMALE X Homozygous WHITE MALE WWOX X wwxx wx, wo= gametes of the female wx, wx= gametes of the male “WWXX, ‘WWox, or ‘the males all homozygous white, the females all homozygous white. That the germ cells in the white female, which I have shown to be heterozygous for color, and which is pre- sumably also heterozygous for the sex determiner, are really segregated in oogenesis into four distinct groups is strongly indicated by the realization of the results of this hypothesis as shown in §§3 and 4. In this segre- gation there is no real ‘‘coupling,’’ the sex determiner (x) being equally distributed among the white and the yellow gametes, but the chances are also equal that any gamete may receive the x factor, and become a male zygote when fertilized, or lack it, and become on fertil- ization a female organism. As would be expected, there are similarities between Colias and Abrazxas* in the method of inheritance of the white female variety in each. The female in both is heterozygous for sex, producing in equal numbers eggs which give rise to males and to females when fertilized by the like sperms of the homozygous male. But there are striking differences between the two forms in inherit- ance, e. g., the dominance of the type color in Abraxas, compared with its dominance in the male only in Colias, white being dominant in the female; females of the type orm that are homozygous for color are found in Colias, but not in Abraxas, in which all the type females are heterozygous, just as are all the white females of Colias that have hitherto been bred. The segregation of the * Doncaster, L., 1908, Rept. Evol. Committee Roy. Soc., IV, p. 53. 274 THE AMERICAN NATURALIST [Vou. XLV color and sex determiners in the grossulariata female and the white female Colias, both of which are hetero- zygous in these two respects, takes place presumably by quite different methods. Other differences or similarities will doubtless come to light when the white male of Colias is bred. | The notation which I have here used to express the gametic constitution of Colias applies equally well to Abraxas, assuming that maleness is dominant and that in gametogenesis of the heterozygote for color and sex, viz., the female glossulariata, GLOX, the male determiner, X, accompanies into one gamete the determiner for high color, G; while the determiner for the undeveloped color, L, is coupled with that for the undeveloped (female) sex, viz., O. This seems to me to be a more plausible way of expressing the combinations demanded by the results than that there is a ‘‘repulsion’’ between the determiner for femaleness (which is assumed in this view of the case to be dominant) and that for the dominant strong color, G, as suggested by Bateson and Punnett. On the other hand, it is true that their assumption that in Abraxas the male is a homozygous recessive may be applied equally well to Colias. However, I am constrained to adopt the view that the male in both is a homozygous dominant for the following reason: Dominance in the male postulates the presence in all the sperms and in half the eggs of a chemical substance which in double quantity in an oosperm so stimulates it that the male characters, both primary and secondary, one by one make their appearance; while in single quan- tity (introduced by the sperm only) a lesser stimulus 1s given, and the organism develops in lesser degree along different lines into the female form. This hypothesis carries within itself an ‘‘explanation,’’ feeble though it be, of the male form and color pattern, as well as of those of the female. It is in harmony with the fact that the intenser color of the male butterfly or moth, gener- ally, represents a more advanced condition in the evolu- tion of pigment than the paler colors of the female. If, on the other hand, following the interpretation of No. 533] INHERITANCE IN COLIAS PHILODICE 275 Doneaster’s results given by Bateson and Punnett, 1908, and by Castle, 1909, viz., that the male is recessive and the sperms contain no sex determiner, which is presumed to be present in half of the eggs only, then we must imagine that a single quantity of this determiner raises one oosperm to the female condition, while, in the entire absence of it, it is understood that another oosperm pro- ceeds to the development of the frequently more complex organs and generally brighter colors of the male. In using a modification of the convenient notation for sex-limited characters devised by Wilson and modified by Castle to express the parallelism between recent dis- coveries in cytology and Mendelian segregation, I do not wish to imply that the symbol X, as applied to Colias, refers to any sort of chromosome. Nor is there, so far as I know, any cytological evidence as to the dominance or recessiveness of the homozygous male condition in the possibly large class of cases like Abraxas and Colias in which the female is presumably heterozygous for the sex determiner. As Castle, 1909, has shown, there are two categories of cases in sex inheritance: viz., (A) those in which the female is assumed to be a homozygous dominant for the sex-determining factor (XX), while the male is a hetero- zygote, producing two sorts of spermatozoa that are not only physiologically but presumably even morpholog- ically different. This category is illustrated cytologically by the extreme case of Anasa, in which one set of sperms, the male-producing, contain only four chromosomes each, while the other, the female-producing, have five, the num- ber characteristic of all the eggs. The second class of cases (B) is that including Abraxas and Colias, in which the peculiarities of their inheritance can be explained by assuming that the female is heterozygous for the differ- ential sex factor, producing two types of eggs, one des- tined, when fertilized by the sperm of the homozygous male to produce only males, the other only females. Furthermore Castle, following Bateson and Punnett, 1908, regards maleness as recessive, the oosperm contain- 276 THE AMERICAN NATURALIST [ Vou. XLV ing only one sex factor, viz., that brought in by the sper- matozoon. The field represented by class A has naturally been well explored by cytologists, for in spermatogenesis the odd chromosome was discovered, and there it is expected; moreover the study of spermatogenesis is attended with less difficulty than oogenesis. Hence comparatively few observers have paid any attention to the behavior of the chromosomes in the maturation of the egg, and cyto- logical evidence of the occurrence of possibly dimorphic eggs in the second class of cases is lacking, though Bal- zer’s 1908 observations on oogenesis in the sea-urchin, mentioned by Wilson, 1909 b, indicate that something may be done along this line. The cytological evidence bearing upon the Lepidoptera, so far as it goes, however, indicates that the male is morphologically homozygous. There is no dimorphism of spermatozoa, the same number of chromosomes being found in all the spermatids. There is, however, a hetero- chromosome, interpreted by the various observers as a pair of equal idiochromosomes, associated with the plas- mosome in the growth period. According to Dederer, 1907, and Cook, 1910, it ultimately becomes indistinguish- able from the other chromosomes, though in the butterfly and the moth examined by Stevens, 1906, its large size made it visible through the maturation mitoses, in both of which it divides into equal parts. Thus, in the seven moths and one butterfly (Euvanessa antiopa) examined by these observers, there is cytological evidence, if the chromosome theory of sex determination be assumed, that the male is homozygous. Unfortunately we have no exact information, so far as I am aware, as to oogene- sis in butterflies. If it should be shown that in Lepidop- tera there is a visible dimorphism of ova as regards the number of chromosomes, the cytological interpretation of sex determination would receive an interesting and important confirmation. : If such visible dimorphism should be discovered, it would be most interesting to see what bearing it has, if any, upon the question whether the homozygous male 1s No. 533] INHERITANCE IN COLIAS PHILODICE 277 dominant or recessive. If the latter be indicated, then we may find that a suitable designation of the gametes of Abraxas and Colias would be that suggested by Wilson, 1909 b, viz., for the male YY and for the female XY, Y being the small synaptic mate of X, which is the large odd ‘‘female-producing’’ chromosome. On the other hand, if the male is dominant a state of affairs that is exactly the reverse might be expected, viz., an absence of a chromosome, or an abnormally small one, in half of the eggs would be the visible sign of future femaleness. If these conditions should be realized, we might be able to identify the ‘‘equal idiochromosomes’”’ already found in the spermatogenesis of butterflies with my XX of the male, the corresponding chromosome in the male-producing type of egg being X, the female- producing ova either lacking the chromosome altogether or having one of reduced size. Dr. Castle, in a recent letter to me, expressed the opinion that the well-known anabolic tendency of the female, especially in reproductive activities, renders it extremely probable, on the other hand, that the female- producing gamete in every case of disparity should have the larger chromatic equipment. This seems to me very plausible, and it may well be that the findings of cytology in reference to this question can never do more than demonstrate the presence of this constant anabolic tend- ency in the female-producing gametes. The appearance of the large X chromosome in the female-producing gamete of the Hemiptera may be, therefore, only the visible expression of a sex tendency already established, as Morgan’s observations on the cytology of Phylloxera indicate. But the demonstration of this anabolic tendency, even In the unfertilized gamete, does not mean necessarily the presence of a sex determiner that is absent or deficient in the male-producing gamete, and hence the dominance of femaleness. It is just as reasonable to assume that the constant katabolic tendency of the male, evinced possibly by deficiency in chromatin at the start and certainly by the presence of horns, high colors and elaborate plumage 218 THE AMERICAN NATURALIST [ Vou. XLV in adult life in many animals, is due to the excess of some hormone in a gamete which thereby becomes male-pro- ducing, in other words, to a dominance of maleness. On the other hand, in the absence of such an excitant, the recessive condition of femaleness would result, with a constant tendency towards quiescence, towards the accu- mulation of reserves of food to nourish the offspring, and the absence in the adult of the brilliant colors, horns and all the well-known and highly specialized secondary sexual characters of the male. If it should be proved that maleness is dominant in lepidoptera in which the female is sexually heterozygous, may it not be true, on the other hand, that femaleness is dominant in the forms in which the male is hererorygo ni for sex, as in Castle’s class A? I see no inconsistency in these two antithetic categories, but should expect to find in the latter either that the female, and not the male, is the more variable, active and progressive, as in the bee, or that, as in hemiptera, both sexes are in external appearance and in habits much alike. In brief, I have tried to point out in this discussion that a different interpretation from that of Castle may be applied to the case of Abraxas, and of Colias also, viz., that these cases, and others that may fall into the same category, differ from those of the well-established class A of Castle in that one is the exact reverse of the other, the female in class A being a homozygous dominant for the sex determiner, whereas in class B the male is a homo- zygous dominant, and not a homozygous recessive as has hitherto been assumed. The view here set forth not only accounts for the facts of Mendelian inheritance in these two insects equally as well as the other, but has the added advantage of harmonizing with the facts regard- ing the secondary sexual characters in lepidoptera and birds.” The high colors and elaborate plumage of the * The recent experiments of Goodale, however, described in the Biological Bulletin, Vol. 20, No. 1, December, 1910, show that the removal of the ovaries from the Rouen duck produces a gradual tendency toward the as- sumption of the male plumage which is not in accordance with the view that No.533] INHERITANCE IN COLIAS PHILODICE 279 male are dominant characters eventually produced in the adult, according to my view, by the presence in the oosperm of a double quantity of a male-producing enzyme or similar substance. This hypothesis does not depend upon cytology for its support, though it is not impossible that future discoveries in oogenesis may be found to be in harmony with it. 6. DIMORPHS If complete separation of the yellow- and the white- bearing gametes should fail to occur in the oogenesis of the white female of Colias, in the differential division of an oocyte destined after fertilization to become a female individual, then the right wings of the future butterfly might be white, the left yellow, or vice versa. Such an individual, captured by Mr. J. H. Rogers, Jr., of Med- ford, Mass., is figured in Psyche, Vol. X, Pl. X, Fig. 4. A similar specimen of Colias edusa, the right wings being white, is figured by Fitch, 1878, in the Entomol- ogist (No. 178, pp. 49-61). Fitch shows also a female with the fore wings white and the hind wings yellow. A gynandromorph might be produced by similar failure in the separation of a gamete containing the sex de- terminer from one lacking it. Various combinations of color and sex are theoretically possible in one individ- ual, if we assume that imperfect division of the gametes may occur in gametogenesis. The discovery of these combinations in nature, or their production by artificial disturbance of the ova, is well within the limits of possi- bility. The production of a dimorph with one side yellow and one white is easily explained if we assume, for example, that the determiners for yellowness and for whiteness, after synapsis, reside in a single bivalent chromosome, which fails to divide differentially in oogenesis, but passes over bodily into one of the gametes, the egg. If the male in birds owes his more brilliant plumage to the addition of some- thing to the female type. 280 THE AMERICAN NATURALIST [ Vou. XLV the first cleavage completes the differential division of the bivalent chromosome, instead of dividing it length- wise, the right and left dimorphism is easily understood. Or we might postulate the suppression altogether of the differential oogenetic division of the egg of a white heterozygous, w(y), female of C. philodice which nor- mally results in the separation of color potentials, but it is questionable whether under such conditions the egg would develop. Again, the theory of Boveri, 1902, that a gynandro- morph is produced if a spermatozoon (sperm nucleus) unites with one of the two nuclei in the two-cell stage, instead of with the original egg nucleus; or that of Morgan, 1907, that two sperms enter, one uniting with the egg nucleus and (in the bee) determining the female half, while the other gives rise to the male half, may be applied to these dimorphs. According to Boveri’s view, for example, we have to assume in the case of Colias that a ‘‘white’’ sperm from a heterozygous yellow male enters a ‘‘yellow’’ egg containing no sex determiner, and after awaiting the precocious division of the egg nucleus, unites with one of the two nuclei thus produced, and determines the character of the white, or hybrid, half of the resulting female organism. 7. Precocity OF THE MALES Males of Colias philodice, as in certain other lepi- doptera, not only appear in the fields earlier than the females in the spring, summer, and autumn broods, but also, in every family of this species that I have raised, a very large proportion of males emerge from the chrysalis early in the period during which eclosion takes place. Thus, as shown by Table VII, in brood a, 1909, 28 males emerged from the chrysalis at the beginning of the period of eclosion, while only 3 females emerged during the same time, and, of the first half of the brood to pupate, 26 proved to be males and only 5 females. In general, 82 per cent. of the first half of the four broods for which data are here presented to reach the pupal No.533] INHERITANCE IN COLIAS PHILODICE 281 stage were males, only 18 per cent. females. The re- maining individuals of these four families, constituting the second half of each in reaching the pupal stage, were, on the other hand, largely females (66 per cent.), only 34 per cent. being males. These facts led me to entertain the idea that the eggs which are to become males may be laid before the female- producing ova. To test this hypothesis, I segregated the successive batches of eggs laid by seven females in 1910, and reared the larve of each successive batch separately, to see if the lots laid first by each female would contain a larger proportion of males than those laid later. It will be seen from Table VII that in fam- ilies b, c, d and e there was in each case a slightly larger proportion of male eggs in the first laying than in the batches laid subsequently, but in families g, i and k ex- actly the reverse is true, the last lots of eggs laid by each female (viz., 3g and 4g, 2i, and 3k and 4k) containing more males than females. It is evident, therefore, that the male-producing ova are not laid on the average earlier than those that are female-producing, but that the larval period of the male is shorter than that of the female. In consequencè of this fact it is not surprising to find that when a brood of caterpillars is exposed to any ad- verse conditions such as starvation, an excess of male butterflies, as Mrs. Treat long ago found, will result, for the simple reason that many females, exposed to adverse conditions during a longer period of growth than that of the males, have been eliminated, while the more pre- cocious male caterpillars survive in greater numbers. This will explain, I believe, the excess of males in my cultures, 507, or 55 per cent. of the total number being males, 412, or 45 per cent. being females. There is no evidence, however, of any differential death rate be- tween the yellow and the white females. Neither is more precocious in larval development than the other, and tae diseases appear to strike each with equal viru- ence, 282 THE AMERICAN NATURALIST [ Von. XLV TABLE VII PRECOCITY OF MALES IN Colias philodice | Firat Half of E Second Halfto| Brood to Pupate| Dates of Pu- | Pupate Dates of Pu-| First Eclo- Year | Brood | ———————_ pation Pele pation sion ss |99 $s | 99 : July 2-9 1909| a | 26 | 5 |June me 13 | 18 Te 20 ]| gå July 1 98 3 190 b 14 3 | July 2-4 5 12 |July 4- 1909 c 10 1 uly 7-10 3 9 |July 11-13 1910 a 13 4 |July9-13| 5 | 11 July 13-19 eae Total 633 $139 9 268 8 509 9| cs eg gts per cent. of first half of all Thirty-four per cent. of the second broods in reaching the pupal stage are| half of broods in reaching the pupal ales stage are males Year Brood | Males | nas Per Cent. of Males | Brood | Males 1910 | 1b | 33 | 12 |73fromistbatch| Ig | 16 | 13 | 55 2b 7 3 | 70from 2d batch 2g 18 23 44 7 3b 9 56 from 3d batch | 3g 21 19 52.5 4b 30 28 |51.7from4thbatch 4g 8 5 61.5 le 20 13 | 60.6 li 13 22 2c 4 3 |57 2i 14 22 38.8 ld 12 10 |54.5 1k 8 11 42 2d 7 6 |53.8 2k 15 12 55 3d li 12 | 47.8 3k 13 10 56.5 4k 10 5 66.6 le 13 3 | 81.2 2e 18 12 | 60 | 3e | 27 12 169 | l Oi a aa These investigations are by no means finished, and any one who should chance to capture any unusual speci- men of this species, or of any closely allied to it, show- ing melanistic or other aberrant tendencies, would con- fer a great favor on the writer of this paper by mailing to him the specimen alive in a metal box lined with moist filter paper sewed firmly against the perforated sides. A white male is, of course, especially desired. _ In conclusion, the writer wishes to express his hearty thanks to his friend Mr. P. W. Whiting, an accomplished student of butterflies, for his kind and efficient coopera- tion in the field work connected with these studies and in the laborious processes of preparing specimens for detailed examination. The friendly counsel of Dr. W. E. Castle has been also of great value to the writer m No. 533] INHERITANCE IN COLIAS PHILODICE 283 entering this to him new but extremely fascinating field of investigation. REFERENCES TO LITERATURE Balzer, F., ’08. Verh. d. deutsch, Zool. Gesell. [Reference from Wilson, 1909. | Bateson, bi and Punnett, R. C., ’08. The Heredity of Sex. Sci., Vol. 27, No, 698, m 785-787, May Ak 1908. Boveri; T? Ueber mehrpolige Mitosen als Mittel zur Analyse des Se Vah phys. med. Gesell. Würtzburg, XXXV. [Ref. from Castle, W. E., ’09. A e View of Sex Heredity. Sci., Vol. 29, No. 740, March 5, pp. 395- sa 10. On a Modified PE Ratio among Yellow Mice. Sci., Vol. 32, No. 833, December 16, 1910, pp. 868-870. Cook, M. H, 10. Spermatogenesis in Lepidoptera. Proc. Acad. Nat. Sci. Phila., April, 1910, pp. 294-327, Pl. XXII-XXVII Darwin, C, "71. Descent of Man. Vol. I, C LOT: Siicruiateginbels in Philosamia cynthia, Biol. Bull., Vol. 13, No. 2, pp. 94-105. Doneaster, L., and Raynor, ’06. Proc. Zool. Soc. London, Vol. I, p. 125. ` Edwards, W. H., ’68—’93. The Butterflies of North America. 3 vols., oston. Frohawk, F. W., ’01. On the Occurrence of Colias edusa and C. hyale in 1900, and the ge of Rearing the Variety helice from helice Ova. Entomologist, Vol. Jacobson, B10. Béo re achtungen über den Polymorphismus von Papilio 57 Meijere, J. C. H. de, ’10. Ueber K IAS bezüglich a Polymorphismus von Papilio memnon L. 9, und über die raiona undärer Geschlechtsmerkmale. Ze stik f. indukt. Abstamm. u. Ver aire Bd. 3, No. 3, pp. 161-181, March, 1910. Morgan, T. H., ’07. periment Zoology. Chap. 27, p. 407. New York. -» 789. The Butterflies of the Masters United States and anada. 3 vols., Cambridge, Mass. Stevens, N. M., ’06. Studies in Spermatogenesis, II. Pp. 33-58, Pl. VIII- XV. Carnegie Inst. Washington Wilson, E. B., Recent Denarni on the Determination and Heredity of Sex. ha Vol 29, No. 732 , January 8, 1909, pp. 53-70. Abraxas, Sci., Vol. 29, No. 748, April 30, 1909, pp. 704-706 Note on the Inheritance of Horns and Face-colour in Sheep. Jour, Agri. Sci., Vol. I, Pt. 3, p. 364. NUCLEUS AND CYTOPLASM IN HEREDITY? PROFESSOR MICHAEL F. GUYER UNIVERSITY OF CINCINNATI Tat there is a physico-chemical basis of heredity and that it is, if not exclusively, at least fundamentally bound up in the proteins of the germ-cells, we know for certain. If there is anything else than this physico-chemical basis we do not know it. But.even should there be, it is incon- ceivable that it is not subject to physico-chemical agen- cies and limitations, and we are bound, therefore, to con- tinue our search for these material factors as long as we can unearth new facts or arrive at new generalizations. Before undertaking a discussion of the germ-cell, how- , ever, I wish to call attention to certain chemical facts that are frequently overlooked or slighted by the biologist. In chemical reactions we have not only to take into account the initial chemical substances and such external factors as pressure, temperature, etc., but in many in- stances we must reckon also with the quantitative rela- tions, especially the concentrations of the various sub- stances, and the velocities of their reactions, since altera- tions in either of these factors may profoundly modify the end-products of the reactions. A very simple ex- ample of quantitative relations is seen in the combination of carbon and oxygen. If much oxygen is present, CO, is formed, if little CO, and these are two very different substances, particularly when physiologically considered. Or, when chlorine acts upon methane, CH,, depending upon purely quantitative relations and physical condi- tions, any one of four different substitution products ranging from CH,Cl to CCl, may be secured. The questions of quantitative proportions and of veloci- ties are of especially great significance in a sequence of * Read before the American Society of Naturalists, at Ithaca, December 30, 1910. 284 No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 285 reactions where a number of associated substances are concerned and where certain of the materials, before they can become active, must await the outcome of the reac- tions between other members of the mixture. For ex- ample, when hydrochloric acid is passed into an alcoholic solution of hydrocyanic acid, provided there are suff- cient quantities of these three ingredients, five definite crystalline end-products of their interaction result. First the hydrochloric and hydrocyanic acids combine to form amido-formyl-chloride, which then adds another mole- cule of hydrocyanic acid. This product next. reacts with one molecule of aleohol which to this point has been inac- tive. The result is what we may call end-product one. Then end-products two, three and four, respectively, are formed by the successive additions of a single molecule of alcohol to separate molecules of a part of the immedi- ately preceding end-products. The fifth end-product is not so directly related to the others. It is elaborated chiefly through the interaction of hydrocyanie acid and water, but this interaction can not take place until water is released through dehydration of some of the other compounds. There is little doubt that such sequential reactions as these may be taken as simple models of what goes on on a tremendous scale in the developing germ- cell, It is a well-known chemical fact, moreover, that when two or more progressive reactions are going on simul- taneously, a quickening or retardation of the velocity of either, with the consequent precocious development of certain stages in the sequence, may lead to a partial or complete deflection of the original trend of the reactions and the formation of entirely different end-products than would otherwise have resulted. And velocities may be varied greatly by such factors as temperature and cata- lytic agents. But what has all this to do with the germ-cell? Simply this, the substance of the germ-cell, in so far as we know it, is of materials such as proteins, carbohydrates and fats which we have no reason for doubting are subject 286 THE AMERICAN NATURALIST [Von XLV to the same fundamental laws of chemical behavior whether they exist in living matter or in non-living matter. If in the comparatively simple cases of associ- ated simultaneous reactions with which we are acquainted in non-living matter, relative velocities may so modify the results, we can readily realize of what tremendous importance regulation of this matter must become in living protoplasm where doubtless vast numbers of chemical reactions and interactions are going on at the same time. In fact, could we locate such a time regula- ting factor in the germ-cell it would seem that we had accomplished a long stride toward an understanding of the controling and coordinating mechanism which in- sures the appearance of just the proper substance at the right time in morphogenesis. It would constitute a quali- tative as well as a quantitative regulator, for by deter- mining quantity at any given time it determines what the next chemical reaction will be,and hence in the very doing of this, it necessarily conditions the chemical outcome of that reaction. As we have seen, temperature and cata- lytic agents are important factors in modifying the velo- cities of reactions in ordinary chemical processes, and inasmuch as under normal conditions of development the temperature factor is a fairly constant one, we are left to face the question as to whether in protoplasmic phe- nomena there is anything to correspond to catalyzers. Such substances we find in the enzymes. While the method of enzyme activity is not positively known, the consensus of opinion of those who have studied them most seems to be that they act by catalysis. For instance, both catalyzers and enzymes are effective in very minute quantities; neither appears among the end-products of the substances acted upon, but exists 1m- dependently and in exactly the same quantity as at the beginning of the reaction; external conditions such as temperature affect their activities similarly; and lastly, the rate, that is, the velocity of the reaction concerned, depends upon the amount of the catalyzer or enzyme present. When we have explained the phenomena of No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 287 catalysis, therefore, we have at the same time doubtless gone far toward explaining the action of enzymes. We know that different ferments act differently on the same substance and that the same ferments may act on different substances within certain limits. To realize the truth of the first proposition we have only to compare the results of the butyric, the lactic and the alcoholic fer- mentations of grape sugar. As to examples of the same ferment acting on different substances, we may point to the fact that some varieties of yeast will act readily on d-fructose, d-glucose, and d-mannose. They will not act on d-galactose, however. Furthermore, none of the other known aldose hexoses and ketose hexoses are acted upon by yeasts. In the case of yeast, then, where a given enzyme acts on more than one substance, the molecular configuration of the respective substances must be closely similar. This seems to be a general rule. We do not find the proteolytic enzyme trypsin attacking anything but proteids, although it operates on different kinds of proteids. Even oxidizing ferments are not exceptions in this respect, for certain of them will yield oxidations in some compounds and not in others that are readily oxidizable under the influence of a different oxidizing ferment. But granted that in living protoplasm ferments play the important réle of velocity regulators and consequently of conditioners of both quantitative and qualitative results, where should we look for them in the germ-cell? It is now a matter of common knowledge that probably many ferments are closely associated with nuclear activity and presumably originate within the nucleus. The present tendency is to regard the dissolution of the nuclear mem- brane from time to time as a means of distributing sub- stances to the cytoplasm. Particularly in the case of the germinal vesicle of the egg, upon dissolution of the mem- brane, there is a copious discharge of nuclear material into the cytoplasm, and one would naturally infer that this 1S In Some way a preparation for the subsequent rapid differentiation which will occur. 288 THE AMERICAN NATURALIST [Von XEY Various observers have pointed out the predominant part played by the nucleus in intra-cellular oxidations, operating apparently by means of oxidases. R. S. Lillie has shown that in the indophenol reaction the colored oxidation products in such cells as red corpuscles, and those of liver and kidney, are deposited mainly in and around the nuclei. He further points out? that certain ferments exhibit the properties of nucleoproteids and that they are apparently concerned with later chemical. changes in the protoplasm chiefly oxidative in nature. As far back as 1895 Wilson and Mathews? showed that in the first maturation division of the starfish egg much chromatin is set free in the cytoplasm. In 1902 Conklin* called particular attention to the escape of nuclear material into the cytoplasm upon dissolution of the nuclear membrane in the egg of Crepidula, remarking further upon the large proportion of chromatin that passes into the cytoplasm during every cell cycle, where seemingly it plays some important part in the subsequent changes of the latter. Likewise, F. R. Lillie,® in 1906, pointed: out that an important part in the development of Chaetopterus is played apparently by the great quan- tities of a ‘‘residual substance’’ set free from the ger- minal vesicle. Lyon® in 1904 showed a rhythmic parallel between nuclear division and the production of carbon dioxide by the cleaving egg. And Mathews’ in 1907 sug- gested as probable that the periodic disappearance of the nuclear membrane during mitosis brought about a distribution through the cytoplasm of oxidases which had been synthesized in the nucleus. Wieman in 1910 has shown the existence of alternate phases of acidity and basicity in the process of yolk formation in Leptinotarsa, due to a succession of oxidation processes which occur in * Jour. Exp. Zool., Vol. V, pp. 379-428, 1908. * Jour. Morph., Vol. X, pp. 319-342, 1895. * Jour, Acad. Nat. Sci., Phila., Vol. XII, pp. 1-121, 1902. * Jour. Exp. Zool., Vol. III, pp. 163-268, 1906. * Am. Jour. Physiol., Vol. XI, pp. 52-58, 1904. 7T Am. Jour. Physiol., Vol. XVIII, pp. 89-111, 1907. s Jour. Morph., Vol. XXI, pp. 135-216, 1910. No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 289 the basic-staining food-stream as the result apparently of the discharge of oxidase from the nucleus into the cytoplasm of the egg. Again, it is of great significance that in the embryos of seeds the time of greatest fer- mentative activity in starches and other fermentable bodies coincides with that of maximum size of the nuclei. Many other significant facts might be adduced, but I wish merely to show that there is abudant evidence pointing to the nuclei of cells as sources of enzymes. The idea that among other things the nucleus is con- cerned with enzymic activities in the cell, or, indeed, that the chromosomes themselves are sources of ferments, is by no means a new one. The last few years has seen a steadily increasing tendency to regard them as such. The latest and most outspoken suggestion of this nature, of which I am aware, is the argument that Montgomery,’ brings forward in a recent paper. He says in part, ‘‘The relative constancy of chromatin mass in spermatocytes and spermatids of very different volumes speak strongly for its enzyme nature.’’ Then after pointing out the rel- ative constancy in size between the univalent components of spermatogonia and spermatocytes in Euschistis and reminding us of the well-known fact that, although the egg is many times greater than the sperm, the chromo- somal contribution of each is the same in size and mass, he goes on to say that, ‘‘An enzyme possesses among other properties the power of engendering changes in its medium while still preserving a constant mass.’? And he continues, ‘‘Of all the larger cellular compounds that we know, the chromosomes agree most closely with this defi- nition, and by reason of this constancy of mass alone might be considered enzyme masses.”’ My present thesis, however, while in harmony with all this, is yet different. It is rather just the complement of such a proposition as Montgomery’s for it is an attempt to show reasons why there must be a nicely adjusted Series of such substances in the germ-cells as enzymes. In any epigenetic conception of the germ-cell—and this * Biol. Bul., Vol. XIX, pp. 1-17, 1910. 290 THE AMERICAN NATURALIST (Vou. XLV in greater or less degree seems to be the only plausible one to-day—we are forced, in explaining morphogenesis to postulate the existence of some time-, quantity- and quality-controlling mechanism. The one evident class of substances in the germ-cells which can fulfil the neces- sities of the case are the ferments. For since they will determine the velocities of chemical reactions they must in consequence control the quantitative relations of the cell chemistry at any given unit of time. But from the very fact that where a large number of associated re- actions are going on simultaneously, these quantitative relations at given stages of the chemical interchanges must profoundly influence qualitative results, we can not but conclude that this initial control of velocities must condition the qualitative results. If we regard the chromosomes as centers of such a series of velocity-controllers, or, in other words, as sources of various enzymes, we can at once appreciate the necessity for having them so accurately balanced off in size and particularly in their quantitative relations one to another. For since the velocity of the reaction in a fermentable substance is determined not only by the presence of the ferment, but also by the amount of it, the quantitative relations of the ferments to one another would have to be very accurately maintained. What appears to be in a way a non-chromosomal demonstration of this fact is found in connection with the chloroplasts of plant cells which seem to exercise their functions at least in part through the agency of ferments. As is well known, in cell division these bodies are each carefully divided and handed on to the daughter cells so that a constancy in number and in general re- lationships is maintained. But, it may be objected, what is to be done with those cases of nuclear division in which the mitotic divisions of the germ-cells have been preceded by a series of ami- totic divisions? Wieman?’ has shown that in amitosis the appearance of the division figures is by no means the 1 Loc. cit. No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 291 same in all cases and that where it occurs among germ- cells the mechanism is more carefully adjusted than elsewhere. To quote his own words regarding such divisions in Leptinotarsa, ‘‘Thus in the odgonia and spermatogonia, division of the nucleus is preceded by a very exact division of a large chromatin nucleolus, and as the halves separate surrounded by a clear area, the appearance reminds one very much of the division of a chromosome on a spindle.’’ Other investigators of such amitosis depict in their figures a mechanism which may have the same significance. From preparations of my own showing amitosis in the testes of snails and of Planaria, a similar interpretation could be given. On this enzyme conception, however, constancy in number is not the fundamental necessity. The require- ment is really constancy of equilibrium between chromo- somal constituents. It matters not whether this is maintained in sixteen, eight, four or one chromosome, so long as the balance between the various enzymic foci, or the capacity for the restoration of such a balance, is maintained. The ultimate karyokinetic divisions of such germ-cells as earlier divide amitotically would seem to be the restitution of such a balance so that the proper quantitative conditions exist in the finished germ-cell. Wieman would account for the appearance of amitosis in early gerf-cells on the ground of a reduced oxygen Supply in each individual cell, consequent upon a very rapid increase in cell multiplication. All that is de- manded in the enzymic conception which I am present- ing is the preservation in some way of the general ten- sion of equilibrium so that each enzymic focus can resume its customary activities when the occasion de- mands, or, to express it less teleologically, when the oc- casion permits. This conception would seem all the more cali; since we have had to discard the idea of the continuance of actual chromosomal individuality in favor of that of their genetic continuity as expressed by Wilson. The 292 THE AMERICAN NATURALIST [ Vou. XLV demonstration by Bonnevie,'! that while the identity of the old chromosome is lost in the resting nucleus, never- theless, each new chromosome arises by a kind of endo- genous formation from within the substance of its im- mediate predecessor, is a good point in evidence. That mere number of chromosomes is not of fundamental importance is evidenced by the considerable number of known cases in which closely related species may be characterized by a considerable difference in the number of chromosomes. In my own researches on man and certain birds, I have shown that instead of eight, the expected number of chromosomes in spermatocytes of the second order, only four (disregarding the accessory which may be present) appear, but that they are ap- parently bivalent in nature. As associated with embryonic development we should have to suppose that there are considerable numbers of these initial ferments, which, however, need not all be present in an active condition. Certain ones required for the first stages of development might well be sup- posed in the course of their activities to produce or free others, or activate them at the proper time to take up their part in the progressive chemical activities of de- velopment. It is probable, too, that many of the fer- ments of the fully developed organism peculiar to the special tissues have not existed as such in the germ-cell at all, but have arisen at a later stage in the cells they occupy as the outcome of the metabolic activities of the tissue cell itself. It is a current belief, indeed, that each kind of cell has its own specific ferments whereby it shapes up from the common food supply submitted to it in the lymph the substances necessary for its own intra- molecular assimilation. Many intracellular enzymes are now known to exist and it is probable that proteo- lytic enzymes at least are found within the cells of all living tissues. This is demonstrated by the fact of autolysis, or the self-digestion of living tissues which = Arch. Zellforsch., Bd. I, pp. 450-514, 1908. No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 293 have been taken from the body under perfectly aseptic conditions and kept suitably warm and moist. There is no obstacle in the way of supposing, further- more, that if we regard ferments as of nuclear origin, the cytoplasm of a given tissue may not modify the fer- ment, as it itself takes on the necessary modifications for its own specific functions. We have good evidence that the production of ferments can be modified by even the substratum on which living organisms grow, and such a relation as this, close as it is, is certainly less inti- mate than that existing between nucleus and cytoplasm. For example, molds cultivated upon starch form dia- stase, but if provided with albumin they will produce in- stead a proteolytic ferment. Moreover, by gradually altering their other nutriments, yeasts can be made to utilize after a time various foreign compounds. But granted the necessity of some such set of con- trollers as the enzymes, and locating them in the chromo- somes of the germ-cells, does this not commit us to a rigidly chromosomal theory of heredity? By no means. If, as all evidence indicates, ferments operate as cata- lyzers, then we must not forget that it is the very gen- eral belief among chemists that catalytic agents do not initiate the chemical reactions with which we find them associated, but that they only tremendously accelerate such reactions, or in a few known instances retard them. Since the nature of the building material must determine fundamentally the nature of the thing built, we must look outside the enzymes for much that will determine — the peculiar individual outcome of the developmental processes. Leaving out of consideration for the present other functions the chromosomes may subserve, we might regard them as a sort of gauge for the feeding out of enzymes at the proper rate to bring about proper velocity reactions in the other cellular constituents, and Perhaps regard the whole matter of mitosis and exact- hess in chromosomal distribution as a mechanism by which a quantitative metabolic regulation is maintained. But because chromosomal influences can regulate the 294 THE AMERICAN NATURALIST [Vou. XLV activity of other cellular constituents, there is no war- rant for jumping to the conclusion that they are essen- tially more important than these other constituents. I may repeat in this connection what I have had occasion to say by way of reminder in a former paper, ‘‘A germ- cell in fact should need no special units to generate the peculiar genre equilibrium or idiosynerasy of protoplasm which is distinctive of a particular kind of individual, since such a germ-cell not only is itself already an indi- vidual, but from the very fact of having had the same racial history as other individuals of its peculiar kind (be they germ-cell, embryo or adult) it must likewise as a whole already possess this distinctive idiosyncrasy.’’ That is, the individual proteids of germ-cells—globulins, albumins, nucleoproteids and the like—bear from the very start the stamp of individual peculiarity, wherever they may reside in the cell. And since they constitute at least part of the materials which transform and inter- act and have their actions modified by enzymes, certainly they as much as the enzymes are responsible for the out- come. = Regarding the specificity of corresponding proteins in relation to the natural kinships of living organisms, some very interesting facts are brought to light in the recent voluminous and painstaking researches of Reichert and Brown.'!2 They show, for instance, that in hemoglobin, one of the few crystallizable proteins, the erystals of each species of any genus, while possessing a constant individuality, all belong to the same crystallo- graphic system and generally to the same crystallo- graphic group of the system. These authors further point out the fact that this isomorphism must signify in all probability correspondence in the fundamental chem- ical constitution and molecular configuration of respect- ive hemoglobins. In case of the individual species the difference in the characters of the crystals was found to be as great as with ordinary chemical salts or minerals that belong to an isomorphous group. One is seemingly 2 Univ. Cincinnati Studies, September—October, pp. 1-19, 1909. No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 295 justified, therefore, in inferring from these results that the differences between corresponding proteins in dif- ferent species of a genus are rather to be referred to differences in molecular configuration than in atomic composition of the molecule. The case of hemoglobin is only one of several lines of evidence that might be brought forward as indicating the specificity of proteins. The serum-albumins which constitute the very font of the living molecules of higher animals, differ very decidedly in different species in the readiness with which they crystallize. Or, a foreign serum-albumin injected into the veins of an animal of different species can not take the place of the correspond- ing albumin of the blood of that species, but acts rather as a poison and is quickly eliminated by the kidneys. Lastly, not unduly to multiply examples of protein specificity, may be cited the precipitins which as you know may in general be used to show the degree of re- lationship of allied forms. For instance, when the blood-serum of one species of animal, let us say man, is injected at intervals into some other species, e. g., the rabbit, the serum of the latter acquires the property of producing a precipitate in the serum of the first species, man in this case, but not in the serum of other animals unless they are relatively closely related to the first species. Thus the serum taken from rabbit’s blood after a series of treatments with human blood will produce precipitation in the blood from any human being. It will produce some, though less, precipitation in the blood of the anthropoid apes, still less in monkeys, and none at all in animals distantly related to man. This implies, manifestly, that the more akin forms are, the more nearly identical are their proteins. And from the evi- dence brought forward in connection with the hemo- globins we have seen that we are perhaps justified in re- garding the differences between the proteins of closely allied forms as ones of molecular configuration rather than of molecular composition or constitution. The question may arise in some minds as to whether 296 THE AMERICAN NATURALIST [Vou. XLV there could be sufficient number of configurational dif- ferences in the corresponding protein molecules of dif- ferent species to account for the specificity of the respect- ive proteins. When, however, we consider that to the serum-albumin molecule alone—and it is by no means the most complex protein—estimates assign the capabil- ity of having as many as ten thousand million stereoi- somers, there would seem to be in this factor of configu- ration alone ample possibilities for the necessities of the case. Because of imperfect methods it has in the past been well nigh impossible to tell how nearly chemically iden- tical corresponding proteins of different species are. Reichert and Brown'* point out that what formerly passed current as difference in composition may have been due in reality to contaminations or mixtures. ‘‘ For instance,’’ they go on to say, ‘‘the fact that the egg- white of the egg of certain species remains perfectly clear upon boiling, while that of other species becomes opaque, might be taken as meaning a difference in chem- ical composition, but the difference has been shown to lie in the different amounts of alkali and saline present.’’ Again, ‘‘The centesimal analysis of corresponding al- bumins and globulins have failed to show any positive differences. Oppenheimer states, from the results of a recent study of serum-albumins of man, the horse and the ox, that serum-albumin is a uniform and specific © substance, and that the elementary analyses point to one serum-albumin.’’ This would leave the matter of specificity to be explained solely on the basis of molecu- lar configuration. This brings up the whole question of protein consti- tution and configuration. While this is still pretty much a terra incognita still many interesting facts have come to light, and all of them point to the conclusion that we are in no wise compelled to regard the proteins as out- ***The Differentiation and Specificity of Corresponding Proteins and other Vital Substances in Relation to Biological Classification and Organic Evolution: The Crystallography of Hemoglobins,’’ Publication No. 116, _ Carnegie Institution of Washington, pp. 1-338, 100 plates, 1909. No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 297 side the pale of the well-known principles of constitution, polymerization, stereometry and the like which are known to obtain in simpler organic compounds such as the ‘‘ring’’ compounds or aromatic series, and the straight chain or aliphatic series. It is a familiar fact that the native proteins are read- ily broken down through hydrolysis into simpler bodies which still possess protein characteristics. We may in fact either artificially or in normal digestion get a whole series of stages ranging from complex native proteins to simpler and simpler products. The sequence runs approximately as follows: Proteins. —Meta- or infra-proteins. —Proteoses. —Peptones. —~Polypeptids (a relatively small number of amino-acids linked together). —Individual amino-acids. Three fourths of the albumin molecule, for example, may be made to yield members of the large group of amino-acids. Because of the great abundance of these bodies, and because of their universal presence as degradation prod- ucts of proteins, the conclusion was reached that the protein molecule is essentially built up by a linking to- gether of amino-acid molecules. This suggested the idea that by bringing about such linkage it might be possible to build up molecules of the protein type. To those who are familiar with the recent developments of physiolog- ical chemistry, the fact that the first steps toward this end have already been accomplished is well known. Some seven or eight of the amino-acids (leucin, tyrosin, glycocoll, alanin, aspartic acid, phenyl-alanin, and amido-valerianic acid) had already been produced synthetically before Emil Fischer began his work. Fischer and his pupils have synthesized over twenty new members. But what is still more significant, they 298 THE AMERICAN NATURALIST [ Vou. XLV succeeded in securing linkages of certain ones of these, thus producing polymeric amino-acid compounds called by Fischer polypeptids. Bodies of this same type have been isolated from natural organic substances. These polypeptids resemble peptones in appearance and, moreover, they react in the same way peptones do toward enzymes and various test reagents. One of the artificially synthesized polypeptids, furthermore, is ap- parently identical with one of the known polypeptids found in digestion, and /-leucyl-triglycyl-l-tyrosin, when prepared artificially, seems to have all the properties of the albumoses. The amino-acids possess both acid and basic proper- ties. It is this amphoteric condition that renders link- age possible. The individual amino-acids which consti- tute the units in such polymerizations are frequently spoken of as ‘‘nuclei.’’ Linkage has been obtained not only between similar ‘‘nuclei,’’ but also between ‘‘nuclei’’ of different amino-acids. The results point clearly to the conclusion that the peptones and higher proteins are huge molecules formed chiefly of amino-acid molecules linked together by NH and CO affinities left unsatisfied as a result of processes comparable to dehydration. Such a protein molecule may perhaps be represented as a main chain or ring, of which the respective links are amino-acid ‘‘nuclei.”’ Glycocoll, NH,CH,COOH, for instance, would through dehydration have for its nucleus in such a chain —NH.CH,.CO—. Furthermore, since one H of the CH, of such ‘‘nuclei’’ (e. g., —NH.CH.CO—) i can be substituted by various compounds (acetic acid, buthane, methylparaoxybenzene, ete.) we are led to con- clude that to each link of the protein chain, a side-chain, differing in constitution in different cases, is attached or is attachable by replacement of this hydrogen atom. The well-known instability of living protein would seem No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 299 to be due to the fact that the chemical systems in such a giant molecule are never fully saturated at any one time, so that there is continually an adding and detaching and shifting of side-chains with perhaps at times more funda- mental shifts or replacements in the amino-acid ‘‘nuclei’”’ themselves. Quantitative and qualitative differences of proteins would seem to depend fundamentally on the kind and amount of the constituent amino-acids and second- arily on the chemical nature of the various side-chains. Probably the scheme as outlined is much simpler than the true conditions in the protein molecule, but it will serve as a sort of diagram of the relations which exist there. It is probable, too, that the conditions in differ- ent proteins vary greatly in complexity. The chief point to be emphasized is the fact that the results of many in- vestigators bear out this general conception of the pro- tein molecule. It would seem then that in the light of our knowledge of the complex molecular configuration of the proteins, the substances which appear to be the most intimately concerned with life phenomena, we have, without resort- ing to the idea of mysterious separate entities, ample basis for that peculiar handing on of metabolic energies already established which we term heredity. The mech- anism of heredity would seem to be not so much a local problem of nucleus or cytoplasm as of (1) fundamental species substances, probably mainly protein in nature, together with (2) equally specific enzymic substances which regulate the sequences of the various chemical and physical processes incident to development. As develop- ment progresses, more and more kinds of chemical prod- ucts are released and in consequence an increasing num- ber of chemical reactions are set going. After the germ becomes multicellular such new factors must be reckoned with as the influences, mechanical, chemical, etc., of the various parts of the body on one another. An even with our present meager knowledge of hormones we can see that this may be no inconsiderable factor in modifying the developing organs in complex organisms. 300 THE AMERICAN NATURALIST [ Vou. XLV Looked at this way, the physical basis of heredity could not be. considered a series of equipotent units, but rather it must be regarded as being composed of systems of units of different orders of organization and different degrees of coordination. Alterations in the configura- tion, constitution or relative positions of the unit con- stituents which represent the links of the main protein chain or ring, for instance, would precipitate much deeper-seated changes than would replacement of side- chains by those of different type, and such replacements would, in turn, doubtless appear objectively as differ- ences of greater degree than those resulting from shifts in the composition or configuration of the individual side-chains. Our whole scheme of natural classification, in fact, demands just such a physical basis as is depicted for the structure of the protein molecule. For morphological characters are not all equivalent. In any large group certain characters are more conservative than others and represent more fully the organization, as a whole, while in successive subsidiary groups the characters grade down to less and less inclusive ones until the trivial fea- tures which make up species differences and varietal traits are reached. However, this parallel between the make-up of the protein molecule and the natural classifi- cation of living organisms can be looked on only as a suggestive illustration because in addition to proteins other things often enter into the construction of what we term characters in plants and animals. These characters, indeed, are frequently blends of the effects of numerous influences. But as an example of how changes in different parts of the protein molecule might work out visibly in the organism let us see how such alterations actually work out in simpler and better known compounds. In the familiar benzene ring compounds, for example, there exists (1) the main framework or more stable com- ponent, the so-called ring itself, and (2) innumerable substitution groups which can be attached to the ‘‘ring”’ No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 301 at any one of six places. These substitution groups can be shifted or replaced indefinitely without disrupting the ring itself. Still different effects are obtainable by the union of one or more rings (generally still retaining some of the substitution groups) directly with one another, or through the intermediacy of a third element or radical. In the numerous coal-tar colors, the color, which is one of the most obvious ‘‘characters,’’ does not lie as such in the benzene ring itself, but is determined by the rad- icals attached to the ring, and in certain groups perhaps, in part by the manner in which two or more rings are united. If, for instance, into a molecule of azobenzene a radical of the amido-group (NH,) is introduced, a body is constructed which through salt formation yields a dye. If instead of the amido-group a hydroxyl group is introduced, the result is likewise a dye but one of dif- ferent color. And so a large series of tints may be pro- duced by varying the substitution groups which replace H in the principal molecule. We have constantly increasing evidence, finding re- cent expression, for instance, in a paper of Morgan,"* of a fundamental stereometrical condition of the egg-plasm. And F. R. Lillie has suggested the possibility that a specific polarity and symmetry are characteristic of the ground substance common to all cells of the organism. Stereochemistry is based on the assumption that the combining forces of an atom act in certain definite direc- tions in space. This same conception of orientation must be carried on to the more complex organic units, the stereometrical relations of which, in turn, are but the continued expression, under other conditions, of the original atomic combining forces. And this being true, it seems reasonable to look upon the whole organism as but the further expression of such elemental factors. In view of the facts regarding the closeness of identity of corresponding proteins of nearly related species, we must conclude that between two individuals of the same * Loe. cit. ; = Jour. Exp. Zool., Vol. IX, pp. 534-655, 1910. 302 THE AMERICAN NATURALIST [ Vou. XLV species, set apart by mere differences of sex and minor traits, the basal protoplasmic stereometry and the funda- mental proteid constitution must be in large measure identical, so that bi-parental inheritance, if extending to all the details it has been assumed to embrace, would be largely a matter of duplicating identical protoplasmic constituents. It is an obvious fact, however, that the egg contributes vastly more cytoplasm than the sper- matozoon, and in consequence the developing organism is more maternal than paternal in origin. I have argued this point at some length in a former paper,'® where I at- tempted to show that we are not justified in asserting that the entire quota of characters which go to make up a complete living organism are inherited from each parent equally, but that rather we must restrict our assertion of equal inheritance to the sexual and specific differences which top off, as it were, the more fundamental organ- ismal features. I further pointed out that since the actual manifest physical things contributed equally by each parent were the chromosomes, we might legitimately look to them as the chief source of the factors which determine individual differences. We know that a single reduced or haploid set of chromosomes is sufficient for normal de- velopment, both from the fact of artificial parthenogene- sis, and the fertilization of non-nucleated egg fragments ; hence the egg must contain all the possibilities of a new organism. But the only measurable things contributed by the sperm-cell are the individual characters of the male line. We may infer then that the chromosomes of both male and female origin work together on or with the other germinal contents of the fertilized egg, and these are pre- dominantly of maternal origin. Or to phrase it as I have in a former paper:'7 ‘‘Nevertheless, we can see how the veneer of individual traits may be equally of maternal and paternal origin if, to express it crudely, we look upon cytoplasm and chromatin, respectively, as responsive * Science, June 28, 1907, pp. 1006-1010. * Univ. Cincinnati Studies, September—October, pp. 1-19, 1909. No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 303 mechanism and inciting agent, the character of the re- sponse depending both upon the constitution of the cyto- plasm and the material (enzymes? nutritive substances?) emanating from the nucleus.’’ If we consider that the supplying of the proper amounts and kinds of ferments is one of the important functions of the chromosomes, then we may suppose that in bi- parental inheritance each set of chromosomes is opera- ting, probably catalytically, on a series of fundamental cell constituents that are largely common to both lines of ancestry; and that slight constitutional or configura- tional differences in corresponding enzymes bring about individual differences such as we recognize in the adult. We have already seen that different ferments within cer- tain limits may act on the same substance and yield dif- ferent results; consequently, in the intrusion into the egg of slightly altered enzymes in the chromosomes of the male, we should expect corresponding structural modifications to result. It is not the intention to imply, however, that all char- acter changes must be chromosomal in origin. Any influ- ence which could effect constitutional or configurational changes in other essential constituents of the germ-cell would doubtless produce corresponding alterations in the adult. It is probable that not only changes of nuclear origin are reflected on to the eytoplasm, but that, con- versely, cytoplasmic alterations may affect the nuclear constituents, for we have already seen how even the sub- stratum may modify the enzyme factors in entire organ- isms such as molds and yeasts. Furthermore, there is no reason apparent why if the differences, no matter how produced, are modifications in the fundamental consti- tution or stereometry of the material affected, they should not persist permanently in the new germ-cells. It would seem, in fact, that in the permanent effects of Such reciprocal influences as here depicted for nucleus and cytoplasm, we might be able to account in large measure for the accumulations which have step by step been grafted on to the primitive protoplasm in its epi- 304 THE AMERICAN NATURALIST [ Vou. XLV genesis toward the complex conditions of to-day, or in other words, in its racial evolution. Moreover, it is con- ceivable upon this basis how in later stages of phylogeny, as new chemical configurations or new chemical substances were developed, some of these could bridge back into relations with more primitively established substances and thus bring about ontogenetic short-cuts in develop- ment, or how, on the other hand, these abridgments - might result in part from alterations in the more primi- tive molecular configurations. Racial reversions would be interpretable, at least in part, on the ground of the suppression of recently added processes or materials rather than on the reassertion of independently existing germs which had become latent. The question arises, does not the very fact of the exact- ness with which the chromosomes are divided show that they are of greater fundamental importance than the cytoplasmic substances? Such a conclusion does not necessarily follow. The cytoplasmic substances of con- cern in development exist in the unfertilized egg appar- ently largely in a neutral or relatively inactive condition. The requisites are that these cytoplasmic substances be of a certain constitution and that there be a certain mini- mum amount of each. As insuring the presence of this indispensable minimal quantity there probably exists more or less of a surplus, but a surplus of this kind would not necessarily alter the result, as on my suppost- tion the necessary quantitative conditions which deter- mine the directing of the chemical reactions in the devel- opmental processes are not brought about in this initial resting substance, but in the products of its fermentation, and the quantity of these at any given unit of time will depend upon the quantity of the ferment. Thus it is evident that by having the series of ferments accurately apportioned as we seemingly do in the chromosome, there can be an adequate quantitative and therefore qualita- tive regulation of the chemical processes without the con- troller being considered of fundamentally shane im- portance than the substances controlled. No. 533] NUCLEUS AND CYTOPLASM IN HEREDITY 305 I do not wish to be understood as maintaining that the nucleus or the chromosomes have no other than enzymic functions. We know that the nucleus contains highly complex proteins and it would seem improbable that all of them are concerned exclusively with matters of fer- mentation. There is some evidence, however, that the ferments themselves may be of the nature of nucleo- proteids. If this is true it is possible that under certain conditions they operate as ferments and under other con- ditions as building or other necessary materials. My chief desire has been, without entering into a discussion of the manifold functions the nucleus may perform, to point out one obviously necessary function, the control of velocities in cell chemistry, that is explicable on the basis of enzymic activities, and to bring forward reasons for inferring that these have their sources in the chromo- somes. In conclusion, then, recalling the fact of the inconceiv- able number of stereoisomers that the proteid molecule may possess, and the fact that to make up protoplasm diverse proteins, at least, and various enzymes probably themselves of proteid nature, are required, we would Seem to possess in the chemistry of relatively known chemical substances in nucleus and cytoplasm an ade- quate basis for interpreting the mechanism of heredity without resorting to other more hypothetical entities. Before we embrace out of the void such new phantoms as “‘psychoids’’ or ‘‘entelechies’’ is it not incumbent upon us to strive still further to expand our knowledge of protein constitution and configuration? In the mean- time it would seem best to look upon the organism, whether germ-cell or adult, as but the expression of an extended cycle of processes which are due to the intrinsic Properties of the chemical constituents of protoplasm. Such an attitude at least has the merit of keeping within hailing distance of tangible facts and processes. A COMPARATIVE STUDY OF THE STRUCTURE OF THE PHOTOGENIC ORGANS OF CERTAIN AMERICAN LAMPYRIDÆ F. ALEX. McDERMOTT anv CHAS. G. CRANE HYGIENIC LABORATORY, U. S. Pusiic HEALTH AND MARINE HOSPITAL SERVICE, WASHINGTON, D. C. Or the great amount of work that has been done in the production of light by living forms, not a little has been devoted to the structure of the photogenic organs. The organs whose anatomy and histology have been the sub- ject of most of the researches have been those of the Lampyridæ. Although perhaps twenty-five species of these widely distributed insects occur within the borders of the United States proper, but little work has been done on the anatomy and histology of their luminous organs. The late Dr. Wm. H. Seaman (1) made some observa- tions on Photinus pyralis, the insect which is so common in the parks in Washington in the summer, and a near relative of the pyralis, Photinus marginellus, has been made the subject of an extensive study by Miss Town- send, at Cornell (2). Wielowiejski (4) mentions having studied two American species, but fails to give their names. With these exceptions, however, the American Lampyride seem to have been neglected in the matter of histologic studies of the photogenic organs. Of the for- eign Lampyride which have been studied, the principal species are Lampyris noctiluca, Phausis splendidula, Phosphenus hemipterus and Luciola italica, all Euro- pean species, and all belonging to different subgroups from each other and from the American insects. Many studies have also been made upon the cucuyo, Pyrophorus noctilucus Linn., the large tropical elaterid firefly. It has been thought worth while, therefore, to attempt some further study of the photogenic organs of such species of Lampyridæ as are accessible here, having 12 view especially the determination of the similarities and differences between them and between them and other No. 533] PHOTOGENIC ORGANS OF LAMPYRIDÆ 307 species which had been studied previously. The two species most common here (Washington, D. C.) are Photinus pyralis Linn. and Photuris pennsylvanica Deg., and the majority of our studies have been made on them. The classification relationships between these in- sects and the others that have been studied may be seen from Oliver’s recent catalogue (3). As already stated, Photinus pyralis had been studied to some extent by Seaman, and its near relative, P. marginellus, by Towns- end, but so far as we have been able to find, no studies have been made on any species of Photuris.! A large number of slides have been made, containing transverse, longitudinal and oblique sections of the two insects above mentioned, and a few transverse sections of Photinus consanguineus. With these slides compara- tive studies of the structure of the photogenic organs have been made. The most essential result of these studies is that in these three species the structure of the photogenic organs is practically identical, and very sim- ilar to that described for some of the other species of Lampyride which have been examined. Many of the drawings given by Townsend of the structures in Pho- tinus marginellus may represent with equal faithfulness the corresponding structures in Photinus pyralis and Photuris pennsylvanica; our slides of Photinus consan- guineus were not entirely satisfactory, but so far as could be seen, the structures in this insect are identical with those in its larger congener, pyralis. In all three insects the luminous organ is divided into two distinct layers, the inner one being white and opaque, and serving as a reflector, and the outer being yel- lowish and translucent, and containing the actual photo- genic mechanism. The photogenic organs, as brought out by prior studies, are penetrated from the interior of the insect outward, by innumerable tracheæ, which ramify and anastomose within the true photogenic tissue, and unite within, above the reflecting layer, to 2 Since this was written, it has been noted that Watasé (9) made a few ; Tvations on the structures in Photuris pennsylvanica, but makes only a brief reference to them. 308 THE AMERICAN NATURALIST [ Vou. XLV 1. Cross-sections at about the middle of the fifth REDEFINE — of keg Photinus ag and (b) Photuris pennsylvanica, I, intestine; L, photo genic tissue; M, muscle fibers; R, reflecting layer; S, spiral reni T, tr racher to se iciaaite ie. form larger trachee; the latter lie nearly flat against the inside surface of this reflecting layer, and run diag- onally outward, finally uniting almost at the spiracle with the breathing tracheæ, with which they are iden- tical in appearance. The spiracles are on the dorsal side of the abdomen, one near either edge of each seg- ment, and are furnished with some valvular arrangement at their orifice; the details of this structure have not yet been clearly made out. The arrangement of the smaller trachee and tracheoles is much the same in all three species. The tracheæ pass through the reflecting layer and the photogenic tissue perpendicularly to the surface. These tracheæ are furnished with chitinous hairs on the No. 533] PHOTOGENIC ORGANS OF LAMPYRIDZ 309 interior as far as the point where they enter the reflect- ing layer; the presence of these hairs in trachee beyond this point and in the fine tracheoles, has not been ob- served. In their passage through the photogenic tissue, the trachee are surrounded by the structure referred to by Miss Townsend as the cylinder, a cylindrical mass of cells, sharply differentiated from those of the surround- - 2. Oblique section near edge of Photinus pyralis. L, photogenic tissue ; R, reflecting layer; T, trachea to photogenic organ; BT, trachea leading to ther organs. (Both of these figures are intended only as outline drawings, and no attempt has been made to show all the internal organs, or any great number of trachee.) ing tissue, through which the trachea passes almost cen- trally. Within this cylinder the trachea throws off the numerous small branches, which at the edge of the cyl- inder break into the very fine tracheoles which pass into the photogenic tissue and anastomose between the cells with tracheoles from adjoining cylinders. The appear- ance of the large tracheæ above the luminous organ are Shown in Fig. 2, drawn from an oblique section, the line of the cut being nearly parallel to the line of the larger tracheæ near the edge of the abdomen. At the lower end, Just next to the superficial chitin covering the luminous segments, the main trachea subdivide into the large num- er of branches whose tracheoles radiate into the photo- genic tissue, usually recurving slightly, so as to penetrate the tissue a short distance from the chitin. The entire system suggests that the air is drawn in through the breathing tracheæ, and forced through the 310 THE AMERICAN NATURALIST [Vou. XLV fine passages in the true photogenic tissue, where the oxygen of the air is consumed in a biologic oxidation. In the sections of pyralis there are clearly seen bundles of muscle fibers on either side of the center line of the insect, which pass completely through the abdomen, almost vertically, and are attached to the exterior chitin at the top and bottom. At about the same ‘point, other muscle fibers pass inward from the point of maxi- mum width at each side; these fibers have not been traced to their full extent, but they appear to pass upward and toward the center near the dorsal side of the insect. These fibers are indicated in Fig. 1, a, at M. No similar muscle fibers have been observed in Photuris, although short lengths of muscle fiber passing vertically through the abdominal cavity have occasionally been noted, and these may be fragments of similar muscles to those in the Photinini. The corresponding muscles of Photinus marginellus are clearly shown in Fig. 1 of Miss Towns- end’s paper. Externally, the lower terminations of these bundles of muscle fibers appear as non-photogenic spots on the ventral surface of the luminous segments. It may be well here to call attention to certain differ- ences between Photuris and Photinus, as shown by the cross-sections of the insects. While there is a general similarity of outline in the cross-sections of the two species, the section of Photuris is generally a little flatter, and the ventral curvature of a somewhat larger mean radius, than in Photinus. Another difference has been very marked in our sections. While the thickness of the reflecting layer is about the same in both species, the laver of true photogenic tissue is much thinner, both actually and in comparison with the reflecting layer, 1m Photuris than in Photinus; this difference is clearly seen by reference to Fig. 1, a and b. This difference may be somewhat significant when considered in connection with the slight differences in the quality of the emitted light, and in the modes of emission of the two species. (See reference No. 8.) In Photinus there are two peculiar organs each consisting apparently of a thick-walled, No. 533] PHOTOGENIC ORGANS OF LAMPYRIDZ 311 chitinous tube, coiled into a nearly cylindrical spiral, represented in partial section by S, S, in Fig. 1, a; these two organs appear to be glands which empty into a com- mon duct which could be followed to the posterior ex- tremity, and it seems possible that they are a portion of the male generative system, as they were not found in the female pyralis, although no spermatozoids were seen. The direction of rotation of the spiral was the re- verse on the left side of the insect from that on the right. These organs were not found in the Photuris, although globular, glandular structures were found in approxi- mately the same portion of the latter insect. This struc- ture is shown in Fig. 2 of Miss Townsend’s paper on Photinus marginellus, and in Fig. 1 of Seaman’s (Photinus pyralis) ; the latter erroneously referred to it as the intestine; in our studies, the intestine of both Photinus and Photuris was seen as a nearly straight, thick-walled tube, indicated in section by I in a and b of Fig. 1. The above remarks apply to the male insects. The two sexes in Photuris are almost indistinguishable exter- nally; all those which we sectioned appeared to be males. In Photinus pyralis, however, the female differs mark- edly from the male. The luminous organ in the male occupies the entire ventral surface of the fifth and sixth segments of the abdomen, and the posterior portion of the fourth segment. In the female, the luminous appa- ratus is visible externally as a small, rectangular yellow Spot, occupying about one third of the ventral area of the fifth segment of the abdomen. This organ obtains its air supply from a large trachea which extends along its forward edge, and apparently connects with the spiracles on the dorsal edges of the segment. In its finer struc- ture, the photogenic organ of the female pyralis appears to be exactly like that of the male, as is to be expected. That the photogenic process is an oxidation is scarcely to be doubted, in view of the work which has been done already. The work of one of us (McD.) with Professor 312 THE AMERICAN NATURALIST [ Vou. XLV Joseph H. Kastle, of the University of Virginia, is of especial interest in this connection (6). Our histologic methods presented no particularly new features. Most of our specimens were killed in hot 70 per cent. alcohol, stained entire in acid carmine, and mounted in paraffin. To secure proper penetration of the stain, it was found necessary to clip off the tip of the abdomen, or to slit the dorsal chitin. Osmic acid prepa- rations were used a number of times, and in the sections of Photinus consanguineus, which were otherwise un- satisfactory, one per cent. osmic acid gave very good results for the fine tracheolar structure. For the study of the tissues under the dissecting microscope a good treatment was found to be to allow the detached, fresh luminous segments to soak in a mixture of equal parts of ten per cent. caustic soda and ten per cent. formaldehyde solution for three or four hours. This treatment left the tissues of both the reflecting and the active layers of the same gross appearance, though without entirely de- stroying the cellular structure; after being treated thus, the trachee and tracheoles can be seen as silvery white tubes and threads, on a background of dull, pale yellow, and may be followed down to the point of anastomosis. It seems possible that the reflecting layer fulfils a two- fold purpose—that of reflecting the light outward, and thus increasing its intensity in the desired direction, and of protecting the insect itself from its own radiations. It has recently been shown by Coblentz (7) that the pyralis and other Lampyride contain a fluorescent ma- terial, and a number of observers have shown that fluorescent materials injected into a living animal show a higher degree of toxicity when the animal subsequently is exposed to light than if it be left in the dark. To conclude: We have found that (a) the structure of the photogenic organs in Photinus pyralis, Photinus consanguineus and Photuris pennsylvanica is practi- cally the same, and very similar to the structures of the corresponding organs in some of the other species of Lampyridx that have been studied; (b) the trachee from No. 533] PHOTOGENIC ORGANS OF LAMPYRIDZ 313 the photogenic organs connect near the breathing spiracle with the tracheæ which supply the other organs, and that they closely resemble the latter trachee in structure; c) the view that the photogenic process is an oxidation is borne out by the structure of the photogenic organs. We wish to express our appreciation of the assistance of Director John F. Anderson, of the hygienic labora- tory, and Dr. Norman Roberts and Mr. Geo. F. Leonard, of that laboratory, and we are indebted to Dr. E. A. Schwarz and Mr. H. S. Barber, of the U. S. National Museum, for their kindness in supplying entomologic information, and to Professor W. A. Kepner, of the Uni- versity of Virginia, for criticism and advice. No attempt will be made here to give a complete list of the references to the literature of even the histology of the luminous tissues; so far as the latter branch of the subject is concerned, it is pretty thoroughly covered by the bibliography given by Miss Townsend, and the most complete bibliography yet published of the whole subject of physiologic light is contained in Mangold’s extensive and interesting review cited as reference No. 5, below. 1. Seaman, The Luminous a of Insects. Proc. Amer. Soc. Micro- scopists, 1891, Vol. 13, pp. . perre The Histology of ii Light Organs of Photinus marginellus. ER. NAT., 1904, Vol. >? aag 127-151. bo w e a: D 5 Fe ae =F P> an’s Genera Insectorum, apaes A - Wielowiejski. E zur ah der Leuchtorgane der Insekten. ool. Anz., 1889, Vol. 12, pp. 00. - Mangold. Die e von fjal. Winterstein’s Handbuch der ver- gleichende Sean Vol. III, 24 Half, pp. 225-392, Jena, 1910. - Kastle and MeDer Some Observations on the Production of Light by the Firefly. pel our of Physiol., 1910, Vol. 27, pp. 122-151. über ei n der Feuerfliege herriihrende fluoreszier- ende Substanz. Physikal. Enue, 1909, Vol. 10, pp. -956. 8. McDermott, A Note on the Light-emission of some American Lampyride. Canad, Entomol., 1910, Vol. 42 2, pp. 357-363. 9. Watasé. On the Physical Basis of Animal an hese Biological Lectures delivered at Wood’s Holl, 1895, pp. 101-118. a or ~q c © EA ® = = z © ct Hi i2 B © SHORTER ARTICLES AND DISCUSSION A NEGLECTED PAPER ON NATURAL SELECTION IN THE ENGLISH SPARROW In referring to Professor Bumpus’s paper, ‘‘The Elimination of the Unfit as Illustrated by the Introduced Sparrow, Passer domesticus,’’! as neglected, I do not intend to imply that it is unique in this respect. Several other important quantitative studies of natural selection, for instance papers by Weldon, Di Cesnola and Pearson, are in the same class. Indeed, the im- pression gained by reading papers commemorating the birth of Darwin and the publication of the ‘‘Origin of Species by Means of Natural Selection”? is that the majority of biologists have little interest in natural selection as a scientific problem. The chief reason for this is probably the great development of exper- imental breeding during the last decade—a development which is a great source of satisfaction to biologists, but which has tem- porarily brought the study of evolution to a very one-sided stage of development. At the time this lecture was published the statistical methods which are now considered the most suitable for dealing with such problems were not in the hands of many biologists. Re- cently in connection with some other work I had oceasion to throw Dr. Bumpus’s data? into statistical constants. These are published in the hope that they may suggest to some unoceupied biologist the collection of further quantitative data on the sev- eral problems presented by the introduced sparrow. The characters dealt with are the following: (1) Total length in millimeters from tip of beak to tip of tail; (2) alar extent, the distance in millimeters from tip to tip of extended wings; (3) weight in grams; (4) length of head in millimeters from tip of beak to the occiput; (5) length of humerus in fractions of an inch; (6) length of femur in fractions of an inch; (7) length of tibio-tarsus in fractions of an inch; (8) width of skull in frac- tions of an inch; (9) length of sternum in fractions of an inch. 1 Eleventh lecture before the Marine Biological Laboratory, Woods Hole, 1898; published in Biological Lectures from the Marine Biological Labora- tory, 1898. Boston, Ginn and Co., 1899. 2 Fortunately all the measurements were published. 314 No. 533] SHORTER ARTICLES AND DISCUSSION 315 Three classes of birds were distinguished—adult males, young males, and young and adult females. We draw the following conclusions from the comparison of the means in Tables I-III with their probable errors.* TABLE I AVERAGES FOR ADULT MALES Character Survived Porisbod Difference Total length........... 159.0571+.3154 162.0000 +.3253 —2,9499+ 4531 Alar extent 247.6857 +-.4333 247.3750+.4716 + .3107+.6404 Weight (in grams) 25.4685--.1420 26.2708+-.1966 — .80238+.2424 Beak and head....... 31.6143+.0709 31.6708+.0824 — .0565+.1095 ngth, humerus +.00 .7279 2 | + .0101+.0038 Length of femur. 8-+.00 .7061+.0027 | + .0107+.0037 bio-tarsus. ......... 1.1353--.0041 1.1202+-.0051 +. .0151+.0065 Width of skull....... 6025.0016 .6033+.0017 Í +. Keel of sternum..... .8576+.0042 .8458+.0045 + .0118+.0062 TABLE II AVERAGES FOR YOUNG MALES Character Survived Perished Difference Total length........... 159.6875 +.4978 162.2499+ .7291 | —2.5621+ .8828 lar extent 246.8125+-.7936 247.9167+1.2976 | —1.1042+1.5213 Weight (in grams ).| 25.4938-+.2040 26.2667+ .38208 | — .7729+ .3801 and head........|. 31.8688-+.1190 31.3249+ .1138 4 5439+ .1646 Length, humerus. 7416+-.0039 .7347+ .0055 | + .0069+ .0067 Length of femur..... .7162+.0046 7153+ .0050 | + .0009+ .0068 Tibio-tarsus........... 1.1367 +.0091 1.1398 .0071 | — .0026+ .0115 Width of skull....... .6078+.0024 5993+ .0035 | + .0085+ .0042 Keel of sternum ..... .8514+-.0060 8427+ .0064 | + .0087+ .0088 TABLE III AVERAGES FOR ALL FEMALES Character o Perished Difference Total length........... 157. 3810.4774 | 158.4286+.4859 | —1.0476+.6811 Alar extent............ 1.0000 +.6009 | 241.5714+.7142 | — .5714+.9333 Weight (in grams).. pr 6190.1531 25.3357 +.2054 | — .7167+.2561 Beak and head... 31.4333--.1047 31.4786-+.1068 | — .0453-+.1495 ngth, humerus .7283+-.0024 260-+.0082 0023+. Length of femur..... .7148+.0029 7098+.0036 bio-tarsus........... 1.1436-£.0042 1.1310+.0043 | + .0126+.0060 Width of skull .6001=.0019 '6016--.0031 | — .0015+.0036 Seel of sternum..... 8193.0043 82074.0037 | — .0014+.0057 * For the individual comparisons those differences less than the prob- 1e error will be considered of no significance, those between one and two times their probable errors as possibly significant, and those over thrice their probable errors as probably significant. 316 STANDARD DEVIATIONS FOR ADULT MALES a Character THE AMERICAN NATURALIST TABLE IV Survived Perished [VoL. XLV Difference Tot:.1 Seii ENN | 2.7666+.2230 2.3629 4-.2300 +4037 -+.3208 LINE GRECO. eeii | 3.8005+.3064 3.4255+.3335 +. nee 4529 Weight (in grams)... 1,2451+.1004 1.4276+.139 1715 Beak and head....... | .6220+.0501 5982.0582 4- “0238, 0768 ngth, leup 0196+.001 .0230+.0022 .0034-+-.0027 Length of femur.. 0222+.0018 9+.0019 +.00. 3.0026 i0-tarsus........... Bs 55.0029 0370.0036 —.0015+.0046 Width of skull ...... | 0317.0011 .0123+-.0012 -+.0194+.0016 eel of sternum | 0366.0030 .0325-+- .0082 +.0041+.0044 TABLE V STANDARD DEVIATIONS FOR YOUNG MALES Character Perished Survived © Difference Aa Tength iei- | 2.9521+.3520 3.7444+.5155 — .7923+ .6242 Alar extent 4.7066+.5612 6.6641+.9175 —1.9575+1. 0753 abep (i r wea 1,210] +.1448 1.6474+.2268 — 4373+ . pose ead., 7060.0842 .5847 +.0805 + 1213+ T Length, erus.. 0234+-.0028 0282+-.0039 — .0048+ .0048 Length of yrende E 0272+-.0032 8+.0 + .0014+ .0047 Tibio-tarsus........... 0537+ .0064 0365+. 0050 + .0172+ .0081 Width of skull ...... .0141+.0017 .0180+.0025 — .0039+ .0030 Keel of sternum..... .0356+.0042 .0331 +.0046 + .0^254 .0062 TABLE VI STANDARD DEVIATIONS FOR ALL FEMALES Character Survived Perished Difference Tota a a ee 37+. 3.8119-+.3436 | — .5682+.4817 Alar extent............ 4.0825+-.4249 5.5 —1. +.6600 Weight ‘Gn grams) 00+.1082 1.6112+.1452 — 5712+. Hee nd head.. ..... 7114+.0740 8381+.0755 — 1267+ Length, humerus. 0160.0017 + — pri Length of femur 0197+.0021 .0279+.0025 — .0082+.0033 ibio- tarsus... .0287 +. 36+.0030 — ,0049+.0042 Width of skull....... .0128+.0013 .0245+.0022 — 0117+.0026 l of sternum...... .0292+-.0030 .0286 +.0026 + .0006+.0040 In all three series the individuals which survive are shorter than those which perish. The probable errors support in a very satisfactory manner the conclusion, ‘‘ that when nature selects, through the agency of winter storms of this particular kind of severity, those sparrows which are short stand a better chance of surviving.’’ For weight the results for the three series are also consistent in sign, and even when taken individually indicate No. 533] SHORTER ARTICLES AND DISCUSSION 317 with a considerable degree of probability that the heavier birds are the least able to withstand the vicissitudes of the February sleet and snow. In all three series the length of the humerus is longer in the birds which survive, and in the group of adult males the difference is perhaps statistically significant. The same is true for the length of the femur, but the results are again insignificant except in the adult males where they are per- haps statistically trustworthy. In the adult males and in the adult and young females the length of the tibio-tarsus seems to be longer in the survivors, but the result is insignificant for the young males. If selective elimination be a reality in nature one would not expect all of the characters of a series of individuals which per- ished when exposed to a given set of unfavorable conditions to differ from the same characters in the individuals which survive, and this for the simple reason that variations in many characters may not be of vital importance to the individual—in short, not of selective value. he constants seem to me to justify no conclusion concerning the length of the sternum. For alar extent all three differences individually considered are insignificant; taken comparatively two are negative and one positive in sign. Apparently varia- tions in the spread of wing have under the particular conditions* no significance in determining the chances of survival. The young males which survived have longer skulls (tip of beak to the occiput) than those which perished, and the difference seems to be significant in comparison with its probable error, but in the other two classes of birds the differences are not merely statisti- cally insignificant but negative in sign. Tables IV-VI show the standard deviations and their probable error, These are essential in calculating the probable errors of the means and in testing the hypothesis of a reduction in varia- bility by selective elimination. Bumpus has discussed this ques- tion in detail in his lecture, but to me it seems that the standard deviations as given here do not justify any final conclusions con- cerning the relation of selection to variability : the probiem is too complicated and the data are too few. As in other evolutionary problems we need more measurements. When these are available “Were the eliminative agent, for example, a severe northerly wind of protracted duration, the alar extent might then enter in as a factor of considerable selective value. ”? 318 ; THE AMERICAN NATURALIST [Von. XLV not only type and variability but correlation® will be open for investigation. Looking at the tables of constants, the cautious biometrician will hesitate to say that Professor Bumpus has proved his point. The data available are too scanty to justify dogmatic assertions. But the work is so suggestive and the results so convineing that it is difficult to understand why zoologists have not followed it up by other studies of a comparable nature. To be sure, opportun- ities of this particular kind do not occur every winter, but there are other sources of elimination active in nature, and one of the most important tasks before those interested in the problems which Darwin pointed out to biologists, is to determine whether the individuals which survive are able to do so because of certain structural peculiarities, while those which perish are eliminated because they are in the degree of development or in the correla- tion of their parts structurally unfit. J. ARTHUR HARRIS. 5 Compare — Bumpus’s suggestion on this point, the arguments of Brooks in his ‘‘ Foundations of Zoology,’’ Lectures VI-VIII, and the hypothesis of ale in Journ. Exp. Zool., 2: 425-430, 1905. NOTES AND LITERATURE BIOMETRICS AN INTRODUCTION TO STATISTICAL METHODS IN spite of the great development of biometrie work, and of the application of statistical conceptions and methods in a num- ber of fields of science other than biological, during the last decade there has been produced up to the present time no fully satisfactory introduction to the elementary principles of modern statistical methods. The books which have appeared in this field have been, broadly speaking, either (a) too technical and ad- vanced in their treatment, or (b) compilations of formule with so little in the way of guiding principles as actually to lead any but the already expert into many difficulties, or (c) incomplete, incorrect and superficial at vital points, or finally, (d) have appealed to a very limited class of readers by develop- ing the subject in direct relation to a narrow field of science only. This need for a comprehensive, elementary and sound introduc- tion to statistical methods is admirably met in a recently pub- lished book by Yule. The subject is treated under three main heads as follows: (I) The Theory of Attributes, (II) The Theory of Variables, (III) The Theory of Sampling. In the first part the author deals with the logical basis of statistical theory, a field which is essential to a proper understanding of the subject, and in which he is, by the extent and character of his original investigations, qualified to speak with unique authority. Successive chapters in this portion of the work deal with Notation and Terminology, Consistence, Association, Partial Association, Manifold Classi- fication. The second part of the book takes up the discussion of fre- quency distributions and their physical constants, and the ele- mentary theory of correlation and its applications, ending with an account of multiple and partial correlation. Here we are dealing with matters of immediate practical importance in the application of statistical methods to all kinds of scientific prob- ms. It would be difficult to say too much in commendation of the author’s method of treating these subjects. No knowledge * Yule, G. Udny, ‘‘An Introduction to the Theory of Statistics,’’ Lon- don (Chas. Griffin & Co.), 1911, pp. xiii + 376. 319 320 THE AMERICAN NATURALIST [ Vou. XLV of mathematics beyond algebra up to the binomial theorem is presumed, yet the subject is developed in such a simple, lucid and at the same time thorough way as to give the reader a real and adequate grasp not only of the technique of the methods, but also of their origin and significance. Numerical examples drawn from a wide range of materials are given at every stage and worked out in detail. Particular attention is paid to guiding the unwary beginner around the numerous pitfalls which beset the statistical pathway. Chapters are devoted to the methods of arranging data in the form of frequency distributions, deter- mining centering constants (arithmetic, geometric and harmonic means, mode, median, ete.), variation or ‘‘dispersion’’ measuring constants, and coefficients of correlation. The treatment of cor- relation is particularly comprehensive and practical. The last section of the book deals with the general subject of ‘‘probable errors.’’ The theory of fluctuations in statistical measures due to random sampling is developed first in relation to the theory of attributes and then in relation to the more com- plex theory of variables. The discussion of the simple sampling of attributes leads up in a straightforward way through the point binomial to the normal curve of errors, and the normal correlation surface. Each chapter throughout the book is followed by a short list of selected titles of original papers, and a series of practical problems to be worked out by the student. Appendices give short bibliographies of calculating tables, tables of functions, ete., and general works on the mathematical theory of statistics and the theory of probability. A list of answers and hints in regard to the problems and a full index complete the volume. Altogether the book is a notable one. Those who are familiar with Yule’s paper ‘‘On the Theory of Correlation’’ (published in the Journal of the Royal Statistical Society in 1897 ) which has become one of the classics of biometrie literature will be prepared to welcome the present work. It is marked throughout by the same clearness, directness and appreciation of the diffi- culties of the beginner which distinguished that memoir. For the non-mathematical student desirous of obtaining a sound working knowledge of the elements of modern statistical theory this book will be of the greatest value. In the field which it covers it is without a peer. i RAYMOND PEARL. SECOND EDITION, NOVEMBER, 1910 AMERICAN MEN OF SCIENCE A BIOGRAPHICAL DIRECTORY EDITED BY J. McKEEN CATTELL A Biographical directory requires revision if it is cng bac Nearly a third of the names in the oe en fe che whch appeared in te ie feo ion ve in nearly every case been revised. The amount of work required to pr the revision as great as that given to the first edition. There has been no Sister pd ce general plan of work. Greater strictness has observed 7 in par k as and as possible. There are of course even to ri requests for the information needed. The Da rate hai aare ws po stars have been added tothe tained places on the list. editor’ s object in 3 the conditions on whi pan object This directory should be in the hands of all those who are directly ond interested in wink (1) Men of Science will find it indispensable. m indram borky hares ome ite tnd the like of their fellow worker, but also an invaluable summary of the research work of the country, ek (2) Those interested in n science, even though they may not be professionally engaged in research work, _ ee TaN eh oi Tare and vaian Wem ne bak SE iad Ue aa ere $: mene of l > d th ; “0 4 £ r ( me g ga Fe cs 2. 1e i, rill ; as veal), Päikese? Newspapers and periodicals will find it to be frequently. {) Libraries will fi a Sind the book to be a necessary addition 1 to their 1 iia wail inte oma fate Although the increased in size ira paper and bound it is sold at the Same prioe aa tho tein. Price: Five Dollars, net, THE SCIENCE PRESS ei Meie i The > American Naturalist established in 1867, Devoted to the Advancement of the ae Sciences AON adr Reto Ge Pe of =a Evolution and H CONTENTS OF THE NOVEMBER NUMBER seat’ of Skin leh gmane in Man. meena C. AVENPORT and C LES B. DAVENPO The aie Sense = the he Honey Bee—Can Bees PERE Colors? Jom Shorter Articles Ser tame : The Arithmetic of the Product Moment Method of pene amg pane Coeffi- cient of Correlation: Dr. J. ARTHUR EERE a Notes and Literature: Schlosser on Fayûm Mammal ATTHEW. The Ophidian aan Grayia: Professor T. D, A, COCKERELL, CONTENTS OF THE DECEMBER NUMBER — of Skin Pigmentation in Man. GERTRUDE C. DAVENPORT and CHARLES B. DAVENPORT. Spawn and Larva of a e EREE TR ts Pro- fessor W. H. PIERSO The Inheritance of Sizes a Shapesin Plants. Professor E. A. EMERSON. Shorter Articles and Discussion: The Modification of Mendelian Inheritance p- ne rnal Conditions, Professor T. D. A. COCKER Notes and Literature: Heredity, he W. J. SPILLMAN, Index to Volume XLIV. CONTENTS OF THE JANUARY NUMBER Somatic Alteration: Its Origination and Inheritance. Dr. D. T. MacDoveaL. The Nature of Graft- hybrids. Professor DOUGLAS HOUGHTON CAMPBELL. A Double Hen’s Egg. Dr. J. THOMAS PATTERSON, Notes and Literature: Heredity, Dr. W. J. SPILLMAN. CONTENTS OF THE FEBRUARY NUMBER The —— of the sgn — of Pure Lines to Sex mited 2 heritance Sexual Dimorphism, Profeess T. H. Mor RGAN Pure Linesin afte Study eo Genetics in Lower Organisms. Professor H. 8. JENN. Some Effects of ee upon Growing Mice, and = r e of Such age neces onan a Subsequent Gen- The Mendelian 1 San Dai. ane agar ER SHINE- "iay on geez aktiva weg reae te ~ ess of Barred and Self- colored Fowls, Dr. RAYMOND PEARL. — Considerations eraen the Ph nic Funt- tion in Marine Organisms. F. ALEX. MCDERMOTT, ape — and Discussion : Com rei Correla- n Cases where Symmetrical — are con” et Professor H. S. JENNIN CONTENTS OF THE MARCH NUMBER The Genotype ——— of Heredity. Professor W. JOHANNSEN. The Gen H and Hybridi e it enu y zation. Pro- Notes on Gundlachia and Ancylus. Dr. WILLIAM HEALEY DALL. Notes and Literature: Mimicry, Dr, Franx E. Lurz. CONTENTS OF THE APRIL NUMBER Genetical Studies on Oenothera. II. Dr. BRADLEY MOORE DavVIs. The a hag of Maize. Dr, GEORGE HARRISON SHUL Notes ad Literature: Is the Female Frog Heterozy n regard to Sex Determination ? Professor gou H. cra The Mutation Theory. Dr. R. R GATES. ————— aa Single N Number 35 Cents NATURALIST will he s Yearly leae $4.00 THE SCIENCE PRESS Garrison, N. Y. Sub-Station 84: NEW YORK Lancaster, Pa. | Pa. VOL. XLV, NO. 534 ~~ JUNE, 1911 THE AMERICAN NATURALIST A MONTHLY JOURNAL Devoted to the Advancement of the Biological Sciences with Special Reference to the Factors of Evolution CONTENTS Page Inheritance of Fecundity in the Domestic Fowl. - DR. RAYMOND PEARL. 321 The Biometric Proof of the Pure Line Theory. - The Influence of Changed External Conditions on the Development of Two r THOMAS H. MONTGOMERY 364 AUSTIN HOBART Dr. J. ARTHUR HARRIS. 346 Si of Moths. Shorter Articles and Disoun The he Ontogeny of aye a Genus. ea V. Notes and Literature: Recent Contributions to a PEE of the Extinct Amphibia, Dr, Roy L. MOODIE 37 I - ee ae - 372 THE SCIENCE PRESS LANCASTER, PA. GARRISON, N. ¥. NEW YORK: SUB-STATION 84 The American Naturalist intended for publication and books, etc., intended for review should be MSS. sent to the Editor of THE AMER oe research work bearing on the pro | be given preference in publica Arti tion are especialy w elcome, and wil ICAN NATURALIST r Garrison- -on-Hucson, New York. blems " beisen evolu- ndrea reprints of a are supplied to een tre of charge. e hu Further repent will be supplie The ents and THE SCIENCE PRESS Lancaster, Pa. NEW YORK: Garrison, N. Y. Sub=-Station 84 Entered as second-class matter, m 2, 1908, at the Post Office at Lancaster, Pa., under the Act of ngress of March 3, 1879. THE BULLETIN—Por bargains in Ethnolograph- ical and Pre-historic Specimens. Books on Natural History, Science, Travel, Voyages, etc. See THE BULLETIN post free for 3 cent stamp. 4 Duke St., Adelphi—London—England TENTH EDITION. THE MI ciara ime COMSTOCK PUBLISHING CO., Ithaca, N. Y. e GN Ante Sa EE BIRDS’ EGGS W. F. H. ROSENBERG, 57 Haverstock Hill, London, N. W., England begs to announce the publication of a new Price List (No. 15) of Birds’ Eggs, con- taining over 900 species from all parts of the world. This Catalogue is systematic- ally arranged, with authors’ names, indica- tions of localities, and an index to families. It will be sent post free on application, as will the following lists: No. 11, Birds’ Skins, (5,000 species) ; ee 12, Lepidiiphers: (5,000 species) ; No. 3, Coleoptera ; No. 14, Mammals. In course of ESS : New Price List of Reptiles, Amphibians and Fishes. Largest Stock in the world of specimens in all branches of Zoology. Marine Biological Laboratory WOODS HOLE, MASS. oology: INVESTIGATION Facilities for research fi Coin an ate labore JUNE 1 TO OCTOBER 1 to ries šio in Posh paye es are available oe beginnen ms got a h areetan g DT aberi the go . The fee for such & table is $50.00. iv ction INSTRUCTION Courses of laboraecred in inver- JUNE 28 TO AUGUST &, 1911 tute eones Oi Siea a0 s 0 Te is also o fered. one d Material SUPPLY OZY- animals for ciast DEPARTMENT OPEN THE ENTIRE rts and Mosses. and all information ad . M. GRAY, Curator, Woods Hole, ng The annual announcement will be sen ication sent on app! Hole, the Di Director, Marine Biological Laboratory, Woods He THE AMERICAN NATURALIST VoL. XLV June, 1911 No. 534 INHERITANCE OF FECUNDITY IN THE DOMESTIC FOWL! DR. RAYMOND PEARL MAINE AGRICULTURAL EXPERIMENT STATION THERE are under discussion at the present time two general views regarding certain fundamental points in heredity. Each of these points of view has its zealous adherents. On the one hand, is what may be designated the ‘statistical’? concept of inheritance, and on the other hand, the concept of genotypes. By the ‘‘statis- tical’’ concept of inheritance is meant that point of view which assumes, either by direct assertion or by implica- tion, that all variations are of equal hereditary signifi- cance and consequently may be treated statistically as a homogeneous mass, provided only that they conform to purely statistical canons of homogeneity. This as- sumption of equal hereditary significance for all varia- tions is tacitly made in deducing the law of ancestral inheritance, when individuals are lumped together in a gross correlation table.? The genotype concept, on the other hand, takes as a fundamental postulate, firmly grounded on the basis of breeding experience, that two ! Papers from the Biological Laboratory of the Maine Experiment Sta- tion, No. 25. This paper was read p the meeting of the American Society of “orga at Ithaca, December, 1910. *For a more detailed discussion of this point see a paper by the present Koad cities “*Biometrie Ideas and Methods in E. their Signifi- ce and Limitations,’’ in the Revista di Scienza (in pres 321 322 THE AMERICAN NATURALIST [Vou. XLV sorts of variations can be distinguished, namely those (a) that are represented in the germinal material and are inherited without substantial modification, as in ‘‘pure lines,’’ and those (b) that are somatic and are not inherited. By anything short of the actual breeding test it is quite impossible to tell whether a particular variation observed in the soma belongs to the one cate- gory or to the other. As I have tried to emphasize in other places, it is both to be expected on this view of in- heritance, and is also the case in actual fact, that the somatic manifestation or condition of any character is a most uncertain and unreliable criterion of the behavior of that character in breeding. Finally under the geno- type concept, of course, the whole array of facts brought out by Mendelian experiments find their place. Now while certain adumbrations of the genotype con- cept have long been current in biological speculations in regard to heredity, this general view-point owes its grounding in solid facts primarily to Johannsen’s work with beans and with barley. It is to be noted that in these cases, as well as in most of the investigations of the pure line theory which have followed Johannsen’s work, the organisms used have been such as reproduced either by self-fertilization, or by fission, or by some veg- etative process. This brings us to the consideration of a question of great importance, both theoretical and prac- tical. In cases of diccious organisms, where a “pure” pedigree line in the sense that such lines are found in beans or in Paramecium by definition can not exist, has the genotype concept any bearing or significance? Ina general way it obviously has. Probably no one (except possibly some of the ultra-statistical school) could be found who would deny that in general a distinction is to be made between variations having a gametic and those having merely a somatic basis. But specifically how far has the genotype concept any application in case of ‘‘non-selfed’’? organisms? Johannsen in his ‘‘Ele- mente’’ has thoroughly analyzed Galton’s material and No. 534] FECUNDITY IN THE DOMESTIC FOWL 323 shown that it is capable of a satisfactory and reasonable interpretation on the genotype hypothesis, and East and Shull have gone far in the analysis of genotypes in maize. This, however, is only a beginning. There is the greatest need for careful, thorough investigations of the inheritance of characters showing marked fluctuating variation in organisms having the sexes separate. Here lies one of the crucial fields in the study of inheritance to-day. Through the brilliant results in Mendelian di- rections and from the study of really ‘‘pure’’ lines we are getting clear-cut ideas as to the inheritance of quali- tatively differentiated characters, such as color, pattern and the like, on the one hand, and in regard to the in- heritance of quantitative variation in self-fertilized or non-sexually reproducing organisms, on the other hand. But beyond all these lie the difficult cases where in diœ- cious forms quantitative variations must be dealt with. If these can be cleared up and brought harmoniously into a general scheme or view-point regarding inherit- ance, we shall have gone a long way in the solution of this world-old biological problem. For some four years past the writer has been engaged in a study of the inheritance of fecundity in the domestic fowl. The problem presented here is an important one from the practical as well as the theoretical standpoint. If definite and sure methods of improving the average gg production of poultry by breeding can be discovered it will mean much to the farmers of the nation. At the Same time egg production is a character in some ways well adapted to furnish definite and crucial data regard- Ing inheritance. Variations in egg production are read- ily measured, and can be directly expressed in figures. The general results of this study of the inheritance of fecundity may be said, in a word, to be, so far as they go, in entire accord with the genotype concept, and not to agree at all with the ‘‘statistico-ancestral’’ theory of in- heritance. Indeed, so ill is the accord here that the chief exponent of the latter doctrine has recently attempted to 324 THE AMERICAN NATURALIST [ Vou. XLV throw the whole case out of court? by asserting that fecundity is not inherited in fowls, and that the present writer’s investigations show essentially nothing more than that. It will be the purpose of this paper to pre- sent some figures sufficient to indicate with some degree of probability, I think, first that egg production in fowls . is inherited, and second that it is probably inherited in accord with the genotype concept, in spite of the fact that we do not and can not here have ‘‘pure lines’’ in the strict sense of Johannsen’s definition. In the present paper, owing to limitations of space, the whole of the data in hand obviously can not be presented. Only a few illustrative cases can be given here. Before entering upon the discussion of the evidence it is necessary to call attention to two points. The first is in regard to the unit of measuring egg production used in the work. For reasons which have been dis- cussed in detail elsewhere‘ the unit of study has been taken as the egg production of the bird before March 1 of her pullet year. This ‘‘winter production”? is a better unit for the study of the inheritance of fecundity than any other which can be used practically. All records of production given in this paper are then to be understood as ‘‘winter’’ records, comprising all eggs laid up to March 1 of the first year of a bird’s life. It may be said that the ‘‘normal’’? mean winter production of Barred Plymouth Rocks (the breed used in this work) is fairly indicated by the 8-year average of the Maine Station flock. This average November 1 to March 1 production is 36.12 eggs. This figure is based on eight years con- tinuous trap-nesting of the flock with which the present * Pearson, K., ‘‘Darwinism, Biometry and some Recent Biology, ,” Biometrika, Vol. 7, pp. 368-385, 1910. * Bull. Me. Agr. Exp. Sta., No. 165. U. S. Dept. Agr. Bur. Anim. Ind., Bul. 110, Part II. * It should be said that up to and including the winter of 1907 only the November 1 to March 1 records are available as a ‘‘winter’’ record. Since that time the small number of eggs laid before November 1 (on the average two or three per bird) are included in the ‘‘winter’’ totals. These, then, give, as stated, the total production up to March 1. No. 534] FECUNDITY IN THE DOMESTIC FOWL 325 work was done, carried out before these investigations were begun. In the second place it is desirable to call attention to some of the difficulties which attend an attempt to analyze the inheritance of the character egg production. The most important of these is the fact that this char- acter is not visibly or somatically expressed in the male. A male bird may carry the genes of high fecundity, but the only way to tell whether or not this is so is to breed and rear daughters from him. All Mendelian workers will agree that it is sometimes difficult enough to unravel gametic complexities in the case of characters expressed somatically. It is vastly more difficult when only one sex visibly bears the character. In the second place a very considerable practical difficulty arises from the fact that egg production is influenced markedly by a whole series of environmental circumstances. The greatest of care is always necessary, if one is to get reliable results, to insure that all birds shall be kept under uniform and good conditions. Further, on this account, it is neces- Sary to deal with relatively large numbers of birds. Some of the important conditions to be observed in work on fecundity have been discussed elsewhere? and need not be repeated here. Turning now to the results we may consider first THE EFFECT or SELECTION FOR FECUNDITY IN THE GENERAL POPULATION On the ‘‘statistico-ancestral’’? view of inheritance it would be expected that if fecundity were inherited at all this character would respond to continued selection. That is, it would be expected, if the highest layers only were bred from in each generation, that the general flock average would steadily, if perhaps slowly, increase and that any level reached would be at least maintained by continued selection. In 1898 an experiment in selecting for high egg production was begun at the Maine sta- “Me. Agr. Exp. Sta. Ann, Rept. for 1910, p. 100. 326 THE AMERICAN NATURALIST [ Vou. XLV tion. In this experiment only such females were used as breeders as had laid over 150 eggs in their pullet year (corresponding roughly to an average winter production of 45 or more eggs) and the only males used were such as were out of birds laying 200 or more eggs in the year. This experiment was continued until the end of 1908. The selection, be it understood, was based on the egg record alone, and no account was kept of pedigrees or of genotypes. Every female with a record higher than 150 eggs in the year was used as a breeder regardless of whether her high fecundity was genotypic or phzno- typie. - The results of this selection experiment covering a period of nine years have been fully reported elsewhere.‘ Here it needs only to be said that the net outcome of the experiment was to show that there was no steady or fixed improvement in average flock production after the long period of selection. There was no permanently cumulative effect of the eight (in the last year) genera- tions of selected ancestry. So far from there having been an increase there was actually a decline in mean egg production concurrent with the selection, taking the period as a whole. During parts of the selection period, however, as for example the years 1899-1900 to 1901-02, inclusive, and the years 1902-03 to 1905-06, inclusive, an improvement from year to year was to be noted, but in | each case the flock dropped back in intervening years. This is an important point, the meaning of which is now clear. The flock average from year to year depended largely upon whether the breeders of the year before had had their high fecundity genetically represented or only somatically. In some years the selection was fortunate in getting nearly all the breeders from good (i. e., ‘thigh production’’) genotypes or from good combinations of genes. In other years just the opposite thing happened: the high layers chosen as breeders came from low geno- Dept. Agr. Bur. Anim. Ind., Bul. 110, Parts I and II, 1909 and 1911. “Bettechr. f. indukt. Abst. a. Verb: -Lehre, Bd. 2, 1909, pp. 257-275. No. 534] FECUNDITY IN THE DOMESTIC FOWL 327 types or combinations of genes. The general upshot was that while the selection of high layers merely as such was systematic year after vear the result attained in the general flock production was entirely haphazard and uncertain. This is exactly what would be expected on the genotype hypothesis, but not on the ‘‘statistico- ancestral.’’ TABLE I MEAN WINTER (NOVEMBER 1 TO MARCH 1) EGG PRODUCTION DURING THE SELECTION EXPERIMENT Mean Winter Year Production 1899-1900 aise os so sien es ee en aa eee a 1.03 WE erie. eee yee E Ces vey secu een 37.88 DOU TAOS EE ARA E Aaa SRR S ea 45.23 BPO OG ns bec ak oak ce cdg E hee E 26.01 BOOP Oh ok tin dks abe Oe E N E Ss 26.55 TIOE UG oss fin hin oe ee ae LE eet oe wee 35.04 ROME ENN vos ce ine bos a cee MERWE IN G whos ee 40.66 ROOT i ea eS on ais le ee dey is 21.44 PIED pik eaea s OAS Cs ba IRS os Ue ES bees 15.92 The actual course of the average winter egg produc- tion (not hitherto published) during the period is given by the figures of Table I and shown graphically in Fig. 1. Certainly the first line of evidence, derived from a long-continued experiment, involving more than 2,000 individuals, gives no support to the ‘‘statistico-ances- 5 2 9 mM 5 Ty Pai \ 23 s g $ Y p : ‘ 15 f 7 99-00 œo o-o2 0203 03-04 04-05 05-06 0607 07-08 i YEAR Fig. 1. Diagram showing the course of average winter egg poaae during the period covered by the mass selection experime 328 THE AMERICAN NATURALIST [ Vou. XLV tral’’ theory and indeed is in flat contradiction to one of the most fundamental tenets of that faith. Let us next consider the question, ARE SoMATICALLY EQUAL VARIATIONS IN FECUNDITY OF EQUAL HEREDITARY SIGNIFICANCE? In the spring and summer of 1907 were reared 250 pullets, all of which were the daughters of hens that had laid approximately 200 or more eggs in the first year of their life. This group of mothers was reasonably homo- geneous in respect to records of egg production. All had laid about the same number of eggs. Their daughters were, however, far from a homogeneous lot with respect to egg production.’ It is plain from the results obtained in that experiment that the egg record of a hen is a most unreliable criterion of the probable number of eggs which her daughters will lay. This is demonstrated by examination of individual cases. Thus consider the two mothers nos. 253 and 14. Their winter production records were nearly identical (65 and 66 eggs, respect- ively). Their daughters’ average winter productions were 23.87 and 2.40 eggs, respectively! Certainly it seems reasonable to conclude that the gametic constitu- tions involved in the breeding of 253 and 14 were quite different, though both these hens laid the same number of eggs. Again, take birds no. 386 and 911. One had a winter record of 55 and the other of 52 eggs. Yet their daughters’ average winter productions were, respect- ively, 4.88 and 27.33 eggs. Many more instances of this kind could be brought forward. Taken together, the whole evidence shows beyond the shadow of a doubt that the presence of high fecundity in an individual, and that factor which makes high fecundity appear in the progeny, are two very different things, either of which may be present in an individual without the other. We plainly have here the basis for the distinction of phenotypes and genotypes just as in beans. * Full details regarding this experiment have been published as Bull. 166, Me. Agr. Exp. Sta., 1909. See particularly Table I. No.534] FECUNDITY IN THE DOMESTIC FOWL 329 THE INHERITANCE or Ece PRODUCTION IN PEDIGREE LINES Let us now consider some of the evidence that such things as genotypes of fecundity really exist in fowls. We may first examine some repr tative pedigrees covering four generations and showing the occurrence of high and low fecundity lines. As a typical example of a high fecundity pedigree line in which the high fecundity is genotypic, line D5D39 may be considered. In the presentation of this and other pedigree tables the following conventions are adopted. The band numbers of the birds are in bold-faced type, and following the band number of each female, her winter egg record is given in italic figures enclosed in parenthesis. The band numbers of males are given in italics. PEDIGREE LINE D5D39 F & 10 (100) [ $ F255 (48) X 9'564 QG 12 (16) 303 (64) X 7563 ( £232 (69 555—0 ee ọ (69)X #555 — ae oo L 2 D39 (62) X 7 D5- i 18 (61) B47 (69) X 7562 911 (36) 248 (67) 9 r Hy dia | 287 (65) xX r554 4 Mean = 62 363 (74)X a567 ` Mean=61 kas se B 74 This line is shown graphically in Fig. 2. Little comment on this pedigree line is necessary. We See a certain high degree of fecundity faithfully repro- duced generation after generation. Different males were 330 THE AMERICAN NATURALIST [ Vou. XLV dah 3U > ~ v oo w PERCENTAGE FREQUENCY EIN 90910 855 945 1035 1125 45 #135 2s üs 405 495 585 079 76.5 WINTER EGG PRODUCTION Diagram showing range of variation and mean fecundity in The main polygons of variation give the distribution oss-hatched areas Fic. 2. generation of line D5D39. areas of fecundity in the general flock in each generation. nace the pedigree line, and the heavy dotte represent the mean fecundity of the line in each generati The moet gee tee these No. 534] FECUNDITY IN THE DOMESTIC FOWL 331 used with different females, but in every case the males used were from high fecundity lines and were believed to carry this quality in their germ cells either in homo- zygote or heterozygote condition. In marked contrast to the last example let us consider the low fecundity line D61D168. It is a troublesome mat- ter to propagate the low fecundity lines, because of the difficulty of getting a sufficient number of eggs during the early part of the breeding season. The line D61D168 is of interest not alone as an illustration of a typical low line, but also because there appeared in it a mutation, or something very like one. We will consider here only the main line and not the mutant. PEDIGREE LINE D61D168 r 9 E231 (25) x 7552 Q F283(32) x 75738—0 22 ug 419 (9)X #561 Q F165 (nx seo ~~ é 209(38) X 7555—0 = Mean=9.67 50( 20 313(26)< 7554 Merrie ier X g 550 abies F D168(33 D61 9( 18) ean== (33) 7 D614 163 (93 200( i2) 141 (0) 116( 28) 151( 77) 24( 23) _Mean=17.5 US ESS (ZS This line is shown graphically in Fig. 4, in which the mutant and its progeny are also shown. A low line in which no mutant has appeared, but in which also the mean production is not so low as in line D61D168 is D65D366. Since the egg production has not been so low in the early part of the breeding season with this line it has been easier to propagate it. ” This was the mutant referred to. Its progeny will be considered later. See p. 335, 332 THE AMERICAN NATURALIST [Von. XLV 15 AN 1907-08 | E N 0 er 30 3 81s Meta ts 2 : i [e] ; ; Pian ree 30 15 Fe SOO oad Betyg. 2 te 30 T / 15 1910-11 — r 45 13.5 225 1.5 405 495 58.5 O75 765 855 945 103.5 1125 WINTER EGG PRODUCTION Fic. 3. Diagram showing range of variati generation of im teuen Significance of lines and cross F v on and mean fecundity in e hatching as i ach n No. 534] FECUNDITY IN THE DOMESTIC’ FOWL 333 PEDIGREE LINE D65D366 Q F309 (OD) ( 9 B239(24) x 553 | eatp (9634 (4) 216(41)X F569 + A 164 (5) Q F301 (7) 224 (43) X 71554 | 223 (14) 2442) 2 D366(33)X g7 D65- | 2G 65 (28) Q F242(27) 209 (33) : 854(15)X 0'551 Hisao]. 967 (25) 502 (27) 544 (8) 331 (31) X 7552—0 since Xg560 f 344117) { X S528 1 9 FITUS) | Mean=26 Mean=33.4 Mean=22.33 This line is shown graphically in Fig. 3. In the examples thus far given we have had to do with pedigree lines in which a given degree of fecundity re- appeared from generation to generation with practically no change. In two instances quite certainly, and pos- sibly in several others, a new and distinct variation has suddenly appeared within a line and thereafter bred true, thus presenting the characteristic phenomena of mutation. The most striking instance of this sort oc- curred in line D61D168 and may be given here in detail. The main part of this line has already been discussed (p. 331). It will be recalled that it is a line of low fecund- ity. In 1908 there appeared in it one individual of dis- tinctly higher fecundity than any other bird in the large family of that year. This individual when bred produced only high layers. In the next generation two of these daughters were bred to males known to belong to high fe- cundity genotypes (#554 and 566). One of these matings unfortunately produced no adult female offspring. The _ other led to the production of six adult daughters, all of Which are relatively high layers, with the single excep- tion of G495, which has a record of only one egg, and that record is doubtful. This bird has probably never laid an egg, and almost certainly is pathological. ” Bird died during winter period. 334 THE AMERICAN NATURALIST [ Vou. XLV i AN - J Fag i o À 30| \ \ ~ \ 2. \ : 3 \ g \ k \ z ; g o \ Poaceae | a> F E 15 WA o ka 7 / I / I / 4 - ! 1910-11 / o Tre ls 45 135 225 315 405 495 585 675 765 855 945 1035 112.5 WINTER EGG PRODUCTION . 4. Diagram of pedigree line D61D168. The significance of lines is coy same as in Fi nd 3, except that the mutant line is double cross hatche ye For the sake of PeR E495 and the daughters of D31 are omitted in the 1910-11 generati No. 534] FECUNDITY IN THE DOMESTIC FOWL 335 Leaving this bird out of account because pathological, the mean winter production of the family is 52.8 eggs, very strikingly different from the average (9.67 eggs) of the birds of the same generation in the main low line in which the mutation appeared. Two other daughters of the mutant E248 were mated to ¢D31, a bird known not only to belong to a genotype of mediocre to low fecundity, but to be remarkably pre- potent in respect to this character, so that practically regardless of the females with which he has been mated the get has been uniformly poor in respect to egg pro- duction. Four adult females resulted from the two matings under discussion. They have an average winter production of 23.75 eggs. There are several possible explanations of this result, but the most probable is that we have here simply one more instance of the extraordi- nary prepotency of ¢D31. The last of the daughters of the mutant was mated to a cross-bred male, no. 578, and consequently the progeny can not fairly be compared with the pure Barred Rocks in respect to fecundity. The facts here briefly discussed are shown in the fol- lowing table and graphically in Fig. 4. It is apparent from the table and the diagram that the main line and the ‘“‘mutant’’ line are entirely distinct. Indeed they do not overlap in their ranges even except- ing only the pathological individual G495. The ‘‘mu- tant” pullet E248, for some reason or other, possessed the capacity both to lay a relatively large number of eggs, and the genes necessary to make this quality ap- pear in her progeny. Whether this individual is to be regarded as a true ‘“‘mutation” would appear to be largely a question of definition. In the writer’s opinion the most probable explanation is that E248 is a Men- delian segregation product. That is, let it be supposeđ that both D168 and D61 were heterozygous with respect to degree of fecundity, and were producing in some (un- own) ratio both ‘“‘high fecundity” and ‘‘low fecun- 336 THE AMERICAN NATURALIST [ Vou. XLV PEDIGREE LINE D61D168 (CoMPLETE) F F308(78) x 7 554—0 (26 - (45) 2 (58) 354 (55) Xo" s08| 11; (46) | Q E248 (48)X 7558 < rt Mean of high line=—52.8" 166(49)x @-Ds1 { ? C GR Q G229 (28) 141(61) xr. D31{ O58 UD Mean of D31’s daughters=23.75 172(50) K 7 578— Cross ean of mutant (high) line=56.5 Q E231 (25)X 9552 Q F288(82) T 573 —0 us ete 9)X F551 QF165 (7)X 569 = os 477 (1) 209 (58)X F Mean of main : 313 ook z g VERIS (low) tine=9.67 363 (11) 7'550 $ sige (iow) line=22.0 24 (23) | Mean of main (low) line=17.5 ` dity’’ gametes. Then E248 may be supposed to have originated from the union either of two ‘‘high fecun- dity’’ gametes or one high and one low fecundity gamete. She then would be either a DD or a DR bird, on the as- sumption, which the facts seem to support, and which I have more fully discussed elsewhere,'? that high fecun- dity is dominant over low. “= Omitting G495. See text. 1 ‘í Inheritance in ‘Blood Lines’ in Breeding Animals for Performance, with Special Reference to the ‘200-egg’ Hen,’’ Rept. Amer. Breeders’ Assoc., Vol. VI, 1911 (in press). No.534] FECUNDITY IN THE DOMESTIC FOWL 337 The subsequent breeding history of E248 indicates that it was probably a DD bird, though the reasons for this opinion can not be fully gone into here. The general view, recently emphasized by Nilsson-Ehle,’* that phe- nomena of mutation are, in many cases at least, merely cases of Mendelian segregation has much evidence in its favor. The pedigrees which have been given are merely illus- trations. Many other similar ones might be cited from the records in hand did space permit. In the experi- ments during the past three years the attempt has been made to propagate separately lines of high, medium and low fecundity. In the course of this work it has been found that lines of high fecundity were nearly if not quite as likely to have originated with individuals of a low record of production as with those of a high record. Similarly, many low fecundity lines have originated with individuals which were themselves exceedingly high layers. Indeed one of the highest winter layers which have ever appeared in the stock evidently belonged to a genotype of very low fecundity, since it has never been able to produce progeny of anything but the poorest lay- ing capacity. The breeding history of this bird (D352) is indeed so interesting that it may be briefly discussed here. This bird in her pullet year laid 98 eggs between November 10 and March 1 and made a record for the year of over 200 eggs. She was mated and produced plenty of eggs during the hatching season, but they hatched very badly. Only one female worth putting in the house was obtained. This pullet (E356) made a winter record of only 39 eggs, just about the general flock average. E356 was not mated. Her mother (D352) was kept over and bred to another male the next year, in the hope that as a fowl she might produce more and better chickens than she had as a pullet. As a matter of fact she was again able to produce during the whole * Nilsson-Ehle, H., ‘‘Kreuzungsuntersuchungen an Hafer und Weizen,’’ Lunds Univ. Arsskr., N. F., Afd. 2, Bd. 5, Nr. 2, 1909, pp. 1-122. 338 THE AMERICAN NATURALIST [Vou. XLV breeding season only one pullet worth putting into the laying house. This pullet (F163) made a winter record of but 11 eggs. F163 was bred in 1910, but produced only one daughter worth saving. This daughter, G429, has made a winter record of 18 eggs. It would be hard to get clearer evidence than that afforded by this breeding history that D352 belonged to a low fecundity genotype, in spite of her individual high laying record. THE EFFECT OF THE SELECTION oF FECUNDITY GENOTYPES Let us now consider the bearing of the results so far set forth on the problem of selection. Taking first the question of the effect of selection for fecundity within a population it is plain that if different degrees of fecundity have a genotypic basis, as the facts above presented and a considerable mass of data of a sim- ilar kind, which owing to lack of space can not be given here would appear to indicate, then the results fol- lowing selection will depend entirely upon the genotypic constitution of the population. If high fecundity geno- types are present they can be isolated by selection. If they are not present selection of high laying hens will not change the average production of the flock. The aim of the selection experiments since 1907 has been to discover and propagate separately genotypes of high fecundity and genotypes of low fecundity, all the birds being taken from the same general flock. The re- sults of this work are shown in the following table and in Fig. 5. This table is to be regarded as a continuation of that given on p. 327, supra, which shows the results of mass selection for high fecundity in the same stock. EFFECT OF SELECTION FOR FECUNDITY WITHIN THE POPULATION 1907-08. Mean winter production of general population ........-- 15.92 1908-09. Mean winter production of all high fecundity lines ...... 54.16 1908-09. Mean winter production of all low fecundity lines ....... 22.06 1909-10. Mean winter production of all high fecundity lines ...... 47.57 1909-10. Mean winter production of all low fecundity lines ......- 25.05 1910-11. Mean winter production of all high fecundity lines ....-- 50.58 1910-11. Mean winter production of all low fecundity lines ......- 17.00 No. 534] FECUNDITY IN THE DOMESTIC FOWL 339 oe ey a yen A EA 99-00 00-01 01-02 02-03 03-04 0-05 05-06 06-07 07-08 08-09 0%10 10-0 LAYING YEAR big Showing the effect of selecting high and low fecundity on a geno- typic basis. The solid line denotes means of all “high lines”; the dotted line means of all “low lines.” Up to 1907-08 the attempt had been to increase egg production by breeding merely from the highest layers, regardless of pedigrees. In 1907 and subsequent years the attempt has been to isolate genotypes of high and low fecundity which shall breed true, each to its own type. The results indicate the effectiveness of this method of selection. It should be understood, of course, that only those pedigree lines are included in the high line averages which uniformly in each generation show high fecundity. A similar consideration applies to the low line averages. _ Let us now consider briefly the question of the effec- tiveness of selection within the genotype. According to the ‘‘pure line” concept we should not expect selection of high or low individuals belonging to the same geno- type to produce any effect, except in cases where segre- gation has occurred and the selected individuals are really gametically different, though having the same pedigree. An example of this sort has been given in the case of line D61D168 (cf. p. 331, supra). The ineffective- ness of selection within the line when something of this Sort does not occur is illustrated by line D56D407. In the F, generation in this line there were four birds, of Which three were good layers and one was a poor layer. Two of the good layers and the poor layer were bred. 340 THE AMERICAN NATURALIST [ Vou. XLV Large families were reared in F, and F,. The average results in the three generations are given in the following table. EFFECT OF SELECTION OF GOOD AND PooR WINTER LAYERS IN THE SAME LINE, D56D407 Fi Fə F3 Mean winter record of good layers and their progeny 76.0 46.7 35.57 Mean winter record of poor layers and their progeny 26.0 52.0 36.75" It is evident that selection within the line here was quite without effect. Another example of the same thing from line D31D447 may be given by way of further illustration. In this line there was in the F, generation a family of ten daughters. Of these some were very good and some were poor lay- ers. All were bred. The mean results are shown in the next table. EFFECT OF SELECTION OF GOOD AND Poor WINTER LAYERS IN THE SAME LINE, D31D447 F; Fə F Mean winter record of good layers and their progeny 62.5 23.75 22.00 Mean winter record of poor layers and their progeny 32.0 28.75 14.75 Here again it is plain that selection within the line was without effect. Many more examples of the same sort might be given from the records did space permit. In general there is no evidence whatever that the selection of individuals of different laying records, but belonging to the same fecundity genotype, produces any definite or permanent effect whatever. Discussion AND CONCLUSIONS © Taking into consideration all the facts which have come out of this study, one is led to the following view as to the composition of a flock of fowls in respect to fe- cundity. In the average flock we may presume that there will probably be represented a number of fecundity “If one family of four birds, which ought not in fairness to be included because they were extremely inbred (brother-sister mating) in connec: tion with another experiment, is excluded this average becomes 49.0. No.534] FECUNDITY IN THE DOMESTIC FOWL 341 genotypes, some high, some low, and some intermediate or mediocre. In an ordinary flock these genotypes will be greatly mixed and intermingled. Further, the facts in hand indicate that the range of variation in fecundity within the genotype is relatively very large, nearly as great, in fact, as in the general population. Thus while fecundity genotype means may be and usually are per- fectly distinct, there is much overlapping of individuals in the different lines. In consequence it results that the egg record of an individual bird is of almost no value in helping to tell in advance of the breeding test to what fecundity genotype it belongs. Essentially this same fact has been brought out in all of the work which has been done with pure lines. The only difference in the present case lies in the fact that the range and degree of variation within the line appears to be relatively greater in the case of fecundity than in the case of most char- acters hitherto studied, as, for example, size relations in beans or Paramecium. The most serious difficulty which confronts one in the attempt to analyze the inheritance of a character like fe- cundity lies in the almost inextricable mingling of geno- types in the great majority of individuals. This, of course, is a direct consequence of the manner of repro- duction. The germ plasm of two separate individuals must unite to form a new individual. By prolonging in- cestuous mating one may in theory come indefinitely close to reproductive purity, but in practise even this is extremely difficult, if not impossible of accomplishment on any large scale or through any long period of time. — The fact simply is that a ‘‘pure line” in the strict sense of Johannsen" can not by definition exist in an organism reproducing as the domestic fowl does. This, however, Y no means indicates that the inheritance of fecundity does not rest on a genotype basis, or, in other words, that * Johannsen’s definition is as follows: ‘‘Mit einer reinen Linie bezeichne ich Individuen, welche von einen einzelnen selbstbefruchtenden Individuum > (‘“‘Ueber Erblichkeit in Populationen und reinen Linien,’” P. 9.) 342 THE AMERICAN NATURALIST [Vou. XLV fowls do not carry definite genes for definite degrees of fecundity. We touch here upon an important point; namely, the relation of the mode of reproduction to the mode of in- heritance. As one reflects upon the matter it becomes clear that it is only in the sense of a reproductive line that we can not, by definition, have pure lines in organ- isms where the sexes are separate. It is perfectly pos- sible to have a line of such organisms in which all the individuals are gametically pure with reference to any particular character. For example, it is the simplest of matters to establish a line of horses pure in respect to chestnut coat color. Any individual in such a line mated to any other will never produce anything but chestnut offspring. So similarly with any other char- acter, it is only necessary to obtain homozygous individ- uals in respect to any character in order to form a gametically pure strain with reference to that character. It must further be kept clearly in mind that a repro- ductive ‘‘pure line” (in the sense of Johannsen’s defi- nition) may be made up of individuals not gametically pure (i. e., homozygous). Thus suppose one crosses a yellow and a green pea and then takes an F, heterozy- gote individual seed which originated from a self- fertilized F, individual as the ‘‘single, self-fertilized individual” with which to start a line. The individ- ual which starts such a line arose by self-fertilization and is selfed to produce progeny and would thus fulfil every requirement of a reproductive ‘‘pure line” as de- fined by Johannsen. Yet it would produce both yellow and green offspring. On the other hand, as already pointed out, a line which is not, and from the nature of its mode of reproduction never can be, reproductively ‘‘pure’? may be gametically so (i. e., have none but homozygous individuals with respect to any character). We then see that the fact that in fowls the sexes are separate and we therefore can not have reproductive ‘pure lines’’ gives, per se, no reason to suppose that fe- No. 534] FECUNDITY IN THE DOMESTIC FOWL 343 cundity is not inherited on a genotypic basis. We have to consider the problem of genetic or gametic purity. Do we have homozygote lines in such cases as those dis- cussed in this paper? It plainly is the fact that one can get lines of birds which, broadly speaking, will breed true (perhaps throwing occasionally a few individuals not true to the type of the line) to definite degrees of fecundity. The same thing is true of milk production in dairy cattle, speed in race horses, ete. What are these lines gametically? Theoretically the formation of gametically pure (homozygote) lines with respect to definite degrees of fecundity is simple. Practically it is exceedingly difficult to do this, owing to the fact that (a) the character studied is not expressed in the male, and (b) it is subject to a wide fluctuating variability caused by environmental conditions. The question as to the gametic constitution of the fecundity lines here dis- cussed obviously can not be answered finally now. It is a matter for much further research. One may, however, form a general conception of the probable gametic con- stitution of such lines, which has much evidence in its support. The essential points in such a conception are: 1. Probably no line yet obtained is absolutely pure gametically in respect to fecundity. It represents a mix- ture of a greater or less number of fecundity genes. 2. Lines which breed reasonably true to a definite de- gree of fecundity may in most cases be taken to be made up of individuals bearing a preponderant number of genes of the particular degree of fecundity to which the line breeds true, so that in gametogenesis a great major- ity of the gametes formed carry only these genes. They also carry some genes of higher, or lower fecundity, or both kinds. When individuals of a definite (e. g., ‘“‘high’’) line thus constituted are bred together the ma- jority of the offspring will, purely as a matter of chance, be produced by the union of two high fecundity gametes. It is quite possible that with families of the size obtained with poultry nearly or quite every individual produced 344 THE AMERICAN NATURALIST [Vou. XLV in the line for several successive generations may be of this kind. In the long run, however, it is to be expected that a small number of ‘‘off’’ individuals will appear in the line. These originate by the chance union of two low fecundity genes, or by the union of a ‘‘high’’ gene with a ‘‘low’’ gene of great potency (as in the case of D3l, ef. p. 335). 3. The degree to which such a line will breed true will depend upon the proportion of genes of one type (or of very similar types) present. The higher such proportion the less frequently will the ‘‘off’’ individual segregate out. The practical goal to be worked towards is, of course, to obtain several lines not closely related, but all made up only of individuals homozygous with respect to either high or low or any other definite degree of fecun- dity. Whether a given degree of fecundity is to be regarded as a single unit character, in the Mendelian sense, or, on the other hand, as a complex dependent upon a particu- lar combination of separately segregable unit characters, ean not yet be determined. Every one must recognize the fundamental importance of the investigations of Nilsson-Ehle, Baur and East, which have shown that many characters which at first glance do not appear to conform to any determinate law of inheritance are really complexes, formed by the combination of a number of unit characters, each of which segregates and otherwise behaves in a perfectly regular and lawful manner. There are some facts which indicate that high fecundity is a character of this kind, but it will require prolonged analysis to decide this, because of the numerous practical difficulties which attend the study of fecundity. A great help in this analysis, as well as a contributory line of evidence of much weight in supporting the general conception of the manner of inheritance of fecundity set forth above, is derived from the study of crosses be- tween breeds of poultry in which high and low degrees of fecundity are definite breed characters. Studies of No.534] FECUNDITY IN THE DOMESTIC FOWL 345 this sort carried out at the Maine station indicate that the relatively high fecundity characteristic of the Barred Rock breed is inherited as a sex-limited character. In this respect it behaves like a simple unit character, but this does not necessarily prove that it is not a complex. More data are needed to settle this point. Of much signifi- cance is the fact that, whether simple or complex, fecun- dity is shown by these experiments in cross breeding to be a character resting on a definite gametic basis. In conclusion, I think it may fairly be said that the in- vestigations here reported show in the first place that different degrees of fecundity are inherited in the do- mestic fowl, and in the second place, that in all respects wherein it has been possible, considering the inherent difficulties of the material and the character dealt with, to make the test, the method of this inheritance is in en- tire accord with Johannsen’s concept of genotypes. THE BIOMETRIC PROOF OF THE PURE LINE THEORY? DR. J. ARTHUR HARRIS CARNEGIE INSTITUTION OF WASHINGTON I. [INTRODUCTION On this platform I find myself in a somewhat embar- rassing position. A friend assured me in advance that this symposium would be somewhat analogous to the country parson’s ‘‘praise service,’’ and into this pure devotional atmosphere I must bring a note of agnostic- ism. Agnosticism is a term selected after careful delibera- tion. Johannsen’s propositions are important—if true —and any candid naturalist must hesitate before oppos- ing a new theory which may lead to important advances in biology. Agnosticism is the condition of mind tem- porarily enforced by the results of my own experiments. If one is pledged in advance to the pure line theory many of these observations can be made to confirm Johann- sen’s conclusions. If one is unprejudiced and seeks to fit his theories to his observations, rather than to adjust his facts to his preconceived conclusions, the results are quite as capable of other interpretation. Possibly more extensive work may show clear confirmation of his re- sults. Meanwhile I must withhold final judgment, merely stating that my own work has greatly shaken my confidence in Johannsen’s theory. Here I do not care to dwell upon details of my own experiments. It seems more profitable to try and state the fundamental problems of the pure line theory as they appear to the biometrician and to indicate the methods of work which seem to him necessary to the drawing of sound conclusions. ' From a symposium on ‘‘The Study of Pure Lines of Genotypes,’’ before the American Society of Naturalists, December 29, 1910. No. 534] THE PURE LINE THEORY 347 II. THE FUNDAMENTAL Propositions oF THE Pure LINE THEORY Our symposium has for its subject the Genotype or Pure Line Theory. Some of the speakers have enthusi- astically urged us to replace the words ‘‘pure line theory’’ by ‘‘pure line facts.’’ If this were done there would be little need for this program. Pure line facts are as yet a very insignificant part of biological data. The real occasion for this symposium is the pure line theory— the rank vines which have grown from the nineteen bean seeds which Johannsen planted in 1901. Biologists would have been little interested by the statement that selection within the offspring of a single bean has been ineffective in changing the weight of the seed. It is the daring generalization of the conclusions drawn from these limited experiments—the curt characterization of other researches as of no biological significance or their reinterpretation (from flounder’s fins to. intelligence in school children) in terms of the bean experiments, that forces us to take an interest in these matters. Our first problem is to ascertain what these generali- zations—the elements of the pure line theory as con- trasted with the pure line facts—are. Our second task is to try to ascertain in how far experimental facts sup- port the pure line theory. Davenport? has given a particularly good outline of Johannsen’s theory: The fundamental principle of Johannsen is that an ordinary fre- quency polygon is usually made up of measurements of a characteristic belonging to a non-homogenous mass of individuals; that it is really analyzable into several elementary masses each of which has a “ fre- queney polygon” of its own. In each elementary polygon the varia- tion is strictly due to non-inheritable somatie modifications, selection of extremes of which has no genetic significance. But the selection for breeding of individuals belonging to different elementary polygons, lying, say, at the extremes of the complex, may quickly lead to an isolation of these elementary polygons, the constituent individuals of which reproduce their peculiarities as distinct elementary species. * Davenport, C. B., Science, n. s., 30: 852, 1909. 348 THE AMERICAN NATURALIST [ Vou. XLV We recognize three essential propositions: Proposition 1—Most species or varieties are not homogeneous, but are composed of a large number of minor forms. The series of individuals classified as the same variety or race by the systematist, regarded as homo- geneous material for experiment by the physiologist, lumped together to form a single ‘‘population’’ by the statistician, is designated by Johannsen as a phenotype. This phenotype may generally be analyzed by pure line breeding into many constant and indivisible strains known as genotypes. Systematists have long regarded certain groups as polymorphic. Aster, Rubus, Salix and Crategus at once occur to the botanist and Unio, Salmo and the staphylinids to the zoologist. But the genotype theory seems to regard systematic polymorphism as a much wider phenomenon. ndeed one is sometimes assured that every apparently uniform cultivated variety is a swarm of constant bio- types. Johannsen emphasizes the generality of hetero- geneity. For instance, he says: Ein gegebener Phaenotypus mag Ausdruck einer biologischen Einheit sein; es braucht es aber durchaus nicht zu sein. Die in der Natur dureh variationsstatistische Untersuchungen getundenden Phaenotypen sind es wohl in den allemeisten Fällen nicht !* Again on page 162: In der Praxis wirkt ein Selektion meistens schnell in der beabsichtigen Richtung—eben weil die Bestiinde oder Populationen fast immer sind. One more illustration will suffice: Der oft ausserordentlich grosse Reichtum genotypischer Unterschiede in einer auscheinend einheitlichen Population war von Darwin . . . eben- sowenig in der vollen Tragweite erkannt, als es dem grossen Grandléger der Mikrobiologie, Pasteur, klar sein gies welche bedeutung es hatte, ***¥Elemente,’’ p. 123. *On page 121, he remarks on this point: ‘‘Selbst die shénste ‘typische’ aware that this point has been fully recognized by ‘‘ Biometriker’’ for years No. 534] THE PURE LINE THEORY 349 na iole physiologisch sehr differierende Felersasen in vermeintlich ‘ reinen ’ Hefekulturen koexistieren konnten. Proposition 2.—These genotypes are separated gen- erally by differences which are exceedingly minute. Notwithstanding the constant flood of new species segregated from the classic Linnean groups, necessita- ting frequent supplements to ‘‘Index Kewensis’’ and other works of its kind, many naturalists could hardly understand the small species discussed by de Vries in his great work. Indeed, many laboratory men hardly perceived the usefulness of recognizing species—per- fectly constant, we were assured—so closely related that one taxonomist could not identify the species of another from his descriptions; species so similar that herbarium material was worthless, and only culture side by side could distinguish them. Yet after a lapse of only ten years we find de Vries criticized for not recognizing even smaller divisions than these! Spillman says: ‘‘de Vries overlooks entirely those closely related pure lines, differing frequently only quantitatively, and in a single character. ... They not only do not differ in all their characters as the @nothera mutants do, but their norms present a regular series coming under Quetelet’s law.’’® As examples of these minute differences both he and Lang? quote the ‘‘72 Formen einer Population einer gewissen Heferasse’’ discussed by N acne ce p Jennings says:? The work with genotypes brings out as never before the minuteness of the hereditary differences that separate the various lines. These differences are the smallest that can possibly be detected by refined measurements taken in connection with statistical treatment. Johann- sen found his genotypes of beans differing constantly merely by weights- of two or three hundredths of a gram in the average weight of the ed. Genotypes of Paramecium I found to show constant hereditary differences of one two-hundredths of a millimeter in length. Hanel ***Elemente,’’ p. 318, “Spillman, W. J., AM. Nar., 44: 760, 1910. ‘Lang, A., Arch. f. Induktive Abstamm.- u. T 4: 15-16, 1910. 5 Nilsson- Ehle, H., Bot. Not., 1907: 113-1 * Jennings, H., AMER. Nat., 44: 144-145, in 8 350 THE AMERICAN NATURALIST [Vou. XLV found the genotypes of Hydra to differ in the average number of ten- tacles merely by the fraction of a tentacle. That even smaller heredi- tary differences are not described is certainly due only to the impossi- bility of more accurate measurements; the observed differences go straight down to the limits set by the probable errors of our measures. Proposition 3—These genotypes are rigid hereditary units; by a process of mutation one may give rise to another, but selection within the genotype is incapable of effecting a change. This theory is everywhere so prominent in the writ- ings of the genotypists that discussion or explanation is superfluous. II. THe CARDINAL PROPOSITION OF THE GENOTYPE THEORY Of these three essential propositions of the genotype theory of heredity, the first two might be accepted by Darwinian or Lamarckian or by a member of almost any school. If the proposition concerning the exceeding smallness of the differences be true, the theory might seem to present the greatest difficulty to the de Vriesian,!® for with smaller and smaller genotypes there is a con- stant approach to continuity, but we are assured that continuity is never realized." The third proposition—that genotypic differences are rigid and unchangeable except by mutation—is there- fore the essential one. The most obvious way in which this hypothesis can be tested against concrete facts is to determine the effect of selection upon genotypes. The very heart of the pure line theory is the proposi- “Jennings (AMER. NAT., 44: 145, 1910) tells us, ‘‘The genotype work lends no support to the idea that evolution occurs in large steps, for it reveals a continuous series of the minutest differences between great num- bers of existing races.’’ = Johannsen (‘‘Elemente,’’ p. 356) says in criticism of the Lamarckian theory: ‘‘Die Lamarckismus muss kontinuierlich verschiebbare Typen an- ehmen; wir — aber bei genauer Priifung immer und immer wieder Diskontinuitat. No. 534] THE PURE LINE THEORY 351 tion that selection within the pure line is ineffective.t? The strenuousness with which this has been maintained has even engendered in some minds the opinion that se- lection has no rélé at all to play in evolution or in prac- tical breeding. The attitude of many appears to be that Darwin was quite mistaken when he wrote, ‘‘The key is man’s power of accumulative selection: nature gives suc- cessive variations; man adds them up in certain direc- tions useful to him.” Darwin said, ‘‘If selection consisted merely in sepa- rating some very distinct variety, and breeding from it, the principle would be so obvious as hardly to be worth notice.” Fifty years after this was written we hold a Symposium to celebrate the discovery that selection is after all merely the isolation of distinct varieties! Was Darwin right or wrong? Have all practical breeders except those at the oft-quoted Svalöf station been chiefly occupied in wasting their time for the last fifty years? These are very important questions. The burden of proof obviously lies on the genotypists.'* Much of the evidence offered is most general and not at all unzweideutig. Indeed, when closely analyzed much of the reasoning reduces to a circle of three ares each of one hundred and twenty degrees: : l. Definition—A genotype or biotype is an organic unit, reproducing itself constantly'* except for the transi- tory, non-inheritable modifications due to environmental influence.® It is not capable of change by selection. = Johannsen (‘‘Elemente,’’ p. 137) states the problem: ‘‘Wird Selek- tion von Plus—oder Minus—Varianten innerhalb reiner Linien eine Typen- verschiebung bezw. eine Galton’sche Regression hervorrufen?’’ fiat acceptable, the evidence must be quantitative; the observations must either be numerous enough that variations due to uncontrollable fac- tors will average out, or the experiments be conducted with such refined technique that environmental influences are entirely excluded; the statistical reasoning concerning the observations must be logically sound. "CA biotype is a group of individuals which do not differ from one another in any hereditary quality and which therefore constitute a pure Face.”’—Shull, G. H., Am. Breed. Mag., 1: 100, 1910 *“*Tn a given ‘pure line’ (progeny of a single individual) all detectable 352 THE AMERICAN NATURALIST [ Vou. XLV 2. Observation.—Selection has never been known to produce a change in a genotype. Whenever, as is often the case, selection does result in modification of type this proves that the material considered was impure—that more than one genotype was originally present—or that others arose by mutation, and entirely aep of selection. 3. Conclusion.—It is therefore proved that Br can not modify the characters of a genotype. Johannsen has written a very thick and a very con- vincing-looking book, but if one pins himself down to the task of going from cover to cover he finds that an unfor- tunate amount of the evidence reduces to this kind of reasoning—in short, to no critical evidence at all.’° But behind this citing of examples which are not inconsistent with his theory although they prove nothing concerning it; besides this reiteration of testimony which merely ex- cites in the minds of the court-room spectators suspicions concerning the integrity of the defendant without en- titling the plaintiff to a verdict before an impartial jury," there are certain direct experimental studies variations are due to growth and environmental action, and are not in- herited.’’—Jennings, Proc. Am. Phil. Soc., 47: 521, 1908. ‘‘ The standard deviation and écatiaak d variation express in a pure race mere temporary sagt of no consequence in heredity. If we could make all conditions of and environment the same throughout our pure race, all the echoed nap se that net eS deviation and coeffi- cient of variation would be zero, and t s the positive value of a assistance in determining what e je w Warisata of the progeny.’’ —Jennings, AMER. NAT., 43: 333, ‘* Wenn es gelinge, fiir alle Individuen einer reinen Linie absolut gleiche Lebenslage zu schaffen, miisste die Standardabweichung gleich sein. dany Römer, T., Arch. Rassen- u. Gesells.-Biologie, 7: 437—438, 1910. * For instance, he (‘‘Elemente,’’ p. 162) refers to the fact that Hallet was unable to improve Le Couteur’s wheat, although he had succeeded in ip es seventy other samples from all parts of the world, and explains it by the assumption that in every case the seventy series of wheat were mixtures of biotypes while Le Couteur’s was a pure line. This may be true, but what is it worth as scientific evidence? "In working over the literature of the pure line theory the lover of fair play is sometimes on the verge of losing his patience, for although the experimental data—at least those which are confided to his reader—upon which Johannsen grounds his own theory are very slender, he is unsparing No. 534] THE PURE LINE THEORY 353 which have been adduced in support of the genotype theory. These arguments and the evidence upon which they rest must be examined. For convenience of treat- ment I do this under three propositions concerning selec- tion, which seem so reasonable that I believe few biol- ogists will feel inclined to deny their soundness. They are at least so reasonable that no worker can afford to leave them out of consideration. A, Characters which are not Inherited at all can not be Taken to Prove that Selection in General is Ineffective This is a point of great importance, generally ignored by pure-linists. Biometricians have long known that of the variations of any character whatever not all are in- herited.1* They have also learned that variations in cer- tain characters are not inherited. Suppose now that one takes a character which gives no correlation between its degree of development in in his criticism of the pioneer studies which have made his own work sible. Such bald statements as (‘‘Elemente,’’ p. 285), ‘‘Alle oh ee Schliisse sind aber fiir die eigentliche iisttisthkslisforsehang pee ohne Wert,’ seem to have little of profit to contribute to science. Johannsen’s ipse dixit has been taken as gospel. Woltereck (Verh. Deutch. Kai Ges., 1909: 115) says, ‘‘ Dieses Resultat erschüttert ernstlich die grundlagen der statistischen Variations; und Erblichkeitforschung, wie sie von die Galton- Pearsonschen Schule betrieben wird.’’ A. Lan ng (Verh. Deutch. Zool. Ges., 1909: 24) asserts, ‘‘Die biometrischen Forschung arbeitet mit unreinen material.’’ Römer (Archiv f. Rossen- u. Ges.-Biol., T: 427, 1910) tells us, “‘ Variabilitätstudien sind bis in die neueste Zeit meist an Material ausge- n das aber nach dem jetzigen Stande der Wissenschaft als unrein angesehen werden muss. Soe tritt besonders hervor bei den veilen Untersuchungen der Biometriker > This is one ‘ae the facts which has led the biometrician to discuss prob- abilities while biologists in general clamor for certainty in the individual instance. One of the results of recent experimental work that has been hailed with the greatest enthusiasm is that two individuals may be identical in external appearance and yet produce entirely different offspring: in Short, that some (somatic) variations are and some are not inherited. experimental data collected on this point both by pure line and by Men- elian researches are of high value, but those who hail them as novel simply eb vat ignorance of much of the pioneer work in variation and 354 THE AMERICAN NATURALIST [ Vou. XLV parent and offspring in a population and selects to in- erease or decrease it. He will get no result of selection. If now he takes the same character and selects from the plus and minus variations within a pure line, he will again effect no change by selection. Does either of these cases prove that selection in general is ineffective? Or does the second support. in any way Johannsen’s geno- type theory of heredity? Certainly not. Certain important work of Pearl and Surface seems to - me to deserve mention in this connection.’® These re- searches are sometimes referred to as furnishing evi- dence against the possibility of improvement by selec- tion, and this they do so far as the character with which they have dealt is concerned. In the generalization of their results, however, the greatest caution must be used. From two series of experiments with the same strain of Barred Plymouth Rock fowls they show that there is little hope of increasing the egg-laying capacity by direct selection for fecundity. These results are doubtless of much practical importance. Biologically they are of in- terest in confirming the results of other biometric studies which have shown that for man, horse, swine and mice fertility is very slightly inherited in the population. To consider them as indicating that selection in general is ineffective would be a very grave error, for fertility— so far as we may judge from the statistics so far pub- lished—seems to be a character sui generis in respect to inheritance. To cite these results in support of Johann- sen’s genotype theory of heredity, as has sometimes been done, is absurd. Is it not possible that Johannsen’s results with beans may be due to seed weight being a character which is not. inheritable at all in the population, and which can not, therefore, reasonably be expected to be inherited within the pure line? » Pearl, R., and F. M. Surface, ‘‘ Inheritance of Fecundity,’’ Bull. Me.. Ag. Exp. Sta., 166, 1909. Pearl, R., and F. M. Surface, ‘‘Is there a Cumu- lative Effect of Selection?’’ Zeitschr. Ind. Abstamm.- u. Verebungsl., 2: 257-275, 1909. No. 534] THE PURE LINE THEORY 355 Biologists will agree, I believe, that to test critically the effectiveness of selection in the population and in the pure line, the experimental material must be an appar- ently homogeneous wild species or a garden variety the individuals of which are not differentiated into sub- races by characters other than those under considera- tion. Conclusions drawn from any experiments in which these simple precautions are neglected seem of doubtful value. From Professor Johannsen’s first memoir, that of 1903, we have no reason to suspect that his material is not, so far as the biologist can judge, homogeneous.*! We are told nothing of any vegetative differences seen during the two generations grown in 1901 and 1902. Ap- parently all the numerous reviewers have considered his material perfectly homogeneous except for differentia- tion into genotypes with respect to seed characters. Tn his book, however, one notes with some surprise the casual information (‘‘Elemente,’’ p. 311) that his Pure Line I also has curiously bent seeds, a special ‘‘Ver- halten” in germination and a ‘‘groben Habilus”’ in the vegetative organs. Indeed Johannsen states that from the form and method of germination, etc., of a seed— even though a strong ‘‘minus Abweicher’’—he can gen- erally recognize an individual belonging to Line I. These points should have been made clear at the be- ginning. If Professor Johannsen’s lines really differ in their vegetative characters, so, for instance, that they can be distinguished as they grow in the field, it seems to _ me that their significance for the efficiency of selection is = Surely we can all agree that the population is to be an apparently homo- geneous one, i. e., such that all the individuals would be classified together by a keen taxonomist. If this is not the case, if by definition, ‘‘popula- tion °’ means to the pure linist a mixture of several conspicuously different oa there seems little need for further discussion = Of the seed he says, ‘‘Der Ausgangspunkt dieser kappe war eine gekaufte Partie, etwa Skg, brauner ‘Prinzessbohnen,’ wohl eine der altesten Kruppbohnen unten den vielen Kulturformen von Prisons vul- garis. Die betreffende Ware... war rie REN! wae und so gleich- mässig, wie es überhaupt hier erwartet werden konnte 356 THE AMERICAN NATURALIST [Vou. XLV greatly reduced. We do not know to what extent the dif- ferences in seed weight which give the low correlation in his population are due to the mixture of races slightly differentiated with respect to their vegetative characters. If this differentiation be considerable, the seed weight character with which Professor Johannsen has chiefly worked, may not be inherited at all in the population providing this population be one composed of individuals with the same vegetative characters. It is not sufficient to be assured that these classic beans differ ‘‘nur (oder fast nur)’’ in seed characters; more detailed information is much needed, and until it is forthcoming I must differ from most biologists in my opinion as to the importance to be attached to the conclusions drawn from them. B. Improvement for any Single Character can not be supposed to be Unlimited This is a fundamental consideration too often neg- lected.??. A wheat is selected up to its maximum pro- ductiveness, perhaps by getting the uppermost attain- able limit at one choice from a large field. Then because it can not be made to yield all grain and no stubble we are told that selection can only isolate already existing types. A sugar beet can not be all sugar and the cow can not give pure cream. In arguing for Johannsen’s theory East?* concludes that since Illinois is no longer making progress in high = The principle, however, has been clearly seen by some biologists. For instance, in his ‘‘ Foundations of Zoology,’’ Brooks says 65) sore breeder of domesticated animals or of cultivated plants, who devotes his attention to one or two characteristics, must soon reach a point where no for a single point quickly grows less and less effective, and soon reaches a maximum; but this is no proof of any ‘principle of e ee or anything we except the truth that long ages of natural selection have made the organism such a unit or coordinated whole that no great se continuous change in one feature » possible unless it be accompanied by general or oren Ss e? , E. M., ‘‘The Rôle of Selection in Plant Breeding,’’? Pop. Sci. Mo., Tis apes 1910. ana EI No. 534] THE PURE LINE THEORY 357 and low oil and protein selection in maize, their work has been merely the isolation of pure and constant strains— ‘*sub-races’’—with the characteristics in question as strongly developed in the beginning as we now find them, but continually intercrossing. The case is too compli- cated for discussion in detail, but certainly the fact that the characters can no longer be increased by selection** is no strong argument for the biotype idea. Under its present morphological and physiological organization we have no reason to suppose that the corn grain can be made to contain as much oil as the castor bean. Again Pearl and Surface announce concerning their selection work with corn, e find the results of this experiment or investigation to be very dificult (if not altogether incapable) of rational explanation in accord- ance with the biological implications of the “law of ancestral inherit- ance” and conclude that the results agree better with the genotype theory of Johannsen than with that of the cumulative theory of selection with, of course, the limitations implied by the fact that it is an open fertilized plant What Pearl and Surface have actually done is to take a desirable sweet corn which they for convenience desig- nate as Type I, and attempt—with initial suecess—to improve it for yield in ears and stover, for configuration of ears, and especially for earliness. But this Type I corn is descended from a few ears, the offspring of which have been grown in Maine for fifteen to twenty-five years. The variety originally introduced must have been an “That changes due to selection are at first rapid and then smitty has long been recognized. Indeed, as early as 1869 Hallett stated as two laws of the action of selecti ion, ‘‘ The improvement which is at first ik gradually, after a long series of years, is diminished in amount, and even- tually so far arrested that, plasticity speaking, a limit to the improvement in the desired quality is reached. By still continuing to select the improve- ment is maintained and practically a fixed type is the result.’’ Darwin’s views on this question are partly expressed in a letter of 1869 prised that Hallett has found some varieties of wheat could not be improved in certain a qualities as quickly as at first. All experience shows this with animals. * Pearl, R., and F. M. Surface, ‘‘ Experiments in Breeding Sweet Corn,’’ ‘Me. Ag. Exp. Sta. Bull., 1910. 358 THE AMERICAN NATURALIST [ Von. XLV early one as compared with sweet corn in general, to be able to survive at all in Maine. During the fifteen to twenty-five years the ancestors of the Type I corn were grown in Maine it must have been?® subjected to an oc- casional natural selection, for seed could be taken by the farmers from only plants which had ripened their ears. The somatic organization of some plants is such that they require only a few hours for their life cycle, but so long as sweet corn has the general characteristics of root, shoot and leaf that identify it as Zea Mays it seems reasonable to suppose that there is some limit to the re- duction of the time required for germination, growth and fruiting—an irreducible minimum beyond which selection can not carry it. Surely the fact that Pearl and Surface could not continually reduce the time re- quired for growth while at the same time maintaining a selection for yield of ears and stover may indicate that the irreducible minimum for earliness has been reached in a variety of the physical type they wish to breed. Speaking for myself alone, I must say that the data before us prove nothing against the theory of cumu- lative effect of selection, and they certainly do not fur- nish any critical evidence for the Johannsenian theory. It seems to me that Pearl and Surface again tacitly make this unjustifiable assumption that the modification attainable for any single character is practically un- limited when they consider that their failure to increase egg production by selection is a legitimate argument against the potency of selection. Indeed they say of ‘‘200 egg hens,’’ which lay an egg fifty-five per cent. of the days of the year, ‘‘This figure is of some interest as indi- eating what a relatively small proportion of the theo- retically maximum character is being selected to, when 200-egg birds are bred.’’27 But why, pray, is two hundred and sixty-five and a quarter eggs per year the theoretical maximum? One ~ ™Judging from the account of the difficulties of growing sweet corn which the authors give us. 7 Pearl and Surface, Bull. Me. Ag. Exp. Sta., 166: 55. No. 534] THE PURE LINE THEORY 359 ignorant of the physiology of reproduction in the do- mestic fowl might innocently suppose that even a hen needs a rest. If this be true, may it not be that 200 eggs is about the attainable maximum (the physical or physi- ological limit of the organism) of this variety under the environmental conditions available and that the Maine strain of poultry will not do better than it has? If this is not the attainable limit, why not assume over an egg a day as the theoretical maximum? C. Selection can not in general carry a Character be- yond a Degree Consistent with the Optimum for Maintenance and Reproduction This proposition is perhaps in a sense explanatory of the one immediately preceding. A characteristic is not independent of, but correlated with the other character- istics of the organism, and if it increases or decreases unduly they must also change or the organism be made more or less unfit for survival. Have those who claim to have found selection ineffect- ive been selecting against the morphological or physio- logical balance of the organism, that is in a manner to render the organism less capable of maintenance, growth and reproduction? If this be true their failure to obtain results will be in some measure explained. A possible illustration of this case may be furnished by the work of Pearl and Surface on egg production in the domestic fowl. Their work is again chosen not be- cause of any malicious desire to differ from them?> in interpretation, but because in a brief discussion of the evidence for the genotype theory one must confine his attention to the most important of Johannsen’s sup- porters. The data are: (a) The results of an eight years’ selec- = The criticism presented here must not be interpreted as drawing into question the scientific value of the data or the practical importance of the results of the studies criticized, or be extended to other work of the same authors, but is to be limited to the question of interpretation in relation to the pure line problem. 360 THE AMERICAN NATURALIST [Vou XLV tion for high egg production; (b) a correlation between the egg production of thirty-one individual mothers and the egg production of their daughters, and the compari- son of the egg production of these daughters with that of a large number of pullets of unregistered female parents. We note the following details: 1. During the eight-year selection experiment”? some unfavorable environmental accidents occurred in certain of the laying years. The averages for these years are perhaps too low, and both the actual means and a series of corrected means are given. The corrected means show an insignificant increase, but the unmodified means show a pronounced decrease in mean number of eggs as the result of the eight year selection. 2. In correlating between the egg production of the 31 highly selected mothers and their 217 daughters there is not trustworthy evidence of any relationship between the fertility of the mothers and that of their daughters.” If these constants show any deviation from 0 whatever it is on the negative side. 3. In comparing the daughters of these ‘‘200-egg¢”’ hens with three other series of the same strain but not of such highly selected female parentage, both for winter and spring egg production, it is shown that in five cases out of six the offspring of less highly selected parentage are better layers than those of the less stringently se- lected parents. us all three comparisons indicate that the high lay- ing mothers tend to produce low laying daughters; se- lection to increase egg production actually decreases it. =<‘ The practise in breeding was to use as mothers of the stock bred ın any year only hens which laid between November 1 of the year in which they were hatched and November 1 of the following year, 160 or more eggs. Af e first year, all male birds used in the breeding were the sons of mothers whose production in their first laying year was 200 eggs or more. Since the normal average annual egg production of these birds may be taken to be about 125 eggs, it will be seen that the selection Leas was fairly stringent.’’ Zeit. Ind. Abst.- u. Verebungsl., 2: From a knowledge of the biometrie work of the last poe years this is just the result which one would have expected to get. No. 534] THE PURE LINE THEORY 361 Such a run of results as this can hardly be due to chance.*! They indicate rather the presence of some as yet undetermined physiological factor.*? Candidly viewed and considered in comparison with . other biometric work on the inheritance of fertility and fecundity, I think these experiments can not be held to be strongly opposed to the theory of the effectiveness of se- lection in general. However this may be, they certainly afford no substantiation for ‘Johannsen’s genotype theory of heredity. IV. Summary AND CONCLUSIONS By the genotype theory of Johannsen one understands the following propositions: An apparently uniform population or phenotype is generally not homogeneous, but is composed of a large number of differentiated types, which are to be desig- nated—within limitations to be laid down immediately— as genotypes. Externally, the genotype can not be distinguished from the phenotype. Both may have normal variation curves, but while that of the phenotype may by proper selection be broken up into constituent genotypes, the variation curve of the genotype. can not be modified by selection. In short, the genotype is from the standpoint of heredity a rigid unit. All individuals belonging to the same genotype have the same potencies as parents. Only discontinuous segregations or transformations— mutations—may modify them. “The argument that this observed decrease as the result of selection to merease egg production is due to chance must rest chiefly on one or both o two assumptions. First, that the eight-year selection experiment is abso- lutely untrustworthy because of the accidents which may have affected the number entirely too small to give significant results in the case of a char- acter like fecundity. These admissions would vitiate entirely any conclusion Poenis selection to be drawn from these experiments. = To me it seems that some of Pearl and ate ’s published data are most suggestive of the nature of this factor, but they doubtless have in progress experiments that will throw light on these matters and biologists will await their results with interest. 362 THE AMERICAN NATURALIST [ Vou. XLV The keystone of the pure line arch is the proposition that selection is ineffective except as a means of sepa- rating already existing genotypes. If this keystone- proposition be not sound the whole structure of the theory crumbles. The propositions of the genotype theory are such that scientific proof or disproof is rendered particularly diffi- cult. By theory selection can not effect a change in a pure line; by a slippery process of reasoning in a circle any venules attained by selection are at once discredited by the assertion that the original material was impure. If, on the contrary, any selection experiment is ineffec- tual it is by some process of reasoning quite incompre- hensible to some of us, at once chalked up to the credit of the new theory. If heritable differences appear within a pure line known to be so, these results are also dis- credited by the assertion that the observed change is a mutation or has been produced by the action of the en- vironment. Truly the unbiased investigator is between the devil and the deep sea! The actual experimental data upon which the genotype theory rests are as yet few. Johannsen’s conclusions for beans depend chiefly upon the offspring of only nineteen seeds, and so far as I am aware no other in- vestigator has confirmed his results on Phaseolus. Hanel had only twenty-six original Hydra, and Pear- son’s analysis of his data with more adequate methods than he used, evidences against rather than for the geno- type theory. Jennings gives us the records of only six selection experiments involving altogether only a few actually selected Paramecia. Considering the large en- vironmental and growth factors, his conclusions can not be considered as beyond question.*? The work of Pearl and Surface with poultry and maize seems to me to have “In offering this FERNE I wish to express the highest admiration for Professor Jennings’s two memoirs on variation, heredity and evolution in the protozoa. The co fs of refined statistical with careful experimental methods in the investigation of these organisms marks a great advance in biology. No. 534] THE PURE LINE THEORY 363 no critical bearing on the pure line problem.** This is also true of numbers of other smaller experiments which can not be cited. If one turns from the strictly pure line side of the problem to the more general questions of the ‘‘some- thing’’ or ‘‘Etwas’’ in the germ plasm which determines in large degree the somatic characters of the individual which develops from it, one can only suggest that nothing whatever is explained by giving another name to a well- known fact. Ever since the time of Darwin, and before, we have known that there was ‘‘something’’ in the germ cells which determined the character of the offspring. We have had a dozen different names for this something, and by adding a thirteenth, ‘‘Gene,’’? Johannsen has merely burdened us with another cloak for our igno- rance. Unfortunately biological closets are full of such cloaks, once in fashion—now out. Finally, I must make my own position quite clear. With Professor Jennings’s contention that pure line cul- tures are of fundamental importance in many fields of physiology and genetics, I am in hearty agreement. Like other breeds of facts, ‘‘pure line facts’’ can not be- come too abundant. Indeed, a priori, I am not opposed to the genotype theory. As a theory it is most attract- ive, but one can not accept it without proof on that ac- count. Personally, I am one of ‘‘that last small rem- nant’ who believe that in a problem of this kind the proof must be biometric. This means merely three things. In so far as the nature of the material permits, all the data considered must be quantitative. The data must be numerous enough that biological relationships will not be obscured by the errors of random sampling. The data must be analyzed by logically sound methods. Judged by these standards, I must express the convic- tion that as yet there is no adequate justification for the genotype or pure line theory. “ Naturally, this is purely a matter of interpretation, and does not diminish in the slightest degree the value of the work. THE INFLUENCE OF CHANGED EXTERNAL CONDITIONS ON THE DEVELOPMENT OF TWO SPECIES OF MOTHS PROFESSOR THOMAS H. MONTGOMERY, Jr. UNIVERSITY OF PENNSYLVANIA THESE experiments were carried out in the fall and winter of 1908-09, and their results are not without in- terest even though no marked changes in the insects were effected. I. Attacus cecropia Linn. Cocoons of this large Saturniid were collected in New Jersey in December and January. The controls, kept in their cocoons, were hung out-of-doors exposed to rain and sun until the latter part of April, then placed in a hatching cage in a room at out-of-doors temperature, when they hatched in May. The pupe to be experi- mented upon were removed from the cocoons and kept in horizontal positions unless otherwise specified. A. Experiments with Light Direct Sunlight.—Four pupe, lot no. 86, of which only one was healthy in appearance, were placed in direct sunlight in a warm room (21° C.) on February .5; one of them hatched on February 10 and laid eggs, while the other died. Evidently direct sunlight is not fatal to them. Direct Sunlight behind a Heat Filter—Twenty pup2, lot no. 83, were laid horizontally on their dorsal sur- faces with heads directed towards the sunlight, behind a vertical flat glass jar containing a saturated aqueous solution of alum, in a warm room (21° C.). They were thus placed on January 22, and all hatched in March following. 364 No. 534] INFLUENCE OF CHANGED CONDITIONS 365 Diffuse Sunlight—Five pupæ, lot no. 65, were placed, each vertical with head up in a test-tube in a room that never sank quite as low as freezing; all hatched between May 15 and June 2. Five pupe, lot no. 66, were kept under similar condi- tions, but placed with their heads down in the test-tubes; four of these hatched in May, the fifth being infected by parasites. Three pupe, lot no. 69, were placed verti- cally in separate tubes within a moist chamber near a steam radiator in my private laboratory; two hatched in February, the third died. B. Experiments with the Tracheal Stigmata Covered Twenty pupæ, lot no. 85, were placed in a warm room (21° C.) in diffuse sunlight. On January 22 the stig- mata, of which there are eight easily recognizable pairs counting those of the head, were covered with a gum- arabic solution, but this peeled off and was replaced the next day by pure Canada balsam. It is, however, quite doubtful how efficient the balsam was in excluding air from the respiratory tubules, for it does not adhere very well to the greasy surface of the cuticula. Four of these pupe, lot 85A, had seven stigmata of the right side cov- ered, and all hatched. Four others, lot 85B, had the first pair of stigmata covered, and all hatched. Four others, lot 85C, had the second and third pairs of stigmata cov- ered, and all hatched. Four others, lot 85D, had the sixth and seventh pairs covered, and three hatched. Four others, lot 85E, had the fourth pair covered, and three hatched. C. Experiments with Higher Temperatures Ten pupe, lot no. 68, were placed within a closed and dry glass jar in diffuse light, kept thus at 28° C. for 23 days, then removed from the jar and kept in diffuse light in a room at 21° ©. All hatched, except three that were parasitized. Twenty-two pupe, lot no. 80, were placed on January 366 THE AMERICAN NATURALIST [Vou. XLV 5 in an egg incubator at 39° C., kept there for varying periods, then removed into a warm room (diffuse light, 21° C.) until hatching. These were divided into lots as follows: 80A, 4 pup, incubator 1 day, all hatched in March. 80B, 3 pup, incubator 2 days, all hatched in February and March. 80C, 3 pupe, incubator 3 days, 2 hatched in March. 80D, 3 pup, incubator 4 days, all hatched in March. 80E, 3 pupæ, incubator 6 days, 2 hatched in February. 80F, 3 pup, incubator 7 days, all hatched in February and March. 80G, 3 pup, incubator 28 days, all hatched in March. Twelve pups, lot no. 84, were placed in an egg incu- bator from January 22 to February 18, then removed to diffuse light in a warm room (21° C.) for hatching; the temperature in the incubator was 39° C. until Jan- uary 28, after that 39.5° C. Nine of these hatched in March and April. D. Experiments with Lower Temperature Seven pupa, no. 41, were placed in a tight covered and dry glass jar in an ordinary ice refrigerator from De- cember 7 to March 29, afterwards removed to a warm room; three hatched May 17, the others were destroyed by an accident. Thirteen pupæ, no. 67, were treated similarly; two hatched on May 17, the others were killed accidentally. E. Results of the Experiments The pupe were exposed to unusual external condi- tions: removed from the cocoon, exposed to direct sun- light with and without a heat filter, to diffuse light, to various temperatures ranging from 0° C. to 39° C., with the stigmata covered with balsam, in horizontal and vertical positions. Yet nearly as great a proportion hatched as in the ease of the controls. Higher tempera- tures hastened the rate of development. Further, the pupe so abnormally treated did not differ in coloration from the controls or to no extent that could be measured; No. 534] INFLUENCE OF CHANGED CONDITIONS 367 this result applies to the pattern as well as to the in- tensity of the coloration. For in the controls quite as great a range of color variation was found as in the others. Also the unusual conditions of life did not ap- pear to effect the dimensions of the hatched moths. To decide this I took as the most convenient measurement the length of the fore wing, measured from its point of insertion against the thorax to the most anterior edge of a dark spot placed anteriorly near the apex of the wing; I did not measure to the extreme free edge of this wing, for that portion is very flexible and liable to become folded during the process of mounting the moths. The right wing was measured unless it happened to be misshapen. Only about a hundred moths were pre- served, too few for any statistical study of this wing length, consequently in the following table only the ex- tremes of variation of this length are given (expressed ` in millimeters, and accurate to within a half millimeter). Lot 40 (1g, 59) control ¢ length 67.5 Ọ length 69.0-74.5 Lot 64 (23, 62) control ¢ length 61.0-65.0 2 length 61.0-75.5 Lot 68 (4g, 39) g length 58.0-64.0 2 length 64.0-73.0 Lot 69 (23) g length 66.0-67.0 Lot 79 (3g, 39) control & length 59.5-65.0 ? length 69.0-75.0 Tia 80 (7g, 79) g length 61.5-70.0 2 length 61.0-73.0 Lot 82 (34, 69) control & length 67.0-69.0 Ọ length 66.5-74.0 Lot 83 (124, 69) g length 60.0-68.5 2 length 67.0-73.0 Lot 84 (24, 49) 3 length 67.0-67.5 Q length 68.0-71.5 Lot 85 (124, 39) g length 63.0-69.0 9 length 65.5-73.0 It is probable that this late pupal stage is so advanced in its development that it can not become much modified by external changes. Il. Thyridopteryx ephemereformis STEPH. This psychid is the common ‘‘bag-worm”’ or ‘‘basket- worm.’’ The larva immediately on hatching constructs a bag or cocoon of silk covered with portions of leaves or chips, and increases the size of the bag as it grows and carries it about. At the end of the summer each at- taches its bag firmly to the twig of a tree, and the male 368 THE AMERICAN NATURALIST [Vou. XLV emerges as a winged insect; probably the male does not overwinter. But the female neither forsakes her bag nor acquires wings, she is impregnated by the male within her bag. Each female produces a large number of small eggs but does not oviposit, for she dies within her bag and her dead body becomes a case for the eggs; at her death her viscera change into a soft cottony mass that acts as a further protection for the eggs. Among some 200 cocoons collected on November 24 I found about half a dozen in which the egg case, the degenerate female, was still living. This species is then a very fav- orable insect for obtaining eggs and early embryos in large abundance during the colder season of the year, and should prove a valuable object for experimentation.’ The controls were kept within their cocoons out-of- doors, and hatched in the end of May. In the experi- “ments sometimes the eggs (in early embryonic stages) were removed from the egg cases, sometimes kept in them. A. Experiments with Sunlight Direct Sunlight.—Lot no. 77, collected January 4, con- sisted of egg cases placed in closed dry bottles in the south window of a warm room (21° C.). 77C, kept three weeks in this sunlight, did not hatch; 77A, an untimed period in sunlight, hatched. Lot 78, collected January 4, consisted of freed eggs in corked vials without mois- ture, with similar exposure to the light; they were di- vided into four lots, placed in the sunlight for 3, 7, 10 and 14 days respectively, and all hatched about March 1. Direct Sunlight behind an Alum Heat Filter—Four lots of freed eggs (nos. 51, 53, 49, 50) collected January 4 were used, placed in the sun behind a heat filter for 2, 7, 18 and 28 days, respectively, and all hatched in Jan- uary. “A good popular account of this species is given by McCook: ‘‘ Tenants of an Old Farm,’’ New York, 1885, and this is illustrated with excellent figures. But he makes the common mistake of other naturalists in sup- posing that the female oviposits. See also Howard and Chittenden, circular No. 97, U. S. Department’ of Agriculture, 1908. No. 534] INFLUENCE OF CHANGED CONDITIONS 369 Diffuse Sunlight—A considerable number of lots of egg cases and freed cocoons, collected November 4, were placed in diffuse north light in a warm room (21° C.). and all hatched in January and February. B. Experiments with Colored Light Freed eggs, collected November 4, were placed within vials immersed in colored solutions within larger bottles, the vial passing through the cork of the larger bottle and held by it. The solutions employed were: acid fuchsine in 50 per cent. and 70 per cent. alcohol; Berlin blue in distilled water; safranine O in 95 per cent. alcohol; orange G in 50 per cent. alcohol; eosine in 70 per cent. aleohol; methylen green in distilled water; picric acid in 50 per cent. alcohol; scarlet 12 gm. in 1,000 c.c. water, this last giving monochromatic light.? In Sunlight behind an Alum Filter, then removed to diffuse light in a warm room (21° C.). Lots 43, 46, 47 were immersed in a scarlet solution, as follows: Lot 43 in sunlight 4 days, hatched January 15. Lot 46 in sunlight 28 days, hatched January 9. Lot 47 in sunlight 32 days, hatched January 11. In Diffuse North Light.—The following experiments were made in a breeding room of which the temperature was a few degrees above that out-of-doors. Two differ- ent lots were raised in a fuchsine solution, one in saf- ranine, one in orange G, one in eosine, one in picric acid. All hatched in May. Others were placed in a room at 21° ©. One series were immersed in a fuchsine solution for 7, 18, 28, 35 days, respectively, then removed to ordinary daylight; these hatched in the latter half of January and first half of February. Others were kept continuously immersed in the following solutions: Berlin blue, methyl green, scarlet, and these hatched in the first part of February. *Vide Pennington, W. E., 1897, ‘ʻA Chemico-physiological Study of Spirogyra nitida,’* Publ. Univ. Penna. Contr. Bot. Lab. 1. 370 THE AMERICAN NATURALIST [ Vou. XLV C. Experiments with High Temperatures Freed eggs, from cocoons collected November 4, were placed in an egg incubator at 39° C. for varying periods, then removed to the dark of an ordinarily warmed room (21° C.). Those kept in the incubator for periods of 1, 2, 3, 4, 5, 7 days hatched in February; those kept in the incubator for eight and eleven days did not hatch. Ten unopened cocoons and ten egg cases placed in a dry covered slide box, and ten egg cases placed in a dry closed jar, all at 32° C., did not hatch. Five egg cases placed in a moist chamber at 28° C. hatched December 15 (these had been collected November 24). Four other ege cases, treated like the last but with less moisture, hatched in January. D. Experiments with Low Temperatures Six egg cases were placed out-of-doors in a closed tin box, protected from the rain. They hatched, as was to be anticipated, at the same time as the controls. Fourteen egg cases were placed in a closed jar within an ordinary refrigerator from November 24 until March 29, then removed to a warm room (21° C.); these also hatched at the same time as the controls. E. Results of the Experiments I tried to raise the small hatched larve by placing them upon arbor vite within a moist chamber; but owing to the great time consumed in transferring them to fresh pieces of the food plant, I was obliged to relinquish the attempt, and they all died. Consequently I did not de- termine whether those hatched under the abnormal con- ditions differ from control larve of the same age. The eggs of this species develop into larve under direct sunlight with and without a heat filter, in diffuse light, in all the colored lights employed, at a tempera- ture of 39° C. provided it be not continued longer than seven days, as well as at temperatures at and slightly below freezing. But what seems to be a necessary con- No. 534] INFLUENCE OF CHANGED CONDITIONS 371 dition for development is a certain amount of moisture, for the insects die when subjected to higher tempera- tures within dry vessels. The main effect of increase of temperature seems to be to hasten the rate of develop- ment. Probably it is the relative thickness of the chorion of the eggs that proves their chief protection under changed external conditions. The experiments on this moth and on Attacus would show that the cocoon can have no particular value by ex- cluding the sunlight, for we have found that sunlight is not injurious to the eggs and pupæ. Probably the main value of an insect cocoon is that of protecting against enemies, though it may also be of service in preserving a proper amount of moisture; for cocoons soak up the rain and melting snow, and would retain it for a con- siderable while. SHORTER ARTICLES AND DISCUSSION THE ONTOGENY OF A GENUS In the systematic work of to-day there is noticeable a tendency toward undue magnification of the importance of the smallest units, the species, subspecies, varieties or whatever they are called, to the great detriment of the larger and more important units, the genera, families and higher groups. While there is a very general agreement among systematists as to what consti- tutes a species or a subspecies or variety, the concept of a genus is found to vary widely; we have not yet brought ourselves to see the necessity of bestowing that care upon the genera which we use in the study of species and minor divisions. Yet after all the genera and the families are the units of paramount impor- tance, for they are the units with which the majority of workers must eventually deal. Zoology has become such a vast field that he who would occupy himself with species must of necessity re- strict himself to a very small section of the animal kingdom. t has therefore become essential for us to examine the char- acteristics of natural genera, and to analyze them carefully in order that we may discover certain general truths which will aid us in determining what genera are logical and valid and what are mere artificial aggregations, brought together solely for the sake of convenience. As commonly accepted, a genus is a group of species which is separated from all other similar groups of species by some char- acter common to all the component units, the latter being differ- entiated inter se by the unequal development of the specific, or, more accurately, intergenerie variables. In case a group of species uniformly differs from another similar group in the majority of the characters available for systematic purposes, that group is properly considered a family or a subfamily. Immediately upon its appearance, a genus (at this stage merely a vigorous species) spreads in every direction just as far as it is possible to maintain itself, that is, until it encounters on every side insurmountable barriers. But the conditions found throughout this habitable area are not uniform. This causes many local races to develop, each grading insensibly into all those surrounding it. Thus a genus in its infancy is in reality a well-marked species, differentiated into many geographical races. = , 372 No. 534] NOTES AND LITERATURE BTO These races do not long maintain themselves in their original relationships. There is somewhere within the range of this young genus, normally at or near the center, an area of optimum conditions, where life is easy and there is no severe struggle for existence. Here various more or less aberrant types arise and are able to perpetuate themselves, spreading out in every direc- tion as did the original stock, but never so far, as they are not so well prepared to encounter adverse conditions. Thus in the second stage a genus is in reality a well-marked species, differen- tiated into many geographical races, and in the center of its range being secompanied by several additional closely allied species. After the formation of these several supernumerary species, each usually with a few races of its own, the genus soon reaches maturity. Each of the numerous component forms increases in numbers so that in its own little sphere the struggle for existence becomes acute, and any variation from an arbitrary type is unable to maintain itself. The forms occupying the limits of the range of the genus as a whole (geographical or bathymet- rical) are continually trying to colonize new territory, both from their own initiative and as the result of pressure from behind. This encounter with generically unfavorable conditions induces, in the border forms, a more or less pathological condition, in- ducing great individual variation ; and so we normally find that the species which occupy the outer borders of the area inhabited by the genus as a whole, just as in any species the individuals from the edge of the area inhabited by it, are much more variable than those from any other part. If we take the species of any genus which has reached the stage of maturity just described and arrange them according to the proportionate value of their specific characters, we find in the center a single species, or a group of closely allied species, whose range is coterminal with that of the genus as a whole. This species is, moreover, typically the most variable of any in the genus, and probably is very close to the original stock. The period of maturity being passed, senescence begins to. assert itself. By long existence under fixed conditions the vari- ous component species become, as it were, delicate, and are unable to withstand any changes in their environment. Such changes are, however, of constant occurrence, affecting greater or lesser areas; and therefore discontinuance of distribution creeps in, species being cut off from the main zoogeographie area inhab- 374 THE AMERICAN NATURALIST [ Vou. XLV ited by the genus, one by one, by the extirpation of the inter- mediate forms. It often happens, also, that changing ecological conditions at the center of distribution of the genus, such as the local development or introduction of predaceous forms, or of external or internal parasites, destroys the typical form there, leaving only aberrant types; or they may even obliterate all traces of the genus. Very old genera are thus characterized by having but few species in widely separated localities, each widely different from the others. These are usually (and rightly) regarded as repre- senting a family composed of a few monotypic, or nearly mono- typic, genera. Very often old genera undergo what has aptly been termed an ‘‘explosion’’ of the intergenerie characters, and are then com- posed wholly, or almost wholly, of curious and eccentric species; again a genus in its senescence often is marked by a great devel- opment of certain characters at the expense of others, which usually leads to prompt extinction. In certain localities large numbers of species are remarkable for their eccentric develop- ment, and the exaggeration of certain characters out of all pro- portion to the others, which, so far as we can see, serves no useful purpose. Such localities from a zoological point of view must be considered as old and to have persisted in their present state be- yond the normal life cycle of the genera which have given rise to the erratic species. Just as the life cycle of different animals varies enormously, so does that of species and of genera. Scores of genera belonging to the higher groups of the animal kingdom may arise, grow strong, decline, and finally, with a grand ‘‘ex- plosion’’ of their characters, disappear, before a genus belonging to one of the lower groups, of earlier origin, has reached the summit of its strength. In discussing genera, as well as species, one must always keep in mind that for all animals there are two, and for aquatic animals three types of distribution, viz., (1) geographical, with purely inorganic physical barriers; (2) ecological, with wholly organic barriers, consisting of presence or absence of food and -predaceous or parasitic enemies; and (3) bathymetric, again with purely physical barriers of pressure and temperature, the latter commonly being the more important with lower animals, the former with the higher. Austin HOBART CLARK. U. S. NATIONAL Museum. NOTES AND LITERATURE RECENT CONTRIBUTIONS TO A KNOWLEDGE OF THE EXTINCT AMPHIBIA THE past few months have witnessed an unusual activity among paleontologists in behalf of the extinct Amphibia. There have been several rather extensive papers and an important memoir on the group issued within the last twelve months. It is to be hoped that many other investigators will come to be inter- ested in this group of vertebrates, for it is only by descriptions and discussions that we shall ever attain any adequate conception of the relationships of these highly interesting and important forms. The writer is of the opinion that the present conception is capable of considerable improvement and in order to facilitate this improvement he offers a review of the recent literature on the group. Dr. A. Smith Woodward (1) has described an interesting new amphibian from the ‘‘ Oil Shale, at Airly, New South Wales.” Dr. Woodward locates his form in the genus Bothriceps of Huxley. The skull and greater part of the vertebral column with the ribs and a portion of the right arm are preserved. It is described as a new species under the name Bothriceps major, but as this term had already been used by Lydekker for the reception of the uncertain Petrophryne major of Owent it will be necessary for the Australian specimen to receive a new name, for which the term Bothriceps woodwardi would not be inappropriate. Dr. Woodward allies the form with the Archegosauride, but the reviewer is rather inclined to think that the Tuditanide would be its nearer relatives. This is the third form described from the Hawkesbury formation of New South Wales. Further search will undoubtedly reveal other Paleozoic amphibia. It will be noticed in this as in so many other Paleozoic localities where fossil amphibia are found, that nearly every new specimen repre- sents an unknown form, thus indicating the diversity and age of the group. The known species from the Hawkesbury formation are: Bothriceps australis Huxley, Bothriceps woodwardi and Platyceps wilkinsoni Stephens. Dr. S. W. Williston has described in some ken ne the *Cat. Fossil Amphb. and Rept. Brit. Museum, Pt. IV, p. 375 376 THE AMERICAN NATURALIST [ Vou. XLV remains of Dissorophus multicinctus Cope which has recently been recovered from the reputed Permian of Texas. His material greatly increases our knowledge of the genus and of the anatomy of the Permian amphibia. He describes a complete skull, in which, unfortunately, the sutures are not discernible. Nor are the lateral line canals to be found, a fact to be regretted since we shall undoubtedly be enabled to base considerable impor- tance on these structures did they occur. The skull roof is pitted like all other of the Permian amphibia from Texas. A large portion of the carapace is described with its attached vertebre. The dermal shield is broad, continuous and pitted, forming a covering over the thoracic region of the animal. Limb bones, a scapula and a portion of the interclavicle are described. The form is closely related to another animal recently described by Dr. Williston and the two are placed in the new family Dissoro- phide. The paper closes with a taxonomic list of the Permian amphibia from Texas for which paleontologists will be grateful. There are three orders, nine families and thirty-four species so far known in the fauna. The same writer has described (3) a nearly complete skeleton of a new temnospondylous amphibian from the Texas Permian. The form is very remarkable in several of its characters. The following are the chief unusual characters of the new genus: a median unpaired rostral opening leading into a palatine vacuity, greatly enlarged antorbital vacuities, temporal fenestre, appa- rent absence of the parasphenoid bone, osseous carpus and tarsus and the possession of short heavy ribs borne on transverse pro- cesses. The skeleton is greatly similar to that of Eryops, but the skull shows decided differences. The temporal fenestra is not homologous to the superior temporal fenestra of reptiles, but it is rather to be considered as a greatly elongate and closed epiotic notch. The median unpaired rostral opening is similar in structure to the one found in the skull of Dasyceps bucklandi Lloyd from the Permian of Kenil- worth, though in the present form the opening is much further forward and smaller. The antorbital vacuities in the present form, on the other hand, are much larger than the same openings in Dasyceps. Dr. Williston was able to make out the complete anatomy of the skull and has figured it in three views. The most remarkable feature of the palatal structure is the apparent absence of the parasphenoid. The vertebral formula is 22 for the presacral vertebræ, an uncertain number of caudals and a single sacral. The No. 534] NOTES AND LITERATURE Sti sacral rib is much like that of Eryops in which the structure takes a very unusual form for a rib. The phalangeal formula for the foot is 1, 2, 3, 4, 2. The complete number of digits in the hand is not preserved. The carpus has nine possibly eleven osseous elements, and the tarsus has twelve osseous elements. he paper is well illustrated. There is a restoration of the skeleton of Trematops milleri and an outline drawing of the scapula of Eryops latus Case. The new genus Trematops is the type of a new family Trematopsidæ in which the form described by Cope as Acheloma cumminsi is doubtfully associated. The same writer (4) has redescribed from more complete material the species named by Cope as Diplocaulus limbatus from the Permian of Texas. The paper is based on several more or less incomplete skeletons. These include several additional features to our knowledge of the anatomy of the peculiar Diplo- caulidæ. Limbs have heretofore been unknown in the group although their presence has been suspected from the presence of pectoral girdles preserved with some specimens. Dr. Williston, however, for the first time actually describes well-formed limb bones for the group. The humerus is very remarkable in that it has an epicondylar foramen, a character known in only one other amphibian, Acheloma. The complete morphology of the skull with the exception of some features of the palate are made out and represented in two plates. The clavicular girdle, man- dible, vertebre and limb bones are represented in other plates. The paper coneludes with remarks concerning the relationship: of the group to which Diplocalus belongs and associates the Oklahoma Permian form Crossotelos with the Diplocaulus. He remarks that in the Microsauria the capitulum of the rib is always attached intercentrally and suggests that Diplocaulus must be retained among the Microsauria. The same writer (5) has given an extensive paper on new Permian forms in which he describes a new genus and species of amphibia under the name Cacops aspidephorus. This form he locates in the family Dissorophide. The paper opens with a rief discussion of the ‘‘Character of the Permian Beds of Northern Texas,’’ ‘‘Conditions of Fossilization’’ and ‘‘ Asso- ciated Vertebrates.” The form described in the paper is repre- sented by a skeleton which is remarkably complete ‘‘with no more plaster in its construction than was necessary to cement the freshly broken parts . . . save of many of the phalanges. .. .”” It was so complete and well preserved as to be capable of being mounted like a recent skeleton which has been well executed by 378 THE AMERICAN NATURALIST [Vou. XLV Mr. Paul Miller with remarkable success. A photograph of the mounted skeleton is given in one of the plates. There are four skulls. The most remarkable feature of the dorsum is the presence of a closed otic notch which resembles a temporal fenestra. In none of the skulls was it possible to determine the sutures and the structure of the skull had to be determined more by topographic features. The structure of the palate is of the stegocephalian type, though remarkable in some of its features, such as the large size of the palatal openings. The vertebre were preserved practically complete and the vertebral formula is—presacral, 21; sacral, 2; pygals, 6, and chevron caudals, 15 or 16. Fifteen of the vertebral spines are elongated and expanded and serve to support a carapace of shield-shaped, seute-like plates which overlap shingle-like. They greatly resemble in structure the dermal plates of Dissorophus. A dis- cussion of the ‘‘carapace in allied forms’’ is given and the dermal elements of Aspidosaurus, Zatrachys, Dissorophus are discussed. Plates are suggested by the expanded neural spines of Euchiro- saurus and Eryops. The vertebral column is fully discussed. This includes some unusual features, such as two sacral vertebrae and a well-pre- served atlas which is composed of a single piece. The writer discusses also the significance of the hypocentra and pleuro- centra, one of the most perplexed questions in connection with the extinct amphibia. The pectoral girdle consists of the fused scapula-coracoid, a cleithrum, clavicles and interclavicle. The humerus and its use in diagnosis is discussed at some length. Among the material studied are many humeri, some of which suggest unknown forms of amphibia. Two new families, the Trematopside and Dissorophide, are proposed and the characters given. The paper closes with a discussion of the restoration of Cacops and the description of a peculiar form of reptile in which the vertebre are intermediate between what is known in temno- spondylous amphibia and reptilia. The same writer (16) in a discussion of the faunal relations of the early vertebrates, presented before Section E of the American Association in 1909, gives the relations of the American Permian and Carboniferous amphibian faunas with those known else- where. He reaches the conclusion that the Permian fauna is especially isolated. In his discussion of the Microsauria he says, “Tt has been assumed on entirely insufficient evidence that they too were all amphibians’’—and later: ‘“We may be assured that some of them before the close of the Pennsylvanian were inhabit- No. 534] NOTES AND LITERATURE 379 ants of high-and-dry land regions where fleetness of movement, rather than obscurity, preserved them from their enemies, crawling reptiles in everything save some insignificant technical details of their palates.’? This has been recognized by many students of the fossil amphibia and Gadow placed them in a new group which he has called Proreptilia, but his classification does not seem to have been accepted. Dr. Williston says further, “Specialization of the microsaurs had reached the extraordinary extent of snake-like limbless forms.’’ These snake-like forms have been usually associated in another order, the Aistopoa, but the reviewer has shown elsewhere that the group is a hetero- geneous one and is made up of specialized microsaurian forms of diverse relationships. Dr. E. C. Case (6) has described three, perhaps four, new forms of amphibia from the Permian of Texas. The forms as a whole are very insufficiently described. One species, Trimerorhachis alleni is described in ten lines and no figure given. This manner of descriptions should be subjected to the severest criticism as it imposes many heavy burdens on the shoulders of succeeding workers. The new genus Tersomius is not defined at all. While we may not doubt that the genus is new, judging from the single outline figure, yet it would have been much better, for those who are not so well acquainted with the Permian fauna as is Dr. Case, had he given in what ways it differs from the other amphibia. He allies the genus with Trimerorhachis at least so far as resem- blances are concerned. The new genus and species are given in fifteen lines of less than ten words each. A new form, Gymnarthrus willoughbi, is much better described. Its relations are uncertain. Dr. Broom allies it with the amphibia, but Dr. Case does not regard the form as such. He . remarks its close alliance with Cardiocephalus sternbergii, which is amphibian. If @ymnarthrus is not amphibian it is certainly a very remarkable amphibian-like reptile. Dr. E. B. Branson (7) has described and figured, in an excel- lent photograph, footprints of possible amphibians from the Mississippian rocks of Giles Co., Virginia. Five well-preserved tracks are represented in the figure. The author proposes the new specifie name Dromopus aduncus and gives a list of the amphibian footprints known from the Mississippian. The most notable attempt on the part of paleontologists, to elucidate an entire amphibian fauna, is that of Armand Thevenin (8) in the most important memoir on fossil amphibia for many months. The National Academy of France awarded him a prize 380 THE AMERICAN NATURALIST [Vou XLV for the presentation of the memoir. The paper was published in successive issues of the Annales de Paleontologie and in complete form contains sixty-three quarto pages and nine photogravure plates, illustrating all that is known of the Paleozoic amphibian fauna of France up to the present. The author divides the amphibian forms into four groups: the Phyllospondyles, which is a subordinate group of the ‘‘Stegoce- phales’’; the Temnospondyles; the Aistopodes, and the Micro- saurians, which unfortunately he ranks in with the reptiles, and describes under this heading a form which a few years ago he had concluded was a rhynchocephalian. Dr. Williston was more inclined to regard it as a Cotylosaurian. Whatever reptilian group it belongs to the reviewer is unable to say, but he is quite certain it is not a Microsaurian Dr. Thevenin discusses, under the heading, Phyllospondyles, the forms Protriton fayoli Thevenin, P. petrolei Gaudry, and Pelosaurus laticeps Credner. The second group consists of Actinodon brevis, A. frossardi and Euchirosaurus rochei. The Aistopodes are represented by a single new form which is unnamed. e specimen strikingly suggests the snake-like amphibians of Ohio and Ireland. There are no true representa- tives of the Microsauria known in France. Our author discusses some general questions in regard to the amphibia, such as—‘‘the relations of the Autun amphibia to those of other countries,’’ ‘‘the homologies of the temnospondylous and the phyllospondylous vertebre,’’ ‘‘homologies of the elements of the pectoral girdle,” ‘‘the ancestry of the Stegocephalia” and ‘‘the descendants of the Permian Stegocephalia.” Nothing new is added to our previous knowledge of the com- plex relations of the elements of the temnospondylous vertebra, which is one of the most vexed and most discussed questions in connection with the extinct amphibia. His homologies of the elements of the pectoral girdle are the ordinary interpretations. The ancestors of the Stegocephalia are possibly the crossopte- rygian fishes, although this is no new conclusion nor does our writer claim this. Perhaps the crossopterygians will do as well as anything. At least they will serve until we find what the real ancestors were. In a discussion on ‘‘the descendants of the Permian Stego- cephala’’ he concludes that the branchiosaurian forms were the ancestors of the modern Urodeles and that the Temnospondylia gave rise to some of the reptiles, possibly some of the Cotylo- sauria. Our author, on a later page, gives the stratigraphic dis- No. 534] - NOTES AND LITERATURE 381 tribution of the amphibians and reptiles of the Permian of France. His final conclusion is that the diversity of the reptiles and amphibians shows that the groups had arisen long previously and the existence of similar forms in Europe and America would indicate some land connection of the two continents during the Permian. Dr. Friedrich von Huene has redescribed the skull of Dasyceps bucklandi (Lloyd) (9) from the Permian of Kenilworth. This skull was previously studied by Huxley, but rather inadequately described. After a careful description of the elements of the skull Dr. Huene locates the form in the family Melosauride, although the form has characters which are unusual for the other members of this group. He discusses the character and significance of the ‘‘facialgrube’’ or internasal opening, which has been described in another Permian form by Williston. Huene finds the same opening occurs in many living urodeles and lists nineteen species in which the opening has been described. He says that it has also been observed in certain members of the Permian Microsauria described by Fritsch from Bohemia. Its Significance is possibly the same as in the living amphibia, that of receiving the glandula intermaxillaris. Since this gland in living land-dwelling amphibia secretes a sticky substance used in capturing insects, Dr. Huene suggests that perhaps Dasyceps also captured insects. This may, of course, have been possible, but to the reviewer it suggests a greater activity than could be expected of such a sluggish creature as Dasyceps undoubtedly was, since it would require many insects to feed an animal three or four feet long and it would be necessary to secure them in some quantity. Dr. Huene suggests that the insects ‘‘im Perm und Carbon sehr bedeutende Grössen erreichten’’; such was undoubtedly the case with a few species, but the great majority of insects of the Carboniferous and Permian do not greatly exceed the modern insect fauna, so that Dr. Huene’s argument on that score is not a good one. Dasyceps was probably a land animal and Dr. Huene thinks this is indicated by the presence of the internal opening which occurs only in the land-inhabiting forms among recent Amphibia. Perhaps the analogy may be carried so far. Dr. R. Broom (10) compares the Permian amphibian fauna of North America and Europe and finds little similarity. He regards the American types as more highly developed. He divides the Permian amphibia of North America into four groups. 382° THE AMERICAN NATURALIST [ Vou. XLV He discusses again the relationship of Lysorophus and in his discussion quotes the reviewer as saying what he did not say. The point of the reviewer’s criticism of the reference of Lysoro- phus to the Urodela was not the presence of ribs nor yet the snake-like character which Dr. Broom explains in a very elemen- tary way, but it was the character of the ribs. Their long, curved condition is unknown among other Caudata and the reviewer does not feel satisfied that Lysorophus is a Urodele even though limbs should be discovered. Dr. Broom suggests for the newly described Gymnarthrus of Case an amphibian relationship. The dorsum of the skull shows characters, however, which apparently ally it with Pariotichus. Dr. Broom’s other essay (11) on practically the same subject matter gives the additional suggestion that the American and African amphibia are ‘‘two different modifications of the same earlier fauna.’ Mr. Robert Dunlop (12) has given some interesting notes on Carboniferous and other Paleozoic amphibia of Scotland con- tained in the Kilmarnock Museum before they were destroyed by fire. His notes are accompanied and illustrated by two excel- lent half-tone plates of photographs of type specimens of Loxomma, Pteroplax and Anthracosaurus, all of which is very welcome information. Jaekel (14) has proposed a new classification for the Chordata _ which he calls Vertebrata. He divides the ‘‘Stamm’”’ into three subgroups Tetrapoda, Pisces and Tunicata, and makes no allow- ance for amphioxus. He proposes two new classes of ‘‘Tetra- poda,” Hemispondyla and Microsauria with the ordinary classes Amphibia, Reptilia, Aves and Mammalia. The forms he groups in his new class Hemispondyla are the branchiosaurs and a new _ group which he calls Sclerocepholi. Dr. Jaekel has made several bad blunders in this classification. The first one is to separate the branchiosaurs from the Amphibia, to which they belong with- out the slightest shadow of a doubt. The next one is the alliance of Amphibamus to the Branchiosauride, to which it is not so closely allied as it is to the Cotylosauria. Amphibamus is far removed in structure from the Branchiosauride. His next error is the inclusion of Acanthostoma in the same group with the Branchiosauride. Their structures do not indicate relation- ships at all. His class Microsauria is wholly untenable, as Dr. Williston well says (17). The group which we call Microsauria now will un- _ doubtedly require revision and it looks as if it were going to get No. 534] NOTES AND LITERATURE 383 it, but that the animals now included in that group represent a class distinct from all other vertebrates I, for one, will not for a moment concede. The fundamental error made by Dr. Jaekel, as the reviewer sees it, is the attempt to base a classification of vertebrates on a single character. This has always failed in the the past and must, in the nature of the case, fail in the future; since classifica- tion, if it is to mean anything, must take into consideration the entire organization. The paper is full of many other smaller errors, errors of knowledge and errors of judgment. One of these errors is relating such widely distinct forms as Cerater- peton and Diplocaulus. The same author has given a study of the limbs of the oldest vertebrates in which (15) he attempts to sustain his classification, but his facts and arguments are not at all convincing and the paper is little ‘more than a republication of parts of the essays of other investigators. Dr. Williston (17) has recently published another essay on the Permian fauna of Texas in which he gives especially a study of the vertebre and adopts the view of Cope as to the ultimate fate of the elements of the rhachitomous vertebra. He regards Eosauravus copei Will. (Eosauravus punctulatus (Cope)) as allied to Hylonomus and for that reason ‘‘the oldest known reptile’’ is a microsaur. Just what his reasons for this alliance are he does not say. In the present imperfect state of our knowledge of Hylonomus and its Canadian brothers such a refer- ence would be very uncertain. In the last paragraph he records the interesting discovery of limbs in Lysorophus. A general review of the above essays shows that more than half of them represent pioneer work, that is, descriptive and classifi- catory investigations. Five of the essays bear more largely on the faunal relations as exhibited by the Amphibia. One gives us new light on the significance of a structure found in the ancient forms. This is where work is greatly needed. ir knowledge of the ancient amphibian fauna will increase as time goes on but the greater part of the pioneer work is already done. The way is now open for some good investigations on the struc- ture of the ancient Amphibia and the meaning of these characters as interpreted in the light of modern comparative anatomy and embryology. REFERENCES 1. Woodward, A. S. On a new Labyrinthodont from Oil Shale, at Airly, New South Wales. Records of the Geological Survey of New South Wales, Vol. VIII, Pt. IV, 1909, pp. 317-319, Plate LI. 384 THE AMERICAN NATURALIST [ Vou. XLV 2, a. S. W. Dissorophus, Cope. oe of Geology, Vol. XVIII, No. 6, pp. 526-536, Plates I-III, a Williston, E W. New or Little-known gane Vertebrates, Trematops, : New bss Journal of Geology, Vol. XVII, No. 7, pp. 636-658, 909. 4. Williston, S. W. The Skull and Extremities of Diplocaulus. Transac- tions of the oe Academy of Science, Vol. XXII, pp. 122-131, Plates I-V, 1909. 5. Williston, S. W. Cacops, Desmospondylus, New Genera of Per Vertebrates. Bulletin of ep Geological Society of America, Vol. a, pp. 249-284, Plates 6-17, 1910. 6. Case, E. ©. ner ew or Little- ot epe and Amphibians from the Permian (?) of Texas. Bulletin of the American Museum of Nat- ural History, Vol. XXVIII, Actes XVII, pp. 163-181, Figs. 1-10, 910 ¢ Beans, E. B. Amphibian Footprints from the Mississippian of Vir- ginia. pana = Geology, Vol. XVIII, No. 4, pp. 356-358, 1 figure, 1910. 8. Thevenin, Armand. Les plus mt EPAM de ma Annales Paleontologie. iam V, pp- Plates I-IX 9. EAS Friedrich von. Neubeschre P ie des pe mish ae Dasyceps bucklandi (Lloyd) aus Kenilw nl yer logische und Pale- ontologische Abhandlungen a geb E. Koken. Neue Folge, Bd. VIII, Heft 6, pp. 325-337, Pls mi "XLIV-XLY, 1910. 10. Broom, R. A Comparison of A mech Reptiles of North America with those of South Africa. Bulletin of the American Museum of Natural History, Vol. XXVIII, Art. XX, pp. 197-234 m, R. On th u to those of other Parts of the Wor g Transactions of the Royal Society of South Africa, Vol. I, Part 2, pp. 473—477. 12. CENE Robert. The Fossil Amphibia in the Kilmarnock Museum , 13. Schönfeld, G. Bericht über einen neuen Stegocephale aus d sichischen Rothliegenden und a entwickelungsgeschicht. Stellgang er Stegocephalen. Isis, Dres PP-, F 14. Jk Otto. Ueber die Klassen "Ge Tetrapoden. Zoologischer An- r, Band XXXI . 7/8, pp. 15. Tail. fo Ueber die ältesten Gliedm n Tetrapoden, ae ungsberichte des Gesellschaft Natw SE Freunde zu Ber No. 10, pp. 587-615, with 20 figures, 1909. 16. Williston, S. W. The Faunal Relations of the Early Vertebrates. Jour- nal oF Geology, Vol. XVII, No. 5, pp. 389-402, 1 figure. Published also in ‘‘Outlines of Geologie WAA: Willis and Salisbury, PP- 163-175. 17. Williston, S. W. New Permian Reptiles: Rachitomous Vertebræ. Jour- nal of Geology, Vol. XVIII, No. 7, pp. 585-600, 1910. Roy L. MOODIE. UNIVERSITY OF eee The American Journal of Science Established by Benjamin Silliman in 1818 The Leading Scientific Journal in the United States Devoted to the Physical and Natural Sciences, with special reference to Physics, and Chemistry on the one hand, and to Geology and Mineralogy on the other. ins, Editor: EDWARD S. DANA. es Editors: Professor GEORGE L. GOODALE, JOHN TROWBR nnar, W. G. FARLOW and WM. M. DAVIS yn eg e; rnas sors A. E. RILL, HENRY S. WILL S and L. V. PIRSSON, of New ; Professor 0. F. BARKER, of Philrdelphia ; promeaer JOSEPH S. AIMES, e imore; MR. J. S. DILLER, of Washington. Two volumes annually, in monthly numbers of about 80 pages each. This Journal ended its first series of 50 volumes as a quarterly in 1845; its second series of 50 volumes as a two-monthly in 1870; ge its third series as a monthly ended Dec- ember, 1895. A Fourth Series commenced in 1 Subscription price, $6 per year or 50 cents a BEET postage prepaid in the United States ; $6.25 to Canada ; $6,40 to Countries in the Postal Union. Back numbers at reduced prices. s&@Ten-Volume Indexes, Vols. I-X XI-XX, fourth series, price one dollar. Address The American Journal of Science New Haven, Conn. SECOND EDITION, NOVEMBER, 1910 AMERICAN MEN OF SCIENCE A BIOGRAPHICAL DIRECTORY EDITED BY J. McKEEN CATTELL A Biograph iphical d Sen requires revision if it isto maintain its usefulness. owen pi of the names in the cena edition are new á hie mata which a Babe in the ge division have in revised. The Chaat of work required i o ore re the revision has been as great as that given to the free ‘dion. There has been no : : ni scope t : - s eet sciences, and for this — a fos — included in the first edition have been omited. Efforts st been exerted to as com ssi men will not ethods that wer R e, and stars sane been added to the #1 subjects of research in the case pi 269 new men who iene al pl plases on. on the li list. pa The editor’ s object in selecting this group of scientific men has been to make a study of — which scientific research depends and so far as may be to improve these conditions. ‘There are printed in appendiz the two statistical studies that have been made. — From the Preface to the Second Edition. The seeond edi 5 Te; ition $ the Directory extends to more than 600 pages and contains more than 5500 sketches. ks well Printed on all rag paper ae bound in buckram with leather label. ae, the work has been sed in size by more than 50 per cent., it is roid at the same price as the first edition Price: Five Doliars, net, Postage paid THE SCIENCE PRESS GARRISON, N. Y. LANCASTER, PA. SUB-STATION 84, NEW YORK CITY. The American Naturalist A Monthly Journal, established in 1867, “tired to the A with Special Reference to th dvancement of the a Sciences of Organic Evolution and Heredity CONTENTS OF THE DECEMBER NUMBER aap i = = Pigmentation in Man. GERTRUDE C. and CHARLES B. DAVENPORT. Spawn aa: Larva of coma Jeffersonianum. Pro- fessor W. H., PIER: The Inheritance of Sizes a Shapesin Plants, Professor R. A. EMERSON, Shorter Articles and Discussion : The Modification of rome erat ee cae aa External Conditions. Notes and Literature: Heredity, Dr. W. J. SPILLMAN. Index to Volume XLIV., CONTENTS OF THE JANUARY NUMBER Somatic Alteration: Its Origination and Inheritance, Dr. D. T. MacDougat. The Nature of ones ers Professor DOUGLAS Ho UGHTON CAM A Double Hen’s Egg. Dr. J. THOMAS PATTERSON. Notes and Literature: Heredity, Dr. W. J. SPILLMAN, CONTENTS OF THE FEBRUARY NUMBER The Application of the ne — of Pure Lines to Sex Limited Inheritan Sexual Dimorphism, Professor T. H. reed igs sort in ah ye yg a ofa Genetics in Lower Organisms. Siac’ cach of Temperature acca Growing Mice, and soars Seago o — Effecis in a Subsequent Gen- Dr. Francis B. SuMN The Mendelian 1 Ratio aF Blended Inheritance. SHINK- A Solf. Diis on ths ERAN Conspicuousness of Barred colored Fowls. Dr. RAYMOND PEARL. Some Caia — the eee ic Fu tion in Marine Organism fA ~ MeDants ore: stage 9 — s goi “Diseuani on: Oom mating Pos ger on in Cases w metrica 1" Ty abl monly used. Professor H. 8. JENNIN ‘gether: CONTENTS OF THE MARCH NUMBER The Gen mony Be ae aps of Heredity. Professor W. Jon The Pcie See Hypothesis and Hybridization. Pro- E. M. East. Notes on — and spre EALEY D Notes and Literature: Mimicry, Dr. Frank E. LUTZ. Dr. WILLIAM A CONTENTS OF THE APRIL NUMBER Genetical Studies on Oenothera. I. Dr. BRADLEY Moore Davis. The Genotypes of Maize. Dr. GEORGE HARRISON HULL, Notes and Literature: Is the Female Frog Heterozy- gow s in regard to Sex Determination ? Professor T. H. Morean, The Mutation Theory, Dr. R. RB. TES, ENTS OF THE MAY NUMBER The Inheritance of Polymorphism = = in Colias hilodice. Professor JOHN H. GEROULD. ‘cae and Cytoplasm in Her aes Professor MICHAEL F. GUYER. ¢the Phowr A eee Study of the Structure 0 the Pho enic Organs of “certain Anes can Lampr" rida. ee McDermotr and C G. CRAN aban Bhorter 4 Articles and iga g awa yo we Benoa i in the English Sparrow- etrics, Yule’s Introdue- Notes and aiea: Biom Dr RAYMOND LAS Theory of Statistics. Single Number 35 Cents The NATURALIST will be sent t OO O Yearly Sina $4.00 sh THE SCIENCE PRESS Sub-Station 84: NEW YO i ea E four months f iU: One , Pa VOL. XLV, NO. 535 i JULY, 1911. THE AMERICAN NATURALIST A MONTHLY JOURNAL Devoted to the Advancement of the Biological Sciences with Special Reference to the Factors of Evolution CONTENTS Pag Germ-cell Determinants and Their Significance. Professor R. W. HEGNER - 385 em II. Further Observations on the Pose of the Sauropodous Dinosaurs. OLIVER P. We a ee ee ee ee HI. Shorter Articles and Correspondence: Computation of the Coefficient of Cor- relation. Professor H. a JENNINGS. Note on Batrachoseps attenuatus Esch. Dr. C. V. BURK - - = « - - < - 413 IV. Notes and Literature: Some Recent Studies on m and Correlation in Agricultural Plants: Dr. RAYMOND PEAR On Sex-chromosomes in Hermaphroditism : Professor W. E. CASTLE. aa on Ichthyology : Presi- dent DAVID STARR JORDAN. Some Recent zew on er Plants: Pro- fessor DOUGLAS HOUGHTON CAMPBELL - -415 THE SCIENCE PRESS LANOASTER, PA. GARRISON, N. Y. NEW YORK: SUB-STATION 84 e American Naturalist MSS. intended for publication and books, etc., intended for review ee be sent to the sages THE AMERICAN NATURALIST, Garrison-on-Hudson, New York. research work bearing on = presna of organi evolu- tion are especially welcome, and will be given preference in publicati _ One hundrea reprints of contributions are coal te authors free of charge. Further reprints will be supplied at cost. ~ Subscriptions and advertisements —_— be sent to the publishers. The subscription price is four dollars a year. reign postage is fifty cents and Canadian postage t -five cents Jaitional.. "The charge for single copies is thirty-five cents. The rtising rates are Four Dollars for a page. = THE SCIENCE PRESS | - Pa. Se Garrison, N. Y. NEW YORK: Sub-Station 84 Bae pepe A EE : : pril 2, 1908 ter, Pa., under the Act ot ao s s = prae 1879. TENTH EDITION. THE MICROSCOPE, THE AMERICAN NATURALIST VoL. XLV July, 1911 No. 535 GERM-CELL DETERMINANTS AND THEIR SIGNIFICANCE! PROFESSOR R. W. HEGNER UNIVERSITY OF MICHIGAN Investications of the origin of the germ cells in a num- ber of animals have brought forth certain phenomena which indicate that these cells are determined as such at a very early period in embryonic development, and that im some cases the material which apparently determines the germ cells is visible at this time. Conclusions can be drawn from these observations which are of consider- able interest. The frequently repeated statement that the germ cells are derived from the mesoderm or from the entoderm 1s of course erroneous in those instances where the germ cells can be identified before the formation of the germ layers, and it seems probable that the primary cell differ- entiation, i. e., the separation of the germ cells from the Somatic cells, takes place at an early period in the em- bryonie development of even those animals where this has not been actually observed. A few of the most pro- nounced cases of the early differentiation of germ cells are briefly described in the following paragraphs and Several general conclusions arrived at from this evidence. The best known example is Ascaris, as described by * Contributions from the Zoological Laboratory of the University of Michigan, No. 135. From a paper read before the Research Club of the University of Michigan, November 9, 1910. 385 386 THE AMERICAN NATURALIST [Vou. XLV Boveri (792). The first cleavage division of the egg of Ascaris results in two daughter cells, each containing two long chromosomes (Fig. 1, 4). In the second division the chromosomes of one cell divide normally and each daughter cell receives one half of each (Fig. 1, B, s). The chromosomes of the other cell behave differently ; the thin middle portion of each breaks up into granules matin in A, two-cell stage dividing; s, stem-cell, from which arise the germ cells. B, the same from the side, later in the second cleavage, showing the two types of mitosis and the casting out of chromatin (c) in the somatic cell. ©, resulting four-cell stage; the eliminated chromatin at ¢. D, the third cleavage, repeating the foregoing process in the two upper cells. (Fig. 1, A) which split, half going to each daughter cell, but the swollen ends (Fig. 1, B, c) are cast off into the cytoplasm. In the four-cell stage there are consequently two cells with the full amount of chromatin and two with a reduced amount. This inequality in the amount of chromatin results in different sized nuclei (Fig. 1, C); those with entire chromosomes (s) are larger than those that have lost the swollen ends (c). In the third division No. 535] GERM-CELL DETERMINANTS 387 one of the two cells with the two entire chromosomes loses the swollen ends of each; the other (Fig. 1, D, s) retains its chromosomes intact. A similar reduction in the amount of chromatin takes place in the fourth and fifth divisions and then ceases. The single cell in the thirty-two-cell stage which contains the full amount of chromatin has a larger nucleus than the other thirty-one cells and gives rise to all of the germ cell, whereas the other cells are for the production of somatic cells only. The primordial germ cell of Ascaris, therefore, con- tains two entire chromosomes; every other cell possesses two chromosomes which have lost part of their substance. In other words, the germ cells possess a certain amount of chromatic material not present in the somatic cells. There is also an early differentiation of the germ cells in the fresh water crustacean, Cyclops (Haecker, ’97). According to Haecker, ‘‘Aussenkérnchen”’ arise at one pole of the first cleavage spindle (Fig. 2, A, ak); these A Fig. 2. Origin of the primordial germ cell in Cyclops. (From Haecker.) A, the first cleavage division, showing the “ Aussenkérnchen” (ak) at one pole of the spindle. B, the thirty-two-cell stage; the primordial germ cell (Kz) con- tains all of the “ Aussenkérnchen ” (ak) are derived from disintegrated nucleolar material and are attracted to one pole of the spindle by a dissimilar Influence of the centrosomes. During the first four cleav- age divisions the granules are segregated always in one cell (Fig. 2, B, kg); at the end of the fourth division these ‘‘Aussenkérnehen’’ disappear, but the cell which con- -tained them ean be traced by its delayed mitotic phase and is shown to be the primordial germ cell. In this ease, as in that of Ascaris, the primordial germ cell and the germ cells derived from it possess certain nuclear materials not present in the somatie cells. The 388 THE AMERICAN NATURALIST [Vou. XLV latter seem to be limited because of their absence to the performance of vegetative functions, and the germ cells appear to have the power of reproduction because of their presence. A recent paper by Elpatiewsky (1909) deals with the early embryonic development of the arrow worm, Sag- itta. This investigator finds that, at the stage when the two pronuclei are in the center of the egg, a body appears at the vegetative pole lying near the periphery (Fig. 3, Fic. 3. Origin of the primordial germ cells in Sagitta. (From Elpatiewsky. ) A, egg with conjugating pronuclei; polar bodies (I Rk. and II Rk.) and “ beson derer körper” (æ) embedded in cytoplasm. B, thirty-two-cell stage; the jt mordial germ cell (G) contains the “besonderer vo et eye estas cell t it is also dividing e Epei cell has already divided (E). D, two germ cells (G) resulting from the first division of the primordial germ cell; each con- tains part of the “ besonderer soning (@).. A). This body, which is termed ‘‘besondere körper,” consists of coarse granules which do not stain quite so deeply as the chromosomes. During the first four cleav- age divisions the ‘‘besondere kérper’’ does not divide, but is always to be found in one blastomere. In the fourth division the blastomere which contains this body divides unequally; the larger cell is destined to produce the entoderm (Fig. 3, B, E); the smaller cell, which con- No. 535] GERM-CELL DETERMINANTS 389 tains the ‘‘besondere kérper’’ is the primordial germ cell (Fig. 3, B, G). The first division of this primordial germ cell (Fig. 3, C) results in two daughter cells, one of which obtains a larger portion of the ‘‘besondere kör- per’’ (X) than the other (Fig.3,D). This is interpreted as the differential division, the cell which possesses the larger amount of the divided ‘‘besondere kérper’’ giv- ing rise to the male germ cells, the other to the female germ cells in the hermaphroditic adult. The ‘‘beson- dere kérper’’ now gradually becomes paler and finally disappears. Buchner (710) and Stevens (’10) have con- firmed Elpatiewsky’s observations. The origin of the ‘‘besondere kérper’’ was not determined. of the egg. B, an egg containing a number of cleavage nuclei. ©, the cells which come under the influence of the nucleolus (N) become the primordial germ cells (G) . These investigations show the germ cells of Sagitta to be similar to those of Ascaris and Cyclops in that they contain a darkly staining material not present in the Somatie cells. In Oophthora and other parasitic hymenoptera, Syl- vestri (’09) finds that the nucleolus of the germinal vesicle passes to the posterior end of the egg during maturation (Fig. 4, A). Here it remains until the cleay- age nuclei reach the periphery (Fig. 4, B). The cells which then come under the influence of the nucleolus be- 390 THE AMERICAN NATURALIST [Vou. XLV come the primordial germ cells (Fig. 4, C, G) and give rise to the germ glands of the adult. The similarity between this process and that described for Ascaris, Cy- clops and Sagitta is obvious. Finally in chrysomelid beetles the primordial germ cells are differentiated at a very early period (Hegner, 09). At the posterior end of the eggs of Calligrapha multipunctata and allied species there is a disc-shaped mass of granules which stain like chromatin. I have called this the pole dise (Fig. 5, A, gc. d). When thecleavagenucleireach the periphery of the egg they fuse with the super- ficial layer of cytoplasm everywhere except at the posterior end; cell walls then appear and a blastoderm is formed. ---gn When the cleavage nu- clei which reach the pos- __khpt terior end of the egg en- counter the pole disc y. granules they gather these granules about ~m them and continue their migration (Fig. 5, B, gc); cell walls are formed, and they finally come to lie entirely out- side of the egg (Fig. 5, C, gc). There are sixteen cells ged. which separate from the P egg in this manner, and : bie they take out of the egg with them practically all of the pole dise granules (Fig. 5, C, pd.g). These sixteen cells divide to form thirty-two ; in- this division apparently one half of the granules con- tained in each cell pass to each of the daughter cells (Fig. No. 535] GERM-CELL DETERMINANTS 391 5, D, pd.g). A second division results in sixty-four cells; this number is constant until a late stage in embryonic development. These sixty-four cells have been traced through the early embryonic stages. First they migrate back into the egg through a ‘‘pole cell canal” (Fig. 5, C, pe) sit- Fic. 6. Origin of the primordial germ cells in chrysomelid beetles. A, longi- tudinal section through a freshly laid egg of Call.grapha bigsbyana, showing pole (ge.d) at posteri longitudinal section through the pos erior end of ultipunctata eighteen hours after deposition in e disc granules within the primordial ge ge). O, udinal section i i lon: through the posterior end of an egg of C. bigsbyana twenty-four hours after depo- Sition, ion the primordial germ cells containing pole dise rennin (pd.g). D, a mordial germ cell of C. multipunctata in anaphase of mitosis E p= disc pes a (pd.g) are apparently equally ae ae at either end. germ cells (gc) and neighboring mesoderm cells (m) and ectoderm cells a son an embryo of Ọ. multipunctata. F, longitudinal ae ih rough an egg of Lep- tinotarsa decemlineata twenty-four hours after the posterior er had been killed with a hot needle, any preventing the pole disc granules (pd.g) from taking part in the developmen No germ cells are formed sensi AR with C). bl= blastoderm ; vita spea cell; ge =germ cell; ge.d=germ cell So gn = conjugating pronuclei ; khbl = keimhautblastem ; m = mesoderm p= posterior end of egg; pbl = pseudoblastodermic seers p.c = pole oat fae pd.g = pole disc granules; vm =vitelline membrane; y= yolk. 392 THE AMERICAN NATURALIST [Von. XLV uated near the posterior end of the ventral groove. Then they separate into two apparently equal groups, one on either side of the embryo, which are soon recognizable as the two germ glands. Because only those cells which gather in the pole dise ‘granules become germ cells, I have called these granules ‘‘germ cell determinants’’ (Hegner, ’08). This term has been objected to by Wieman (710) because ‘‘the term im- plies the attribute of certain potentialities that these granules have not been shown to possess’’ (p. 180). The morphological evidence is, I believe, strong enough to warrant the use of the term; recent experiments, how- ever, add to the convincing facts already published (Hegner, ’08, 09). It is possible to show that if the pole disc is prevented from taking part in the development of the egg, no germ cells will be produced. Attempts to extract the pole dise by means of pricking the freshly laid egg and allowing them to flow out were only partially successful (Hegner, ’08). A new method was later em- ployed which absolutely prevented the cleavage nuclei from encountering the pole disc. In these experiments the posterior end of the egg was touched with a hot needle and that portion containing the pole dise was killed. In every instance the development continued and in the eggs so far examined the blastoderm formed normally over all of the surface except at the posterior end; here it was built at the end of the living substance as shown in Fig. 5, F, bl. No germ cells were produced. I conclude from this that the pole dise granules are necessary for the formation of germ cells, and that they are really ‘‘germ cell determinants.’’ Of course it might be argued that some other substance lying at the posterior end of the egg is responsible for the differentiation of the germ cells, but this seems highly improbable. Wieman (710) states that in Leptinotarsa signaticollis, a species I have not studied, ‘‘the granules are not all taken up by the cells in their migration and the greater part of them remains behind after the cells have passed through’’ (p. 186). This is certainly not the case in the many eggs that I have examined, and a reexamination No. 535] GERM-CELL DETERMINANTS 393 shows that only a few of the pole dise granules remain in the egg after the germ cells are formed, as was clearly pointed out in a former paper (Hegner, ’09, Plate II, Fig. 16). The origin of the pole dise granules is not known. It seemed to me probable that they came from the nucleus of the egg just before maturation and consisted of nuclear material. This conclusion was reached (1) because these granules stain like chromatin, (2) because in many insects the nucleus of the oogonium casts out chromatic material (Nebenkerne), and (3) because the substance which deter- mines the germ cells in Ascaris, Cyclops and Oophthora is of nuclear origin, and in one case (Ascaris) is chro- matin. Wieman believes that ‘‘the granules of the pole dise consist of particles derived from the food stream of the ovum that form an accumulation in the protoplasm in its posterior part’’ (p.187). This possibility was pointed out in a former paper (Hegner, 09, p. 274), a fact Wie- man seems to have overlooked. It was also suggested in the same place that if the granules are derived from the nurse cells they probably come from the nuclei of these cells. The pole disc granules gradually disappear after the germ cells are formed. It may be of interest to mention the results of opera- tions performed upon eggs in which the germ cells had already differentiated at the posterior end (Fig. 5, C). Such eggs, when touched with a hot needle, continued to develop, and produced embryos and larve without germ glands. This I believe is the earliest stage on record on which surgical castration has been performed. The visible presence of germ cell determinants in the primordial germ cells of the animals described above sug- gests two possibilities as to their importance: (1) They may represent idiochromatin, i. e., germ plasm, or (2) they may influence the metabolism of the cells and thus determine their character. 1. The history of the germ cells in chrysomelid beetles illustrates in a remarkable way the theory of germinal continuity as expressed by Weismann (’04). Weismann believes with Niigeli that ‘‘there are two great categories a 394 THE AMERICAN NATURALIST [Vou. XLV of living substance—hereditary substance or idioplasm, and ‘nutritive substance’ or trophoplasm, and that the former is much smaller in amount than the latter’’ ( Weis- mann, 704, Vol. I, p. 341). The idioplasm of the germ cells he calls germ plasm, a substance which is ‘‘never formed de novo, but it grows and increases ceaselessly ; it is handed on from one generation to another like a long root creeping through the earth, from which at regular distances shoots grow up and become plants, the indi- viduals of the successive generations” (Vol. I, p. 416). ‘This splitting up of the substance of the ovum into a somatic half, which directs the development of the indi- vidual, and a propagative half which reaches the germ cells and there remains inactive, and later gives rise to the succeeding generation, constitutes the theory of the continuity of the germ plasm (Vol. I, p. 411). Accord- ing to this theory, the body or somatic cells serve only to protect, nourish and transport the germ cells which con- tain the germplasm. Later the germ cells separate from the body and develop into new individuals and the body subsequently dies. In the eggs of chrysomelid beetles the germ cells are - formed at an extremely early period in embryonic devel- opment. They separate entirely from the embryo and come to lie in a group at the posterior end; at this time germ cells are quite distinct from somatic cells. Later the germ cells migrate back into the embryo, where they are protected, nourished and transported until they be- come mature, leave the body and give rise to a new gen- eration. What particular part of the germ cell represents the idioplasm or germ plasm? is a question of fundamental importance. Weismann recognizes the chromosomes as the germ plasm and has built up a complex theory as to the constituents of these bodies. The present discussion is not concerned in any way with the structure of the germ plasm as conceived by Weismann, and the writer does not wish to become involved in a consideration of idants, ids, determinants and biophores. The theory of dichromaticity (Dobell, 09) may aid in answering this No. 535] GERM-CELL DETERMINANTS 395 question. This theory holds that the chromatin of the germ cells is of two kinds—(1) idiochromatin, which is for reproductive purposes, and (2) trophochromatin which performs vegetative functions. In many Protozoa these two kinds of chromatin are separate throughout the life cycle. For example, in Paramecium the micro- nucleus is thought to represent the idiochromatin, the macronucleus, the trophochromatin (Calkins, ’09). Dur- ing conjugation and the subsequent reorganization of the nuclear apparatus the macronucleus breaks down and disappears, whereas the micronucleus gives rise not only to new bodies like itself, but also to new macronuclei. In most animals idiochromatin and trophochromatin are supposed to be contained in one nucleus and are in- distinguishable except in a few cases during the differ- entiation of the germ cells at an early developmental period of the egg. One is tempted to interpret as idio- chromatin (1) that part of the chromosomes of Ascaris which is lost by the somatic cells (Fig. 1, B,c) but retained by the germ cells, (2) the nuclear material which is pres- ent in the primordial germ cell of Cyclops (Fig. 2, B, ak) but is absent from the somatic cells, (3) the similar sub- | stance in the primordial germ cells of Oophthora (Fig. 4, n), (4) the ‘‘besondere kérper’’ in the egg of Sagitta (Fig. 3, x), and (5) the pole dise in the eggs of chryso- melid beetles (Fig. 5, A, ge.d). One difference between these substances and the germ plasm as Weismann conceives it should be pointed out. In the cases cited above the material interpreted as germ plasm is only in one instance chromatin, and in this animal (Ascaris) it does not constitute the entire chro- matin as Weismann’s theory demands. If these extra nuclear bodies really represent the iodioplasm our loca- tion of the germ plasm must be transferred from the chromosomes to this material. 2. The second theory mentioned above, namely, that the extra material possessed by the germ cells determines these as such because of some fundamental principle of metabolism, seems more plausible than the theory just outlined. It is worthy of note that the primordial germ 396 THE AMERICAN NATURALIST [ Vou. XLV cells of several animals belonging to widely separated groups are supplied with extra nutritive material. This is true in the Diptera, Chironomus (Weismann, ’63) and Simula (Metschnikoff, ’66), in the Lepidopteron, Endro- mis (Schwangart, ’05), in the Elasmobranchs (Beard, 702), in the Teleosts (Eigenman, ’92), in the Amphibia (Nussbaum, ’80), and in the Reptilia (Allen, ’06). It has already been pointed out (Hegner, ’09) that the pole dise granules may be nutritive material. ‘‘That the pole-cells need special means of nourishment is doubtless the case, for contrary to the condition in the blastoderm cells, they are at an early period entirely separated from the yolk, and later use up energy in their migration” (p. 275). If this is true, and as Wieman (710) claims, the pole dise granules are derived from the yolk stream, our germ cell determinant hypothesis is not weakened, but gains a distinct argument in its favor. It is interesting to note in this connection that two of the foremost investigators of the relation of the acces- sory chromosomes to sex determination are inclined to believe in the quantitative hypothesis, i. e., that the egg which is fertilized by the spermatozoon containing the accessory develops into a female because there is more chromatin present, and that this plus amount influences the metabolism of the cell and its descendants (Wilson, "10; Morgan, 710). This hypothesis suggests the theory of sex advocated by Geddes and Thomson (’89), that ‘‘the female is the outcome and expression of prepon- derant anabolism, and in contrast the male of prepon- derant katabolism” (p. 132). In Sagitta (Elpatiewsky, ’09), however, it is the male primordial germ cell and not the female that acquires the larger part of the ‘‘beson- dere körper.” Although neither of the two possibilities advanced in the foregoing pages may be correct, nevertheless it seems certain that the peculiar bodies in the primordial germ cells of the animals described above should be named ‘‘germ cell determinants.” In any event, the attention of investigators ought to be directed toward the problem of discovering the origin and complete history of these No. 535] GERM-CELL DETERMINANTS 397 bodies, since their bearing upon theories of heredity is of fundamental importance. LITERATURE Allen, B. M., 1906. The Origin of the Sex-Cells of Chrysemys. Anat. .„ 1902. The Germ Cells. Part I. Raja Batis. Zool. Jahrb., 16 Boveri, T., 1892. Die Entstehung des Gegensatzes zwischen den Geschlechts zellen und den somatischen Zellen bei Ascaris megalocephale. Sitz. - Morph, Physiol. München, Bd. 8. eae, P., 1910. Keimbahn und Orogeno von Sagitta. Anat. Anz., . 35, Calkins, G. N., 1909. Protozoology. Philadelphia. Dobell, C. C., 1909, Chromidia and the Binuclearity Hypothesis: a Review and a Criticiam. Quart. Tous Micr. Sc., Vo 2. On Bigenan, C., 1892. the Precocious Segregation of the Sex-Cells of icrometus aggregatus. Journ. Morph., Blpatiswicy, W., 1909. Die Recor naieh anc Er bei Sagitta. Anat. 35. NZ., f Geddes, P., and Thomso , J. A., 1889. The Zvolaution of Sex. London. Haecker, V., 1897. Die ALEE von Cyclops. Arch. Mikr. Anat., Bd. 49 Hegner, R. W., 1908. The Effects of Removing the Germ Cell Determi- nants from the Boyt of Some Chrysomelid Beetles. Biol. Bull., Vol. 16. he R. W., 1909. The Origin and rhs History of the Germ Cells in some Pon is cD ourn. Morph., Vol. 20. Matson E., pi Rinlinviliipiedhé the Advancement of the — Sciences Factors o ity with Special Reference to the f Organic Evolution and H CONTENTS OF THE JANUARY NUMBER Somatic Alteration: Its Origination and Inheritance, . D. T, MacDoucat, The Nature of Graft- hybrids. Professor Dove.as HIUS A Double Hen’s Egg. DR. J. THOMAS PATTERSON, Xotes and Literature: Heredity, Dr, W. J. SPILLMAN, CONTENTS OF THE FEBRUARY NUMBER The opena of the Conception of Pure Lines to Sex me y the wees and to Sexual Dimorphism, Pro THM SaN Pure i in eis PSE o tics in L Professo bam = denies Some Effects of rof Such upon Growing Mice, and = Persistence = par ‘Effects In a Subsequent Gen- ration, Dr. s B. SUM The Mendelian 1 Af aoee Blended TR SHINK- Organismā Daai on em “Relative Comptia = or and Self- colored Fow . RAYMON Pome Considerations concerning aa ‘Photogenic Func- n Marine iien s. F. ALEX. MOTT, Shorter Articles Discussion : Computi a p tion in Caen pe Symmetrical abies are com” only used. Professor H. S. JENN CONTENTS OF THE MARCH NUMBER The reie Conception of Heredity. Professor . JOHANNSEN, was” Geno Hypothesis and Hybridizati P cen EL pos y ization. TO- a on Gundlachia and Ancylus. Dr. WILLIAM HEALEY DALL. Notes and Literature: Mimicry, Dr. Franx E. LUTZ, CONTENTS OF THE APRIL NUMBER mare pe on Oenothera. I. Dr. BRADLEY Moo. The ane = Maize. Dr. GEORGE HARRISON HUL Notes and Literature: Is the Female Frog Heterozy- gous in regard to Sex Determination ? Professor T. H. Morcan. The Mutation Theory. Dr, R. R GATES, CONTENTS OF THE MAY NUM The Inheritance of Polymorphism and Be r Coli Philodice. Pro Jace oun H. GEROU 3 2 ee in Sontag "poi A eee sot of the Structure of the hes ei oe rtain — Lamp Shorter peie sm apa cease oe ana. occas; A Neglected Paper on Natural Selection in the 1 parr . Dr. J. ARTHUR HARRIS, Siem wii otes and Literature: Biometrics, Yule's Introduc- ot the Theory of Statistics, Dr, RAYMOND CONTENTS OF JUNE NUM reece! a — in the aes pir Dr The poate Froot tot the Pure Line Theory. Dr. J. A be ged External Conditions 0B the Dev — of Two Species of Moss. Pro- fessor Tuomas H. MONTGOMERY. Shorter pE mi Discussion : The Ontogeny of # Genus. AUSTIN HoBART CLA Notes and Literature: Recent Contributions to & Knowledge of the Extinct Amphibia, Dr, ROY L. MOODIE. gee Number 35 Cents NATURALIST “11 Lio ——— ot Yearly Subscription, $4.00 One Dollar four months sue THE SCIENCE PRESS Sub-Station 84: NEW YORK er, Par ile VOL. XLV, NO. 536 ped t “ his, AUGUST, 1911 THE AMERICAN NATURALIST A MONTHLY JOURNAL Devoted to the Advancement of the Biological Sciences with Special Reference to the Factors of Evolution CONTENTS Page Pg baft Di bad Girė A Comparative Microscopic Study o i with special bee to the Questions of Color Inheritance among Mulattos. 449 Professor H. E, Jor - A Coefficient oat veal: ee for Students of Saget oh Dr. £ - 471 ArTHuR HA The Adaptations of the Primates. Professor F. B. LOOMIS Oe ye kas Jean Marchant, an Eighteenth Century Mutationist. Dr. HENRI Hus - - 493 Notes and Literature: Notes on Heredity. Dr. W. J. SPILLMAN THE SCIENCE PRESS LANOASTER, PA. GARRISON, N. ¥. NEW YORE: SUB-STATION 84 The American Naturalist intended for Pate and books, etc., intended for review should be MSS, sent to = Editor of THE A icles containing research MER sie NATURALIST, Garrison- -on-Hudson, New York work bearin n the problems - ‘organo evolu- tion are especialy welcome, and will be given sreterinoét in publica reprints of a are supplied to authors pe of charge. e hu F arier Fariáta “will be supplied at c Subscriptions and advertisements nin in sent to the publishers. The subscription price is four dollar: a yea Canadian postage twenty-five ph additional. thirty-five cents. Lancaster, Pa. postage is fifty cents and char eign "The arge ür sep copies is The advertising rates are Four Dollars for a pa THE SCIENCE PRESS Garrison, N. ¥. NEW YORK: Sub-Station 84 Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of Congress of March 3, 1879. THE BULLETIN—For bargains in Ethnolograph- ical and Pre-historic Specimens. Books on Natural History, Science, Travel, Voyages, etc. See THE BULLETIN post free for 3 cent stamp. 4 Duke St., Adelphi—i en¢on—England TENTH EDITION. THE MICROSCOPE, tion to Microscopic Meth thods and to Histology, = ane Rury aac po of Saas Universit is now 350 large octavo and above and fally szara pe e ea Price $2.00, postpaid. ` COMSTOCK PUBLISHING CO., Ithaca, N. *- o a BIRDS’ EGGS W. F. H. ROSENBERG, 57 Haverstock Hill, London, N. W., England begs to announce the publication of a new Price List (No. 15) of Birds’ Eggs, con- taining over 900 species from all parts of the world. This Catalogue is systematic- ally arranged, with authors’ names, indica- tions of localities, and an index to families. It will be sent post free on application, as will the following lists: No. 11, Birds’ Skins, (5,000 species) ; No. 12, Lepidoptera: (5,000 species) ; No. 18, meaty, No. 14, Mammals. I of preparation: New Price List of Repic eaS and Fishes. Largest Stock in the Ss of specimens in all branches of Zoolog Marine Biological Laboratory WOODS HOLE, MASS. Facilities a research in Zoolo ey; INVESTIGATION ests bore JUNE 1 TO OCTOBER 1 Earlene bles instruction Courses of laboratory ver- ectures are ofe erpryologY, a INSTRUCTION JUNE 28 TO AUGUST &, 1911 Tax ogy rse cts of Biolo Par anhe also offered terial —Presery: for cae Zoology— of animals SUPPLY be all ty PS DEPARTMENT OPEN THE ENTIRE Zool YEAR Ma rts and Moem ké all information ag GEO. M. GRAY, ciee Woods we? nual announcement will be as ae the! e Director, Marine Biological Borban e on apia THE AMERICAN NATURALIST VoL. XLV August, 1911 No. 536 A COMPARATIVE MICROSCOPIC STUDY OF THE MELANIN CONTENT OF PIGMENTED SKINS WITH SPECIAL REFERENCE TO THE QUESTION OF COLOR INHERIT- ANCE AMONG MULATTOS PROFESSOR H. E. JORDAN UNIVERSITY OF VIRGINIA INTRODUCTORY, WITH A STATEMENT OF PROBLEMS NVOLVED Tue comparative histologic study of pigmented skins was undertaken with the hope of discovering evidence that might throw more light on the problem of color inheritance among the descendants of crosses between whites and negroes. The primary point at issue, until very recently, was whether human skin color in inherit- ance conformed more closely to the alternative (Men- delian), or the blended scheme; or perhaps to the ances- tral (Galtonian) scheme. The studies of G. C. and C. -B. Davenport show conclusively that there is a measure of segregation among the individuals of the third gen- eration, hence a Mendelian-like inheritance. This study, suggested to me by the above-mentioned investigators, is, consequently, more especially an at- tempt to test, as critically as microscopic data will allow, the theory of discrete unit characters in color inherit- ance (discontinuity theory; segregation theory), as op- posed to the theory of continuity of the pigment char- acter with interruption of the pigmentation process at : 449 450 THE AMERICAN NATURALIST [Vou. XLV various stages. The study involves two incidental prob- lems: (1) Source of the epidermal pigment, (2) cause of degree of coloration of skin. These matters must re- ceive consideration first. Then follows a discussion of the bearings of the determined facts on the question of color inheritance in crosses between whites and blacks, or mulattos. Material—The material studied comprises 18 pieces of skin taken from near the mid-line of the abdomen, including 5 pieces from full-blooded negros of varying grades of color, 6 pieces from mulattos of various shades, 4 pieces from brunets, 2 from blonds, and 1 piece Fic. 1. Camera lucida drawing of unstained section of negro skin (speci- men No. 14), showing the distribution of the pigment granules in the epidermis. Nuclei drawn according to their appearance in stained preparation. Dermal eleidin and keratohyalin granules of the stratum granulosum not shown. x 750. of pathologically pigmented skin; also a melanotic sar- coma, several pieces of pigmented skin of incomplete history, sections of infant’s scalp and eyelid of new- born mulatto. Individuals were classified as blond or brunet according to color of hair. Mulattos were so ad- judged in most cases from general appearance, ù. e- absence of distinct negro features, e. g., thick lips, flat nose, etc., though in several cases the individual con- No. 536] COLOR INHERITANCE 451 tributing the skin admitted to me being ‘‘mixed.’’ All ut one piece of mulatto skin were tested by the Bradley color top before embedding. In only one case (No. 2) could the individual contributing negro skin be ques- tioned by myself in regard to purity; the remaining four pieces are so classified on the assurance of pathol- ogist and surgeon. Four pieces were tested with the color top. Sections were cut both in celloidin and par- affin, and studied unstained and stained (with 1/12 oil immersion lens) in carbol fuchsin or the hematoxylin and eosin combination. The following is a list of the specimens and the essential points of interest regarding each: 2. Drawing of section of light brown mulatto skin (specimen No. 3) showing distribution of the pigment granules in the rete mucosum Malpighi. Darker and lighter mulatto skins differ from this only in the greater and smaller number, respectively, of melanin granules. 0. 1. Pure negro; from postmortem by Dr. H. Marshall (professor of pathology, University of Virginia) ; Zenk- er’s fixation; color (matched by mixing yellow, white, red and black on the Bradley color top)—yellow, 4 per cent. ; ; white, 8 per cent.; red, 22 per cent.; black, 66 per cent. 2. Pure negro woman; from abdominal operation by Dr. Harvey Stone (adjunct professor of surgery) ; fixed in 95 per cent. aleohol; a shade lighter than No. 1. 3. Light mulatto woman; abdominal operation by Dr. Stone; 95 per cent. alcohol fixation; color—yellow, 10 per cent.; white, 18 per cent. ; ; red, 22 per cent.; black, 50 per cent. (Fig. 2). 452 THE AMERICAN NATURALIST [Vou. XLV 4. Light mulatto; abdominal operation by Dr. Stone; 95 per cent. alcohol fixation; a shade darker than No. 3. 5. Blond; abdominal operation by Dr. Stone; 95 per cent. alcohol fixation. 6. Pathologically pigmented (grayish yellow) skin of white individual; postmortem (thyroid adenomata) by Dr. Marshall; 95 per cent. alcohol fixation. 7. Brunet; abdominal operation by Dr. W. H. Good- win (adjunct professor of surgery); 95 per cent. alcohol fixation. Fig. 3. Drawing of section of skin of blond (“cross between blond and brunet ’"—specimen No. 13). In lighter blonds (e. g., specimen No. 5) fewer basal cells contain still fewer granules; in darker brunets the basal cells contain -more granules. x 750. 8. Light mulatto male (age 80); postmortem (acute eystitis) by Dr. W. Thalhimer (instructor in pathology) ; Zenker’s fixation; color—yellow, 10 per cent.; white, 17 per cent.; red, 30 per cent.; black, 43 per cent. 9. Pure negro; abdominal operation by Dr. Stephen H. Watts (professor of surgery)—95 per cent. alcohol fixation. 10. Brown mulatto boy; abdominal operation by Dr. S. H. Watts; 95 per cent. alcohol fixation. 11. Brunet; abdominal operation by Dr. S. H. Watts; 95 per cent. alcohol fixation. 12. Brunet female (gray); postmortem (goitre and old age) by Dr. W. Thalhimer; Zenker fixation. 13. Dark blond male—age 30 years—(‘‘cross between blonde and brunet’’); from postmortem (pyemia and perinephric abscess) by Dr. W. Thalhimer; 95 per cent. alcohol fixation (Fig. 3). 14, Pure negro; abdominal operation by Dr. S. H. No. 536] COLOR INHERITANCE 453 Watts; 95 per cent. alcohol fixation; color—yellow, 5 per cent.; white, 7 per cent.; red, 19 per cent.; black, 69 per cent. (Fig. 1). 15. Pure negro; abdominal operation by Dr. S. H. Watts; 95 per cent. alcohol fixation; slightly lighter than No. 14. 16. Brunet; abdominal operation by Dr. S. H. Watts; 95 per cent. alcohol fixation. 17. Mulatto; abdominal operation by Dr. S. H. Watts; 95 per cent. alcohol fixation; color—yellow, 14 per cent.; white, 14 per cent.; red, 35 per cent.; black, 37 per cent. 18. Mulatto; abdominal operation by Dr. S. H. Watts; 95 per cent. alcohol fixation; a shade lighter than No. 17. Ranged according to degree of coloration, judged macroscopically, the 17 pieces of skin (omitting AN 6) take the following order: Negro 9, 14 (B, 69), 1 (B, 66), 2, 15. Mulatto 4, 3 (B, 50), 8 (B, 43), 17 (B, 37), 10, 18. Brunet 7, 11, 12, 16, Blond 13, 5. This list could have been enlarged indefinitely, but more material was not deemed requisite to the demands of this study. Number 6 took rank between numbers. 18 and 7. Sources of Epidermal Pigment.—There are obviously three distinct views which might be held—and as a mat- ter of fact have been held—in regard to the origin of the pigment of the skin: (1) In the epithelial cells of epi- dermis; (2) in the connective tissue cells of the dermis, and secondarily transferred to the cells of the rete mucosum Malpighi; (3) in both the epithelial and con- nective tissue cells. The second view more especially further involves the question as to whether the pigment arises in the connective tissue cells as a result of cellular (secretory) activity or whether the cell extracts the pig- ment (fully formed or unsynthesized) from the blood; also the manner of the transference of the pigment to the epithelial cell. There are two works which bear upon this point more directly: (1) That of Karg (1888) who ably supports the second view on the basis of findings from a microscopic 454 THE AMERICAN NATURALIST [Vou. XLV study of white skin transplanted to a negro and vice versa, i. e., negro skin transplanted to a white individ- ual; (2) that of Meirowsky (1908), who studied the origin of melanic pigment in the skin and eye and, more espe- cially, on the basis of experiments with pigeons and find- ings in pigmented skin kept alive for several days in a paraffin bath at a temperature of 56 degrees, urges the first view. These two masterly papers advance dia- metrically opposing views. Karg unqualifiedly put aside the idea that pigment may be formed within the epi- dermal cell. He says, ‘‘Es gelang so, festzustellen, das es nur eines Modus der Entstehung des Pigments giebt. Aus der Lederhaut dringen pigmentirte Zellen in die Epidermis, veristeln sich hier weit und geben ihr Pig- ment an die Epithelzellen ab,’’ p. 370. Meirowsky de- scribes conditions more in conformity with our later ideas of cellular physiology. According to him, the melanie granules are passed out of the nucleus of the epidermal cell into its cytoplasm, i. e., pigmentation is an intracellular process, both dermal and epidermal. Historically Riehl (1884) appears to have been the first to describe the condition afterwards interpreted by Karg as supporting his view of epithelial pigmentation, viz., invasion of processes of pigmented connective tissue cells among the epidermal cells. Riehl studied more especially the pigment of hair. He did not generalize, however; and while he thought it improbable, he did not regard it impossible, that pigment may arise out of the protoplasm of epithelial cells. He seems rather to incline to the third view above stated. Moreover, on the basis of findings in a study of three cases of Addison’s disease, where the adventitia of blood vessels was richly infiltrated with pigmented cells coincidentally with a hemorrhagic aggregation of red blood cells, he concludes that the pigment arises from the hemoglobin of the red cells. Aeby (1885) likewise describes the transporta- tion of pigment to the epidermis by wandering cells. Ehrmann (1885-86), who studied the pigmented epi- thelium of amphibia, differs from Riehl and Aeby only in holding that the pigmented stellate cells of the cutis No. 536] COLOR INHERITANCE 455 are non-motile. He describes a network of pigmented cutis cells (chromatophores) connected with the epi- dermis by processes through which the pigment granules stream into the epithelial cells of the rete mucosum Malpighi. He also concludes that the pigment arises from the hemoglobin of the red cells, since the pigment cells are most abundant in the vicinity of the blood ves- sels. The reception of the pigment granules by the epithelial cells he regards a phagocytic process, the epi- thelial cells being described as corroding the processes and assimilating the contents as part of their own organi- zation. The important observations of Jager (1885) on pigment spots in dog and rabbit after inflammation, and those of R. Krause (1888) on apes, are in substantial accord with those of the afore-mentioned investigators, more particularly in regard to the secondary origin of pigment in the epidermis, and the primary source of the same in the blood. For further information regarding literature of pigment cells, more particularly in the lower vertebrates, the reader is referred to the splendid article by Karg. It remains to outline more fully Karg’s position as _ repr tative of the second view above stated. Pig- mentation of epidermis and its appendages (hair, ete.) is a secondary process. The pigment is transferred to the epithelial cells through cells which have their origin in the cutis (i. e. chromatophore, Ehrmann). They wander into, or, remaining on border line between epi- dermis and cutis, send processes into the intercellular spaces of the epidermis. They end in the epithelial cells (capable of a certain amount of ameboid motility) to which they surrender their pigment through process of absorption causing streaming from process to cell. These pigmented cells are wandering cells of the nature of connective tissue cells. They are thought to obtain their pigment from the blood. This, however, is not re- garded as hæmoglobin since no red corpuscles are ever seen in these cells. There is here (i. e., in transplanted human skin) no network of pigment cells as described by Ehrmann in amphibia. Nor can the pigment cells be to THE AMERICAN NATURALIST [Vor. XLV pigmented leucocytes since they have no similarity to the latter, nor are pigmented leucocytes found in negro blood. In the cutis he claims to be able to see all the transition stages between pigment-free connective tissue cells and such as are strongly pigmented. On the basis of his findings he thinks it appears only reasonable that in the pigmenting skin (transplanted white) the earlier unpigmented granules (Reinke’s trophoplasts; Alt- mann’s bioblasts) of the chromatophore take on a dark color (by as yet unknown ways) and that this is connected with the presence of blood; and that it can not, however, be regarded as the product of fragmentated red cor- puscles. Melanin may be hemosiderin, but is the product of a living cell. His final position thus approaches some- what to Meirowsky’s, though they still differ as to the cell that elaborates the pigment for the epidermis. It is important in this connection to note that while Karg de- seribes numerous branching pigmented connective tissue cells (wandering cells) among the epidermal cells of the transplanted white skin, he failed to find such in the nor- mal skin of the negro. He thinks it probable that such may be found in the skin of the negro embryo. Meirowsky in his monograph also gives a very com: plete review of the literature to date. He uses experi- mental methods mainly. His findings supporting the first view may be briefly summarized: Pigmentation is pos- sible in the epidermis without the agency of ‘t melano- blasts,” or even any aid from the cutis. There are both epidermal and cutis melanoblasts, and they are inde- pendent of each other. Pigmented epidermal cells are capable, under certain stimuli, to assume irregularly branching forms (filling the intercellular spaces) simu- lating mesodermal chromatophores (so-called melano- blasts) which might have wandered into the epidermis. This is probably the correct interpretation of Karg’s figures. He brings forth cytological evidence to show that the pigment arises from a reddish nuclear sub- stance (a ‘‘pyroniuroter Kernsubstanz’’)—he does not commit himself as to its chemical nature—which passes into the cytoplasm and gradually assumes the yellowish ` No. 536] COLOR INHERITANCE 457 brown color of the melanie granules. These results from experimental procedure are confirmed by findings in the retina of the calf embryo, where the identical stages in pig- ment formation are observed. He thinks it more probable that the pigment has origin in an albuminous substance (this is in complete accord with the work of Chittenden and Albro—1903) of the nucleus than that hemoglobin has any contributory rôle. The earlier stages in such a process can be observed in carcinoma without progress to the final stage of pigmentation. This indicates that we are probably dealing with a nuclear substance, which, shed into the cytoplasm, under the influence of an oxidation enzyme, becomes a melanic substance. As bearing on the point of the origin of pigment my own observations are as follows: No undoubted branching pigmented cells can be seen among the colored epidermal cells in any of my specimens. Occasionally a process of a pigmented connective tissue cell of the cutis is seen to extend for some distance into the rete mucosum Malpighi (Fig. 1). But their number seems very much too meager to supply the pigment of the many colored cells of the epidermis. There is a nice correspondence between the relative abundance of pigmented cells in the dermis and epidermis of the several specimens of skin. In light skins there are few of each type in each layer; in darker skins there is a decided increase in both. But this proportional increase is as reasonably interpreted as due to the same cause influencing both layers, as that the increased number of pigmented epidermal cells demands an increased number of cutis melanoblasts. Moreover, when one con- siders that there is a continual exfoliation of the super- ficial layers and a replacing of the same from the lower layers, the number of epidermal pigmented cells in colored skins seems out of all proportion to the number of the cutis melanoblasts which are supposed to furnish the pigment. The pigment cells of the cutis are most abundant along the border between dermis and epidermis and along the capillaries of the vascular papille. This first point would seem to indicate the function ascribed to them by Karg 458 THE AMERICAN NATURALIST [Vou. XLV and others, but when one assumes (as all investigators agree in doing) that pigment formation is somehow related to the blood as source of nutrition or supply, the segregation of the pigmented cutis cells at this level becomes intelligible on other grounds, t. e., necessary closeness to the capillaries of papillae. Only those layers of the epidermis next the border line (7. e., next the capil- laries) have the pigment granules of the prevailing size and color for normal pigment cells. Thus my evidence points to a dependence of both cutis and epidermal cells upon the same source (the blood of the capillaries) for a sine qua non of pigment formation, and an independence of each with respect to the other as a necessary source of supply or even as an aid to pigmentation. That the blood constituent is not hemoglobin the arguments of Karg and the observations of Meirowsky seem con- clusively to prove. That it is not an iron-containing element (e. g., hemosiderin) I have demonstrated by the method of testing with potassium ferricyanide as used by Brown (1910) for the liver. Chemical analy- sis by Abel and Davis (1896) also has shown that the melanie pigment of the hair and skin of the negro is free of iron. The evidence at hand, as furnished by Chittenden and Albro, von Fiirth, Spiegler, Gessard, Riddle, Meirowsky, and others seems to render it very probable that in vital melanogenesis we are dealing with a proteid substance (tyrosin; trophoplast; chromo- gen) acted upon by an enzyme or oxidase (tyrosinase) and that one (probably the former) is supplied by the cell (nucleus) and the other by the blood. The fact that the pigment granules, in epidermal cells that are not packed with them, are segregated in the distal portion, indicates that they are responsive to the influence of light. However, the further discussion will not be com- plicated by a consideration of this possible factor. The following discussion will accept as well supported the position that pigment is formed in the epidermal cells— the analogous formation of pigment in ganglion cells gives further support—by virtue of a cellular metab- No. 536] COLOR INHERITANCE 459 olism made possible by close association with the nutri- tive source, i. e., blood vessels.! 1A specimen of leucoderma from a dark negro obtained at autopsy through the kindness of Dr. W. Thalhimer gives further evidence in toa of this position. Macroseopically, the two leucodermic areas, abou m. in diameter, and bilaterally symmetrically placed over the ae ces perfectly normal Sn for their very much lighter color (grayish yellow). A section through the transition area shows the following his- Pii conditions: (1) The leucodermic area contains a comparatively very large number of pigmented cells in the corium. (2) In the epidermis the comparatively small amount of pigment present is confined to some of the columnar cells of the stratum Malpighi. (3) The corium of the pigmented skin has only a moderate amount of pigmented cells. (4) In the epidermis of the normal skin, the columnar cells are laden with melanie granules, and all of the more superficial layers of cells contain a considerable amount of melanie pigmen (5) The PAPA granules are everywhere the same in respect to ki and size. e epidermal cells appear identical, except for the legers abundance > the pigment granules, in the two regions. In v of the above facts, it seems clear that lack of pigment in the an a the leucodermic patch is not due to a dermal deficiency (supposing the derma to be its source of supply) » nor to an inability on or) © small am ets speak in favor of the epidermal origin of the epidermal pigment, and indicate a local inability on the part of the epidermal cells to manufacture in normal quantity (for this individual) the melanie granule rmore, interpreting melanogenesis in terms of tyrosin and tyrosinase, the ts indicate a local ciency or inhibition 0 e these factors more reasonable to supp f es tg m columnar cells of the epidermis of the leucodermie areas were for some obscure cause (nervous?) unable to elaborate the granules (‘‘pyroninroter Kernsubstanz,’’ Meirowsky) which, ace the influence of an oxidase probably everywhere ghee turn mel recent work of c An Diala on Cutaneous Pigmenta- tion in Normal and Pathological SS Journ. Path. and Bact., 15: 3, 1911) in the main also confirms Mei owsky’s findings regarding t the spaces’’ (p. 314). The ‘‘pyronin-red’’ substance he interpre ably the TEENY proteid portion of the complex granules from which the lipoid portion has been dissolved by his (Meirowsky) method of preparing his airt 7 i, e., use of alcohol (p. 316). In origin melanotic m is then a lipochrome, the melanin being the chromatic proteid a 460 THE AMERICAN NATURALIST [Vou. XLV Cause of Degree of Coloration.—Theoretically at least six possibilities are conceivable: (1) Number of pigmented epidermal cells, or, indirectly, (2) number of pigmented connective tissue cells of cutis, (3) number of pigment granules in pigmented epidermal cell, (4) coloration of granules in pigmented epidermal cells, (5) numbers 1 and 3, (6) combination of 1 or 3 and 4, or a combination of the three. It is also theoretically possible that a difference in the size of the granules may play a part in determining the degree of coloration. In view of what was said under the previous heading, number 2 can be disregarded and there remain five plausible possibilities. Which ‘‘ possibility ”’ or set of possibilities expresses the reality will become clear from the description of the several types of skin. In anticipation of ensuing results it may be said in brief that there seems to be only one factor in skin coloration, viz., the number of the pigment granules, a greater num- ber of granules of course involving a greater number of cells—with a small and possibly negligible variation in size. DESCRIPTION or NEGRO SKINS The several samples of skin will be described in the order of their degree of coloration as seen from the sur- face. This agrees almost absolutely with the degree (i. e., quantity) of the pigmentation factor, as will appear below. its separation from the complex lipoid granules. In several points my aie E may be true that the pigment granules of the cutis cells are slightly coarser, I can not regard them as ea than those of the epidermis. (2) In no case have I been able to discern melanie granules in the nuclei (studied in unstained preparations). (3) I find a greater relative amount of cutis pigment in the specimen of leucoderma, whereas Dyson reports no pigment in the cutis of his two samples of leucoderma. (4) Consequently (and for still other reasons) I can not accept his position that eutis pig- ment is secondary to epithelial, i. e, that pigment passes from epidermis via lymph channels to the derma where it is supposed to be taken up by wandering cells. If this were true the cutis underlying the more highly pean epithelium should contain relatively (to leucodermic area) more pigment. But just the reverse is the case in my ‘specimen. In fact the cone of pigment in the epidermis would seem to be too meager to supply the cutis pigment present under the leucodermic patch. All the evidence indicates that the path of pigment transportation is from depth to surface of epidermis. No. 536] COLOR INHERITANCE j 461 All the samples being taken from the same body region, there is a very close correspondence in the number of epidermal layers involved. Number 9: Here all the epidermal strata of cells are pigmented. In the basal layer the cells are packed with the yellowish brown spherical granules to such an extent as partially to obscure the nucleus and cause the cell to bulge. In the more superficial layers, the granules are massed distally and more scattered proximally, the nucleus appearing very prominently. In the cutis, chro- matophores are abundant, their granules being similar in shape, size and color to those of the basal epidermal layer. These cells are always in the near vicinity of the blood vessels. In the upper layers of the epidermis the pigment granules become darker, finer and frequently of oblong shape. These several differential features are ascribed to the several factors of desiccation, pressure and keratiza- tion that the containing cells undergo in their passage to the exfoliating surface. The same explanation probably (at least to some degree) applies to the invariably darker bacillary pigment granules of the shaft of the hairs ap- pearing in all of the sections of this first group.? In sec- tions of child’s scalp, however, both shaft and bulb contain - the same yellowish-brown granules as found in the skin, the only appreciable difference being a considerable ir- regularity in shape and size. It must be noted here that not all the basal cells (though in negro skin the excep- tions both in dermis and epidermis are rare) are equally packed with granules. The optical effect of a small num- ber of granules is a lighter shade of brown than that given by a denser mass of identically colored granules. The objection may be raised that discrimination or lack of discrimination between a darker and lighter pigmented granule is the result of an interpretation where the ‘‘ per- sonal equation ’’ may factor largely. It is not denied that it is difficult to satisfy oneself absolutely that the granules dividing cells of the choroid, epidermis, connective tissue generally, and bone marrow contain absolutely black pigment granules. 462 THE AMERICAN NATURALIST [Voi XLV of the various basal cells of the same and different samples of skin are of identically the same color (the color difference between the skin pigment granules and those of the sections of attached hair is decided enough) but all possible caution was observed to offset the personal factor. In short, when the writer after much study was still some- what undecided as to a definite stand, the slides were shown to three different professors, all with long training and much experience in the use of the microscope. These men were asked to answer the following questions with respect to the basal cells of the eighteen samples, and this without knowledge of what the other men had written: Are the pigment granules of the samples of skin of the same color or of different colors? Name the color or colors? The replies were uniform in recognizing only one type of granules, and in describing it as ‘‘ yellowish brown,’’ ‘‘ brownish yellow ’’ and ‘‘a dark golden or yellowish brown—amber—somewhere between brass and copper,” respectively. One man noted the slightly darker shade of the granules of the more superficial cells. Number 14 (Fig. 1) : All the layers are again pigmented, but there is a slightly smaller amount in the upper layers than in number 9. The pigment cells of the cutis are here somewhat more numerous, showing many anastomosing processes, and forming in places a network of pigmented strands just beneath the scarf skin. A variation in amount of pigment in different regions of the basal layer is again evident. The deeper color of number 9 as com- pared with 14 seems due to the greater amount of pigment present in the superficial layers. Numbers 1 and 2 are very like the foregoing except that there is a slight decrease in the number of greatly packed basal cells. Number 15: This sample shows a quite appreciably smaller number of pigment cells in the epidermis and a yet more pronounced decrease in the corium. Description or Murarro SKINS Number 4: The number of granules in the great major- ity of the basal cells of the epidermis is somewhat less No. 536] COLOR INHERITANCE 463 than in the last of the foregoing group. Cells also now appear in the lower layers with only relatively scattered granules. There is an almost complete absence of gran- ules from the cells of the superficial layers. There are decidedly fewer pigmented cutis cells. Hair in section again shows the same sort of pigment granules as above described. This statement holds good for hair wherever they have appeared in these sections (not seen in the samples of blond skin). Number 3 (Fig. 2): The pigmented basal cells of this specimen contain still fewer granules than in number 4. The very small number and size of the chromatophores of the corium is striking. Here again only the cells of the rete mucosum Malpighi contain the granules. Number 8 is almost identical with number 3. Number 17 is like number 8 with slightly less pigment in the rete mucosum Malpighi; but here the superficial layers again contain a considerable quantity of melanin granules and the melanic cutis cells are more abundant than in numbers 3 and 18. Numbers 10 and 18 are both very like number 8, show- ing only a slight decrease in the quantity of pigment. This seems the best place for a brief discussion of the foregoing facts. What is the fundamental cause of the difference in the degree of color of the skins described? Plainly, I believe, a difference in the abundance of the pig- ment granules in the basal cells of the rete mucosum Mal- pighi. A densely packed mother cell of this layer gives rise to two daughter cells of very similar constitution which are only slightly altered as they pass to the upper layers. Hence in skins where the basal cells manufacture much pigment, the entire rete mucosum Malpighi, formed of its descendants—the factors of desiccation and cornification not being able at once to produce a very appreciable des- truction or modification of the granules—will appear pig- mented. But between negro and mulatto skin there does not seem to be any apparent difference as to the number of epidermal cells producing pigment, but only as to the quantity of pigment produced by the same basal cells, an initial greater quantity determining a secondarily persist- ing greater quantity, and thus an apparently greater num- 464 THE AMERICAN NATURALIST [Vou. XLV ber of epidermal pigmented cells. Probably also, the cells generally of the rete mucosum Malpighi retain in small degree the property of the basal mother cell to produce pigment granules. DESCRIPTION OF SKIN OF BRUNET Number 6 (pathologically pigmented skin) : Here one is unable to determine just how much pigment is due to the normal (‘‘ midway between blond and brunet ’’) and how much to the pathologic condition (‘‘thyroid adenomata’’ —‘‘gall stones, but not jaundiced’’). The specimen con- tains fewer granules than number 18 and more than number 7. Number 7: Here the granules are very few in number and confined exclusively to the basal layer. There is great variation in the number of granules held by the basal cells. Only very rarely does a small pigmented cutis cell appear in the sections. Number 11 has still fewer granules than 7 and Number 17 shows only occasional cells of the basal layer slightly pigmented (few melanin granules). Number 16 is more like the blond skin to be described. There are very few granules in only occasional cells; and no pigmented cutis cells are seen in the section. This skin could not be told from blond. Description oF BLonp SKIN Number 13 (Fig. 3): Here the layer of distinctly pig- mented basal cells is fairly complete. The section is very like number 7 of the brunet series. : Number 5 is almost identical with number 16. The pig- ment granules are very. rare; and only a few to a cell. It must be emphasized at this point that the melanie granules of number 5 are in point of shape, size and coloration indistinguishable from those of number 9. _ There is a continuous gradation in color (and the num- ber of pigment granules) from negro to blond skin with an overlapping at the extremes. The melanic granules of the specimen of melanotic sar- coma were of the same yellowish-brown color (perhaps of a trifle lighter shade), but of very irregular shape and No. 536] COLOR INHERITANCE 465 with great variations in size (the larger ‘‘granules’’ may be the result of fusion of smaller masses). STATEMENT OF RESULTS AND RELATED Facts The facts whose interpretation is sought in terms of some principle of heredity are these: (1) The degree of skin coloration is due to the variable number of pigment granules in the cells of the rete mucosum Malpighi involv- ing incidentally a variable number of more superficial cells. (2) The pigment granules (melanic) of all skin (albinos excepted) are identical in size (practically), shape and color (without qualification). (3) The ascend- ing scale of morphological conditions paralleling a pro- gressively deepening grade of pigmentation may be described as follows: (a) few cells of basal layer pig- mented with few granules—blonds, (b) more cells contain- ing more granules—brunets, (c) a more or less complete basal layer of cells with many and very many melanic granules (mulattos), (d) the cells of basal layer packed and distended with pigment granules; the cells of the more superficial layer also with very many granules. Or, restated and explained, (4) The progressive in- crease in progressively darker skins both in the number of granules and in the number of the pigmented cells. That these two facts are related to each other as cause (number of granules) and effect (number of pigment cells) is strongly indicated by the fact that in light mulatto and brunet skins, where only the basal cells are distinctly pigmented, the number of granules per cell in general de- creases with the progressively lighter shades. (5) The results recently published by the Davenports showing a segregation of the original skin colors (grand- parental colors) among the individuals of the third gen- eration, 7. e., children of mulatto parents. 6) The accumulation of the pigmented cells near the border line between the dermis and epidermis or in the vicinity of the blood stream. (7) Melanin formation is an intracellular metabolic process going on independently and in a measure propor- tionately in both dermis and epidermis. This seems demonstrated by the researches of Meirowsky and others, 466 THE AMERICAN NATURALIST [Vou. XLV and the view is indirectly supported by the comparative findings above described, viz.: (a) absence of connective tissue cells (chromatophores) among the epidermal cells (admitted by Karg for normal negro skin), (b) compara- tive rarity of pigmented processes from cutis cells, and (c) an apparently disproportionate number of chromato- phores as compared with the epidermal pigment cells. (8) The agreement between the cytologic facts of Meirowsky and the chemical results of Chittenden and Albro, and others, that the antecedent of melanin is some form of proteid. (9) The observations of Meirowsky showing a passage of granules from the nucleus to the cytoplasm as the initial step in melanogenesis, and a progressive colora- tion of these granules to a final stage of yellowish-brown pigment. (10) The production of artificial melanins (‘‘mela- noidins’’—Schmiedeberg) by Chittenden and Albro and others from ‘‘antialbumid’’ and various proteids, and the results described by Spiegler and Riddle and others indi- eating an interaction of a chromogen (tyrosin compound) with an oxidizing enzyme (tyrosinase) in the formation of melanin. Discussion There appears, then, proximately to be only one factor in skin-pigmentation, viz., the number of granules of identical shade (yellowish-brown), incidentally the num- ber of cells containing such granules. The granules would seem to be the result of intracellular activities (Meirowsky) and to have origin in cell proteids (Chit- tenden and Albro, and Meirowsky). In terms of chro- mogen and oxidase, the granules may be thought of as tyrosin which under the influence of tyrosinase from the blood or tissues generally becomes melanic. The melanogenetic process seems to stop at the same point in all grades of colored skin, from negro to blond— in hair of the same it may possibly proceed slightly further, though even this seems doubtful in view of ap- pearances in the hair bulb. One seems to be dealing, then, with a continuous process, i. e., the production of melanie No. 536] COLOR INHERITANCE 467 granules; and the numerical point at which the process stops determines the color of the skin. But thought of in terms of greater and lesser ability for tyrosin production (intervention of a tyrosinase-producing factor would modify the results, but not fundamentally alter the mechanism of inheritance) ; and attributing such factor to a specific cell-organ which may be a chromosome or part of such (‘‘teleomorph’’—Spillman), the segregation noted by the Davenports becomes as intelligible as other Mendelian phenomena. From the standpoint of the num- ber of granules some mulatto skins are certainly different from a ‘‘blend’”’ between negro and white, and this is true in the direction both of more and of less—from the stand- point of the amount of pigment some mulattos are iden- tical, on the one hand, with negros and on the other with brunets—and evidences a measure of segregation of “strong melanogenesis’’ and ‘‘slight melanogenesis.”’ A plausible interpretation of Karg’s experimental find- ings might be made on the basis of a larger and smaller amount of tyrosinase in negro and white blood, respect- ively—or more likely perhaps on the basis of more and less pronounced stimulation by negro and white blood, respectively, to tyrosin production. The occurrence of melanotic sarcomata in albinos and white horses forces the assumption that in both cases tyrosinase is present, as in ordinary colored animals. In albinos there is evidently an absence generally of tyrosin in usual events (production of tyrosin locally ac- cords better with our present knowledge than a hypoth- esis of local tyrosinase production). If Spiegler’s view represents the veritable condition, viz., that in white horses there is present a white melanin—rendered quite doubtful by Gortner’s recent work—the end-result of an oxidation process of tyrosin, the presence of melanotic tumors in white horses may be explained in the same way as in ordinary cases, as shown by the work of Gessard. Accordingly, when one considers the question of color inheritance among crosses between ordinary white indi- viduals and albinos, two factors (at least—these most prominent and apparently most important; a ‘‘multi- plicity of units’’ or factors may be involved in color-in- 468 THE AMERICAN NATURALIST [Vou. XLV heritance as the Davenports suggest) appear to be in- volved, i. e., a tyrosin-producing factor and a tyrosinase- producing factor, one at least a function of the epidermal cell, and both having as likely a chromosomal representa- tive (a ‘‘teleomorph’’—Spillman) as any other cell organ or function. The tyrosin-producing factor is probably generally absent in albinos, locally appearing abnormally in tumor cells, hence two albinos can never produce col- ored offspring, as amply shown by the results of the experimental breeders. The observations of Stedman, reported by Bateson (p. 227) ‘‘to the effect that an albino negress married to a European had children all mulattos’’ does not neces- sarily imply that the factor determining the blackness of the negro (tyrosin production) was carried by the albino. Mulattos are frequently so classified on the basis of marks other than color of skin. Many mulattos are no darker than many white brunets. In the above case the factor controlling tyrosin production may very well have been contributed by the father alone. This instance does not necessarily controvert the assumption that albinos lack the factor of tyrosin production. Moreover, crosses between albinos and pigmented in- dividuals result in families where albinism greatly pre- ponderates, as shown in the recent “dissertation” by Stainer. In crosses between whites and blacks one deals apparently more especially with the factors of great and small capacity for tyrosin production—tyrosinase being probably of more general distribution. J udging from the pedigrees published by Stainer, absence of capacity for tyrosin production (albinism) in man behaves more like a dominant character (or at any rate, not like a pure recessive) to the presence of such capacity. This is not in accord with the results of the Davenports, which seem to indicate that ‘‘internal conditions that lead to deeper pigmentation dominate over the weaker conditions’’s similarly as regards color of hair and eyes, ‘‘the more pigmented condition tends to be dominant over the less pigmented’’ according to the earlier investigations of Holmes and Loomis as well as the more recent work of the Davenports. Nor does it accord with the results of No. 536] COLOR INHERITANCE 469 the experimental breeders with lower mammals. It may, of course, be found that all mammalian albinos have the white pigment (melanin) described by Spiegler for white horses. Such a result would seem to correlate a number of apparently discordant facts. It would obvi- ate the further assumption of an ‘‘antioxidase’’ sug- gested by Gortner, and render more intelligible the non- recessive behavior of human albinism.* No theory of color-inheritance is satisfactory that can not embrace all the facts of albinism, and such is the present state of affairs. In crosses between whites and negroes there is gener- ally a partial dominance of the deeper pigmented condi- tion over the lighter in the second (mulatto) generation; the third generation showing a measure of segregation of the original colors. The partiality and incomplete- SEE: of dominance and segregation may be due to a ‘‘myriad’”’ other factors modifying and obscuring more or less the final results.* Seeing that we are dealing with only one kind of col- ored granules, the apparent segregation noted in the families of mulatto parents does not here seem to be due to a condition of unstable equilibrium in the chemical constitutions of the parental melanin and an attempt at readjustment to an original state of greater stability, as suggested by Riddle. The apparent continuity of the melanogenetic process, as seen in the continuous numerical gradation of the same colored pigment granules where a graded series of skins is examined, rests, in fact, where single families of mulattos are considered, upon discontinuities or discrete ? In the second part of Davenport’s paper on ‘‘ Heredity of Skin Pig- furnishes the most cogent argument yet offered for the recessive nature of albinism. t Professor L. W. Lyde, in an article on ‘‘ Climate and Racial Skin Color’’ (Contemporary Review, February, 1911), states his conclusion that ‘‘pig- ment is latent in all humans and depends for its development on relative um. Influence of a Tropical Climate on Europeans,’’ Eugenics Review, April, 1911) believes that ‘‘there is no reason for assuming that a dark com- plexion is due to climate.’’ 470 THE AMERICAN NATURALIST [Vou. XLV ‘‘unit characters’’ controlling conditions of a more and a less numerous production of melanic granules, which conditions conform more or less closely to an alternative mode of inheritance. LITERATURE CITED 1. Abel, J. and Saps W.: 96. The paap of the Negro’s Skin and Exp. Med., Vol. I, p. 2. Aeby, "4 185. Die Her kunti des RR im Epithel. Centralblatt für die medicinischen Wissenschaften, 23, S. 273. 3. Bateson, W. ’09. Mendel’s Principles of Heredity. Cambridge, Uni- He a - H own 10. Changes in the Hemosiderin Content of the Rab- bit’s Liver during Autolysis. Jour. Exp. Med., Vol. 12, No. 5, p. 623. . Chittenden, R. and Albro, Alice H. ’99. Formation of Melanins or Melanin- like peipei from Proteid Substances. Amer. Jour. Phys s p. 291 i Hasenpert, Gertrud C. and Charles B. ’07. Heredity of Eye-color in Man. Science, n. s., 2 . 289. 709. Heredity of Hair Color in Man. Am. Nat., Vol. 43, No. 508, 93. or a P. 710. Heredity of Skin Pigmentation in Man. Am. Nart., Vol. 44, No. i. -. 2 (m — M). .67449 = 4$ -WN À This is the formula which we are seeking, the probable error of the difference between the mean for any family and that for the whole population. By calculating (m — M)/Em-m for every family we should have a criterion of its superiority or inferiority—the individual prepotency of the parent in question—relative to the average condition in the series to which it belongs. _ Tocher has pointed out advantages in using (m—M)/om—s instead of (m — M)/Em-m, but this 18 merely a matter of convenience. The significance of the ratios can be tested by tables of the normal curve. (b) Case of Characters not Measurable on a Quantitative Scale For characters not quantitatively measurable two methods of treatment are available. The first consists m testing the divergence of a family from the general popu- lation on the basis of the relative frequency of a given character. The second consists in testing the deviation No. 536] COEFFICIENT OF INDIVIDUAL PREPOTENCY 4i7 of a family from the population with respect to the dis- tribution of a character. At present the second of these methods seems of little practical importance for our purpose because of the relatively small numbers of indi- viduals available in breeding experiments, even with plants, and because of the arithmetical routine. Consider the first method. Let N be the number of individuals in a population due to P parents. Let X be a character common to all but appearing in different inten- sities (say from 0 development to the greatest possible intensity) in the several individuals, not measurable but capable of division into m classes. Let s,,85,8;-°-: Sm be the classes and Y, Ys» Ys * +Y, be the frequencies in the popu- lation as a whole. Now if a single family of n members be observed the probability of an individual belonging to any class, say sa is y,, /N = p, while the probability of its not belonging to that class is (1—p)—gq. The actual number of individuals with character s, in the family should be np—y,, , while the frequency for the m—1 remaining classes within the family will be given by Yur Yas Yn >*> You providing (a) that the family is not differentiated : from the population, e. g., that there is no individual prepotency in the sense that we have used the term, and (b) that n is so large that the probable errors of random sampling are negligible. In actual work (b) can never, or almost never, be realized. Our problem is to determine whether differences between the theoretical class frequencies, ys’, and the actually observed class frequencies, ys”, in the family are to be regarded as due to chance merely or whether they are so large that they can reasonably be considered as indicating a differentia- tion of the family from the population to which it belongs. In short, our problem is to test (7,— 7.) against its probable error. Pearson has shown that the standard deviation of (ys” — y’) for any grade is a ; a—t \ =, —¥,) = npg (1 -y 1) 478 THE AMERICAN NATURALIST [Von. XLV and Tocher has pointed out that as a test for significance of divergence we may use either of the three ratios (a) (ys” — ys’)/Vnpq(N —n)/(N — 1). (b) 100{(ys”/n) — p}/V1002pq(N — n)/n(N — 1). (c) 100{(4s” /ye’) —1}/100Vq(N — n) /np(N —1). The significance of these ratios can be judged from the tables of the probability integral." III. RECAPITULATION / Individual prepotency is here used to designate the superior capacity of certain parents for producing off- spring of any desired character. The conception is most general, and does not imply a similarity in soma between parent and offspring, but the prepotency of the parent is judged entirely by the offspring it produces. The term is used merely to describe a long-known phenomenon, and no theoretical explanation is suggested. Various breeders have tried to obtain a measure of individual prepotency in its present significance. The purpose of the present note is to point out certain bio- metric formule, in use for other purposes for several years, which seem well adapted for this purpose. They at least obviate several of the objectional features of some of the methods which have been employed. Their applicability in practical work will probably be limited by the arithmetical routine, but in experimental studies their importance may be very considerable.” Illustration of their application will be published soon. CoLD SPRING HARBOR, N. Y., May 19, 1911. TOf course a statistical formula is not applicable to cases not covered by the assumptions on which it was developed. - It seems unnecessary to discuss these here. Those using the formulæ should familiarize themselves with the limitations laid down by Pearson and Tocher in proposing the formule. THE ADAPTATIONS OF THE PRIMATES PROFESSOR F. B. LOOMIS, AMHERST, Mass. THE development of the primates has taken place in regions of comparatively high temperatures, especially in tropical and semitropical climate. This is chiefly due to their arboreal adaptation, which keeps them where the trees throughout the year offer food either as fruit, leaves, blossoms, insects or small animals. The first primates are yet to be found, but they doubtless lived either during the last of the Cretaceous or in the earliest Eocene; for during the Lower Eocene of the Wa- satch epoch there suddenly appear in America two well- distinguished families of primates, the general feeders or Notharctide, and the fruit eaters or Anaptomorphi- dæ.! Between these no intermediate or ancestral group is known, but the wide divergence in form would indicate a considerable time element for development. The gen- era Anaptomorphus and Pelycodus appear in America as a part of the wave of migration which introduces for the first time representatives of the modern groups of mammals. Somewhat later the primates appear in Eng- land and France, apparently part of the same original stock but differing slightly as a result of independent development. The original group of primates or ancestral stock seems to have been a large-brained arboreal insectivor, some- what similar to the tree shrews (Tupaiide). Appar- — ently their home was to the north in the Hudson Bay *The considerable group termed Proglires by Osborn and including Mixodectes, Microsyops, Cynodontomys, Indrodon, Olbodotes, ete., all hav- ing in common the gnawing adaptation and a very primate-like set of premolars and molars, are now assigned by Matthew and Osborn to the Insectivora. 479 480 THE AMERICAN NATURALIST [Vou. XLV region or further north in the forest areas; and under the decidedly tropical climate which is evidenced by the palms and ferns, crocodiles and primates themselves.” From this ancestral center the first primates, along with other groups, migrated in all directions possible, climate and land bridges being considered. This opened three paths, one south into America, a second southeasterly into England and France, and a third southwesterly into Asia, thence ever southerly across China and India and along the Indo-Madagascar isthmus (or chain of islands) to Madagascar and Africa. —> General Feeders >— Fruit Eaters Fie. 1. Diagram of the radiation of the primates in the Eocene. The first primates, as indicated, separate into two groups, first a group of long-headed (dolicocephalic) gen- eral feeders with unspecialized teeth, which probably took fruit, leaves, insects and small animals: and second a group of short headed (brachycephalic) fruit-eaters with crowded and rather high pointed teeth. These are * For a discussion of the climate see Wortman, Amer, Jour. Sci., 1903, Vol. 165, p. 417; and Wieland, same journal, Vol. 166, p. 401, 1903. No. 536] ADAPTATIONS OF THE PRIMATES 481 the first adaptations of the primates and it probably took some time to arrive at the degree of difference found in the Wasatch of North America. The fruit-eating brachycephalic group includes Anap- tomorphus of American Eocene, Necrolemur and Micro- cherus of the European Eocene, and Tarsius, now living in southeast Asia. During the Eocene the climate was progressively colder, becoming at least temperate by the ayeuwg Je4yjsaouy / 6 \ J 2 £ 7 “ee, ¥ Aone | aes reso (Af) &, Sage a Anthropoidea (Ey) (A) (Af) ay wa Cebidae (S Amer) Fic. 2. Genealogical relationships of the Eocene primates. end of the Eocene in the northern parts of America. This climate acted to force southward all the primates _ of the north and also several other groups, so that dur- ing the Lower Eocene we have the whole primate group pushing down, the Anaptomorphidæ all over North America, the Microchoeridæ on to what there was of Eu- rope and the ancestors of Tarsius? on to eastern Asia, * Earle, Amer. Nart., Vol. 31, pp. 569-575 and 680—689, 1897. 482 THE AMERICAN NATURALIST [Vot XLV which through the lower and middle Eocene was sepa- rated from Europe. The fact that Tarsius is confined to islands possibly explains why it has remained in so primitive a condition in many ways, though specialized in the limbs which are as yet unknown in any others of this group. The general feeders are a larger and more abundantly preserved group. It includes the Notharctide* of North America (to which belong Pelycodus and Notharctus) ; the Adapiide of Europe (including Adapis and Plesa- dapis); the Homunculide® of South American Miocene (including the genera Homunculus, Pitheculites, Ho- munculites) and lastly the living lemurs of southern Asia, Madagascar and Africa. All have the dentition %, 1, %4, 34 — 40, and long heads, and apparently ate both vegetable and animal food. The group originated like the foregoing in northern America and migrated south- ward, driven by the change in climate. The earliest known forms are those in the Wasatch of western Amer- ica, and they are likewise the most primitive. Though preserved only in Wyoming and New Mexico, they prob- ably occupied pretty much all of our western plains coun- try, then forested. South America seems to have been isolated from early Eocene times, so that some repre- sentatives of this group probably got into that continent by early Eocene times, i. e., the radiation over Nort America must have been pretty rapid and general by lower Eocene times. Those in North America after the separation of South America flourished for some time, being especially abundant in the Wind River and Bridger epochs, but with the cold of the Uinta epoch they were crowded south and finally exterminated in North Amer- ica, never more to be widely distributed on that continent. *See Osborn, Bull. Amer. Museum Nat. Hist., Vol. 16, pp. 169-214, 1902, ro he mea and Loomis, Amer. Jour. Sci., Vol. 171, pp. 217- *See Schlosser, ‘‘Die Affen, Lemuren, Chiropteren, ete.,’’? des Euro- päischen Tertiiirs, Theil 1, s. 19-54, 1887. *See Ameghino, Anal. d. Museo Nac. d. Buenos Aires, Vol. 15, PP- 424—429, 1906. No. 536] ADAPTATIONS OF THE PRIMATES 483 Those in South America quickly differentiated by the loss of the first premolar, making a dental formula of %, 1, 3%, °4—36. With this also goes a deepening of the ramus of the lower jaw, a shortening of the face, and a tendency to develop the occipital region so that it over- hangs widely the foramen magnum. This group of forms is termed the Cebide. On becoming successfully adapted to the South American continent and during the long isolation of that area, these early forms have gradually adapted themselves in various directions, often paralleling old world types. In size they have developed, the largest forms having a body of 27 to 28 inches in length, and legs as long, making a height of 41% feet, which is a good-sized monkey. They have always re- mained arboreal with opposable thumbs and a prehensile tail; but they vary from the slender spider monkeys to the robust and powerful woolly monkeys (Lagothrix). From the above has been specialized the family of mar- mosets (Hapalide), by the loss of the last molar (mak- ing the dentition %, 4, 3%, %2 — 32), by the development of a broad nasal septum, the loss of the prehensile character of the tail and opposability of the thumb. The southwesterly wave of migration crossed the Beh- rings connection and moved down the easterly part of Asia across the Indo-Madagascar isthmus and into Africa. This isthmus or series of islands sank at or toward the end of the Eocene, leaving lemurs stranded all along the area occupied by the isthmus. Those on the islands and especially on Madagascar have remained very much as they were, adapting themselves in minor ways, but being always arboreal. Some peculiarities must have developed very early for they are common to the group, like the having of the lower incisors project- ing forward (proclivous), the lower canine small and like an incisor, while the first premolar acts as a canine tooth. Then the fourth digit of the hand is longest, and the second one of the foot is clawed instead of having a nail. In this lemur group we know only the immediate 484 THE AMERICAN NATURALIST [Vov. XLV ancestors of the living forms, and as yet no record has been found of the forms intermediate between those on America and the living types. In the case of the Mada- gascar form, Chiromys or the aye-aye, we have a repre- sentative of the group which has adopted a gnawing habit to get grubs, etc., under the bark, and a great change has resulted in the dentition, by which the first incisor has become specialized into a rodent-like gnawing tooth and there is a reduction in the teeth so that the formula is only %, °/o, 1/o, 4=18. The easterly wave of migration is represented by sev- eral species of Adapis found in the middle and upper Eocene of England and France. Apparently. the prog- ress of this easterly migration was slower, so that they reach Europe considerably later than the same latitude in America. The primates are not in the front wave of immigration on the European side, so that it is possible that the forested condition was not as favorable. The Adapiide in Europe, small primates with a long low skull and the ancestral dental formula %, 4, %, *4—=40, the teeth being very generalized. With the close of the Eocene the first adaptive radia- tion of the Primates was complete, and they had achieved an almost world-wide distribution. At the end of the period the North American contingent was extinct, the South American group was isolated, the Asiatic and African forms were scattered on islands and on the Afri- can continent, and the European contingent was located in central and southern Europe, or what land there was at that time in those regions (see Fig. 3) ; and it is among these that the next act in the great primate drama took place. The Oligocene period is one in which there was a grad- ual rising or emergence of continental areas so that the southern part of Europe was an archipelago, which to- ward the end developed into a long peninsula, extending from the present Asia Minor (see Fig. 3). During this period the change in the Adapiide is but little known, but No. 536] ADAPTATIONS OF THE PRIMATES 485 during that time they shortened the skull and lost the first two premolars, and made a considerable increase in size. In Europe their remains are very scarce and con- fined to the Lower Oligocene when it was the true Adapis which was holding over from the Eocene. Schlosser has just reported some primates from the Fayûm formations of Oligocene age in northern Africa. These he gives new generic names, Meripithecus, Parapithecus and Proplio- pithecus, assigning the first two to the Cercopithecide and the last to the Simiide. They seem from the de- scriptions to be primitive members of the Cercopithecide, which would indicate that the change to the modern type by the loss of the first two premolars was accomplished in the early Oligocene, perhaps in Africa as the two areas are in connection at the time across Gibraltar.’ At the beginning of the Miocene the European primates had the dental formula %, 4, %, #432, a shortened face, and a shortened tail, but were still arboreal forms. Dur- ing the Lower Miocene two divisions arise, the one adher- ing to the quadrupedal gait, the heavy jaws and longer snout: the other acquiring the bipedal gait, and shorten- ing the face with a corresponding broadening of the teeth. In both divisions there is a tendency to come down to the ground. - The former group is the Cercopithecide in its broad sense, or ‘‘old world monkeys’’; while the latter are the Simiide or apes. The Cereopithecide seem to run back to some such form as the Oreopithecus, found in northern Italy, and present- ing dental characteristics resembling the baboon, but at the same time having a shortened face suggesting the Simiide. A second form belonging to this group is Mesopithecus found in considerable abundance in the Lower Pliocene of Greece. This form seems to be in- termediate between the macaques and langurs, resem- * For geography see Matthew, Bull. Amer. Museum Nat. Hist., Vol. 22, p. 364, 1906. For the Fayum Primates, see Zoologischen Anzeiger, Bd. 35, for March, 1910, and Matthew, Amer. Nat., Vol. 44, Nov., 1910, p. 700. 486 THE AMERICAN NATURALIST [Vot XLV bling the former in the stout limbs, the latter in its den- tition. Considering the different subfamilies it would appear that the Cercopithecide originated in southern Europe, that it was fairly successful, and that as a result of this, the family adapted itself in three directions; first one group left the trees and took to life on the ground, giving rise to Cynocephalus and Macacus; the second group became leaf feeders, and developed a pouched stomach and for some reason also disproportionally long hind limbs, giving rise to Semnopithecus and Nasalis of Asia and Colobus of Africa: while those remaining in the trees and changing but little are Cercopithecus and Cer- cocebus of Africa. The differentiations took place in the Miocene and are fundamentally based on food supplies. Those forms which had developed strength enough to defend them- selves, their fore and hind limbs being approximately equal in length, and their food including insects, lizards, frogs, etc., as well as all sorts of vegetable life, like leaves, fruit, blossoms, etc., came down from the trees. The terrestrial forms which continued to live in the forests make the genus Macacus, or macaques, which during the Pliocene spread pretty well all over Europe, even up into England, and also into western Asia where they still live. In the Pleistocene some representatives of the genus went with the great wave of migration from southern Asia into Africa, but they have become extinct in that continent except for one species, the Barbary ape. Those members of the group which left the woods and took to the more open country developed great strength and powerful jaws and are the baboons (Cynocephalus ) These too originated in southern Europe and migrate during the Pliocene eastward into Asia, and during the Pleistocene on down into Africa, to which continent they are now confined. The second subfamily of the Cercopithecide are the langurs (Semnopithecus, Nasalis and Colobus) which, while remaining largely arboreal, have specialized as her- No. 536] ADAPTATIONS OF THE PRIMATES 487 7 Fic. 3. Diagram of the radiation of the quadrupedal old world monkeys. — c j line ~— Cynocephalus Macacus line SP te ee eee Semnopithecus üne 488 THE AMERICAN NATURALIST Von.. XLV bivors among the primates, feeding exclusively on leaves; in response to which they have developed a stomach of several pouches comparable to that of a sheep or cow. Like the preceding subfamily, they originated in southern Europe and during the Pliocene moved over into south- ern Asia, where the langurs and the nasal monkeys live to-day. Some members of the group, however, moved during the Pleistocene in Africa where the thumb was much reduced, which feature distinguishes the genus Colobus. of eo rays Cree Macacvs (Eu) (A) (AN oF s 8 2 V: F Eoo e Cercopithecus (Af) & G & f 2 Pleistocene Recent Miocene | Pliocene Fic. 4. Genealogical relationships of the quadrupedal old world monkeys. Lastly the unspecialized subdivision of the family, the Cercopithecus genus, followed the same lines of migra- tion and reached Africa where they now live, having spread over the major part of the continent. Turning back to the early Miocene, we find that there No. 536] ADAPTATIONS OF THE PRIMATES 489 was another group of primates which tended to come to the ground, and these tended to assume a more or less upright position, with a bipedal gait. The hands thus free to take hold of objects, were free to develop a deft- ness and adaptability, which seems to be the key to the progressive development of the apes. It seems however that this handling of objects (food, sticks, stones) began before they left the trees and was really the cause of taking the bipedal gait. The climbing offered an ever changing grasp and carrying food to the mouth was a nat- ural starting point; so that, with the front paws used as hands, there is a good reason for exempting them from the heavy work of locomotion. Contributory to this idea is the eolith development. These crude flaked flints* begin back in the Miocene at least, and as Penck?’ sug- gests the only known primate which might be suggested as an eolith-maker is Dryopithecus. It seems highly probable then that the hands had begun to be used as such before the first apes came to the ground and that this specialization of the hand was the cause of the upright position and bipedal gait. Of course the varied exper- ience resulting from taking up all sorts of objects and using them for different purposes tended to develop the intelligence, and that furthered handling, the two acting and reacting on each other. In the early Pliocene of southern Europe three divi- sions of the simian group have already arisen,” one group remaining arboreal, or more probably reverting again to the trees, a second group developing great mus- cular and skeletal strength, the third group developing especially the brain and central nervous system. The first of these groups, i. e., the retrogressive or aboreal group, is represented in the Upper Miocene of southern Europe by Pliopithecus, a form ancestral to the modern gibbons, and one which during the upper Miocene 3 See MacCurdy, Amer. Anthrop., Vol. 7, n. s., pp. 425-479, 1905. ° Science, Vol. 29, n. s., p. 359, 1909. See Schlosser, Zoologischen Anzeiger, Vol. 22, p. 289, 1900. THE AMERICAN NATURALIST [ Vou. XLV 490 apes. >— Gorilla Simia Fie. 5. Diagram of the radiations of the bipedal primates or — > Anthropithecus Bat to Mao No. 536] ADAPTATIONS OF THE PRIMATES 491 and lower Pliocene spread over a large part of Europe (France, Germany, Switzerland). However when the colder climate of the Pliocene developed, the European contingent was exterminated, and only those in Asia have survived as the gibbons of to-day. The second group which developed especially strength is rare both in prehistoric and recent times: but it seems gi oe pir picus phop? Anthropithecus — Chimpanzee . — Dryopithecus —Anthropithecus SNAJP Gorilla S; N i Miocene Pliocene Pleistocene Recent Fic. 6. Genealogical relationships of the anthropoid apes. to have originated first in southern Europe though no representatives have yet been found. In the Pliocene however the genus Simia has been found in southern Asia where its representatives still remain as the orang utan. During the Pleistocene representatives of the group seem to have reached Africa where they have per- sisted as the gorilla. The third group has at its base Dryopithecus, the middle and upper Miocene ape which ranged over a con- 492 THE AMERICAN NATURALIST [Vou. XLV siderable part of Europe: but at the end of the Miocene this form became extinct and with it the last of the chim- panzee line in Europe, the next repr tative being Anthropithecus, the true chimpanzees, found in the Pli- ocene of India. This first chimpanzee makes a slightly closer approximation to man than the living species. During the Pleistocene the wave of immigration into Africa included Anthropithecus, which has survived only on that continent. In 1896 Dubois found in Java in beds now generally called early Pleistocene, the top of a skull, a femur, and a few fragments of a transitional form which is in many ways like the apes and in others like man. This he called Pithecanthropus erectus, and it stands as either a very high grade ape or as a low grade man, the latter being the usual designation. If not the actual ancestor of man, it is at least a typical stage in his development. From the distribution of Anthropithecus and Pithecan- thropus it seems certain that man originated in southern Asia, at least by the beginning of the Pleistocene: and that he radiated from there westward across Europe where such remains as the Heidelberg jaw, and those of the Neanderthal type have been found so widely. He probably also migrated easterly into North America, and thence south with the Pleistocene fauna into South Amer- ica where very primitive remains have recently been de- scribed by Ameghino as Diprothomo platensis, and Homo pampensis. Remains have been strangely scarce 1 North America, though the fauna, with which early man usually associated is present in various parts of the continent. JEAN MARCHANT; AN EIGHTEENTH CENTURY MUTATIONIST! DR. HENRI HUS UNIVERSITY OF MICHIGAN Jean Marcuant was the son of Nicholas Marchant (died, Paris, 1678), director of the Jardin du Roi, the principal author of the famous ‘‘ Mémoires pour servir a l’histoire des Plantes,’’ published in 1676 under the aus- pices of l’Académie royale des Sciences and edited by Dodart.? The name Marchant is perhaps most familiar in con- nection with the genus Marchantia, which, though not unfrequently attributed to Linnæus, was named by Jean Marchant in honor of his father,’ when, because of his discovery of the ‘‘flower’’ and ‘‘seeds,’’ he removed it from the genus Lichen, under which it formerly had fig- ured as Lichen petreus stellatus.* 4 AGENDY EE the Botanical Laboratory of the University of egn No. s, D., as trois premiers botanistes de Leora ae des Sciences, Dodart et les deux Marchant,’’ Bull. Soc. bot. France, 35: 285, 1888. In this paper Clos a to great length in aien petik to show that Nicholas Marchant was the chief aii to the work just referred to, though a ae reference to a statement on the a of the academy and which curiously enough seems to have escaped him, would at once have settled the matter beyond question. In “Table en a des matières contenues dans 1’Histoire et = Mémoires de l’Académie Royale des Sciences,’’ publiée par son ordre, 1: 200, 1666-1698, 1778, one reads: ‘‘ Marchant [M. Nicholas] a fourni a le Botanique des Mémoires pour servir à l’Histoire des Plantes. T. 4, p. 122 * Marchant, J., ‘‘ Nouvelle découverte dee fleurs et des graines d’un anp rangée par les botanistes sous le genre du Lichen,’’? Mém. de Phoil. . d. 8c., 1713, pp. 229-234. ‘‘Nous établirons pour cette plante un nouveau genre que nous appellerons Marchantia du nom de feu M. Marchant, mon père, qui le premier eut l'honneur d’occuper une place de botaniste dans cette Académie, Se le Roy en 1666 créa sith Compagnie. ’’ *Caspar Bauhin, Pin. 493 494 THE AMERICAN NATURALIST [Vou. XLV It was indeed a discovery far beyond mediocrity and indicative of excellent powers of observation. For since the days of Cxsalpino but little advance seemed to have been made as to the organs of reproduction. Even one of Linneus’s first papers, ‘‘Preludia Sponsaliorum arborum,’’ deals with the sexes of plants, and, attracting the attention of Olaf Rudbeck (1729), secured for the ‘‘father of botany’’ in spe, the position of assistant.” Darwin, in the introduction to his ‘‘Origin of Species,” points to Buffon as the first transmutationist, though, as he says, it was the views of Lamarck which first at- tracted general attention. But long before their time, when Buffon was but a boy of twelve, Jean Marchant had made some very pertinent observations on the sud- den origin of species. Believing genera to have been created as such, a view expressed in his early days by Linnæus himself in his ‘‘Systema Nature,’’® he was able to see new ‘‘species’’ originate suddenly. He had at his disposal a garden, probably already used by his father,’ who was an ardent collector and introduced the seed of many foreign plants, growing and describing them.’ His observations were made upon Mercurialis annua, the dog’s mercury, a plant long known as possessing certain reputed virtues.!° ° Wittrock, Veit B., ‘‘ Nagra ord om Linné och hans pees för den ater a Vetenskupes: ”? Acta Horti Bergiani, 4: No. 1, sane omne est naturale, in primordio tale creatum cP the ‘‘ Jardin du Roi.’’ ‘Il [Nicholas Marchant] faisoit campagne,’’ Hist. de 1’Acad. roy. d. Se., 1666-1686, 1: 200, Paris, 1733. <‘ Hist. de 1’Acad. roy. d. Se.,’’ 1680, p. 307 °’ This term is also pe for M. perennis, in which case the name French BPA is given to M. ua. * It was supposed Der ie juice of species of PANTENE especially of the ARR ll M. tomentosa, had the power to determine the sex of children, according to whether the mother drank the juice ng the male or of the female plant. sins the true sex of the plant was not known, as also is apparent from Marchant’s paper. Thus boy’s mercury was the name applied to a female plant of M. oe girl’s mereury the name given to the plant ser staminate flow Also, staminate flowers repeatedly have been observed on the PeR via (f. ambigua, Duby, "Bot, Gall.,’’ 1: 417), No. 536] JEAN MARCHANT 495 In 1715 Marchant noted in his garden" the appearance of a laciniate form of Mercurialis annua, which he desig- nated Mercurialis foliis capillaceis. The next year, in the same part of the garden, this plant reappeared, being represented by four individuals. There appeared further two plants, the foliage of which, though also of a laciniate character, was sufficiently different to permit of their being readily distinguished. To these plants he applied the name Mercurialis foliis in varias et inequales lacinias quasi dilaceratis. The description of the leaves at once leads us to recognize this plant as a typical laciniate form, especially his reference to ‘‘a large num- ber of leaves, which, because of their irregular outline, resemble mere remnants of leaves torn or gnawed by caterpillars,’’ curiously enough, the same expression which I used to describe the appearance of the leaves in the flowering shoot of Arctium minus laciniatum. In an attempt to explain the successive appearance of these two new forms of Mercurialis we could assume ` that both forms had been created in 1714, but that the seed of one had germinated in 1715, while that of the other remained dormant for a year. It is, however, far more probable that but one plant of the first laciniate variety had been allowed to grow up, and that in the next year special orders were given to the gardener who had the care of this particular portion of the garden, to allow to grow all seedlings which in the least resembled those of Mercurialis. Still another possibility is that the first form, created in 1714, gave rise, in 1715, through a second mutation, to another, less laciniate form, which appeared in 1716. After reporting these plants in 1719, Marchant makes no further mention of them, but de Candolle!? refers to them, under Mercurialis annua, in the following terms: “Presumably the Jardin Royal. ‘‘Jean Marchant avait, ainsi que son père, le titre de directeur de la culture des plantes du Jardin du roi.’’ Michaud, ‘‘ Biog. Univ.,’’ 26: 486, 2d ed. mt A a O 263.707. 496 THE AMERICAN NATURALIST [Von XLV Mem, de tAcad 1729: FL 6 Per 64. Xj KA \ DA l và ; xf Wf S SARTA Fre, 1. Monstrose occurrit: 1° laciniata, foliis laciniato-dissectis. © In Gallia (Marchant). — March. in Act. Acad. Paris, 1719, p. 59. t. 6. — 2° capil- lacea (Guep. “ Flore Maine-et-Loire,” ed. 3, p. 401), foliis ad lacinias auguste lineari-lanceolatas, lineares v. capillaceas integras reductis. © In hortis Andegaviæ et ad Issy-l’Evéque (Guepin, J. e., Gren. et Godron, No. 536] JEAN MARCHANT 497 Mem. de lAcad. 1729. F a pag. by Mercuriahe altera, folar in vartas ct murquates lacinias y the “ e Frc. 2. “Flore de France,” Vol. 3, p. 99). — Marchant in Mém. de l’Acad. Paris, 1719, p. 64, t. 6 This discovery on the part of Marchant is particularly interesting for various reasons. In the first place be- 498 THE AMERICAN NATURALIST [Vou. XLV cause it is the second historical case of mutation on rec- ord. Further, because of a certain analogy with Spreng- er’s discovery of Chelidonium majus laciniatum. Sprenger, an apothecary of Heidelberg, cultivated a large number of plants in his garden as was the custom in those days, and, about 1590, observed there a type of Chelidonium majus formerly unknown to him and to which he gave the name Chelidonia major foliis et flori- bus incisis. Tournefort, in his ‘‘Schola botanica,’’ men- tions three forms, Chelidonium majus vulgare ©. B., Ch. majus foliis quernis C. B. (Ch. folio laciniato J. B.) and Ch. majus foliis et flores minutissime laciniatis Hort. reg. par. The second of these was identical with the variety discovered by Sprenger, the third a form which had originated in the Paris Botanic Garden and was dis- tinguished from the other by the greater reduction of the leaf blade. Of the laciniate varieties of both Cheli- donium majus and Mercurialis annua there exist there- fore two forms. While those of Chelidonium majus, and especially the first, are fairly well known and may be en- countered in almost any botanie garden, it is not so in the case of either of the laciniate varieties of Mercurialis annua. At least I do not remember seeing them, nor do I recollect the occurrence of their names in the seed-ex- change lists annually published. In at least one instance the claim has been made that one of the laciniate varieties of Chelidonium majus orig- inated de novo" a claim which to Korschinsky™® appears to lack foundation. No such claim has been made for one of the laciniate varieties of Mercurialis annua. How- ever, as seen from de Candolle’s account, given above, at least one of the laciniate forms was reported as occur- ring both in the botanic garden at Angers and also at Issy- IEvéque. These two towns, the first in Maine-et-Loire, “Roze, E., ‘‘Le Chelidonium lacimiatum Miller,’’? Journ. de Bot., 9: 296, 1895. “Clos, D., ‘‘Réapparition de la Chélidoine à feuille de Fumeterre,’’ Compt. rend. 115: 381, Paris, 18 5 Korse rechinsky, 8 y + Hpeabocedesis und Evolution,’’ Flora, 89: 240, 1901. No. 536] JEAN MARCHANT 499 the other in Saône-et-Loire, lie more than 200 miles apart. It is possible, perhaps, that the seed was brought from Paris to Issy-l’Evéque and from there transported accidentally to Angers. Or is it more prob- able that the variety originated a second time? It isa question which forever must remain unanswered, though it is not improbable that a French student, who most readily commands the means of research in this direc- tion, could, by delving into historical records, perhaps throw some light upon the subject. The work of Jean Marchant gives evidence, not only of great exactitude, but also of excellent powers of observa- tion. Thus, a few years after his discovery of the lacin- iate forms of Mercurialis annua, he was able to give an account, and, as I believe, the first, of a myxomycete, of the ‘‘flowers of tan.’’!® As is evident from the description and drawings, he was dealing with a Fuligo. He recognized the vegetable character of the organism, but unfortunately places it with the sponges, giving it the name Spongia fugax, mollis, fora et amoena,in pulveri coriari nasceus. In this case, as in the other, Marchant’s ambition did not lead him beyond a mere, apparently most accurate, descrip- tion, something decidedly pleasing when we remember that in the same year Jean Marchant was elected to mem- bership of the Académie (1678), there appeared Father Kircher’s Mundus subterraneus," and that one of his contemporaries was de Maillet (1656-1738), who ‘‘de- rived birds from flying fishes, lions from sea-lions, and man from l’homme marin, the husband of the mermaid!’’ Believing genera to have been created as such, Mar- chant did not go beyond this point, but realized that species were derived from preexisting ones. His own * Marchant, J., ‘‘Observation touchant une végétation particuliére qui nait sur l’écorce du chêne battue, et mise en poudre, vulgairement appelée du Tan,’’ Mém. de Math. et de Phys. de l'Acad. roy. d. Sc., 1727: 335, Paris, 1729. ‘ " Osborn, H. F., ‘‘From the Greeks to Darwin,’’ 109, 1908. ‘‘The worthy priest describes orchids giving birth to birds, ete.’’ 500 THE AMERICAN NATURALIST [Vot XLV words should be quoted here. As far as I am aware, there exists but one account of any length of Marchant’s discovery, that of Korschinsky.!S Godron refers to him’? as does de Vries.” Since the original papers are rather inaccessible to the majority, it was deemed of interest to give here a trans- lation of Marchant’s two articles dealing with his dis- covery. The translation of course has been made as literal as possible. The first paper is merely a résumé of an address made before the academy by Marchant. The second gives a detailed account as published in the Mémoirs. -Ox THE Propuction or New Spectres or Prants” In the month of July, 1715, Mons. Marchant noted in his garden a plant which he did not know, and which attained a height of from five to six inches.?? It per- sisted until the end of December, when it dried up and died. He believed to be able to class it only with the genus to which the mercury belonged; and since it was entirely new and thus far had not been described by authors, he called it Mercurialis foliis capillaceis. The following year in the month of April, and in the same place where this plant had been, he saw appear six others, of which four were quite similar to the former, and two others sufficiently different to make another species of mercury, which he named Mercurialis foliis in varias & inequales lacinias quasi dilaceratis. It per- sisted until the end of December, in which respect these two species are different from the common mercury, which, though annual like these, does not last as long. * Loe. cit. » Godron, ‘‘ De l’espéce,’’? 1: 160 (not seen). * De Vries, H., ‘‘Die Mutationstheorie,’’ 1: 136, 1901. tí Mercurialis annua laciniata ist 1719 von Marchant als neue Form entdeckt worden; sie ist seitdem samenbestandig geblieben.’’ _ ” “Fist. de 1’Acad. roy. d. Se.,’? 1719, p. 57, Paris, 1721. =“ The French ‘‘ pied’? is sist to 1 foot 14 inch of our measure and 7 divided into 12 ‘‘pouces,’’ each ‘‘pouce’’ being divided into 12 ‘ Professor W. E. CAST NODAR Ichthyology: President DAVID STARR Some eH eta me on Fossil Plants : Professor Dovetas HOUGHTON CAMPBELL. Single Number 35 Cents The NATURALIST will ha 2: Oo ai Yearly Subscription, $4.00 One Dollar four h THE SCIENCE PRESS N. Y. Sub-Station 84: NEW YORK Lancaster, Pe VOL. XLV, NO. 537 SEPTEMBER, 1911 THE AMERICAN NATURALIST A MONTHLY JOURNAL Devoted to the Advancement of the Biological Sciences with Special Reference to the Factors of Evolution CONTENTS Page I. Inheritance of the “Eye” in Vigna. Dr. W. J. SPILLMAN. - - - - -513 - 524 Il. Heredity of Hair Form among the Filipinos. Dr. ROBERT BENNETT BEAN TII. The Zoogeography of the East Indian Archipelago. Dr. P.N. VAN KAMPEN- 537 Shorter Articles and Discussion : Biometric Arguments regarding the Genotype Con- cept. Dr. RAYMOND PEARL. On the Formation of nas nag and Contingency Tables when the Number of Combinations is Large. Dr. J. ARTHUR HARRIS. Acquired Characters defined. - 561 Notes and Litera : The Present Day e and Sady of Animal Pr A Useful aa Bibliography. V. L. K. THE SCIENCE PRESS LANOASTER, PA. GARRISON, N. ¥. NEW YORK: SUB-STATION 84 The American Naturalist MSS intended for publication and books, etc., intended for review should be sent to the Editor of THE AMERICAN NATURALIST, Garrison-on-Hudson, New York. Articles containing research work bearing on the problems of organic evolu- tion are especially welcome, and will be given preference in publication. hundrea reprints of contributions are supplied to authors free of charge. One Further reprints will be supplied at cost. Subscriptions and advertisements should be sent to the publishers. subscription price is fsur dollars a year. i anadian postage twenty-five cents additional. The postage is fifty cents and The charge for single copies is thirty-five cents. The advertising rates are Four Dollars for a page. THE SCIENCE PRESS Lancaster, Pa. Garrison, N. Y. NEW YORK: Sub-Station 84 Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of Congress of March 3, 1879. THE BULLETIN—Por bargains in Ethnolograph- ical and Pre-historic Specimens. BULLETIN post free for 3 cent stamp. 4 Duke St., Adelphi—London—England BIRDS’ EGGS W. F. H. ROSENBERG, 57 Haverstock Hill, London, N. W., England begs to announce the publication of a new Price List (No. 15) of Birds’ Eggs, con- taining over 900 species from all parts of the world. This Catalogue is systematic- ally arranged, with authors’ names, indica- tions of localities, and an index to families. Tt will be sent post free on application, as will the following lists: No, 11, Birds’ Skins, (5,000 species) ; No. 12, Lepidoptera, (5,000 species) ; No. 13, Coleoptera ; No. 14, Mammals. In course of preparation: New Price List of Reptiles, Amphibians and Fishes. Largest Stock in the world of specimens in all branches of Zoology. Books on Natural History, Science, Travel, Voyages, etc. See THE TENTH EDITION. THE MICROSCOPE, an introduction to Microscopic Methods and to Histology, by SIMON HENRY GAR, —— Catv oer 350 e octavo and above gures 1 and fully revised edition. Price $2.00, postpaid. COMSTOCK PUBLISHING CO., ithaca, N. Y. Back or Current Numbers of any American or Foreign technical or trade journal and magazine fur- nished on short notice at moderate rates; all kinds of Government and State Reports in stock. Clippings 0u special subjects furnished promptly. Large stock of American Naturalist, Science and Popular Science on hand. Magazines, Books and Papers of all kinds bought Special Subscription Price List on BEU Information concerning any periodical furnished free of charge. A. W. CASTELLANOS 259 Armstrong AVe., Jersey City, N.J-, U.S.A. THE AMERICAN NATURALIST Vou. XLV September, 1911 No. 53 INHERITANCE OF THE “EYE” IN VIGNA DR. W. J. SPILLMAN U. S. DEPARTMENT OF AGRICULTURE CERTAIN races of the cowpea (Vigna unguiculata) have the seed coat completely pigmented, others have no pigment, while others have pigment confined to cer- tain areas. In this paper the fully pigmented races are referred to as having solid color, those without pig- ment as white, while the pigmented area of the partially pigmented :seed coats is called the ‘‘eye.’’ This eye, when small, is always confined to the region of the hilum, and when large always surrounds the hilum. It varies widely in size and form, as shown in the accom- panying illustrations, which show the principal types of eye. In a the pigmented area is confined to two patches on opposite sides of the hilum. In b the area surrounds the hilum except at its micropylar end. OEE Fic. 1. Forms of the “eye,” or pigment area, in the seeds of the cowpea. In c the hilum is completely surrounded, but there is a broad indentation at the lower end of the pigmented area. In d the eye covers nearly the entire ventral! 1 The terms ventral and dorsal are used in this paper to describe respect- ively the side of the seed showing the hilum and the opposite side. 513 514 THE AMERICAN NATURALIST [ Vou. XLV surface of the seed, but has a characteristic notch at the micropylar end (lower end in the figure). In e the area has extended over the micropylar end of the seed. In this form and the next there are usually some isolated spots of pigment in the non-pigmented area. In f the pigmented area covers all except the dorsal portion of the chalazal end of the seed. Between the forms a and f there is nearly a complete series of connecting links represented in the material in my possession, but in the present paper I shall not attempt to deal with this whole series, because the genetic relation between some of its members is not yet worked out. Forms a-c will here be considered as one, under the name ordinary eye or small eye. Forms e and f will be treated as one, under the name Holstein, from the color pattern of a variety hav- ing this name. The evidence thus far available indi- cates that form d, which is here called large eye, is al- ways heterozygous between Holstein and small eye. Perhaps other of these forms of eye are also hetero- zygous; material which it is hoped to obtain from the crop of 1911 (F,) will probably determine this point. There is some evidence that forms a, b and c are the DD, DR and RR of a Mendelian pair. The same is prob- ably true of e and f and a type intermediate between them. Form g of Fig. 1 represents a genetically dis- tinct type of eye. In it the pigmented area surrounds the hilum, but the micropylar end of the area has the margin very indistinct; fine dots of pigment extend over the micropylar end of the seed. In races of this type that are black pigmented the pigment extends over the whole surface of the seed, but it is much denser in the region of the eye than elsewhere. The reduced pigmen- tation in such cases varies from very dark, rendering close inspection necessary to detect the eye, to very pale, making the eye as conspicuous as it is in buff or red pig- mented races. Whether these variations in intensity of pigmentation have genetic significance is not yet deter- mined, but investigations now in progress will, it is oe hoped, settle this point. This type of eye, in which an ill-defined area of pigment extends over the micropylar No.537] INHERITANCE OF THE “EYE” IN VIGNA 515 end of the seed (see g, Fig. 1), is here called the Watson Eye, from a variety known as Watson’s Hybrid, which has this type of eye. This variety, as well as the Hol- stein variety, is said to have originated from crosses between black varieties and Black Eye. My investiga- tions show that both these color patterns occur in the second generation of all crosses of this kind, as will be seen later, thus confirming the supposed origin of both these varieties. The Watson type of eye has not heretofore been recog- nized. But since the essential difference between it and other eye types was noticed, the writer has found in the collection of cowpea varieties in the office of the agros- tologist several races having this type of eye. The ex- istence of this type was discovered as a result of the discovery that in several crosses between eyed and com- pletely pigmented races the ratio of fully pigmented to individuals with eyes in F, was 3:1 in certain cases and 9:7 in others. The latter ratio suggested that two genetically distinct types of eye were present in certain crosses. A careful study of the rather extensive mater- ial at hand revealed the types of eye already described, and a tabulation of the statistical data relating to the inheritance of each of them revealed the relations be- tween these types which are discussed in this paper. STATISTICAL Data Cross: Small Eye X Solid Color (fully pigmented) In all, 21 crosses of this kind were made. In all cases F, was solid color. It happened that in classifying the individuals of the F, generation, before the types of eye present were fully recognized, Large Eye and Small Eye individuals were not separated in 5 of the 21 fam- ilies. In the remaining 16 families, taken together, the proportion of the various types of color pattern in F, was as follows: Solid Watson Large Small Color Eye Holstein Eye e Number of individuals in the 16 families taken together ..... 127 41 12 34 12 Highest expectation ........... 126 42 14 28 14 516 THE AMERICAN NATURALIST [Vou. XLV on the assumption that the ratio between these various types is 9:3:1:2:1. The reason for assuming this ratio will be given later. In the 5 families in which the large and small eyes were not separated, and in which the corresponding ex- pectation is the ratio 9:3:3:1, the results were: Solid Watson Large and . Color Eye Holstein Saali Eye Number of individuals in 5 families CARER OVO o ea e aaa 100 30 34 16 Highest expectation .............. 101 34 34 11 In both these groups the actual numbers are in close accord with the theoretical numbers called for by the assumed ratios. Cross: Holstein X Solid Color Six crosses were made between various types of the Holstein pattern and solid-colored races. The F, plants all produced seeds that were fully pigmented. In F, taking the six families together, there were 75 individ- uals bearing fully pigmented seeds and 25 with Hol- stein seeds, or exactly the ratio 3:1. Cross: Watson Eye X Solid Color Two of the 4 Watson parents used in these matings proved to be heterozygous. The other two, taken to- gether, gave 56 solid color and 13 Watson individuals in F,. F, in all these cases was fully pigmented. The 3:1 ratio for the number of individuals in F, is 52:17, which agrees fairly well with the experimental results. Cross: Small Eye X Holstein Only one cross of this character was made. It gave F, with large eye; in F, there were 3 Holstein, 4 Large — Eye and 1 Small Hye individuals, which is near the ratio 1:2:1 for these three types of color pattern. MENDELIAN FACTORS INVOLVED We may bring together here the various ratios found in the above classes of matings. For convenience let us designate the various color patterns as follows: No.537] INHERITANCE OF THE “EYE” IN VIGNA 517 ee aged te pigmented seed coat). a Wats ae er ais stds ig = Large Eye. H = Holstein. Crosses Rati EXS skera oi. HXS 3:1. wxs 3:1. EXH 1:2: These ratios point clearly to the following conclu- sions: 1. Types E and S differ from each other in two fac- tors each of which exhibits the phenomenon of domi- nance or partial dominance. 2. Types H and S differ in one factor which shows dominance. 3. Types W and S differ in one factor which shows dominance. 4. Types E and H differ in one factor, the heterozy- gote being intermediate between the parental types. It is not difficult to formulate, in terms of the cus- tomary hypothetical factors, an hypothesis that readily explains these facts. In fact, I have been able to formu- late four such hypotheses. In each of these a certain set of factors is assumed, each factor being assumed to have the power of producing a particular effect on the color pattern of the seed coat. But such explanations offer no suggestion as to the real nature and modus operandi of these factors. Unfortunately we do not know much concerning the complex chemical processes that go forward in the living cell. But enough is known to show that differences such as are exhibited by the various color patterns here considered may be due to Some such cause as slight difference in the rate at which some enzyme may be produced in the cells of the seed coat. For instance, it is known that an enzyme may cause certain sugars and chromogens to unite to form a glucoside, thus removing the chromogen from the sphere of action of an oxidizing enzyme that might con- 518 THE AMERICAN NATURALIST [Vou. XLV vert it into pigment.? Another enzyme, or, under ap- propriate conditions, the same enzyme that brought about the synthesis of sugar and chromogen into gluco- side, may hydrolyze glucoside into sugar and chromo- gen. It is also known that the rate at which a chemical reaction determined by an enzyme goes forward de- pends on the amount of enzyme present.. It is there- fore readily seen that the rate at which a given enzyme is produced in the cell may determine whether or not chromogen shall be available for conversion into pig- ment. It is also highly probable that a principal func- tion of the chromatin of the cell is to produce the enzymes which govern at least the rate of many of the metabolic processes in the cell.* It is certain that environmental conditions during ontogeny determine whether pigment shall be produced in a given tissue, even when the potentiality of pigment production is known to be present. Thus, in seeds that have pigmented coats, pigment may not occur, say in the endosperm. Causes similar to those which determine the particular tissue to be pigmented may also deter- mine what portion of that tissue shall be pigmented. The sugar in the pigment cells of the seed coat is pre- sumably transported there by osmosis from cells some distance away. Other materials necessary to the reac- tions may be brought from other parts of the organism, and some of the products of a reaction the accumula- tion of which might retard the reaction may be trans- ported to other parts of the organism as they are pro- duced. We thus have to deal with an exceedingly com- plex problem, many of the elements of which can not even be conjectured in our present ignorance of cell metabolism. It would therefore be idle to attempt to formulate a definite theory of the processes involved. A slight change in the permeability of certain cell walls, _a change in the size of certain cells, the substitution of ***On the Formation of Anthrocyanin,’? M. Wheldale, Jour. of Gen., 1, No. 2. * Guyer, M. F., ‘Nucleus and Cytoplasm in Heredity,’’? AMER. NAT., May, 1911.. No. 537] INHERITANCE OF THE “EYE” IN VIGNA 519 a slightly different group of atoms for a single side chain in a molecule of chromatin or one of the constit- uents of chromatin, or any one of numerous other changes, might determine whether pigment is to be formed in a given cell, or whether it may be produced at all. But if the change which is responsible for the dif- ference between two related organisms is a change in any permanent organ of the cell, then the difference in question will be hereditary. If it occurs in the material of a chromosome, or any other cell organ that behaves as a chromosome does in the reduction division, the dif- ference in question will Mendelize. Let us suppose, merely for purposes of illustration, that the difference between fully pigmented seed coat and the Holstein coat pattern is due to a difference in the rate at which a particular chromosome manufactures a particular enzyme under given conditions. Then when these two patterns are crossed we should get the usual phenomena of monohybridism, with the ratio 3:1 (or 1:2:1). Thus the Holstein pattern is not necessarily due to the ‘‘loss’’ of a ‘‘factor’’; it may be due to some such cause as a difference in the quantity of an enzyme produced by a particular chromosome. The use of such expressions as ‘‘presence of a factor’’ and ‘‘absence of a factor” in what follows is therefore not meant to imply the presence of a morphological entity in one race and its absence in another. It rather means that in one race some cell organ, probably a chromosome in Men- ` delian inheritance, performs a certain function differ- ently, or under different conditions, in the two races. Since the phenomena of Mendelian inheritance point clearly to the physical behavior of some cell organ, I prefer to think of the symbols used in expressing the genetic constitution of a type as representing the bodies, differences in the functions of which give rise to the character ‘‘pair.’’ Thus the symbol W in what follows may be considered as representing a cell organ which, under certain con- ditions, performs a certain function in such a way as to account for the difference between Small Eye and Wat- 520 THE AMERICAN NATURALIST [ Von. XLV son Eye, while w represents the same cell organ, or rather the corresponding organ, in another variety, which does not perform this same function in the same way under similar conditions. The symbol w need not imply that the power of performing any function is lost. It is here meant to imply only that the function is not performed in such manner as to produce the effect that W would have produced had it been present. Another way of putting it is that W represents a function per- formed, while w represents that the function is not per- formed. With this understanding of what is meant by a Mendelian ‘‘factor,’’ we may proceed to examine the hypotheses which explain the statistical results given above. The behavior of the cross: small eye X solid color in- dicated that these two types differ in two factors which are transmitted independently of each other. Let us represent these factors as they appear in fully pig- mented peas by W and H, and in small-eyed peas as W and h. This implies, according to my conception of Mendelian factors, that some cell organ (W), probably a chromosome, performs a certain function in certain races of peas that is either not performed, or is per- formed differently, by the corresponding organ (w) in another race of peas. A similar remark applies to H and h. With reference to these two factors, the formula of fully pigmented peas may be written WWHH, and of small-eyed peas wwhh. We have already seen that the Holstein type differs from the type with small eye in one factor; also from fully pigmented in one factor. Hence it must have either the formula WWhh or wwHH. That is, in one of the two factors concerned it is like Solid Color, in the other, like Small Eye. We may therefore take wwHH as the formula of the Holstein type. In an exactly similar manner we arrive at the formula WWhh for the Watson type. We may bring , these formule together for purposes of comparison. : (1) Solid Color H (2) Watson Eye : WWhh. (3) Holstein : wwHH. (4) Small Eye : wwhh. No. 537} INHERITANCE OF THE “EYE” IN VIGNA 521 These formule give the ratios previously assumed in dealing with our statistical data for the F, generation, provided we assume W and H dominant respectively to wand h.t An interesting deduction from them is that the cross between the Watson and the Holstein types should give the same ratio in F, namely, 9:3:1:2:1, that the cross between solid color and small eye gave. In the original crosses, the cross Holstein Watson was not in- cluded, but it is included in crossings now being made. These formule give some hint as to the nature of the effect produced by the factors W and H. Comparing (4) and (3), the factor H has the effect of enlarging the pig- mented area from the small eye type to the Holstein type (see b and e, Fig. 1). Comparing (2) and (4), W is seen - to have the effect of enlarging the pigmented area of the Small Eye type, changing it to the Watson type (see g, Fig. 1). W and H together, even in the hybrid WwHh, have the effect of spreading pigment over the whole seed- coat. We have seen that the heterozygote between Holstein and Small Eye, which has the formula wwHh, is inter- mediate between the parent forms. In earlier pages this type has been designated Large Eye. It therefore ap- pears that H duplex enlarges the pigment area about twice as much as H simplex does. Whether the same is true of the factor W has not yet been determined. As the ratio 9:3:1:2:1 is somewhat unusual, it may not be out of place to illustrate the manner of its occurrence. In the cross: Small EyeX Solid Color, F, has the formula WwHh, and is solid color. Generation F, is as follows: F, OF THE Cross SMALL EYE X SoL Color. == wwhh X WWHH Solid Color Watson Holstein Large Eye Small Eye 1 WWHH 1 — — —- 2 WWHh 2 — cake 1 WwWhh — 2 = — — 2 WwHH 2 oe tie! a wk 4 WwHh 4 ew aa wen al 2 Wwhh _ 2 — —- — 1 wwHH — saan Í — ave 2 wwHh — sas jae 2 Mee 1 wwhh — — = oa 1 9 3 1 2 1 t The dominance is only partial in the case of H and h. 522 THE AMERICAN NATURALIST [Von XLV Should it later prove possible to separate the two Watson F, types, WWhh and Wwhh, as we can the corresponding Holstein types, we should then have the interesting ratio $2 251-251. There is evidence in my material that the Holstein and Small Eye types are still further influenced by one or more additional factors. This matter is now under inves- - tigation. These additional factors appear to act in a manner entirely similar to the factor H, but have less effect. Taking the formula for Small Eye and adding a factor I, similar in effect to H, we should have the three types wwhhII, wwhhli, and wwhhii, all three of which appear to be distinguishable, giving probably the three . types a, b and c of Fig. 1. This point will be studied in F, of the original crosses, of which a vast quantity of ma- terial is growing, as well as in new crosses now being made. OTHER HYPOTHESES In the above hypothesis the factors W and H were assumed to have the property of enlarging the pigmented area. The facts can also be explained by assuming that w and h have the power of reducing the pigmented area. In fully pigmented races, on this hypothesis, these factors are absent. When w is introduced the pigmented area is reduced to the Holstein pattern. h, without w, reduces it to the Watson pattern, while w and h together produce the Small Eye pattern. It is necessary in this case to assume W and H dominant, respectively, to w and h, as in the first hypothesis. We may also assume a factor W for Watson Eye, a factor, E, allellomorphie to W, for Small Eye, and a third factor, S, which tends to enlarge the pigmented area, converting Small Eye into Holstein, and Watson Eye into Solid Color. Here, W must be epistatic to E. Under these assumptions the formulæ of the principal color patterns would be: Solid Color : WWSS. Holstein EESS. No. 537] INHERITANCE OF THE “EYE” IN VIGNA 523 A fourth hypothesis is as follows: H =a factor for Holstein pattern. E =a factor for Small Eye pattern. S—a factor which enlarges the pigmented area, converting E into Watson, and H into Solid Color. H is here epistatic to E. All four of these hypotheses are in complete agreement with the statistical data, and lead to exactly the same types, and the same ratios, in all generations. If we look upon the symbols as representing cell organs, differences in whose functions are responsible for the phenomena observed, then these four hypotheses are identical. They all provide exactly the same set of cell organs. The differences between the hypotheses are found only in the nature of the functions which these cell organs are supposed to perform. As we know nothing definite about these functions, the hypotheses are essen- tially identical within the range of present knowledge. The functions assumed for the bodies represented by the symbols of the first hypothesis seem to the writer to accord more nearly with our meager knowledge of cell chemistry. The author desires to acknowledge his indebtedness to Mr. G. W. Oliver, who not only performed the cross- fertilizations planned by the writer, but freely made available similar extensive material of his own, which is included with my own in these studies. Mr. Oliver also grew the F, hybrids. He is also under obligations to Mr. J. W. Froley, Mr. E. D. Carmack and Mr. W. R. Hum- phries, who grew the F, plants and made the necessary field notes; to Mr. E. P. Humbert, who rendered valuable assistance in classifying the F, material, and to Pro- fessor C. V. Piper, agrostologist of this bureau, who fur- nished seeds of the varieties used as parents in these investigations, and made many helpful suggestions con- cerning probable ‘‘factors’’ present in the varieties. HEREDITY OF HAIR FORM AMONG THE FILIPINOS ROBERT BENNETT BEAN, M.D. ASSOCIATE PROFESSOR OF ANATOMY, THE TULANE UNIVERSITY OF LOUISIANA Warme connected with the Philippine Medical School in Manila, in the year 1909 two of my pupils, Maria P. Mendoza and Manuel Ramirez, became interested in the heredity of hair form through the work of Gertrude C. Davenport and Charles B. Davenport! and they collected the records of 36 families, largely Chinese-Tagalog crosses, although two families were Negritos who had married Filipinos. They tested the hair form by making sections of hair dipped in thick celloidin hardened in 70 per cent. alcohol and cross-sectioned with a hand microtome, after which the sections were examined under the low power of the microscope, the measurements being made with an ocular micrometer. They divided the forms of hair into— Straight, with diameters of 100: 90 or over, Wavy, with diameters of 100: 70-90, Curly, with diameters of 100: 60-70. They decided that wherever a union occurred between individuals with straight and wavy hair the straight hair predominated. It seems expedient, however, to consider the hair form in single families as wel] as en masse. In any discussion of heredity it seems necessary to consider individuals rather than the mass, except in the formulation of laws that take into consideration prob- able errors in the mass. In Table I, showing the result, in mass, of crossing different hair forms, we should expect, if Mendel’s laws prevail, where wavy and straight are _ crossed or where.curly and straight are crossed, to get an = equal number of offspring with straight and curved hair z AMERICAN Naturauist, Vol. XLII, April, 1908. No. 537] HAIR FORM AMONG FILIPINOS 525 TABLE I COMPILED By MENDOZA AND RAMIREZ 2d Sa PE Children Expectation No. F eae Me MF M | s P ante c w S c s Cie ye S, | | tie sa si 4 iD] R 3i as | | pas a 3 is | RR 6 wW | on | s {Ss sjal 1 SR È 3 Pee a i be ioe Rk | BR a c lw Poe he hoe te 1 o COR DR! 53 | 1.7 S iie. A E | ed i R DR 1 10 | -¢ a ee ipni pR gs] Se RR e 1 | 1 RH] R 2(?) Reis oig i 1) DR DR ea 7 o eee Wag oe ili DR niis] 35 : {5 3 a| i 2 1 1 [DR(?)) R | 2 2 15 P i Ili ti R 1g] ià C C wis a 2 1 1i DRR]? 2 mie © {8(C)| 4 DR | DR | 4 0 18 se $| É DR I6] l6 : : 9 g ea wt g 8 20 x 3 BIE 4 21 ~ ee TOR 6 ae IW 22 W z Yvi? 4i i T OR | DR 88| I7 23 : o ; 1 1 1 | R | DR|15 | 15 24 5 sigil 6 ni DD]: 2 a a Se 5 | RADR 25| 25 #3]. bik | OR a5 | 25 8l oii Hig A y; ete PRT is L5 526 THE AMERICAN NATURALIST [Vou. XLV | "n | Parent Children Expectation No. | —| -E gers ba 3 | FF M a | FM MM | “3 bod 8 Tone S e| a ia 28 c | s | ¢ 1 ¢ | BR | DBT 416 [18 | S C Mie F 2 i DR} BR iIi Sirag)? 3 | BR | DR/| 25 | 25 e 8 hf. ele ilija ET KI eK 32 £l g 4 DD | BR| 4 3, = 2 1 1 DE. B |2? 2 34 T aitai 1s m n ee e = a Yi ég] Ss imirt i arp SiS signar S do meai T Y - : Wo. o 3 + 4 DE | PRE 3 1.25 eli "Ie ey sizing o AO ee cath A Nw) = ate 13 R | DR |15 |15 40 sijil] i| E] f i: Soa Uiz 0i 1 5 * DEE C 3 ulo =) 6) 2 DR | DR | 1.5 | 5 eee Bier ba ie oe ee ee wsi vw ‘ M.S oe | @ tt | 1.) 2. DR nnn t 2 eigi iois n 5 | 35 46 L al? iit o oe F= father, M= mother, D= dominant, R= recessive, S= straight, W=>wavy, C=curly, K=kinky. where the curved variety is a dominant heterozygote and the straight is a recessive homozygote. Such is the case, for of the 31 families examined in which wavy is crossed with straight or curly with straight, there are 157 chil- dren of whom 84 have straight hair and 73 have hair curly = or wavy. Where the wavy and the curly are crossed, if each is a heterozygote with the straight recessive, we should expect to find three children with the curv ed No. 537] HAIR FORM AMONG FILIPINOS 527 proportion, for of the five families where wavy hair is crossed with curly, there are 33 children with hair of the curved variety and eight with straight hair. When individual families are considered, however, there seems to be an alteration of the dominance in some eases. Straight hair is dominant in some families, whereas in others the wavy or curly is dominant. In Table I straight hair appears to be dominant in the families num- bered 3, 6, 11, 18, 24, 25, 26, 27, 31, 34, 38 and 45. In the first three of these families where both parents have straight hair, and particularly in family 6, curly hair is evidently recessive. If the gametic composition of the two parents in family 3 is DrX Dr, in which D = straight hair dominant and r= curly hair recessive, then the expected proportion of offspring would be 3 straight to 1 curly haired child. The number of curly-haired children is 3 instead of 1, which is greater than it should be. But if the straight hair be recessive in this family, the gametic composition must be rrXrr where r is straight recessive, and all of the children should have straight hair, which is not true. In any case the number of curly-haired chil- dren is in excess of what it should be. The curved form of hair seems to be dominant in a greater number of families than is the straight, as in families numbered 1, 4, 10, 12 to 16, 22, 28, 29, 33, 35, 36, 39, 41, 43 and 44. In family 1 the proportion of curly- haired children is in excess (excessive dominance) of what it should be even though the curved be dominant. The same is also true for families 4, 10, 13 and 39. There is, on the other hand, imperfect dominance (recessive dominance) of the curved hair in families 14, 15, 16 and 29; or, in other words, there are fewer children with curved hair in these families than would be anticipated if the curved form be dominant. When kinky hair and straight hair are crossed, as in families 17 and 32, there results an intermediate form— in one family all the children have curly hair, in the other all have wavy hair. The result of crossing two of these 528 THE AMERICAN NATURALIST [Vou. XLV intermediate forms is not known unless such a cross is represented by families 8, 12, 22, 37, 42,43 and 44. There is a tendency toward segregation in these families, but no kinky hair appears. Judging from this cursory analysis of the families under consideration, the inferences are that in the cross- ing of different grades of hair-form there is a tendency towards segregation or alternate heredity, a modified Mendelism, for there is not a perfect dominance, nor the opposite, for either of the extreme grades of hair form, kinky (woolly) or straight. It would be of interest to know what relationship the curly, wavy or kinky hair of the European bears to the same sort of hair of the Negrito. Presumably all of the curly or wavy hair in the families under consideration is of European origin except in families 17 and 32 of Negritos, but in no instance can the previous inclusion of Negrito blood be absolutely excluded. Some of the wavy or curly hair may therefore be due to previous crosses of the Filipino and Negrito. Presumably, also, all of the straight hair is derived from the Filipino, but again the European or Chinese straight hair can not be excluded. It may be that the straight hair of the Chinese and of the European would each have a different re- action in heredity with the curved variety of both the European and the Negrito. Unfortunately, observations to determine this may not be possible owing to the mixed condition of the population at the present time. 3 It may be well to tabulate the families in which the parents are alike and unlike in hair form and to note the kind of hair of the children; in order to determine some- what exactly the relative behavior of the different forms of hair in heredity. The following crosses will be con- sidered: Group I, straight straight; group II, curly X curly; group III, wavyXwavy; group IV, straightx _ curly; group V, straight < wavy; group VI, curly x wavy- In this and subsequent tables F. = father, M.= mother. C.— curly. W.= wavy, S. = straight, EF. = No. 537] HAIR FORM AMONG FILIPINOS 529 TABLE II Group I, STRAIGHT X STRAIGHT | Children Parents No FF | EM M.F. M.M s | C. W. STE Oe y M. 3 — |} — — —- 3 aS S 6 Bones S S 1 2 S S 11 Cc S S S 1 1 | S S 19 S S S A . 3s S 20 S S Si B | 4.558 S 21 S S G o Boa Oi 9S S father’s father, F.M. = father’s mother, M.F. = mother’s father and M.M. — mother’s mother. Here are three families (3, 6, 11) in which the parents appear to be simplex or heterozygotes, and three (19, 20, 21) in which they appear to be duplex or homozygotes, and the straight hair is dominant to the curved variety. The parents in the first three families would have a gametic composition of Drx Dr in which D represents the dominant character, straight hair, and r represents the recessive, curved hair. In any cross of this nature an equal number of children with straight hair and with the curved variety would be expected and the expectation is realized. The parents in the second three families would have a gametic composition of DD DD, where D repre- sents the dominant character, straight hair, and only straight-haired children should result from such unions. The gametie composition of the last three families could be rrXrr, where r is recessive straight hair, but if straight hair acts as the dominant character in the first three families there is no reason to suppose that it would be recessive in the last three. TABLE III Groups II anp III, CURLY X CURLY AND Wavy X Wavy Children Parents No. F.F. | F.M M.F. | MM, C. W. sS. F. M. 42 — — — — 2 Cc C 22 — Ww — — 2 4 1 WwW WwW 43 — S (0 ao 1 1 2 Ww Ww 44 WwW C S wW 2 1 2 W WwW Total 7 6 5 Jf 530 THE AMERICAN NATURALIST [Vou. XLV Family 42 is apparently duplex or homozygote with a gametic composition of ceXce, but there is nothing to indicate whether the curly hair is dominant or recessive. The wavy haired families behave as if simplex or hetero- zygote, producing both straight and eurly-haired children. If wavy hair is produced by a combination of straight and curly it should behave as it does, representing the heterozygote of the F! generation, and in the next genera- tion (F?) there is a return of the two original forms of hair, curly and straight, and a continuance of the wavy. TABLE IV Group IV, STRAIGHT X CURLY Children Parents Children No F.F. | F.M. | M.F. | M.M. | paze Wii & | Be E N wees 7 = p C S 1 3351 ¢ 14 C Y S S Ce + a | 1 15 C WwW S S GIB Ei 1 16 c wW S S CiS 21 Fi 18 a 8 S S 3| 8 C 23 S Cc C S C Fi 1 26 S S S — 1 67°81 4 27 S S C S 3i S C 28 S Ç S S 1 21A iO 29 C © S S OS 2 1 30 S S — — 2 Sip 2, 33 -— —_ ini EIB Aa 1 39 — — S Cc sic 3 40 — — sone sew | a. 7.6 t 1 45 BER m C — | 2 8| B C Total | Pe lg | H o LL In some straightX curly families, curly hair seems to be dominant, as in 14, 16, 29, 33 and 39. Whereas m others straight hair seems to be dominant, as in 7, 18, 26, 27, 28, 30 and 45. It will be noted that whatever the character of the father’s hair, that character is dominant except in family 39. It is also to be noted that straight hair is dominant when three of the grandparents have straight hair, whereas curly hair is dominant when only two of the grandparents have curved hair. If the hair form of the dominant parent is duplex no recessives. are to be expected, but if the dominant parent is simplex half No. 537] HAIR FORM AMONG FILIPINOS 531 of the children will have little or no dominant hair form. Neither of these expectations is met, except in the families where curly hair appears dominant and there the latter hypothesis fits the facts—14 children have curly hair, 13 have not. We may be dealing with a case of dihybridism in which there is one masked character.? The two pairs of characters may be: straight dominant to its absence (curly), and straight dominant to wavy. We should expect on this supposition to get straight, wavy and curly in the proportions: straight 9, wavy 3 and eurly 4. The actual number of children with the different kinds of hair is straight 33, wavy 6 and curly 21. The results of crossing straight and wavy hair may profitably be considered at this point. TABLE V GROUP V, STRAIGHT X CURLY Children | Parents Children No F.F F.M. | M.F. | M.M CIW FB | F. | MCW 8 1 W S S S wis 3 1 1 2 — — wi — 3 4i WIIS 4 ee — WwW S S W 1 5 2 5 — — — WwW 1 LES iW. 9 S — — S 1 1; 8 W 10 C S S S WIB 4 1 2 13 —— — S W wis 2 2 1 24 S S — — Ji Si 8 WwW 25 S S — = 5IS |W 31 S S S S T 1 4-80 W 34 — a = — } 1 5}; wis 35 == ae — ma wis 1 3 2 36 — — S S 1 2 8158 WwW 38 S S — w 3 TIS W : 41 os S S Cc wis 1 5 6 46 — aan C S 3 4-85 W Total att | 13°) 171 44 The father is again prepotent, but there are more ex- ceptions in the straight wavy cross than in the straight Sstion rox va % Qyserver the only changes made were concerned with the blossoms at the points in question. One blossom at least in each experiment served as a control. See the accompanying triangular diagram. Experiment No. 1—July 26. Time of observation, one hour—from 9 to 10 a.m. At (a) a normal blossom was growing in situ as a control. At (b) a normal blossom was ine in situ as a control. At (c) petals only of a blossom were carelessly pinned to a stem. An Elis plumipes once alighted on the petals of (c), but immediately discovered the deception and flew away. The species of Melissodes in no instance alighted. These bees, it would seem, possess rather keen discerning ‘powers, since without alighting they quickly detect the difference between a normal and a mutilated flower. In many instances, however, they inspect very carefully No. 538] THE BEHAVIOR OF BEES 611 a suspicious blossom before passing on. From this test it is evident that the petals alone, as used at (c), were quite as efficient in inviting inspection as the normal blossoms themselves. The blossom at (b) received fewer inspections, probably because it was less readily per- ceived among the leaves which nearly surrounded it. The bee visits were distributed as follows: DATA FOR EXPERIMENT No. 1 Distribution by Species — Distribution by Species Total Entrances Points of Triangle Total Inspections Elis plumipes _ Honey Bees Bumble Bees Unidentified Bees Melissodes sp. Elis plumipes Honey Bees Bumble Bees Unidentified Bees N | Melissodes sp. he | SE 0 p O m OO e m A Deo A (c) 1 Experiment No. 2.—July 26. Time of observation one half hour, from 10 minutes past 10 to 10:40 a.m. Blossoms in the triangular arrangement as before. See diagram. At point (a) same blossom with petals removed. At point (b) same blossom entire as a control. At point (c) petals alone as in preceding experiment. DATA FOR EXPERIMENT No. 2 Š% z Distribution by Species z Distribution of Species a © e paunu et anes 38 = : % a z g H = Cl errare E Tr Sot Bot S 1S) al She. | e ee ee ee ee ee ee a = Š Š s els| Bs G Š & S 2e à = = ee =) (a) 4 4 0 (b) 5 50 5 5 ( 62 0 TU ee eo! The removal of the corolla of (a), which in the past experiment received 81 inspections, reduced the number of inspections at once to 4 as compared with 62 inspec- tions of the detached petals at (c). Exactly similar results were obtained by Darwin in 612 THE AMERICAN NATURALIST [Von XLV his experiments with Lobelia erinus. He says: “I cut off the petals of some and only the lower striped petals of others and these flowers were not once again sucked by the bees, although some actually crawled over them. The removal of the two little upper petals alone made no difference in their visits.’’ Experiment No. 3—July 26. Time of observation one half hour, from 10:40 to 11:10 a.m. Blossoms in the tri- angular arrangement. At point (a) 3 petals are replaced loosely. At point (b) same blossom entire as before. Control. At point (c) petals alone as in preceding experi- ments. DATA FOR EXPERIMENT No. 3 EA = Distribution by Species xt Distribution by Specie a E : $ PORET A E N 5 E 2 ee | E ; 2 n| is AYE} e/a lglale + é-}@48 lag y a Š = m| 4! Za a 3 = a; |) se xi i © = Ble BS = 3 = Be go 3 au a & D w | 2A = a a Q 2 ah S41 it te Boat ele ee Se ee S S = = 5 > <4 À x = a IH a p = S SE p (a) | 53 51 2 | 0 (b) 51 1 | oi (e) 67 65 4 | = With these petals now carelessly affixed to the blossom at (a) from which the corolla had been entirely cut away, the number of inspections is at once as frequent as for the control at (b). It appears that color and texture more than normal form and arrangement first direct the bees to inspect the blossoms. _ Experiment No. 4—July 26. Time of observations one half hour, from 11:15 to 11:45 a.m. Blossoms in the triangular arrangement with the following change from the preceding experiment. At (a) cloth petals of an artificial rose are carefully arranged and pinned in position to simulate an open cotton blossom. The color of those petals approximated the creamy yellow of a natural cotton blossom; the tex- ture, however, was very different. At (b) control. Same blossom entire as in the preced- s ing , experiment. - No. 538] THE BEHAVIOR OF BEES . 613 DATA FOR EXPERIMENT No, 4 | || 3 a. Distribution by Species n Distribution by Species e | 3 : © i 3 S oot j ú 3 2 a Po i TRS ee eae Sale gy BE; eo THp = S È -EE-D Z 3 = = coe n Š = +2 rs Š = = Pn ° = 3 Š w) et SS a > Š A N i 2 cs S < | 8121 8a "a A x S| S| £s a 3 3 $ ee be 2 3 3 | 3/8! 3 £ È ` 5 | Hilal Pe = = S qi a |» (a) 6 6 o | (b) 48 45 I 2 9 Lox _© 65 62 | I o | At (c) control. Petals only as in the preceding ex- periment. Bees have been very little attracted by the artificial cloth petals at (a). Although the color is not precisely that of a cotton blossom, several bees gave evidence of having perceived them. The texture, which is that of coarse meshed cloth, is quite unlike that of cotton petals, however, and may have been readily perceived as un- real by the bees. The few inspections were without doubt invited by the color of the artificial petals, since no odors could be considered operative unless of a repel- lent nature. Experiment No. 5—July 26. Time of observation one half hour, from 11:45 to 12:15 a.m. Blossoms in the triangular arrangement with the following changes: At (a) five cotton petals (normal number) are care- lessly placed over the artificial cloth petals. At (b) control. Same blossom entire as in the pre- ceding experiments. At (c) petals alone as in the preceding experiments. DATA FOR EXPERIMENT No. 5 3 á Distribution by Species g Distribution by Species a £ s s +. i = F i ei; Ele) eann tF T: 3 8 = E m] o| 58| a $ 7 ote P ~ iA 22 z S p S| Z| 25 = a 3 Gs 2 = S $ a ke g 3 = a 16189 3 S = 3 |8| 8/3 Eie [g| a Re) al a Ble (a) 48 45 3 1 b) 44 43 1 (e) 50 | 48 1 1 614 THE AMERICAN NATURALIST [Vou. XLV It is now evident that all the blossoms serve equally well to invite inspection. It is plainly indicated that the artificial cloth petals could have possessed little or no repellent odor, although they received very few inspec- tions in the experiment just preceding. It is not improb- able that the different texture of the material revealed the artificial nature of the cloth petals to the bees. Experiment No. 6.—July 27. Day cloudy, showery in forenoon, thus greatly interfering with frequency of visits. Blossoms in the triangular arrangement. At (a) control. A normal blossom pinned in position. At (b) control. A normal blossom growing in situ. At (c) a single petal pinned to a stem. Observations were begun at 9:00 a.m., but rain inter- vened at 9:15. A single inspection was recorded for (c). Observations were again begun at 10:25, lasting for one half hour until 11:05. The blossoms were arranged in the triangle as follows: At (a) control. A normal cotton blossom pinned in position. | At (b) control. A normal cotton blossom growing in situ. At (c) a half opened bud simulated by pinning normal petals together, the calyx being represented by a por- tion of a green cotton leaf carefully wrapped around the base. In this way it was absolutely certain that no unaccustomed odors were introduced. This bud-like ar- rangement prevented all chances of examination of the inner details by bees until they had actually squeezed down between the petals. DATA FOR EXPERIMENT No. 6 ea Distribution by Species Fa Distribution by Species See E S So n S í | a i m E i Fiir |e) fre ig ais s | 3 E a] 3 5 š E |AIM] Za a E ties ae tS 3 s Lye ee ee A elol £ = z = o| 2] «wm ie i g BRS ite es Ne a i. etm |e imal E eo) | [Rae ØF 24 | 0 No. 538] THE BEHAVIOR OF BEES 615 A record of the kind of bees was not accurately kept, but species of Melissodes were almost the only visitors. The blossom at (b) was less visible than those at (a) or (c), both of which were in plain view of each other. The blossom at (b) was not visible either from (a) or (c), so that many bees which inspected (a) and (c) fre- quently failed to perceive (b). Experiment No. 7—July 27. Period of observations one half hour from 11:08 to 11:38. Blossoms in the previ- ous triangular arrangement changed as follows: (a) Normal blossom used in the preceding experiment concealed by fastening the surrounding leaves in such a manner that the blossom would be visible only by bees passing directly over it. (b) Control. Normal blossom growing in situ. (c) Artificially constructed bud as used in latter half of experiment 6. The inspections were as follows: (a) Received a single inspection from a bee flying directly over. (b) Received 12 inspections, two of these being en- trances. (c) Received 40 inspections, none being entrances. Experiment 7 differs from experiment 6 in no partic- ular whatever except in the change which has rendered the blossom at (a) invisible, except from a certain posi- tion. The number of inspections at (b) and (c) re- mained practically constant for each half-hour period. It is interesting to note, however, that (a), receiving 48 inspections in experiment 6, received but a single inspec- tion in experiment 7. A change in surroundings which makes a blossom less visible to the visual sense of bee visitors at once decreases the number of inspections. | Experiment No. 8—July 27. Time of observations ten minutes from 11:38 to 11:48 a.m. Blossoms in the triangular arrangement, with no change whatever from the preceding experiment except in making the blossom 616 THE AMERICAN NATURALIST [Von XLV at (a) again as visible as in experiment 7 by pushing aside the surrounding leaves. (a) Is inspected 15 times, including one entrance. (b) Is inspected 7 times, including two entrances. (c) Is inspected 13 times with no entrances. The blossom at (a) has now become as attractive to the bees as those at (b) and (c) which serve as controls. Experiment No. 9.—July 27. Period of observation 15 minutes, from 10:50 a.m. until 12:05 p.m. Triangular arrangement as in preceding experiments with the fol- lowing changes: (a) Petals of (a) in experiment 8 are removed and artificial crêpe paper petals of nearly the same color are substituted. (b) Artificial blossoms growing in situ as a control. (c) Artificial floral structure used at (c) in experi- ments 7 and 8. (a) Receives only two inspections. (b) Receives 16 inspections, including two entrances. (c) Receives 3 inspections, including one entrance. The artificial nature of the paper petals at (a) was at once perceived by the bees in their passing flights. The few inspections noted were indicated by a momen- tary pause in flight quite unlike the more prolonged hovering movements over the blossom at (c). Experiment No. 10—July 27. Period of observations 20 minutes, from 12:05 p.m. to 12:25 p.m. The same tri- angular arrangement was used as before. The only change from experiments 8 or 9 consisted in placing three real cotton petals carelessly upon the paper petals at (a) in such a way that only part of the paper petals was con- cealed. Blossoms (b) and (c) were left unchanged. (a) Receives 11 inspections. (b) Receives 7 inspections, including one entrance. (c) Receives 21 inspections. Passing bees were at once led to inspect the real petals placed at (a), although these very imperfectly covered _ the artificial paper petals beneath. No very decided re- No. 538] THE BEHAVIOR OF BEES 617 pellent odors can be held to reside in the artificial paper petals which failed to attract passing bees when used alone. Experiment No. 11.—July 27. Observations for this . experiment continued 10 minutes, from 1:26 to 1:36 p.m. The triangular arrangement was used. At (a) a single real cotton-blossom petal is pinned to a stem. At (b) a cotton bud and calyx simulated by neatly wrapping a portion of a cotton leaf around the base of five real petals rolled together. At (c) a normal open cotton blossom growing in situ as a control. DATA FOR EXPERIMENT No. 11 D am OT = a m E) Z | Distribution by Species s Distribution by Species a S So ~ . | a Í = ; ee be baeri g z S S tel sis Pete a ee. pej S N } a: Smee) Eee ae ee a cl eres 2} 3/3) 2} Fe 24] 314) 2 843) 34 - 1s feo ee eee > > @ te | sigts a a = 3 pa & | = 5 lah es ond es ees | | | — @ 12 | 2 bode bees 6 | 8 | 8 ae | () | 3 3 | | | The single petal at (a) is sufficient to invite the in- spection of passing bees, although there is little more than a fraction remaining of the size and color of a normal open cotton blossom. The writer has observed that a partly opened bud, as represented at (c), appears to invite more frequent inspection and entrance than a fully expanded blossom which has been much oftener entered by bees. It is possible that bees in their en- trances leave traces of odors which are detected by later visitors, causing them to pass on in search of fresher blossoms. ; Experiment No. 12—July 27. Period of observations one half hour, from 1:36 to 2:06 p.m. In this test, which practically duplicates experiment 11, two blossoms were used in the same row and on plants about three feet apart. 618 THE AMERICAN NATURALIST [Vou. XLV At (a) a single petal was pinned to a stem. At (b) control. A normal cotton blossom in situ as grown. : The single petal at (a) received 16 inspections, some of which were very thorough, as a number of bees ap- peared to examine the petal intently from all sides. The control blossom at (b) received 26 inspections, including 8 entrances within. In both cases the visiting bees were all species of Melissodes. Experiment No. 13.—July 27. Period of observations one half hour, beginning at 2:06 p.m. and ending at 2:36 P.M. Two blossoms were arranged in the same row as in the preceding experiment. At (a) a perfect cotton blossom was pinned in the same relative position as the blossom at (b). At (b) control. A perfect cotton blossom growing in situ. The blossom at (a) received 7 inspections, including one entrance. The blossom at (b) received 12 inspections, including 5 entrances. Species of Melissodes were the only visitors. ' Experiment No. 14.—July 28. Time of observations one half hour, from 9:15 to 9:45 a.m. Three blossoms were arranged on three consecutive plants in the same row. Throughout this series of experiments for July 28, these positions were unchanged. See the diagram. At (a) control. A perfect cotton blossom pinned in position. At (b) large blossom of a wild convolvulus (white with a deep purple throat) was pinned in position. DATA FOR EXPERIMENT No. 14 Distribution by Species © A Distribution by Species 3 ee i g 3 ; a : — Q -1 a n a mn co eS Tisina: EF S ae: w n $ = am) i) 3 g 3 2 ‘Sn pa j = = Be ae a E 3 biS E ES WG ae ee a ee ee ee 5 = = si S S 3 = Biel E & g = R | 21 a!5 a S 5 al p 6) | B] L 1 (b) 14 12 2 0 (c) 14 "AL No. 538] THE BEHAVIOR OF BEES 619 At (c) control. A perfect cotton blossom pinned in position. Although strikingly unlike a cotton blossom in color and general appearance, the convolvulus blossom at- tracts attention quite as often. It is hardly to be ex- pected that the bees would enter it as frequently as a cotton blossom, if at all, since it is a well-known habit of many bees to confine their visits pretty constantly at any one time to blossoms of the same species of plant. Especially has this been shown true for the honey bee by Hermann Miiller and others. M. H. Mendleson, of California, affords an instance where a single colony out of 200 visited solely mustard flowers, while the rest gathered from sage blossoms alone.? Experiment No. 15.—July 28. Time of observation one half hour, from 9:45 to 10:15 a.m. Blossoms ar- ranged in the same row as before with the following changes: At (a) petals removed from the blossoms of preced- ing experiment. | At (b) same white convolvulus blossom used in the preceding experiment. At (c) control. Same cotton blossom of preceding experiment pinned in position. DATA FOR EXPERIMENT No. 15 = : Distribution by Species | g Distribution by Species o9 s 3 2 R ; pai 5 2 ù% mn g A = e; o ETETE AIT rE E N 5 s Š Š A a Š § alsa Sig | 3 s | Sfel ea) 1 3 1 rn = S 3 3 SSF S Š 2 18181 3 & | R| 2 1.8 | Pia a | § | & |) a] 2 (a) 0 | 4 () | 19 | 17 2 | 0 | () | 20 | 20 0 | | x By the removal of the petals of a cotton blossom as at (a), the blossom no longer advertised itself to the atten- tion of bees, as has been demonstrated in previous ex- periments. In the present test, although the white con- * Mendleson, M. H., ‘‘Gleanings in Bee Culture,’’ October, 1908, 36. 620 THE AMERICAN NATURALIST [Vou. XLV volvulus blossom had completely wilted and collapsed, its noticeable color alone continued to invite inspection by passing bees quite as often as the control blossoms at (c). Experiment No. 16.—July 28. Period of observation one half hour, from 10:15 to 10:45 a.m. The three points in the preceding experiment were used as follows: At (a) a single cotton petal was placed on the blossom of the preceding experiment, from which all the petals had been removed. At (b) a single cotton petal was loosely pinned to a stem. At (c) control. A normal cotton blossom pinned in a conspicuous position. The blossom at (a) receives 8 inspections. The petal at (b) receives 9 inspections. The normal blossom at (c) receives 27 inspections with no entrances. All visitors were Melissodes, except a small bee which inspected (c). As the blossom at (c) was conspicuous from all sides, the writer judged that this fact accounted for the much greater number of inspections given this blossom, since (a) and (b) were visible almost wholly from one side only. In the next experiment this question was further tested. Experiment No. 17.—July 28. Period of observation one half hour, from 10:45 to 11:15 a.m. In this experi- ment the only changes from the preceding consisted in an interchange of material. DATA FOR EXPERIMENT No, 17 2 a Distribution by Species a Distribution by Species =) 3 g i a 3 pej A | | a : ! eli fii pis ts th 8. diag ey oe 3 Sa ae eee lies Meee | i 4 ieS 8s 3 | 2 tele i siaii S OG UE le ee ee ge 2\/8|818 [8iai5 | 6 | § | & lelals raii Å | | EN ELD A (a) 8 8 | 0 | @) | 22 | 21 xa 0 () | 27 | 25 a eee | No. 538] THE BEHAVIOR OF BEES 621 At (a) single detached petal pinned to leaf stem. At (b) control. Normal cotton blossom pinned in position. At (c) the cotton blossom with its single replaced petal at (a) in preceding experiment. In this experiment the more exposed position (c) ap- pears to be of considerable advantage to a blossom lo- cated here, even though its normal appearance is greatly changed by mutilation. The general form and appear- ance of a cotton blossom, as a whole, does not appear to play a very important rôle in initiating the procedure of inspection by passing bees, since a single detached petal receives quite as many inspections as a normal blossom. Experiment No. 18.—July 29. Observations continued one hour, from 8:20 to 9:20 a.m. In this experiment three blossoms were used, as in previous experiments, and arranged on consecutive plants in the same row. A blossom of an Asiatie cotton (Hawasaki) was com- pared with two ordinary American upland blossoms as controls. (a) control. Normal American upland blossom pinned in position. ; (b) Hawasaki blossom entire pinned in position. (c) control. Normal American upland blossom pinned in position. DATA FOR EXPERIMENT No. 18 = | 2 | Distribution by Species | z Distribution by Species z S | 9 ' 3 2 Ba ae | | a er hase | E mo Kd 3 = l 9 s? a krri el tle HEE B E 3 E || e| Zs a oe ee $ o| 38 an 3 Š wiat § N Eon s > 3| og E 3 | 4 S | | 8 se | 3 a3 1613/3" = honed . {=} © = a LRS 8] RAP | & ERMES iii Sarees | | | | | ! (a) | 20 | 19 io ad | | | | Oy) I4 | 12 1 1 0 | | | (c) | 29 | 24 pause. 2 | | L jij Experiment No. 19—July 29. Observations con- tinued one half hour, from 9:25 to 9:55 a.m. 622 THE AMERICAN NATURALIST [Vou. XLV (a) Control. Normal American upland blossom pinned in position. (b) Control. Normal American upland blossom pinned in position. (c) Hawasaki blossom entire (af b) in last experi- ment. DATA FOR EXPERIMENT No. 19 å 4 Distribution by Species 2 Distribution by Species S S ò ==) 3S k =I 3 | rie Baltes Ones Ai he ee ee ee e eB | 8 8 E IA R Se | & s E AIM] 3s Qa lam S = es ag = 3 >! 2 Es egos Pe eat ee tego & Pee ee = > = 5 © ‘a © e S Eo = = À £ = § |Rlals į = a | Rl alo (a) | 11 8 0 | (b) 0 7 1 1 | (c) 2 | Experiment No. 20—July 29. Observations con- tinued one half hour, from 10:50 to 11:20 a.m.. Three blossoms arranged in the same row as for previous ex- periments. (a) Control. Normal American upland blossom pinned in position. (b) Hawasaki blossom entire at (c) in experiment 19. (ce) Control. Normal American upland blossom pinned in position. DATA FOR EXPERIMENT No. 20 A a Distribution by Species 2 Distribution by Species re S o SOE lon : 2 2 = a : 2 n ke] Sa $ ie oe ae ae Bo) Big +i ge a9 =| G g = Se | Š = = Bm Che 4 s = mi 2| dE m Š 2 hi ej 8s an a S a e na = 2 o = > os 2 o ya a e | & a E |e 1 Ra S (a) 20 8 1 1 0 3 1 (b) 10 8 2 (c) 16 6 Lid 8 1 REPRODUCTION IN THE BROWN RAT (MUS NORWEGICUS) NEWTON MILLER CLARK UNIVERSITY, WORCESTER, MASS. 1 Mirodutton oon ov n.d Ns 0 METRE RS Vcd beak DAE LEE PORN oe 623 By Apparatus ss bn ORR E EGG iy os we a Fk ee ee AAS 623 S Raproduction s.. aran aaa ho eek a i es ee hes 624 Oe OES irrena rere eee eis oe r eee Pe Ao ye 625 bi Period. of Gestation ceri vie ck oak eee ca ones anes a= oe 626 C. Mating Habits 6s ee a I N ees 627 d. Eating the Young -nre eoe ve be coe nee eee oe aes eed 630 2. Care: of the: Young irii. Wik incense es Ce cas oe es a ee es 632 Di Growth Os Oe a eee ack pu wees oe dase 632 G- Conelasions iS. wy casa ee kd eke ona ken bd A E 635 In view of the fact that the brown rat is playing such an important réle in the economic field and the medical world at the present time, it is of interest if not of prac- tical value to know the details of its reproduction. The life history of this species as lived in its natural envi- ronment is as yet imperfectly known. It was to bridge this deficiency in our knowledge that data were obtained on a number of rats kept in the laboratory from No- vember 5, 1909, to December 8, 1910. My results may be much different from those of natural conditions, yet I am convinced that the results obtained are not radically different from what actually takes place in nature where shelter is good and food abundant. Two sets of cages of four each were built. The first consisted of revolving cages, a foot wide and eighteen inches in diameter with nesting boxes, measuring 855 inches, suspended from the axles. These cages were similar to those used by Slonaker. The sec- ond set of cages were rectangular, measuring 24 X 24 X 8 inches. These were made of screening, three wires to the inch. For the sake of cleanliness the cages were suspended, as shown in Fig. 1. Newspapers were spread 623 624 THE AMERICAN NATURALIST Von XLV beneath all the cages to catch the excrement and removed as occasion demanded. In all the cages, siphon watering troughs were used; i. e., bottles held upright with their mouths dipping into shallow dishes. Open vessels were objectionable because rats drop their excrement as they climb about in the cage, thus polluting the water. Fic. 1. Showing the plan of the cages. The experiment was started November 5, 1909, with seven pairs of rats which had just been caught. For convenience, each pair was numbered, the first four being in the revolving cages and the other three in the stationary. The pairs numbered 1, 2 and 5 were old (adults), the females apparently having given birth to young. The others were probably in their third to sixth months and had not given birth. Female No. 6, after giving birth to young on December 31, 1909, was severely injured by her mate. The wound was not yet healed September 12, 1910, when she was killed, at which time she was found pregnant. The fol- lowing table is the reproductive record of the other six pairs. No. 538] REPRODUCTION IN THE BROWN RAT 625 REPRODUCTIVE RECORD. FIRST GENERATION = je ie DO 2 SOnlodioOg| © |S D: ie] 2l, = Feb. 4, 1910 | 1] 1 All eaten February 5. Male not taken away. Feb. 26,1910 | 7 | 2 All dead-part eaten February 28. Male way Febru Feb. 27, 1910 | 1 | 3/11] 5 | 6 ale taken aar February 27. Mch. 14, 1910 2| 4 All dead July 21. Male Ls taken away. ch. 22, 1910/3 | 5 Male taken away Marc Mch. 22, 1910 |4 | 6| 7| 4 REE E os Ge All dead April 10. May 4, 1910 li a All eaten May 9. May 16, 1910 |2] 9 May 22, 1910 | 3 |10 May 22, 1910 | 1 |11| 12| 4 | 8 | Male not removed. June 1, 1910 | 7 |12 a oram June 3. Male removed before June 4, 1910 | 5/18 an eaten June 4. Male removed before June 30, 1910 | 7 | 14 All jord, July 6. Male removed before irth. ; July 23, 1910 | 7 |15 All dead July 24. Male removed before birth. Aug. 24, 1910 | 4 |16 All eaten August 27. Male not removed be- fore birth. Sept. 3, 1910 | 5 |17 Sept. 5, 1910 |7 |18|11| 4 | 7 | Male removed before birth.” Sept. 8, 1910 | 2 |19/10 All meee espamegioas 29. Male removed be- fo Sept. 20, 1910 |.1 | 20 Four still alive. Male removed before birth. Sept. 20, 1910 | 3 |21|10| 6 | 4 | Ninestill alive. Male removed before birth. Sept. 26, 1910 | 4 | 22 All eate removed before birth. Sept. 26, 1910 | 5 | 23 All eaten ale not removed before birth Oct. 17, 1910 | 2 | 24 All eaten. Male not removed before birth - 20,1910 | 5 | 25 All eaten. Male removed before birth. Nov. 26, 1910 | 4 | 26 All eaten. Male removed before birth. Nov. 26, 1910 | 7 | 27 Part eaten. Male removed before birth. REPRODUCTIVE RECORD. SECOND GENERATION Females of Litter No. 11 ep, | es bo = Hoo | SS E eis ~~ Female 2333 S £ Oct. 15, 1910. 11-2 | 28 All eaten. Male removed. Oct. 17, a 11-5 29 All eaten. Male removed Oct. 24 1 910. 11-6 30; 9 art eaten — others oved. : Male remov The works of Lantz and Boelter lead us to infer that the brown rat breeds the whole year round. My rats gave birth to young from February to December, inclu- sive, and, since they were seen mating in November, De- 626 THE AMERICAN NATURALIST [Vou. XLV cember and January, it is safe to say that they produce young in January, also, which gives their breeding period as the whole year. The gestation period was found to be 234 to 254 days, counting from the time of the first copulation. If 234 days is the shortest time, then females Nos. 1 and 7 must have mated on February 4 and June 30, respec- tively, the days they gave birth and also ate their young. Frequently a female eats her young and when she does so, she may breed immediately. It is seen from the table that female No. 7, which had the habit of de- vouring her young, gave birth to seven litters at inter- vals of about thirty days; i. e., taking February 2 as the date of conception for her first offspring, she produced, seven litters in seven months. If the young are reared, a second litter may be pro- duced in two months. Some do not give birth so soon, which may be due, largely to the fact that a female is not sure of conceiving at any given copulation. Daniel’s experiment with white mice seems to have some bearing on this point. He found that a female suckling a large litter, carried the second longer, as much as ten days over the average gestation time. According to my data, the brown rat does not conceive before the last ten days of the lactation period, which should not have so marked an effect on the gestation period as is found in the case of the white mouse, which becomes pregnant as soon as a litter is born. The number of young in the eight litters which I was able to count varied from seven to twelve, with an aver- age of 10.5. Boelter quotes Brehm as recording a litter of twenty-one. Lantz gives two cases where seventeen embryos were found in one female and nineteen in another. He also thinks that for temperate regions the average litter is not under ten. From the data at hand, we can conclude that a pair of adult rats is capable of rearing fifty to sixty young per year. Zuschlag’s theo- retical table, as quoted by Boelter, gives sixty-four as No. 538] REPRODUCTION IN THE BROWN RAT 627 the possible number of young at the end of the year, the product of a single pair. In this table Zuschlag assumes eight to be the average number of young per litter, and eight litters per year to be the product of a single pair. On this basis the number of offspring at the end of the second and succeeding years is far in excess to the num- ber computed on experimental data. Rats begin mating, as a usual thing, about five o’clock P.M., and to obtain the period of gestation, males were placed with the females every day at this hour. If a female was in heat, she was removed to a separate cage with one to three males. At first females were left with the males not more than two hours, in which time many copulations had taken place, but in no case did a preg- nancy result. Later, they were left with the males twelve to fifteen hours, and even then, failures to be- come pregnant far outnumbered the pregnancies. I have not observed a single case of a female mating with a male smaller than herself. It is not common for an old female, even when in heat, to chase a young male about the cage as though he were a female, not letting him come near her. This same female, if placed with a larger male, which could boss her, would mate with him at once without any opposition. Mating in this case seems to be, to some extent, dependent upon the domi- nating ability of the male. The number of coitions during a single period of heat is, apparently, great. In one case a female, placed with four males, mated with them in such rapid succession that fifty attempted coitions per half hour would be a conservative estimate. It is impossible to say how many of these attempts were successful, because the rat re- quires such a short time (four seconds being a long time) to perform the act, thereby making the details of the process difficult of observation. The following table is a month’s record of seven fe- — males kept to determine the frequency of the mating periods. Such periods are indicated by an X. It is XLV d : Z NN E yuejuĝojiq | X x $ =x fe que suaig - x X x L'N O ‘Jolaıd-u oN x x x x 8 ON x ‘3 |ad-ujoN | X x x xX GIL ON i xX ‘3əa|d-ujoy X 1x F-1L ON a quBluse tg | 4 PIM x X | TI OK S a a al alm sermo els|t|o|o|»]|e]|z]|r]1e]|o 6g | 92 | zz |o |s | ve | ez | ex |t | oc | or sr j| er heat = JOquiaaon 1aqo~O x WIV S 5i 628 No. 538] REPRODUCTION IN THE BROWN RAT 629 seen that only one pregnancy occurred as a result of each six times of mating. The table indicates, also, either that the mating periods come irregularly, or that the females come in heat about every five days. I am convinced that odor is the final test for sex rec- ognition. When rats are placed together, the males as- sume an aggressive attitude and fight all those that op- pose them. The females may, usually do, for a little while, resist the males; but they soon yield. They then lie on their backs entirely passive, while the males nose them about the head and smell of their genital organs. This attitude is frequently taken by pregnant females. The males in such cases have shown their superiority and the females recognize it. From now on there is little fighting on the part of the males, and afterwards they will often permit themselves to be severely punished by the females without injuring them. Copulation does not usually follow the above proceedings. When males meet, there is a battle royal until one is recognized victor. The conquered, then, tries to elude the stronger, and will not submit to be smelt of as a female. Often males smell of each other, probably de- termining one another’s sex, before beginning the fight. I am sure there is no sex recognition when the fight is begun at once, for the females are treated in the same way. There is very little courtship among the rats. The male is absolutely silent and the female almost so during the period of heat. A pair which had been together for several months, were seen to arouse from their sleep at - five-thirty o’clock, p.m., and begin copulating at once without any preliminaries at all. They had not mated previously during the day. When a female resists the advances of the male, she does so by fighting him away, as shown in Fig. 2, or by kicking him away with her hind foot; or she may lie on her back, as previously described. On the other hand, a ~ female in full heat is the more active of the two. In one 630 THE AMERICAN NATURALIST [Von XLV instance, she was seen to clasp the male. If he is not ag- gressive, she throws herself before him in a crouching position, a procedure which she repeats until he takes notice of her. Again, she may strike his head, as though fighting, until he follows her. A female in full heat is much more active and less pugnacious than at other times. Fic. 2. Rats in a fighting attitude. We find the anomaly of mammals eating their young carried to an extreme in the case of brown rats kept in confinement. My records show a large per cent. eaten, almost fifty, which has much to do with the number of litters per year, consequently the number of young. This infanticide has usually been attributed to the male, but the young are eaten whether the male is or is not No. 538] REPRODUCTION IN THE BROWN RAT 631 present at the time they are born, which throws the blame on the female. In fact, I have no direct evidence against the male. The young, when eaten, may be devoured at birth or any time within the next four days. It has been sug- gested—I think with a little basis for the statement— that disturbing the parents causes them to eat their young. W. T. Hornaday, in a letter, and Bostock state that the large carnivora are kept from all disturbing influences at and about the time they are giving birth. In rare cases some species devour their young if dis- turbed, but more frequently the young are deserted. My rats were in a room which was kept locked and free from any disturbing influences except my morning visits. Frequently, females by themselves, which built their nests in closed boxes, have had their young and eaten them between my visits. In such cases these females could not have been disturbed. On the other hand, I have taken the young from the nests, weighed them, and even handled them without the mother injuring them when replaced. Another suggestion is that the rats have not had enough flesh diet and, when the young are born, eat them to satisfy their desire for meat. Here again I think there is no truth, provided plenty of grain and vegetables are accessible. My rats have eaten their young when on a meat-grain-vegetable diet, as well as when on a grain-vegetable allowance. Dr. Slonaker suggests that the same motive, which leads the rats to eat their dead under natural conditions, might also cause them to eat their helpless young. This is in accord with the rat’s habit of killing off its weak, One of the reasons for eating the young, I think, might be found in the habit that mammals have of thoroughly licking their young, and in many cases eating the embry- onic membranes, even among the herbivorous animals. It seems but a little step from the eating of the placenta to the devouring of the young. There may be some 632 THE AMERICAN NATURALIST [Vou. XLV truth in all these theories, but I doubt if the principal cause has yet been suggested. CARE OF THE YOUNG Before the young are born, the female builds a nest as elaborate as the means at hand will permit. Almost any sheltered nook about buildings is a suitable nesting place for the brown rat. On the farm, in addition to breeding about buildings, it digs its burrows in the field and nests in old straw stacks and grain still in the shock. The nest consists for the most part of a coarse sub- stance such as straw or corn husks, with a lining of a softer material, especially feathers when obtainable. Blue found that where rats use run-ways, they nest in a branch leading off from the main course. This branch is in the form of a Y with the nest in one arm and a storehouse in the other. The young at birth are entirely helpless. The mother gets them all together and then huddles over them for hours at a time. She never lies on her side to let the young suck; she always crouches over them while the young lie on their sides or backs to get hold of the teats. Usually all the young nurse at the same time, and a litter of twelve, at about weaning time, almost holds the female off her feet while they suck. The female can easily be induced to move her young elsewhere by disturbing the nest. The young only a few days old are caught around the body by the mother, but if they are a week or so old, she takes hold of them by the skin. In the latter case they are carried much as a kitten is carried by its mother. In the laboratory the female spends most of her time with the young; some even carry all their food into the nests. Taking food into the nest is a common habit of the rat and must not be considered as a trait peculiar to females with young. GROWTH The young at birth weigh on an average 6.4 grams. The males are a little the larger, measuring in body No. 5388] REPRODUCTION IN THE BROWN RAT 633 length 52 mm. as compared with 49 mm. for the females. They also weigh more, and by the end of the second week are as much as two grams heavier. Very little change is noted in the appearance of the young the first two days, save a gain in size and strength. On the third day a change in color can be detected. The flesh tint is being replaced by a darker hue with a de- cided tinge of blue on the distal portion of the tail. Curiously enough, the tail is the first to show any marked change of color. On the fourth day, the tail is still darker and the flesh color of the dorsal parts has been replaced by a gray with a touch of blue. Up to this time, no hair is noticed except the vibrisse. By the end of the week the body is covered with a thin coat of very fine hair, which gives a delicate bluish-brown color. The coloring of the legs begin about this time with the palms of the feet, the joints of the toes and legs turning a bluish tint. The tail, in the meantime, has darkened al- most to the base and the nose turned nearly black. On the eighth or ninth day, the gray of the adult is noticed on the back of the head and neck, and with this as a center, it spreads until about the eighteenth day, when the whole color is that of the adult with the exception that it is darker and softer. The following is a diary of litter No. 18 for the first twenty-five days. The weights given are the averages for the whole litter. Day Wt. Grams Notes 1 6.1 Flesh color; blind; ears closed; helpless. 2 6.4 Able to right themselves if turned over. 3 7.2 Color changing to bluish tint on the dorsal parts; tail color- ing from the distal end; no hair except vibrisse. 4 8.1 Flesh color limited to the ventral parts; creeping about in t he nest. 5 9.0 Color a bluish-brown on the back and sides. 6 9.8 ESM So 416 Sparsely covered with short fine hair. 9 123 Faint brownish tinge on the back of the head and neck; tail a dull blue except about 4 mm. at its base; joints of the feet and legs, also the palms of the hind feet, blue; under parts, pinkish. 634 THE AMERICAN NATURALIST [Vot XLV 10 = 13.0 Dorsal Re a delicate coin nose, almost black; incisors appear in the upper jaw Ft 138 Heels Gat black; rest of wiag the same as yesterday; one seen sitting up and washing its face 12 14.8 Lower incisors appear (all the incisors appear on the same day in some cases; usually the lower appear a day later than upper). 13 15.1 All the dorsal parts are brown. 10 PTY Eyes beginning to open; one young attempted to bite me. Lt 18.8 All have their eyes open; color the same as that of the adults except that it is darker and softer. 18 -19.8 19 202 20 BLO 21 209 Found the young hungry and when given a dog-biscuit, they ate it greedily. It is probable that they began taking solid food the first or second day after cutting their lower incisors, and judging from their weights, yesterday was the first day they did not have food they could eat 22. 23.0 23 239 24 24.9 25 25.9 From now on there is little change noticed except increase in size. The young have grown rapidly from the first with- out the initial loss of weight as found in human infants, or even a marked retardation as found in guinea-pigs by Minot. During the sixth week the young are weaned. At this time they weigh fifty to seventy grams, and are able to take care of themselves, provided food is abundant. I have seen a number of rats of about this age and size wandering about, in or around farm buildings, appar- ently, in a starved condition. Presumably, these had just been weaned and were unable to find suitable food. If food has been abundant the males, by the end of the sixth month, weigh 230 to 290 grams while the females weigh 170 to 240 grams. The male and female of one of my pairs, kept throughout the year and which I judged to be in their third month when caught, weigh 337 grams and 223 grams, respectively. My old male No. 5 weighs 460 grams and my oldest female, 345 grams. These data indicate that a rat does not reach its full growth before the end of the eighteenth month. No. 538] REPRODUCTION IN THE BROWN RAT 635 Sexual maturity is attained much sooner than full growth. One female of litter No. 11 conceived on the day she was four months old, hence giving birth to her first young in her fifth month. The males are sexually mature as soon as the females, and I have seen some indications that they may be mature early in their fourth month. CONCLUSIONS : The brown rat breeds in every month ot the year. 2. The gestation period is 234 to 254 days. 3. The number of young per litter varies from six to nineteeen with an average of between ten and eleven. 4. Five or six litters may be reared by a single pair in a year. 5. Seven litters were born in seven months by one female and, presumably, twelve would be produced in the course of a year when all the young perish at birth. 6. There is very little courtship among the brown rats. 7. Odor is the primary factor in sex recognition; the aggressiveness of the male is second. 8. Brown rats in captivity eat almost fifty per cent. of their young at birth. Most of the young eaten, if not all, are eaten by the females. 9. Full growth is attained not under eighteen months. 10. Sexual maturity is reached at least by the end of the fourth month in both sexes. REFERENCES Blue, R., 08. The Underlying Principles of Anti-plague Measures. Cali- fornia State Journal of Medicine, a Reprinted in ‘‘ Eradicating lague from San Francisco,’’ 1909, p. > Boelter, W. R., ’09. The Rat Problem, pp. . (note), 88. Bostock, F. R., 04. The Training of Wild Animals, pp. 40, 50. Daniel, J. F., ’10. Observations on the Period of Gestation in White Mice. Journal of Smem Zoology, Vol. 9, No. 4, p. 868 Lantz, D. E., Natural pii of ra Rat. Publie Health and Marine Hospital Service of U. S., pp. Minot, C. S., ’08. The Problem fr ye Growth and Death, p. 94. Slotmiee: J. R., 708. A Description of an Apparatus for Recording the Activity of Small Mammals. Anatomical Record, June, Vol. 2, pp. 116-122. SHORTER ARTICLES AND DISCUSSION DATA, DIALECTICS AND OTHER DIGRESSIONS Some... persons vainly seek by dialectics and far-fetched arguments, either to upset or establish things that are only to be founded on anatomical demonstration, and believed on the evidence of the senses. He who truly desires to be informed of the question in hand, and whether the facts alleged be sensible, visible, or not, must be held bound either to look for himself or to take on trust the conclusions to which they have come who have looked; and indeed there is no higher method ‘of attaining to assurance and certainty—William Harvey, Second Dis- quisition to John Riolon, Jun. To THE EDITOR OF THE AMERICAN NATURALIST: My reasons for asking you to publish the above from your Sep- tember, 1911, issue are two: the text is excellent; the sermon is wide of the mark. The text, the reader will have noted, heads a latest contribu- tion to our knowledge of egg production in the domestic fowl* which Dr. Pearl has been prevailed upon to write up by the con- viction that certain criticisms? ‘‘rest on either a misconception of what our results really are, or else a lack of understanding of the real facts regarding certain of the biological points in- volved.’’ For his ‘‘endeavor, if possible, to remedy this defect in some degree at least’’ those biologists who are thereby instructed in matters of fact will doubtless be grateful. The obvious implication of Dr. Pearl’s quotation from Harvey and of his concluding remarks is that my arguments concerning the genotype concept are of a purely scholastic and ‘‘far- fetched’’ order. In consideration of these implications and in justice to my paper which appeared in your June number may I call your readers’ attention to the following points? First. I certainly did not ‘‘. . . seek by dialectics and far- fetched arguments, . . . to upset . . . things that are only to * Pearl, R., ‘‘ Biometric Arguments —* the Genotype Concept,’’ AMER. NAT., Vol. 45, pp. 561-566, 1911 2 Harris, x Arthur, ‘‘The Biometric “Proof of the Pure Line Theory,’’ moe NAT., Vol. 45, pp. 346-363, 1911 636 No. 538] SHORTER ARTICLES AND DISCUSSION 637 be . . . believed on the evidence of the senses.’ Unfortunately for Pearl’s excellent-in-itself quotation, the genotype theory is not a pickled specimen concerning the structure of which all the anatomists who can crowd around the table will agree. Quite to the contrary, it is a far-reaching generalization of the kind which should not be accepted until it has been shown not only to describe and epitomize the results of great series of actually ob- served facts but to stand every test which can reasonably be ap- plied to it. A careful examination of all the pure line literature known to me had convinced me that in the enthusiasm for the new theory the elementary principles of scientific reasoning were often ignored and matters of plain common sense overlooked. I had frequently found biologists enthusiastically supporting the popular theory without knowing what its essential implications are. It seemed useful, therefore, to ‘‘state the fundamental prob- lems of the pure line theory as they appear to the biometrician’ and to call attention to some of the weak points in arguments in its support. Judging from some of the vagaries encountered in the genotypic literature since then, I fear that my plea for more caution, less assumption and less reasoning in circles in our theorizing about ‘‘sensible, visible’’ facts was rather wasted ort Second. My paper was written before Pearl’s preliminary publication of the results of individual pedigrees in the same number of the AMERICAN NATURALIST, and before his advance statements concerning correlation in the paper just issued. If new and pertinent facts prove that my views were wrong the views will be discarded. When Dr. Pearl has given us all the data and not adumbrations merely—when all the cards are down, face up on the table—it will be time for a eritic to show reasons for differences of opinion or to admit that he was wrong. Until that time it seems foolish—in fact dangerously near dialectics— to squander in argument space that might be used to publish tables of data. In passing, I must remind the reader that our present sore need is not possible illustrations of the genotype theory if valid but critical evidence® for or against it. è Very unfortunately Pearl’s sentence, ‘‘So far as concern sonal opinion of the critical value of the work done in this orator no discussion will be entered upon by the present writer’? cuts tw o ways. Some of those who have not read what I really said will conclude that T- 638 THE AMERICAN NATURALIST [Vö XLV © Third. Insinuations concerning dialectics may perhaps justify a digression concerning data. During the past five or six years some 100,000 countings, weighings, measurements, ete., bearing directly on the problem of pure lines in garden beans—the species on which Johannsen based his studies—have accumulated in my notes. These data have taught me how idle it is to discuss the pure line problem without the most refined biometric analysis of large masses of data. Such analysis necessarily proceeds with disheartening slowness. But I have been able to see no advan- tage in dragging this material through a long series of prelimi- nary papers necessarily based upon uncompleted work. When the data are all in, and arranged in an orderly manner they will be honestly set forth as ‘‘an accumulation of plain, unadorned facts, available to any one’s inspection.’’ J. ARTHUR HARRIS. COLD SPRING HARBOR, September 15, 1911. actually. or tacitly—and quite unfairly—I ar rew in question the accuracy il Pea or trustworthiness of the observations. Others will do Dr. rl the in- J e of thinking that if a criticism was made there was probably some ustification for it. Both will be quite wrong. ve different places I ‘‘The work of Pearl and Surface with poultry and maize seems to me to have no critical bearing on the pure line problem.’’ Critical bearing which I wrote seems to me to convey a meaning quite different from critical value which Dr. Pearl writes. A judge might recognize the critical value of an expert’s observations and yet fail to see that his testimony had any eritical bearing in a case NOTES AND LITERATURE DARWINISM AND HUMAN LIFE PROFESSOR JAMES ARTHUR THOMSON’s recent! book under the title of ‘‘Darwinism and Human Life’’ is most attractive read- ing. Professor Thomson thinks independently and writes fasci- natingly. He gives even the most familiar of subjects new color and atmosphere. The matter of the book was given in 1909 as the ‘South African Lectures,’’ whose ‘‘chief aim was to explain the gist of Darwinism.’’ An endeavor was made to add to the necessarily general and somewhat familiar content of the lectures, sugges- tions of how ‘‘ Darwinism a every-day life, in farm and garden, in city and empire.’ e Darwinian reader interested by this prospect of finding his old wine put into new bottles runs rapidly through the chapters with the familiar headings of What we owe to Darwin, The Web of Life, The Struggle for Existence, The Raw Materials of Progress, Facts of Inheritance and Selection: Organic and Social, nosing for Darwinism and Human Life. And he finds himself rather disappointed at first, for he does not discover as much of the practical interlocking of Darwinism and human affairs as perhaps he felt justified in expecting. But in the last chapter he does find it more obviously and in more abundance than elsewhere and he begins really to read. And lo, when he stops reading he finds that he has read the book, all of it, backwards! And is very glad he has. At any rate, all this is what I did. Professor Thomson is a good selectionist; though not a bad one; that is, not one who has an all other possible evolution factors—phobia. However, Darwinism for him rests on, or is, mostly selection. And it is the possible play of selection in human life, its play among individuals, among societies and among races, on which most of his direct application of evolu- tion knowledge to human affair rests. Hence organic selection, | social selection, eugenics, selection of Utopias, reversed human 1 Thomson, J. A., ‘‘ Darwinism and Human Life,’’ 245 pp., frontispiece (Charles Darwin), 1910, H. Holt & Co., New Yo — 639 640 THE AMERICAN NATURALIST [Vou. XLV selection, and the like, are the subjects of his more concrete ‘‘Darwinism and Human Life” paragraphs. But Professor Thomson recognizes the broader aspect of his subject. He sees that all of Darwinism, in its very broadest sense, has interrela- tion with all of human doing and becoming. And it is this recognition, and the constant suggestion of it, everywhere in his discussion of the familiar subjects of the ‘‘gist of Darwinism,’’ that make even the practised Darwinian reader read with fresh interest the whole of the book; even if he does happen, as your reviewer did, to do it backward! You te STANFORD UNIVERSITY, CAL. The American Journal of Science Established by Benjamin Silliman in 1818 The Leading Scientific Journal in the United States Devoted to the Physical and Natural Sciences, with special reference to Physics, and Chemistry on the one hand, and to Geology and Mineralogy on the other. Editor: EDWARD S. DANA. Associate Editors: Professor GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW and WM. M. DAVIS of Cambridge ; Professors A. E. VERRILL, HENRY 8. WILLIAMS and L. V. PIRSSON, of New Haven; Professor O. F. BARKER, of Philrdelphia; Professor JOSEPH S. AIMES, ts) imore; MR. J. S. DILLER, of Washington. Two volumes annually, in monthly numbers of about 80 pages each. This Journal ended its first series of 50 volumes as a quarterly in 1845; its second series of 50 volumes as a two-monthly in 1870; its third series as a monthly ended Dec- ember, 1895. A Fourth Series commenced in 1896. Subscription price, $6 per year or 50 cents a number, postage prepaid in the United States ; $6.25 to Canada ; $6,40 to Countries in the Postal Union. Back numbers at reduced pric ka-Ten-Volume Indexes, Vols. I-X XI-XX, fourth series, price one dollar. Address The American Journal of Science New Haven, Conn. SECOND EDITION, NOVEMBER, 1910 AMERICAN MEN OF SCIENCE A BIOGRAPHICAL DIRECTORY EDITED BY J. McKEEN CATTELL A Biographical directory ires revision if it is to maintain ts usefulness. Nearly a third of the names in the pany Hie metal a red in the first division have in near every case been revised. The ist fis been as great as that given to the first edition. There has been no Gr $ ; rons taud 1 as conditions on which scientific research depends and so far as may be to um these conditions. 7. an appendiz the two ioa studies that have been made.—From the Preface to the Second Edition. The second edition of the Directory extends to more than 600 pages and contains more than 5500 sketches. It is well printed on all rag paper seg bound in buckram with leather label. Although the work has been in size by more than 50 per cent., it is sold at the same price as the first edition. Price: Five Dollars, net, Postage paid THE SCIENCE PRESS GARRISON, N. Y. LANCASTER, PA. SUB-STATION 84, NEW YORK CITY. The American Naturalist A Monthly Journal, established in with S 1867, Devoted to t e Advancement of the ste Sciences h pecial Reference to the Factors of EA Evolution and Her CONTENTS OF THE APRIL NUMBER Genetical Studies on Oenothera. II. Dr. BRADLEY The we aes of Maize. Dr. GEORGE HARRISON so mad oT Is the Female Frog Heterozy- regard to Sex Determination ? Professor . E aa The Mutation Theory. Dr. R. GATES, CONTENTS OF JUNE NUMBE Inheritance 5 Fecundity in the Domestic ae Dr AYMOND PEARL. The Samaras sion, of the Pure Line Theory. Dr. J. REHAR ay piei an meaner External Conditions on Pro- CONTENTS OF THE MAY NUMBER The Inheritance of Polymorphism a vie in Colias Philodice. Professor JoHN H. GE Nucleus and ps age in aoti. MICHAEL F. GUYER A Comparative Study of the pcan et of the Photo- en e orean of Certain American Pose siet . McDermott and ai eee Soorte Articles and Discussion; A A ‘Paper nN 1 Selection in the English Sparrow. fo da ppan HARRIS ick: a Literature : Biometrics, Yule’s Introduc- o e Theory of Statistics. Dr. RAYMOND Poti CONTENTS OF JULY NUMBER Germ-cell Determinants and Their Significance. Pro- H sor R. EGNER. Further Observations on the Pose of the Sauropodous Dinosaurs. OLIVER P. HAY. Shorter Articles and Correspondence : Computation of the en ae = st Ba ara bs Professor H. $ JEN = No achoseps attenuatus Esch. 5 Borsa fessor THOM - MONTGOMERY. Notes. and Literature: Some Recent Studies on Vari Shorter mre and seh The Ontogeny of a ation and Correlation in Apenes Plants: As To a car ton OND i oka T name ‘AS Notes ad Feito prem t Contributions to a seaininay © Prenden Davip STARR J JORDAN: na of the Extinct Amphi ibia, Dr. Ror Sone Rec ze ent Boo aka = Possit Plants: Professor Las HoucH poean AIEEE CONTENTS OF AUGUST NUMBE Comparative pte e Study of the gece Content of Pigmen Skins, inte kegs refer- ce to the Question of Color Bee e among Mulattos. Professor H. A Coeficient of Song gree: Pre cca for Students of Heredi J. ARTHUR ies a The Adaptations T the Primates. datas F. B Jean 5 Marchant an Eighteenth Century Mutationist. Pasi es and ad Literature : Notes on Heredity. Dr. W.J- > CONTENTS OF SEPTEMBER NUMBER TREE s the “Eye” in Vigna. Dr. W. J. LMA hacker ot r Form among the Filipinos. Dr. ROBERT DENKETI BEAN. The Zoogeography of = East Indian Archipelago. Dr. P. N. VAN KAMP Shorter Articles and Diéa ssion: Biom ments nag: oi ford the Genotype Con RAYM EARL: On the Formation or Go tion anit ‘Coagation —_ when = aby Combinations is L J. ART Acquired Char a BaD ew Notes and Literature: The once and Study of cians Psyehoiogy ri Useful Evo- lution Bibliography. V. — fined. Seago meter Cents Yearly Subscription, $4. 00 Dollar NATURALIST will be sent to new subscribers for four months for One T HE SCIENCE PRESS ooo _Sub-Station 84: NEW YORK Lancaster, Pa. . : VOL. XLV, NO. 539 ~~ NOVEMBER, 1911 THE AMERICAN NATURALIST A MONTHLY JOURNAL Devoted to the Advancement of the Biological Sciences with Special Reference to the Factors of Evolution CONTENTS Page I. The Origin of Species in Nature. Dr. HENRI Hus - - 642 II. Some Experimental Observations concerning the porate of oe Bees their Visits to Cotton Blossoms. H. A. ALLARD > 668 . The Distribution of Pure Line Means. DE. J. ARTHUR HARRIS - -686 IV. Shorter Articles and SERE A Kanor of EI EPET m tions. Dr. J. K. Shaw - 701 THE SOIENCE PRESS LANCASTER, PA. GARRISON, N. ¥. NEW YORK: SUB-STATION 84 The American Naturalist MSS. intended for publication and books, etc., intended for ee should be sent to ney mg of THE AMERICAN NATURALIST, Garrison-on-Hud k es containing research work bearing on the problems of ‘organi evolu- tion are especially welcome, and will be given preference in publica reprints of appeal are supplied to authors pa of charge. Further iama will be supplied at c criptions and advertisements r e sent to the geeni The or Sange price is four dollars thirty-five cents, ear. nadian postage twenty-five itg additional. The advertising rates are Four Dollars for a page. postage is fifty ce and char eign ents "The arge for single copies is THE SCIENCE PRESS NEW YORK: Sub-Station 84 Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of Congress of March 3, 1879. Lancaster, Pa. Garrison, N. Y. THE BULLETIN—ror bargains in Ethnolograph- ical and Pre-historic Specimens. Books on Natural History, Science, Travel, Voyages, etc. See THE BULLETIN post free for 3 cent stamp. 4 Duke St., Adelphi—t ondon—England TENTH EDITION. THE MICROSCOPE, an introduction to Microscopic Methods and to Histology, by SIMON Primi GAGE, of Cornell University. Over 350 la. and above 250 Fg iat es in this new and fully wade edition. Price $2.00, postpaid. COMSTOCK PUBLISHING CO., Ithaca, N. Y. BIRDS’ EGGS W. F. H. ROSENBERG, 57 Haverstock Hill, London, N. W., England begs to announce the publication of a new Price List (No. 15) of Birds’ Eggs, con- taining over 900 species from all parts of the world. This Catalogue is systematic- ally arranged, with authors’ names, indica- tions of localities, and an index to families. It will be sent post free on application, as will the following lists: No. 11, Birds’ Skins, (5,000 species ) ; No. 12, Tegidegaern (5,000 species) ; No, 13, Calera: No. 14, Mammals. In course of preparation: New Price List of Reptiles, Amphibians and Fishes. Largest Stock in the world of specimens Zoology. rb ey an Back or Current Numbers of any American or Foreign technical or trade journal and magazine fur- nished on short notice at moderate rates; all kinds of Government and State Reports in stock. Clippings On special subjects furnished promptly. Large stock of American Naturalist, Science and Popular Science on hand. Magazines, Books and Papers of all kinds bought Special Subscription Price List on request. Information concerning any periodical furnished free of charge. A. W. CASTELLANOS 259 Armstrong Ave., Jersey City, N.J., U.S.A. THE AMERICAN NATURALIST VoL. XLV November, 1911 No. 539 THE ORIGIN OF SPECIES IN NATURE? DR. HENRI HUS MISSOURI BOTANICAL GARDEN UNIVERSITY OF MICHIGAN Amone the duties of the botanist is that of adding to the number of known species, varieties and forms. To reach this end several ways are open. One may, by studying the flora of new or insufficiently explored areas, not only extend the range of species formerly known, but add new species. The work of Setchell and his students, especially Gardner, done on the algal flora of the Pacific Coast, yields an excellent illustration.’ Sometimes it is possible to correlate with such investi- gations work of economic value. The results obtained by agricultural explorers working under the auspices of the Office of Foreign Seed and Plant Introduction of the Bureau of Plant Industry, U. S. Dept. of Agriculture and other? institutions, offer a case in point. Such work, to lead to the desired result, requires men of especial fit- ness, physical, moral and mental. Not only must they possess physical endurance to overcome the hardships with which they necessarily meet. Moral courage, moral strength are necessary to extricate them from the many 1 Read, through the courtesy of Mr. Henry W. Anderson, before the Academy of Science of St. Louis, May 15, 1911. Contributions from the Botanical Laboratory of the University of Michigan, No. 125. 2 Univ. Calif, Publ. Bot., 1° et seq., 1903-1910. o * Hansen, N. E., ‘‘The Wild Alfalfas and Clovers of Siberia, with a Perspective View of the Alfalfas of the World,’’ Bull. No. 150, B P. L; U. 8. D. A., 1909 Aaronsohn, i. «í Agricultural and Botanical Explorations in Palestine, ”” Bull. No. 180, B. P. L, U. & D. A, 1910. ‘ 641 642 THE AMERICAN NATURALIST [Vou. XLV difficult positions in which they are bound to find them- selves and to bring to a successful end the undertaking upon which they have embarked. Further, a thorough knowledge and an excellent judgment are essential. Yet none of these are sufficient unless they be complemented by an all-pervading love for the subject, a devotion which counts all obstacles as naught and persists in the face of difficulties which to most men would appear unsur- mountable. Frank Meyer, agricultural explorer, now, for the second time, exploring eastern Asia—last heard of when entering Thibet—exemplifies this ideal. Such are the men who on their return to the civilized world bring back with them the rare plants which delight the collector and which, in themselves frequently of apparent insignificance, in a few years add immensely to our col- lections. The discovery of Nicotiana Forgetiana* is, in recent years, perhaps the most widely known instance through the part played by this species in the introduc- tion to our gardens of Nicotiana Sandere (N. alata X Forgetiana). But not always is it necessary to draw on the flora of distant countries. By thoroughly going over a well-covered territory one may be able to add new varieties of a more or less well-known species, such as was the case with Oxalis stricta viridiflora which, since the time of publication of the original paper, has been found near Thomson, Ga.* This plant, since the spring of 1909, has established itself in the Ann Arbor experiment garden, spreading rapidly from seed selfsown in the latter part of the same summer. The plants grow in the open as well as in more shaded places and on a light, sandy soil. The varietal character maintains itself per- fectly, as it does in the Missouri Botanical Garden, where, on a recent visit, I saw a bed of it, in the open and on a fairly heavy soil. Though there is a certain amount of variation in the intensity of the green coloring of the * Curtis’ Botanical Magazine, 4th Ser., 1, pl. 8006, 1905. *Hus, H., ‘‘Virescence of Oxalis stricta,” Ann. Rep. Missouri Bot. Gard., 18: 99, pl. 10, 11, 1907. Bartlet t, H. H., ‘On Ozalis stricta viridiflora,’’ Rhodora, 11: 118, June, 1909. No. 539] ORIGIN OF SPECIES IN NATURE 643 flower, due to a variation in the relative amounts of chloroplasts and yellow chromoplasts, the plant appears to have undergone a sufficient number of fairly rigid tests, under widely divergent conditions, and for a suffi- cient number of generations (9), to entitle it to recogni- tion as a distinct, non-pathological variety. The varia- tion in the color of the flower I am inclined to ascribe to differences in light intensity. In a bed, situated in the middle of the experiment garden, i. e., in an open place where no obstruction prevents full and direct illumina- tion, the flowers are uniformly green. When the plants are grown in the greenhouse or in the shade of shrubs, there seems to be a tendency for the flowers to assume a more yellow tinge. But not in a single instance has a flower been observed which possibly could have been mis- taken for one of the species. Again it may be that a form deserving specific rank is discovered, as instanced by the now well-known Capsella Heegeri.? Such discoveries possess an added charm since in these cases the possibility of a recent origin of the new form is not excluded. Sometimes it is possible to prove this experimentally, as was done by de Vries for his evening primroses. An illustration of probably recent origin and of repetition of mutation is yielded by the discovery of a single specimen of the inermis variety of Cynara Cardunculus in Algiers by Trabut.8 But some years ago what would seem to have been the same variety was grown in the St. Louis experiment garden from seed which my father kindly obtained for me from the then director of the Botanical Garden at Buenos Aires and the ancestry of which I understand was to be traced to an individual growing wild in the vicinity.’ " Solms-Laubach, H. zu, ‘‘Cruciferenstudien—I, Capsella heegeri, eine neuentstandene Form der deutschen Flora,’’ Bot. Zeit., 55: 167, pl. 7, 1900. $ Trabut, L., ‘‘Sur une arain inerme du Cynara Cardunculus,’’ Bull. Soc. Bot. Fr., 57: 350, 191 ? Cynara Cardunouhis, a ya of the Mediterranean Tegion, emigrated to South America, where it occupies large areas (Darwin ‘í Animals and Plants under Domestication’’). An analogous case is that of Silybum Marianum, likewise a native of the Mediterranean region and now largely 644 THE AMERICAN NATURALIST [Vou. XLV Generally speaking, the investigation of the flora of a newly discovered or formerly inaccessible region is a matter of choice or, better, perhaps, opportunity. The same is true for the detailed study of our collective species, leading to a recognition of component elements as illustrated by the work of Small on Oxalis,” Rhus, ete. Whoever can devote a part of his time to the study of a genus is able to establish the existence of differences which, formerly ignored and in themselves slight, are of the greatest importance for the tracing of relationships. Unfortunately, perhaps, in such studies there is usually developed a mass of detail so great, so intricate that none but a few specialists are able to recognize the various forms. Hieracium, Aster, Cyperus are names sufficient to strike terror in the hearts of any but the most ardent systematists. But before sucha detailed study can be made there must be gathered, in the first place, herbarium mate- rial. It is not sufficient to possess a single specimen. Material must be collected from various habitats and both during the flowering and fruiting periods. Often there are differences between the leaves of old and young shoots, as in Crategus. Or the leaves of the fruiting branches may be different from those of purely vegetative shoots as in Hedera Helix. Ficus and Pothos both may show leaves of a very different character."’ None would at first sight believe leaves of seedlings or of adventitious shoots of Eucalyptus globulus and those of older specimens to belong to the same species. Many species of Acacia show, during the first few yéars of their existence, well-developed phyllodes which later dis- appear entirely. Sometimes even the leaves on the same naturalized in California where it forms a common and, on account of its leaves, a most conspicuous weed. ‘North American Flora,’’ 25': 25, 1907. “In the greenhouses of the Missouri Botanical Garden may be seen a very fine specimen of Ficus repens which near the ground and up to 4 height of some eight feet, produces the small leaves typical of this species. When, however, the plant makes a bend to twine horizontally along a rod, large leaves are PPAS Other instances are given in de Vries’s ‘‘ Species and a ranea ON d No. 539] ORIGIN OF SPECIES IN NATURE 645 shoot may present a very different appearance. And while this is something not entirely unexpected in hy- brids as in Quercus Leana (Q. coccineaXimbricaria) and Boston fern hybrids, it is a source of astonishment when this occurs in species which we do not believe to be of hybrid origin. Heterophylly is a matter of common observation in the ubiquitous horseradish, Radicula Armoracia. Sterculia diversifolia, commonly cultivated in greenhouses, has leaves which are mostly ovate to lanceolate in outline and which are often entire or vari- ously three- to five-lobed on the same shoot. The sassa- fras, S. variifolium, illustrates the same principle, in fact, derives its name from this feature. The paper mul- berry, Broussonetia papyrifera, shares the irregular lobing of the leaf with its near relatives, Morus rubra and M. alba12 The cut-leaved Persian lilac is very variable in this respect.1* Sometimes seasonal differ- ences are so great as to make collecting throughout the year an absolute necessity,’ an instance so strikingly illustrated by Viola palmata. In woody plants there is not infrequently a return to the nepionic stage in the leaves produced near the base.!* When we remember that heterochromatism'® may add a further complication, it becomes evident that it is not sufficient to be able to refer to herbarium specimens only. The living plant must be studied in its various stages of development. It is by this means only that one may arrive at a true esti- mate of the stability and significance of minute details. 12 Rep. Bot. Dep. New Jersey Agrie. Coll. Exp. St., 325, pl. 21, 1909. Fry, A., ‘‘Note on Variation in Leaves of Mulberry Trees,’’ Biometrica, 1: 258, ‘1902. 13 Masters, M. T., ‘‘ Vegetable Teratology,’’ fig. 177. See also, Schlecht- endal, Bot. Zeit., 13: 559, 1855, and Lloyd, Torreya, 2: 137, 1902. * Cushman, J. A., ‘‘Studies of Localized Stages of Growth in some Common New England Plants,’’ A. NAT., 36: 865; ibid., 37: 243, 38: 819. 3 Jackson, R. T., ‘‘ Localized Stages in Development in Plants and Animals,’’ Mem. Boston Soc. Nat. Hist., 5‘: 89, 1899. » Teorie von Marilaun, A., ‘‘The Natural History of Plants,’’ 1: 149; 646 THE AMERICAN NATURALIST [Vou. XLV Trelease,17 in a recent paper on his favorite agaves, says: ‘f... my own conception of specific identities and differences in the genus oscillates as my study pro- ceeds. ...”’ While ultimately work of this nature, conscientiously carried out, leads to a clear delineation of the characters peculiar to each species, subspecies, variety, etc., there is reached, usually not so very long after the inception of the work and very long before its termination, a period of chaos which, to say the least, is the reverse of stimu- lating and ordinarily sufficient to dampen the ardor of the greatest enthusiast. Thus, in my cultures of local forms of Capsella Bursa-pastoris, the four forms describéd by Shull'® and some of those of Almquist!’ can be recognized readily. But there remain so many whose classification is doubtful at the present time that it would cause Mr. Murdock, associated with me in this investigation, and myself to throw up our hands in despair, were it not for the fact that there have appeared in our cultures extreme forms, so striking and so different from any previously described, that we are in- clined to believe we are possibly dealing with mutations in the sense of de Vries (Fig. 1). Research along these lines requires long and tedious experimental efforts. It is a different matter where the finding of distinctly new forms in an old territory is concerned. Here an element of chance largely enters. By increasing the number of our observations our chances of discovering something new may be increased materially. In other words, directly or vicariously a large amount of ground has to be covered. Yet I am not satisfied that a thorough knowledge of a comparatively small area would not lead to equally satisfactory results. The one requires about as much exertion as does the other and certainly the " Trelease, W., ‘‘ Species in Agave,’’ Proc. Amer. Phil. Soc., 49: 232, pl. 32, 33, 1910. 8 Shull, G. H., ‘“‘ Bursa Bursa-pastoris and Bursa Heegeri Biotypes and Hybrids, ”? Carn. ‘Inst. Publ., No. 112, 1909. 3 Almquist, E., ‘‘ Studien ueber die Capsella Bursa-pastoris (L.),’’ Acta Horti Berg., 4, Wo. 6, 1907. No. 539] ORIGIN OF SPECIES IN NATURE 647 number of observations is identical. But the greater familiarity with the territory and the fact that observa- tion may be continued throughout the seasons offers cer- tain advantages. It is a question which must be decided for individual cases by every investigator along this line, Fig. 1. Appearance of a new form among seedlings of Capsella Bursa-pastoris. x1. since its answer must depend on local conditions, such as proximity to residence, climate, whether arid or temper- ate, etc. A garden, for instance, offers as good if not better opportunities to observe the appearance of new forms as does the field. Here all the requirements for such work are met with, i. e., large numbers grown under highly favorable conditions. Parks, nurseries and espe- cially the larger botanical gardens, where are kept accu- 648 THE AMERICAN NATURALIST (Vou XLV rate records of the histories of the various plants grown, constitute an important field. Thus, in the Missouri Botanical Garden, there are grown every year numerous plants of Bellis perennis ‘‘delicata.’’ In 1906 the seed for these specimens was obtained from a Philadelphia seed firm. The seed was sown and the plants were handled in the usual manner until they finally found their way to the beds used for decorative purposes. It was while the plants, several thousands in number, were in this position, that the attention of Mr. Shelby Jones, at that time a student at the garden, was attracted to a plant because of the abnormal character of its flowers. In the capitula of this specimen the rayflowers had either disap- peared, or, what is more probable, had been replaced by Fig. 2. Capitula of Bellis perennis “ Delicata’’ and Bellis perennis discoidea dise-flowers. The result was a rather striking, maroon-red button, in sharp contrast with the normal heads (Fig. 2). For in the normal form of the variety under cultivation the upper surface of the ligules was either white or rose- colored, while the lower surface was red. In the discoid variety, owing to the tubular nature of the transformed No. 539] ORIGIN OF SPECIES IN NATURE 649 rayflowers, the white coloring disappeared from view. It is an instance of change of color analogous to that of the cactus dahlia, though differing in one striking fea- ture. For in the latter case the brilliancy of the flowers is due to the recurved rayflorets which show the bright color of the upper surface instead of the dull shade of the lower one, exposed to view in the older form of dahlia. The plant was segregated and propagated vegeta- tively, as many as one hundred specimens being on hand at atime. Among these not a single instance of a return to the conventional form was observed. Of course, it would have been far more interesting to note the result of sexual propagation, a proceeding which was rendered impossible by the failure of the flowers to produce good seed. This, however, is quite in accordance with the experience of horticulturists, fide André.2° In this most comprehensive publication the variety discoidea is men- tioned under the name of ‘‘Paquerette vivace var. double a fleurs tuyautées,’’ its slight fertility noted and attention called to the fact that the seed for such plants must be collected from flowers which are almost double. It is added: ‘‘leur tendance a doubler est assez bien fixée pour qu’on ne trouve dans le semis presque plus de fleurs de duplicature imparfaite.’’ In this connection Master’s notes published in the appendix to his ‘‘Vegetable Teratology,” and de Vries’s remarks on Matthiola incana fl. pl. are of considerable interest. Penzig evi- dently refers to a similar form,?? though his use of the term ‘‘ox-eye daisy’’ for this form appears to be con- trary to usage, it being reserved ordinarily for Chrysan- themum Leucanthemum. Similar instances are not rare among Composite and are illustrated by Matricaria (Chamomilla) discoidea and Anthemis tinctoria dis- coidea. Very frequently one is able to find a few tubular flowers among the rayflowers of our cultivated com- 2 Vilmorin-Andrieux et Cie, ‘‘Les fleurs de pleine terre,’’ 5th ed., R.-Ed. André, éditeur, 781, Paris, 1909. 2 De Vries, H., ‘‘Plantbreeding,’’ 238, Chicago, 1907. 2 Penzig, O., ‘‘Pflanzen-teratologie,’’ 2: 59, Genoa, 1894. 650 THE AMERICAN NATURALIST [Vou. XLV posites, especially Cosmos and Coreopsis. Some of my earliest experiments, undertaken to study the effect of selection on such deviations, yielded negative results. De Vries? describes and illustrates a fistulosa variety of Dahlia variabilis which appeared among his cultures. Among chrysanthemums such ‘‘quilled’’ forms are not rare. ‘‘Knterprise’’ is a variety which illustrates this type and which not infrequently is met with at chrysan- themum exhibitions. Next to gardens, cultivated fields, i. e., places where a large number of individuals of the same species or variety are grown, offer the best opportunity for the discovery of new or aberrant forms. One only has to be reminded of the case of the beardless Anderbeck oats.?* In a similar manner there was found, among the numer- ous alfalfa plants which occupy a large portion of the acreage of the garden, a white-flowered specimen, the flowers of which, on fading, became almost yellow. Since then there were found in a neighboring field, a group of such white-flowered plants. Color varieties are among the deviations most fre- quently noted, perhaps because the resulting change is particularly easy of observation. They occur both as to flowers, as for instance in Lobelia syphilitica alba, Tri- folium pratense album, ete., and as to fruits, as instanced by Gaylussacia resinosa leucocarpa,® Vaccinium penn- sylvanicum leucocarpum, V. corymbosum atrococcum f. leucococcum, V. canadense chiococcum.® White straw- rries of course are well known. Yellow-fruited vari- eties exist of Ilex myrtifolia?" as well as of Ilex opaca and I. verticillata. The Californian holly, Heteromeles arbutifolia, also has a yellow-berried variety.2% Calli- *De Vries, H, ape ? 1: 480, fig. 134. * De Vries, H., loc. ci * Porter, Thos. C., Oppe to our Native Flora,’’ Bull. Torr. Bot. Cl., 16: 21, 1889, * Deane, W., ‘‘ Albino Fruits of Vacciniums in New England,’’ Rhodora, 3: 263, 1901. OE o R., ‘‘Ilex myrtifolia with Yellow Fruit,’’ Torreya, 2: 43, *The American Botanist, 15: 49, 1909. No. 539] ORIGIN OF SPECIES IN NATURE 651 carpa americana and Rubus cuneifolius both possess white-fruited varieties.?° Of Solanum nigrum there exist, besides the black- fruited form, one with yellow and one with green fruit. Atropa Belladonna exhibits the same color forms of the fruit which to us are familiar in the cultivated peppers, viz., red and yellow. Finally, in shrubs which are grown chiefly because of the coloring of the bark, for instance Cornus stolonifera, with a red bark, there may be met with forms which have a yellow bark. So numerous are references to color varieties in our literature and these variations are met with so often that their frequency suggests facility of origin perhaps parallelled only by that through which dwarf forms are produced. For Several years, on collecting trips in the vicinity of St. Louis, Mo., all patches of Lobelia cardinalis were exam- ined with particular care for white-flowered plants, since they were known to exist in various parts of the United States. Though these observations did not lead to the desired result, Mr. O. S. Ledman, of St. Louis, was able to find on the peninsula in the northern portion of Pitts- burg Lake, St. Clair Co., Ill., opposite St. Louis and in a locality with which I believed myself to be thoroughly familiar and had searched most carefully on several occasions, some plants of Lobelia cardinalis alba. This would indicate that the formation of the white- flowered variety had taken place recently, though the possibility of transportation from other localities, though improbable, is not excluded. That a white-flow- ered form of Medicago sativa is formed as readily seems likely. Various species belonging to the Leguminose possess alba yarieties. De Candolle speaks of a color variety,” and at first it was thought the specimens under consideration could be classed here. Since then, how- ever, there were found in the same alfalfa field several » Rolfs, P. H., ‘‘ Variation from the Normal,’’ Asa Gray Bull., 8: 75, 0. 190: % Medicago sativa versicolor (Sér. mss.) : floribus luteis ecoeruleisque. M. falcata versicolor Wallr. sched. cort., p. 398. M. lutea-cærulea hort., Prod., 2: 173, 1825. 652 THE AMERICAN NATURALIST [Vou. XLV plants which answer the description of the variety men- tioned in the Prodromus, more or less. It is more than probable that these plants, the flowers of which exhibit a great variation in color, are to be included under the ‘variegated alfalfas,’’ for which Westgate’! suggests the designation ‘‘Medicago falcata X (M. sativa), to indicate that ‘‘the hybrids have been-recrossed several times with ordinary alfalfa and also among themselves.’’ Hybridization, however, does not explain the white color of the flowers. That white-flowered plants and those with flowers of a yellow or cream color are not of rare occurrence is shown by the data furnished by Westgate? who notes such plants among four, respect- ively five of the forms experimented with. On the other hand, in view of the known existence of white-flowered forms the assumption of the formation of the white- flowered plants at Ann Arbor, through recent mutation, does not seem warranted. At the same time, such a possibility is not excluded, for while we do not believe that hybridization can have as its direct result the pro- duction of a new character, or the loss of one,** it 1s not at all improbable that hybridization indirectly may cause a plant to initiate a mutation period. It is ex- pected that the seed gathered from the white-flowered plants will, in the course of a generation or two, yield a pure white offspring. This seems to be the experience of others. Mr. C. V. Piper, of the Bureau of Plant In- dustry, U. S. D. A., in a reply to a recent letter concern- ing alba varieties of Medicago sativa says: ‘‘Some of * Westgate, J. M., ‘‘ Variegated Alfalfa,’’? U. S. D. A., B. P. I., Bull. No. 169, pai 1910. 2 Loc 37, Table I. Roane rites which may suggest themselves but which obviously fall out- side our definition, are in the first place those in which atavism comes into play, such as is the case in a cross between Datura levis and D. feros. (Naudin, Ann. Se. nat., 5 Sér., 3: 155, 1865. Reciprocal crosses of the two white- fovea plants vishded a uniform pei sae with pale purple flowers. (See also de Vries, ‘‘Mutationstheorie,’’ 2: 44, 201.) More recently, through the work of Cuénot, Durham and Nilsson, another, more oe explanation of the reappearance of the purple color can be given. ( L. — ‘*Vererbungslehre und Deszendenztheorie,’’ 1910.) No. 539] ORIGIN OF SPECIES IN NATURE 653 our selections of these white-flowered forms now breed true.’’ In connection with this alba form I wish to call atten- tion to the local distribution of a white-flowered form of Solanum Dulcamara, lately taken into cultivation in the experiment garden. This variety occurs but rarely Mich. Fic. 3. Arctium minus laciniatum at Albion, in the vicinity of Ann Arbor, chiefly on river and lake shores and in swampy places. But at Albion, Mich., it was found in great abundance, growing in moist situa- tions alongside of the species. While it will take some years to determine the constancy of this variety, there is in my mind no question as to the outcome of the ex- periment. It is a variety not always recognized in our floras, though evidently met with from time to time.** But if we give specific rank to mere color varieties, as we do in the case of Datura Tatula and Datura Stramo- nium, why not give taxonomic recognition to the equiva- lent color variety of Solanum Dulcamara, the more so * Collins, F. S., “A Variety of Solanum New to America,’’ Rhodora, 12: 40, 1910 654 THE AMERICAN NATURALIST [Vou. XLV since the addition of such an exceedingly variable char- acter as is pubescence, seems sufficient to bring this about (Solanum Dulcamara var. villosissimum Desv.) ? The last illustration points to the largest source of new or at least imperfectly known species and varieties, i. e., the native flora. The same rule applies here as in cultivated fields: striking morphological differences most readily attract attention. When the plant normally is possessed of large leaves, as is, for instance, Arctiwm minus, any differences in the foliage become especially noticeable. It is through this fortunate circumstance that we owe to Professor Charles E. Barr, of Albion College, Michigan, the discovery of at least a new local- ity for the apparently rare laciniate form of the species just mentioned (Fig. 3). Laciniate forms are of relatively frequent occurrence. We find them not only among the phanerogams and the vascular cryptogams, but even among the alge one can meet with forms which may be interpreted as such, for instance Callophyllis furcata Farlow® and C. furcata f. dissecta Farlow*® (Fig. 4), though of course there is no connection. One of the earliest accounts of the sudden appearance of a laciniate variety is given by Marchant.’”? In 1715, in his garden, he discovered a plant which, though evidently belonging to the genus Mercurialis, was entirely new to him and which did not appear to have been described previously. He named it Mercurialis foliis capillaceis. No seed being collected, the next year the same garden spot was anxiously watched. Six plants made their appearance, four of which possessed the character of the plants which had appeared in 1715. The other two were sufficiently different to be segre- * Exsic. in Collins, Holden and Setchell, Phycotheca Bor. Am. Fase. 18, No. 883, 1901. = Setchell, W. A., and Gardner, N. L., ‘‘Algw of Northwestern Amer- +” 306, Univ. Calif. Publ., 1, 1903. ‘x Marchant, J., ‘‘Sur la prodato de nouvelles espèces de plantes,’’ Hist. de V Acad. d. Sc, 1719, 57, Paris, 1721. Marchant, J., t Cmcrvations sur la nature des plantes,’? Mém. de V’Acad. Roy. d. Bex 1719, 59, pl. 6, 7, Paris, 1721. See also THE AMERICAN NATURALIST, 45: 493, August, 1911. A . 539] ORIGIN OF SPECIES IN NATURE 655 ` b Fig. 4. Callophyllis furcata and C. furcata dissecta. 656 THE AMERICAN NATURALIST [Vou. XLV gated under the term Mercurialis foliis in varias et imequales lacinias quasi dilaceratis. They were remark- able chiefly because of their laciniate leaves. Both forms appeared to differ from Mercurialis annua in having a longer lease of life, since they remained green until the latter part of December. Plants of both continued to appear in 1717 and 1718, propagating themselves. Masters and others have given lists containing several instances of the appearance of laciniate forms, to whose number one readily may add by glancing over the cata- logues annually published by the principal seed houses. It is especially of trees and shrubs (Acer platanoides laciniatum, Betula pendula var. dalecarlia® Rubus fruticosus laciniatus) that cut-leaved varieties are in de- mand. This popularity is equaled only by that which Chelidonium majus laciniatum appears to enjoy in the world scientific, dating from the time when Roze?’ called attention to its history. Laciniation seems to have taken place at least twice in the genus Chelidonium. Of C. japonicum Thumb. occurs a var. dissecta.*® There exist several varieties of Chelidonium majus. Thus we have the broad-petaled form (C. majus latı- petalum) of the Groningen Botanic Garden, the double- flowered variety and the cut-leaved one, the latter par- ticularly interesting because the laciniation extends to the petals. During the last four years I have cultivated s A cut-leaved variety of B. pendula is oa — vide Sanford, S., A neti Cherry Birch,’’ Rhodora, 4: 83, 1902. ” Roze, ‘*Le Chelidonium laciniatum ane Journ. de Bot., 9: 296, 301, ie "1895. “Prain, D, ig Revision of the Genus Chelidonium,’’ Bull. Herb. Boiss., 3: 570, 1 The aeee yield other instances of the reappearance of an abnormality in members of the same genus, for instance pistillody of the amens. The most quoted instance is that of Papaver somniferum poly- cephalum, a variety which was grown more than fifty years ago in the trial grounds of the Vilmorins. Similar varieties were described by von Mohl for P. orientale and by Elkan for P. dubium (Henri van Heurck, ‘‘ Notice sur une prolification axillaire floripare du Papaver setigerum,’’ Bull. Soc Roy. Bot. Belg., 2: 329, 1863) and is said to oceur also in Macleya aids (Le Sourd-Dussiples et Georges Bergeron, ‘‘Note sur un cas de méta- morphose ascendante,’’ Bull. de la Soc. bot. de France, 8: 348, 1861). No. 539] ORIGIN OF SPECIES IN NATURE 657 eight lots of Chelidonium majus laciniatum, two of which have double flowers, the seed being obtained through the exchange lists of various botanic gardens. These eight lots apparently represent five distinct forms. The differences are not great and probably would re- main unnoticed by the casual observer. After constant association with them one can not fail to recognize the differences, however slight. But though slight, the dif- ferences are constant. They consist in the degree of laciniation as well as in degree of hirsuteness. It might be argued that these differences perhaps are due to dif- ferences in external conditions or in the age of the plants or in seasons, etc., just as the leaves produced by Acer saccharinum var. Wieri, in the latter part of the summer at the extremities of the long, slender twigs, have a lamina far more reduced than those formed earlier in the year. Roze calls attention to the fact that in his cul- tures of Chelidonium laciniatum the degree of lacinia- tion of the leaves increased as the season advanced,“ the petals undergoing a similar change. He also noted that the degree of laciniation increased with the amount of light received. While we have noted that the differences between the varieties are most marked in the early part of the year, the fact that our cultures were carried out under uni- form conditions as to soil, light and water supply, and that the various types are recognizable even in the late summer, seems to indicate that these forms are entitled to varietal rank. The specimen illustrated in Fig. 5 ought to set at rest all doubts upon this point. One hardly would care to account for the extreme reduction of leaf surface in this case on the basis of the influence of fluctuating variability. This form, in all probability, is identical with the Chelidonium majus foliis et flore minutissime laciniatis of the Hortus regius (1661), which “ Roze, loc. cit., . “, ,. si l’on suit la plante dans sa croissance, on remarque que les deux ou trois premières feuilles (après les feuilles germi- natives) ont l’apparence de celles de la forme crenatum et les dernières celles de la forme fumarifolium, mais a découpures plus courtes et moins étroites. ”? 658 THE AMERICAN NATURALIST [Vou. XLV originated in the Paris Botanical Garden from seed of Chelidonium majus laciniatum and to which de Candolle afterwards gave the name C. laciniatum fumariefolium.” A second plant in our garden represents an extreme in another direction and is intermediate between the Fic. 5a. Chelidonium majus. species and the variety laciniatum. It probably is iden- tical with the var. crenatum.? Besides these three varie- ties I believe to be able to distinguish two other forms, constant from seed. It is possible that all of them origi- nated through mutation of C. majus laciniatum, as did the variety fumariefolium. It is equally possible that laciniate forms of Chelidonium majus have originated "Frog, 1: 183: “C. majus crenatum Lange, FI. dan. No. 539] ORIGIN OF SPECIES IN NATURE 659 more than once and directly from the parent species. This is the view taken by Clos, who describes the find- ing by P. Barthés of a plant of C. majus fumariefolium in Soréze, Tarn. It is to be noted that these plants bore underdeveloped, seedless pods. This last seems to speak ~ i Fic. 5b. ©. fumariefolium. in favor of the assumption of a creatio de novo, since plants from other stock appear fertile. To this ex- tremely interesting point, a discussion of which falls outside the scope of the present paper, I hope to return in an article shortly to be published. The correctness of the view that the same mutation or at least a mutation in the same general direction may take place in different stocks, must be granted a priori. agi in the support of this belief we find in the ex- os, D., ‘‘Réapparition de la Chélidoine a feuille de Fumeterre,’’ ane rend., 115: 381, Paris, 1892. 660 THE AMERICAN NATURALIST [Vou. XLV istence of several distinct laciniate forms of different species of trees and shrubs. The varieties heterophylla, laciniata, asplenifolia*® and incisa of Fagus sylvatica, Alnus glutinosa laciniata, A. glutinosa quercifolia and A. glutinosa oxyacanthifolia are instances to which numer- ous others might be added. It would seem that the laciniate forms of Mercurialis annua, observed by Marchant, would find a place here.*® The repeated sudden appearance of the same variety has been noted by various authors. Darwin, Korschin- sky,‘ the late director of the St. Petersburg Botanic Garden, and de Vries*S give numerous instances. Thisel- ton Dyer was able to show the repeated formation of at least two new varieties of Cyclamen latifolium.*® Many, from personal experience, will be able to supply other instances. And while most of the cases which come to our notice probably are explainable through accidental transportation of seed or through Mendelian splitting of a hybrid between the species and a retrograde variety or through atavism,” there are others which do not admit # This variety, like cut-leaved varieties of other species, not infrequently shows atavism in certain shoots. See de Vries, ‘‘Atavismus durch Samen und durch Knospen’’ (‘‘Mutationstheorie,’’ 1: 482), and also R. G. Leavitt, ‘‘ Partial one Sy in Leaves of the Fern-leaved Beech,’’ Ehodora, 6: 45, 1904; O. Paulsen, ‘‘Blivende Axelblade hos Boegen,’’ Bot. Tidskr., 24: 281, 1902; A eden ‘t Contributo alla teratologia vegetale,’’ E. 8. Bot., Hai 44, 1902. “This variation in the degree of laciniation within a single species ought to throw some light on ‘‘unit-characters.’’ If we consider—as I believe generally is done—laciniation to be a ‘‘unit-character’’ it would seem that such a ‘‘unit-character’? may be subject to considerable varia- tion, though the degree of variation is constant or at least approximately ch individual case. This relative stability of the varietal character of course does not prevent the extremes of each degree of variation from rschinsky, 8. Tienen gee und Evolution,’’ Flora, 89, 240, 1901. ‘3 íí Species and Varieties nd “ Thiselton Dyer, W ties Cultural Evolution of Cyclamen lati- folium,’? Proc. Roy. se "61: 135, 1897. See also J. Denman, ‘(The Sporting Peculiarity of ihe, Paria Cyclamen,’’ Gard. Chron., 3d Ser., 29: 266, 1900. = Among the plants cultivated in the experimental grounds of D. M. _ Ferry & Co., of Detroit, Mich., is a variety of cabbage bearing the name No. 539] ORIGIN OF SPECIES IN NATURE 661 of such an interpretation. Thus the finding by Mr. W. H. Ransome, of several plants of the four-leaved variety of Fragaria vesca at a point about twenty miles west of Kalispel, Flathead Co., Mont. A new form, which has appeared at various times and which because of the na- ture of the variation is incapacitated from reproducing itself by seed, would from this very fact constitute an ideal illustration of repeated mutation, since a hybrid origin of the individuals which appeared later, is ex- cluded. Such an instance is yielded by the wheat-ear carnation, Dianthus Caryophyllus imbricatus. LANSA FEE = 3S i a A Ge rg EE E UN g a 8 z ALAIS a 8 = apm Se BS a S § =en a : = o 29 aa = S 2 p> | 2 ag = a 25 = 4 3 & | 2/2] 38 3 2 S algi 38 g © S x G 2 5 $ 3 3 z = e | & | § a 5 e | § | § |Blals (a) 26 1 2 4/13 7 3 2 | 2 (b) 33 22 S Fora 8 2 u e W | 14 i 2i 0 be Experiment No. 25—July 30. Time of observations one half hour, from 11:30 to 12:00 a.m. Three blossoms arranged in the same row. (a) Normal cotton blossoms pinned in position as a control. (b) Normal cotton blossom with honey added at base of petals within. DATA FOR EXPERIMENT No. 25 2 g Distribution by Species z Distribution by Species 5 = = ; x k Se E E zanr ii?i’ ph S$ a Š = = Sa 3 9 È fse] 3 2 oe ee eee 4i ije 38 ai i ir a $ Dd = J © a ~ — A ee Be eee ea (a) 22 16 6 4 4 (b) 19 14 4/1 1 fe) | 1 10 3/2 2 re 672 THE AMERICAN NATURALIST [Vou. XLV (c) Normal cotton blossom pinned in position as a control. During the forenoon the weather was dull, so that bees were less frequent: in their visits. Experiment No. 26—August 1. Period of observa- tions one half hour, from 9:00 to 9:30 a.m. Blossoms arranged in same row as follows: (a) Normal cotton blossom with honey at base of petals within. (b) Normal cotton blossom pinned in position as a control. (c) Normal cotton blossom pinned in position as a control. DATA FOR EXPERIMENT No. 26 | Distribution of Species A 2 Distribution of Species a ea S a n Lio] fos} ~ : : w Sa sil a e ieras | & | paa 7 A $ = |M |S] Se a 3 E |M zg zg a Š S or ae £] 3 = mie | g9 3 > 2 Z (812 8% | 4 Š & e 3A = z R] Fay a z 3 4 pe < 2 B f=] = 2 3 3 = TERRE £ $ K] sig É = & = Ñ a| p z 5 (a) 26 22 i A a Se i 1 0 (b) 27 2 1 c) 20 17 i; s 1 Experiment No. 27.—August 1. Period of observa- tions one half hour, from 9:00 to 9:30 a.m. Blossoms are arranged in the same row. At (a) an unmutilated cotton blossom was pinned in position. Portions of cotton leaves were carefully cut out and fastened outside and within the blossom in such a manner as to extend just to the margin of the petals on both sides. In this way none of the yellow color of the petals remained visible. The stamen tube, pistil, etc., projected as in a normal blossom. The blossom was practically without petals, since these were not visible, although such odors as they may have possessed could . still diffuse around the blossom. A drop or two of honey was also added at the base of the petals within in order to make certain that agreeable odors were present, since No. 539] THE BEHAVIOR OF BEES 673 these must now necessarily constitute the sole allure- ment. At (b) normal blossom pinned in position. The tips -of the petals were lightly smeared with honey. At (c) control. Normal cotton blossom pinned in position. DATA FOR EXPERIMENT No. 27 A | a Distribution by Species 2 Distribution by Species = = > m | 3 : pA S S g | n =] 2 E ae ee ae g| 3 2 * iS ae gee 28 A S E TAJA] Se a 3 E Aim] se Sa (an) $3 = a a ag pes S s wi 2 | 6&3 z E 2 me | 2) ed gee 3 ~ Se E 3% $ S.d ua Pes pt 2 cit eee betes z a = Ñ | ajal z = R [Rl ale 1 Re ee | | (a) | 0 | | ay + 3 23 10/1 | 2 | 2 (c) 14 10| 1 | 6 | 6 As shown in previous experiments, the removal of the petals no longer advertises a cotton blossom to the notice of bees. The same results are obtained when the petals are no longer visible, although still attached to the blossom as at (a). It is natural to suppose that the pres- ence of honey would add appreciably to the zone of alluring odors surrounding the blossom. Without the conspicuous corolla to invite inspection, however, the bees are not led to approach sufficiently near to discover the blossom by its attendant odors alone. These results are not in agreement with some of the general conclusions of Plateau in his noteworthy mem- oirs: ‘‘Comment les fleures attirent les insectes.” He states: ‘‘Les insectes visitent activement les inflores- cences qui n’ont subi aucune mutilation mais dont la forme et les couleurs sont masquées par des feuilles vertes.””3 This would follow only when other attractive influences were actively operative, as various odors agreeable to bee visitors. Experiment No. 28.—August 1. Time of observation one half hour, from 9:30 to 10:00 a.m. This experiment makes use of most of the material and the same positions of the preceding, with the changes as follows: * Bulletin de l’Académie royale des Sciences, No. 11, November, 1895. 674 THE AMERICAN NATURALIST [Vou. XLV (a) Outer leaf covering removed from the blossom used at (a) in preceding experiment, thus making the outer surface of the petals visible. Honey at the base within, as before. (b) Normal blossom pinned in position as a control. No honey has been added to this blossom. (c) Normal blossom used in preceding experiment with petals removed. DATA FOR EXPERIMENT No. 28 Distribution by Species è ERREI 3 a Distribution by Species g > 3 : Pe a : % a j| og Se 3 a >z Fe S73 K a $ Sig S g A % = ® | faa] C=] =| & es Q fea) = as 2 x g m Sa 5 Š g m = oa s = Š lej a2 = Š = bh! 2; 8s 3 S 3 eij] 24 r 2 R 2 2) 3a = z 2 a a oA g£ 2 & =| g = a Rg % Bl] = $ 2 $ Sie ee tes es | §$ |] § Rara oe = = R JEJA] p = 8 aka (a) 9 6 0 b) 11 1 (c) 1 1 By the removal of the outer covering of the blossoms at (a), which in the previous experiment received no in- spections, it became nearly or quite as attractive as the control at (b). The blossom at (c), however, no longer afforded means of attracting the bees. In this experi- ment and the previous one the corolla at (a) was concealed with portions of cotton leaves to guard against introducing repellent odors which may have attended the use of any other material. In the course of this experiment the number of bees flying about became much reduced toward ten o’clock, although the day was clear, hot and sunny. The writer was even forced to postpone his observations for the remainder of the forenoon owing to the scarcity of visiting bees. Experiment No. 29.—August 1. Period of observation one half hour, from 2:00 to 2:30 p.m. The material is arranged in the same row. At (a) a cotton bud not due to open until the next morning had its petals quite fully pulled open so as to resemble a naturally opening blossom. No. 539] THE BEHAVIOR OF BEES 675 At (b) a second cotton bud due to open the next morn- ing had its petals partly pulled open. At (c) a normal cotton blossom growing in situ as a control. DATA FOR EXPERIMENT No, 29 h a Distribution by Species z Distribution by Species = 9 ò = B : % FRA a : & 9 A © a n © eee oe te ELS a Pe a ao a 2 £ m Sa a S g = Bn =” as 3 Š h| 2| 8S H 3 = si 21 88 Se ee ee a ee Z S > $ oi 818 = Ss S 318] 8 ñ 5 = A | Ala} e z = À | Ala} ep | (a) 15 5 10 0 (b) 2 0 (c) 24 4 20 14 14 In this experiment it was intended to observe the behavior of bees toward immature buds at (a) and (b) in comparison with fully expanded, mature blossoms. It was assumed that the former, owing to their imma- turity, would perhaps prove less attractive to bees through the sense of smell than the fully matured blossoms. Although not definitely proved, it is reason- able to suppose that the processes of active nectar secre- tion simultaneously attend the unfolding of the petals and the shedding of the pollen. It would then follow that the odor of the unopened buds at (a) and (b) would prove less alluring than the blossom at (c). The yellow petals of the blossom at (a) have served to invite frequent inspection, although at (b) this is not as evident. This difference may depend upon the fact that the bud at (b) was much less conspicuous, since the petals have been only slightly pulled open. Whatever the true explanation, the mature, fully unfolded blossom which serves as a control at (c) has received many more inspections, nearly 60 per cent. of which are actual entrances. Experiment No. 30.—August 2. Period of observation one hour, from 9:30 to 10:30 a.m. Two blossoms were nin in the same row in equally conspicuous posi- ions. At (a) a very clean, thin 5X7 glass plate was sup- 676 THE AMERICAN NATURALIST [Von XLV ported in front of a fully opened cotton blossom pinned in position. The surrounding cotton leaves were then carefully drawn in closely around the plate so as to over- lap the edges and most of the glass surface. In this arrangement, although the blossom was plainly visible through the glass, only a small portion of the glass sur- face remained in view. At (b) control. A fully opened cotton blossom was pinned in position. DATA FOR EXPERIMENT No. 30 h a Distribution by Species PA Distribution by Species K E] 2 aa ORE A A TES 3 : 8 . SoS oe Te eee Se ee fe ee og a $ AE £ pes Š Sl S| g =} n = Š = = = S = fa] Sn aS fi Š m pe S] Š o =e = H $ Š pi 2 a3 S 3 P| a] of = E 2 S |$i¢1s3 13 3 | ale) 2" $ 5 Š 2 elsi a S S 2-7 Si eis = S = Ñ Hilal p a = 3 | a | | amas: POT (a) | 18 | 6 34 0 | | (b) | 37 | 20 714 ) 514 Of those bees which attempted to inspect the blossom at (a), eight flew more or less forcibly against the glass, including two small unidentified bees, one Bombus sp., two honey bees and three Melissodes. One small bee tried persistently several times to fly through the glass toward the blossom just behind it. It is at once obvious that the blossom at (a) invited inspection by passing bees solely through the sense of sight. Experiment No. 31.—August 9. Observations con- tinued one hour, from 9:00 to 10:00 a.m. Two blossoms were used in the same row as before. A box of thin wood was carefully constructed for this experiment. The dimensions were such that both cover and bottom were made of thin clear glass by using for each a 5X7 glass plate. These plates fitted tightly in lateral grooves. The box was about 4 inches deep, 80 that the end of a cotton branch together with its leaves and blossoms could be carefully pushed into a natural position within. By sliding the glass cover into place the box became practically air-tight. The blossom was plainly visible to passing bees, although any attractive No. 539] THE BEHAVIOR OF BEES 677 odors which it diffused could no longer act as allure- ments. When placed in position among the cotton limbs the box and glass were almost completely concealed by drawing in and fastening around it a number of the sur- rounding leaves. This box enclosed the blossom at (a). At (b) control. A natural blossom pinned in position. The blossom at (a) received only two inspections, both by small, unidentified bees. The blossom at (b) received only five inspections, three being by Melissodes and two by honey bees. There were no entrances at (b). Experiment No. 32.—August 9. Observations con- tinued for one half hour, from 10 to 10:30 a.m. This experiment was identical with the preceding, except a third artificial blossom was added. (a) Same blossom enclosed in the glass case as used at (a) in Experiment 31. (b) The same control blossom pinned in ee NE (c) An artificial paper blossom of crepe paper simu- lating a cotton blossom in color. The blossom at (a) received one inspection by a Melis- sodes which hovered in front of the glass a few seconds only. (b) Received eight inspections, which include two entrances by honey bees. The six inspections were entirely by Melissodes. Three bees inspected the arti- ficial paper blossoms at (c), including one each by a Melissodes, an unidentified bee, and a honey bee, which paused for a few seconds over the blossom but did not enter. During the last few days bee visitors have been rather too infrequent for satisfactory work, probably in part owing to the partly cloudy forenoons. Experiment No. 33—August 11. Observations con- tinued for about one half hour, from 9:00 to 9:30 a.m. Three blossoms are arranged in a row. At (a) three petals were placed in position on a leaf which was spread out flat and held between two thin clean 5X7 glass plates bound firmly together. This was placed in position among the branches of a plant and was nearly concealed by overlapping around it a number of cotton 678 THE AMERICAN NATURALIST [Von. XLV leaves. In this arrangement the yellow petal color still remained perceptible to the visual powers of bees. At (b) the detached petals of a cotton blossom were enclosed in a 250-c.c. graduated flask which was stoppered and fastened among the cotton leaves. The yellow petals were plainly visible through the clear thin glass of the ask, At (c) a normal cotton blossom as a control. (a) Received no inspections. (b) Received one inspection by a honey bee which touched the flask. (c) Received three inspections by honey bees. Two bees did not enter the blossom but alighted outside to get at the extra-involucral nectaries. During this experi- ment the bees were too infrequent visitors to make any results conclusive. Experiment No. 34—August 12. Time of observa- tions 45 minutes, from 1:30 to 2:15 p.m. This experiment was an exact duplicate of the previous one. (a) Petals of cotton blossom on a leaf between glass plates as at (a) in Experiment 33. (b) Petals of cotton blossoms enclosed in a flask as at (b) in Experiment 33. (c) Normal blossoms growing in situ as a control. The blossom at (a) was twice inspected by Melissodes. The petals in the flask at (b) were inspected two times by bees, once by a honey bee and once by a small, unidenti- fied bee. The blossom at (c) was inspected six times, in- cluding five entrances. One inspection was made by a small, unidentified bee, four entrances were by honey bees, and one entrance by a Melissodes. For several days it was rather difficult to secure sat- isfactory data, as the bees were much less frequent visit- ors. The species of Melissodes, which were extremely common at the beginning of the experiments finally be- came far less common, although honey bees greatly increased. It is very interesting to note that many honey bees finally began to confine their visits solely to the outer involucral nectaries instead of entering the blos- No. 539] THE BEHAVIOR OF BEES 679 soms. This change of habit seemed to become quite gen- eral at about the same period, for the writer noted it in all parts of the cotton field. Throughout the period of observations the bees by their behavior and varying numbers showed themselves extremely sensitive to atmospheric changes, temperature relations, air movements, moisture, sunshine, ete. At times conditions even too obscure for human perception may have regulated their activities. The composition of the bee fauna to be observed in cotton fields shows much variation, depending upon the time of day, prevailing weather and seasonal influences. The position and expo- sure of the cotton field with relation to various local phys- iographic features, as type of soil, nearness to woods, swamps, hills and other crops also greatly influence the relative numbers and kinds of bees. While carrying on his observations the writer noted that the bees, Melissodes, were exceedingly abundant among cotton grown on certain heavy, red-clay soils. These bees were much less abundant in fields on the lighter, sandy loams in some other localities. Honey bees are especially noticeable near bee trees or domestic hives. The marked abundance of other bees in particular localities likewise probably depends upon the proximity of the plants to their favorite breeding places. Nectar glands are especially abundant on the cotton plant, including the leaves as well as the blossoms. The blossoms are supplied with several sets of nectaries. Cotton blossoms with their abundant supplies of readily accessible pollen and nectar and their open structure exclude few insect visitors. It follows that a consider- able number of species of bees, wasps and other insects are at all times especially common visitors among cotton blossoms. During the time the experiments previously described were in progress nearly 2,000 bees were observed to per- ceive, inspect or enter the blossoms and other material involved. 1,645 of thesevisits were distributed as follows: 1,381 or 83.9 per cent. were by species of Melis- 680 THE AMERICAN NATURALIST [Vor XLV sodes. 130 or 7.8 per cent. were by honey bees. 40 or 2.4 per cent. were by bumble bees. 83 or 5 per cent. were by various unidentified bees. Eleven visits were made by Elis plumipes and one by a large butterfly. The size and yellow color of the petals serve to make cotton blossoms particularly con- spicuous in contrast with their shaded background of dark green foliage. Once visiting insects have entered a cotton field, there is little doubt but that their visual powers almost wholly enable them to discover the blossoms. This is indicated by those experiments where the corolla of certain blossoms has been covered or entirely removed, since following this procedure the re- maining portions of the blossoms were unvisited. The size and general appearance of cotton blossoms do not appear to be of great importance in initiating the process of inspection, since a single petal may receive as many inspections as the control. It is of interest to note in this connection that in experiments 14 and 15 the bees did not discriminate between the white convolvulus and cotton blossoms at least until after closer inspection. When such artificial material as cloth or paper was used, al- though the color more or less resembled cotton petals, the bees were rarely induced to inspect it closely. This discrimination may depend upon perceptible differences in color and texture rather than the presence of repellent odors which the material possessed. Many eminent observers have adduced a great deal of evidence which proves beyond doubt that bees develop keen powers of discernment in their associations with the structural details of different flowers. The actual number of entrances into cotton blossoms is small in comparison with those instances when blossoms have been merely perceived or inspected. The writer’s observations show that of 1,061 inspections of the control blossoms only 129, or 12.1 per cent., were actual entrances. One hundred and twenty of these en- trances were distributed among the several kinds of bees, as follows : 45, or 37.5 per cent., were by Melissodes; No. 539] THE BEHAVIOR OF BEES 681 45, or 37.5 per cent., were by honey bees; 6, or 5.0 per cent., were by Elis plumipes; 16, or 13.3 per cent., were by bumble bees; 8, or 6.6 per cent., were by various small unidentified bees. A single entrance was made by a large butterfly. Although it seems clear that the corolla of cotton blossoms invites the first approach of the bees through their visual sense, it is not so easy to determine the rela- tive importance of the sense of sight and smell involved in their nearer inspections. Just why do so few bees decide to enter? In their careful inspection of a single petal or a suspicious blossom is the sense of sight alone involved? Except for a single Elis plumipes, no bee has ever alighted upon detached cotton petals, although these have served to attract attention quite as often as the control blossoms. It is not unusual, however, for the bees to inspect these structures very intently, almost touching the surface in their movements over them. The bees have just as persistently refused to enter all artificial blossoms or blossoms mutilated by removing a part or all of the petals. In experiment 6 a bee was completely deceived by the unreal structure at (c) made to simulate an expanding bud by the use of actual cot- ton petals and portions of a cotton leaf. In this instance the bee inspected and finally squeezed itself down be- tween the petals. Unless the fresh petals themselves possess a characteristic odor, odors such as might emanate from a normal blossom were entirely lacking and, therefore, could not have induced the bee to enter. Plateau concludes that visual conspicuousness by means of bright colors is of no advantage whatever to blossoms So far as insect visitors are concerned. He claims that if in nature all blossoms were green like the surround- ing foliage, they would be just as readily discovered by bees and other insects in virtue of their odor. The writer’s experiments in the field indicate that conspicu- Ousness in virtue of their position and yellow coloration 1S a very important factor in leading bees and other in- Sects to perceive cotton blossoms. 682 THE AMERICAN NATURALIST [Von. XLV It does not appear that the addition of small quanti- ties of honey either upon the petals themselves or at the base of the flower within appreciably increased the in- spections or entrances, although if a bee chanced to dis- cover this honey, its fondness for it was evinced by its strong reluctance to leave. It is probable that the inspections are largely of a visual nature, though these may be supplemented by certain odors when the blossoms are more closely ex- amined. Many noted observers, especially Miiller, have adduced abundant evidence to prove that the visual power of bees becomes very critical in their behavior toward minute differences of floral structure. The bee Melissodes bimaculata, which is probably by far the commonest of this genus in certain cotton fields at Thompson’s Mills, behaves somewhat differently from other bees in its inspections. Its flight is swift and irregular, and its entrance into a blossom is usually preceded by a more careful examination than that re- sorted to by bumble bees, the common honey bee, or the wasp Elis plumipes. It is the usual procedure for the last to fly straight into a blossom or almost drop into it from above, apparently without troubling itself about any preliminary examination. The bumble bees too are less fastidious in their closer inspection. Many instances are recorded which illustrate the habit of bees to profit by previous successful or unsuccessful experiences. A sort of memory by association is de- veloped so that older, more experienced bees often ap- pear to work among blossoms to much better advantage than younger bees. As an illustration of the influence of previous association upon subsequent behavior, the writer cites the following interesting instance which has come under his observation at Thompson’s Mills, North Georgia. It has been mentioned that the common honey bee sooner or later discovers the outer involucral nec- taries of cotton blossoms and visits them very con- stantly, seemingly in preference to the inner floral No. 539] THE BEHAVIOR OF BEES 683 nectaries. These particular nectaries, although present in our common American cottons, are never found on the Asiatic cottons, Hawasaki, ete. In the writer’s variety tests these foreign cottons have been grown side by side and sometimes intermingled in the rows with the American cottons. The honey bees, in passing from blossom to blossom, visiting each time the outer involu- eral nectaries occasionally met the Asiatic variety. The previous association with the American cottons and their outer involucral nectaries led these bees to visit without success similar structures of the unfamiliar Asiatic variety. The bees quickly recognized their error after alighting and left the blossoms. Is this procedure other than the working of an associative memory? The writer is of the opinion that the honey bees do not dis- cover these extra-floral nectaries until after more or less association with cotton blossoms each summer. This habit of the honey bee appears to become more noticeable later in the season. During the season of 1908 it appeared to be very general. It is a habit which seems to be almost wholly confined to honey bees. These visits of the bees to the outer basal portion of the Asiatic cotton blossoms indicate that the visual powers alone were employed throughout the process. Although the bees first discovered the blossoms by their conspicuous petals, it is evident that they were led to search for outer involucral nectaries on the Asiatic cot- ton blossoms solely by their familiarity with the general form and structure of cotton blossoms. As an illustration of associative memory this behavior of the honey bees is exactly similar to the behavior of certain bees in experiments conducted by Pérez‘, who used scarlet pelargoniums which are not visited by bees, Since those flowers possess no nectar. He added honey to certain flowers which were then visited by bees, and says: * Pérez, J., ‘Notes Zoologiques’’ (Actes de la Société Linnéenne de bition "Vol. XLVII, série V, tome VII, pp. 250-251, 1894). 684 THE AMERICAN NATURALIST [Vou. XLV La couleur écarlate s’était si bien associée dans leur souvenir à idée du miel, qu’elles se passaient à la fin sur des fleurs de cette couleur n’en ayant pas recu, et ne les quittaient qu’ apres s’étre assurées, par un examen scrupuleux et persistant, qu’elles n’avaient rien à y recueillir. A translation of his own words follows: ‘‘Scarlet color and honey had become so closely associated in their minds that they finally alighted upon flowers of the same color which had received none, and would not leave until they had assured themselves by a scrupulous and persistent examination that these flowers had nothing to offer them.’’ Plateau gets precisely the same results when he says: Lorsque l'insecte avait ainsi absorbé le liquide d’un certain nombre defleurs miellees, il lui arrivait de se diriger vers les Pelargoniums non nunis de miel.” . ‘t After the insects had gathered honey from a number of flowers to which it had been added, they were then led to visit Pelargonium blossoms which had not re- ceived it.” These observations are hardly in agree- ment with the rather radical conclusions of Bethe® that bees are devoid of sense impressions, and are incapable of profiting by previous experiences, that their activities are purely reflex, mechanical. Forel, Wasman, Buttel- Reepen, Huber and others have shown, nevertheless, that bees do profit by previous experiences and form habits under certain conditions. Lovell? has shown that once bees have been accustomed to visiting a certain color, they tend to return to it regularly until it is to their advantage to change. Once the bees have entered the cotton fields, it is quite obvious that they are led to discover the blossoms by the conspicuous corolla. It would be interesting to learn just how they find the fields themselves. Although a single cotton blossom does not : * Bulletin de 1’ Academie royale de Belgique, 3e série, 33, January, 1897. - © Bethe Albrecht, ‘‘Durfen wir Ameisen und Beinen parohia Quali- täten zuschrieben?’’ in Arch. f. d. ges. Physiologie, Bd. 70, 1898. "Lovell, John H., ‘‘The Color Sense of the Honey Bee: Can Bees Dis- tinguish Colors?’’ AMER. NAT., Vol. XLIV, No. 527, November, 1910. No. 539] THE BEHAVIOR OF BEES 685 seem especially odoriferous, it is not improbable that a field of well-developed cotton plants may readily adver- tise its location to the olfactory sense of bees by odorif- erous clouds, so to speak, which are wafted away with every air-movement. During a hot, sunny afternoon the combined odors volatilizing from the great numbers of foliage and floral nectaries, the pollen, etc., must be very considerable. Especially during clear sunny days fol- lowing periods of cloudy or rainy weather bees become unusually active and numerous. Many of these visitors have no doubt learned the location of the fields by previ- ous association. THE DISTRIBUTION OF PURE LINE MEANS DR. J. ARTHUR HARRIS CARNEGIE INSTITUTION OF WASHINGTON SEvERAL times recently we have been told that the means of a character in a series of pure lines form a ‘“Quetelet’s Curve.’ Some of those responsible for this assertion seem to attribute a particular virtue to ‘‘ Que- telet’s Law,’’ and to feel that the statement that the means of a series of pure lines form a chance curve fur- nishes uncontrovertible evidence for the genotype theory of heredity. The questions which interest the biologist are, first, whether the statement is true in the sense that it is made on a sufficient body of actual observations, and second, what is the general biological significance to be attached to it, if true. But among these biologists the interpretation of the facts has apparently preceded the demonstration of the existence of the facts themselves. Now while it is not at all unlikely that the means of genotypes—if such entities in Johannsen’s sense of the term do exist in nature— form a chance curve, it by no means follows that con- versely a series of averages which can be arranged in a symmetrical variation polygon proves or even suggests the existence of differentiated pure lines or biotypes. Yet just such differences in means are being accepted and cited without criticism as valid evidence in support of Johannsen’s sweeping generalizations. A case in point is a paper by Roemer? on pure lines in peas. It is with regret that one criticizes Roemer’s *Compare, for example, in this connection: Nilsson-Ehle, Bot. Not., 1907, pp. 113-140; Lang, Zeitschr, f. Ind. Abst.- u. Vererbungsl., Vol. 4, pp. 15- 16, 1910; Spillman, Amer. NAT., Vol. 44, p. 761, 1910; Pearl, AMER. NAT., Vol. 45, P. 423, 1911. a , T., ‘“‘Variabilitätsstudien,’? Arch. f. Rassen- u. Gesellsch.- Biologie, Vol. 7, pp. 397-469, 1910. 686 No. 539] DISTRIBUTION OF PURE LINE MEANS 687 paper. It is an exceedingly laborious Arbeit and appa- rently done with scrupulous care. One who himself has experienced the labor of calculating a few tables of con- stants has sympathy for a worker who has industriously filled pages with them. But the tenability of the geno- type theory is one of the most pressing of current evolu- tionary problems, and all available evidence must be scrutinized. Roemer’s data are chosen for two very excellent reasons, the first of which is that of all of the men who have discussed the disposition of the means of pure lines in a ‘‘ Quetelet’s Curve,’’ he is, so far as I am aware, the only one who has put on record sufficient data for a critical test of his conclusions. If without over- trying the case, as the lawyers have it, we can give the second reason, it is that Roemer’s data and conclusions have been accepted as perfectly valid by genotype specialists. One of them, for example, says: The work is essentially a confirmation, with another plant, of Johannsen’s epoch-making investigations on beans, though it lacks any extensive studies on the effect of selection within the pure line. The essential objective point of Roemer’s research is rather to determine the biometrie characteristics of pure lines as such in relation to the general population. Among the more important general results are the following: 1. The different biotypes in a population arrange themselves in fre- quency distributions in accord with Quetelet’s Law. 2. No relation was found to exist between the variability of the biotypes (i. e., variation within the general population) and variation within the pure lines. Our problem is twofold. First, we have to determine whether Roemer is really justified in regarding his lines as differentiated. Second, we have to inquire concerning the critical value of his data as evidence in support of the genotype theory of heredity. Incidently we shall make the first of these problems serve as an illustration of the use of a coefficient of individual prepotency recently proposed in these pages. Harris, J. Arthur, ‘‘A Coefficient of Individual Prepotency for Stu- dents of Heredity,’’ Amer. Nart., Vol. 45, pp. 471-478. 1911. 688 THE AMERICAN NATURALIST [Vou. XLV Il. THE PROBLEM or DIFFERENTIATION IN RoEMER’S Pure LINES The method of Roemer’s study was very simple. In 1908 a population of pea plants was grown from a sample of ordinary seed. In 1909 the offspring of each of a num- ber of these plants was studied separately, and the means of several characters calculated. By a comparison of selected pairs of these means Roemer concludes that the several lines differ from each other, and by a seriation of all the line means he obtains the Quetelet’s curve. Such evidence as this can not be accepted. Every mean calculated on a sample of individuals is more or less untrustworthy as a measure of the character in individ- uals in general, because of the errors of random sam- pling, and in attaching significance to a series of averages this fact must be fully taken into account. It can not adequately be allowed for by a comparison of selected cases with their probable errors. © First Test. A Comparison of the Variability within the ‘“‘Pure Line” with that of the “Population” One of the tests of the presence of differentiated ‘“biotypes,’’ ‘‘genotypes’’ or ‘‘pure lines’’ within a ‘‘population”’ is the comparison of the intra-line with the population variability. If both be the same there is no justification in the assumption that the population is composed of a number of differentiated pure lines.* If the variability of the population is greater than that of the individual lines it may (or may not) comprise a series of ‘‘genotypes.”’ The reason for this is obvious. The standard devia- tion within the pure line, ø, describes only the differences occurring among the individuals of the group, while ž, the standard deviation of the group, includes also the amounts by which the several lines are differentiated. * This is, of course, under condition that the individuals of the several pure lines are not reared under conditions which tend to increase artificially their variability beyond that of the population. No. 539] DISTRIBUTION OF PURE LINE MEANS 689 Roemer does not give us the population standard deviations for the several characters in 1909 but only the averages, Mi, Ms, Ms, °**, Ms, and the standard deviations Tis Ta, C3, t, os We may approximate the desired con- stants very closely indeed® by the following method. Let there be s samples or pure lines of n4, na, na, +t, Ns individuals each, with means my My, Mz, ***, Ma, e standard deviations o,, o,, ¢ -, os. These form the population S(n) =N, for hick the physical constants x and M are desired. The mean is clearly M — S (nm)/S (n). In calculating the S.D. we may take the first two rough moments, v,’, v⁄, about any point we one and adjust by the familiar formula o?=p,—v,’—v,’. If the moments be taken about 0° v,’— M, and it is at once clear that for the population BF A] (Kum)? TEN a) gogi when S indicates a summation for all groups or lines.’ The population constants have been calculated by these formulæ for all the characters of Roemer’s two large series. He has given population constants, M and 3, for the 1908 series, the parents of the 1909 plants. The two are conveniently laid side by side for com- parison in Table I.8 The data in hand hardly seem to justify detailed comparison with reference to probable * There is no approximation in the formula. The accuracy in practise depends solely upon the trustworthiness of the original m’s and a’s, and upon the number of decimal places retained in the arithmetical routine. * For several advantages in doing this see AMER. NAT., Vol. 44, pp. 693- 699. 1910. ' The application of the formula to Roemer’s data is of course exceed- ingly laborious, involving as it does the determination and summation by pairs of 3,108 squares, and the summation of the products of their totals by the frequencies upon which they are based. The publication of a little tabulated data would have reduced many days’ labor necessary for a critical test of his results to a few hou *The constants for 1908 are tuken from Roemer’s Table I. Those for 1909 are calculated by the formule given above 690 THE AMERICAN NATURALIST [Von. XLV errors. It will be noted at once that for all the char- acters the mean is higher in 1909°—indeed for some characters in the ‘‘Kapital Erbse” it is almost double that found in 1908! With one exception the standard TABLE I COMPARISON OF PHYSICAL CONSTANTS FOR 1908 AND 1909 POPULATIONS Averages Standard Deviations paS Vari- Character and Com- |———_— parison Gelbe Svalöfs Gelbe fanran Gelbe Svalöfs Viktoria- | Kapital- | Viktoria- p reent Viktoria- | Kapital- Erbse se Erbse T Erbse bse Weight of Plant 1908 Population 13.09 7.99 4.250 2.815 32.47 35.23 1909 ae 20.82 20.39 6.568 7.127 31.54 34.95 1909/ ee —— 1.59 2.55 1.545 2.531 97 .99 Length of 1908 Popa 114.96 | 78.96 12.985 | 12.575 11.30 | 15.95 1909 = 136.81 | 158.42 16.331 | 20.163 11.93 | 12.73 1909/1908 Ratio 1.19 2.01 1.257 1.603 1.06 .80 Thickness of Stem 1908 Population 24.03 | 20.50 2.766 2.081 11.50 | 10.15 1909 di 24.05 2.390 2.933 9.48 | 12.19 1909/1908 Ratio 1.05 3:47 .864 1.409 .82 1.20 Number of Pods _ 1908 Population 4.59 5.62 1.364 1.805 20.72 | 82.12 1909 6.48 11.54 1.987 4,263 30.66 | 36.95 1909/1908 Ratio 1.41 2.05 1.456 2.361 1.48 1.15 Weight of Pods 1908 Population 9.71 5.76 3.192 2.184 | 32.80 | 37.95 1909 s 11.28 | 10.26 4.290 4.407 | 38.03 | 42.98 Pad /1908 ie 1.16 1.78 1.343 2.017 1.16 1.13 um 1908 Population 19.64 | 24.02 6.267 8.418 | 31.91 | 35.05 1909 26.60 | 45.11 9.162 | 18.633 | 34.44 | 41.31 ese boy Ratio 1.35 1.88 1.461 2.213 1.08 1.18 Weight of Seeds 1908 Population 7.63 4.56 2.569 1.883 | 33.67 | 40.21 1909 8.56 7.26 3.511 3.376 40.99 | 46.49 1909/1908 Ratio 1.12 1.59 1.366 1.841 1.22 1.16 deviations in 1909 are higher than those in 1908. Mean and standard deviation are generally closely correlated, and this doubtless accounts for the greater variability of the 1909 series. Possibly, however, the 1909 plants °” Roemer states that conditions for growth in 1909 were superior to those in 1908, No. 539] DISTRIBUTION OF PURE LINE MEANS 691 were grown under conditions more heterogeneous than those to which the 1908 plants were exposed. In the second case, the S.D. might be directly raised, 7. e., heterogeneity in the crop may be merely a reflection of heterogeneity in the substratum. ere is no way of determining whether Roemer’s cultural conditions were more heterogeneous in 1909 than in 1908, but it must be noted that in ten of the four- teen cases the coefficient of variation is higher in 1909. Two ratios are to be examined, Mean Pure Line Variability — Parental Population Variability’ Mean Pure Line Variability General Population Variability’ Consider first the ratio of the mean pure line to the pa- rental population variability. If the offspring of the individual parents are differentiated we should expect to find the mean variability of the pure lines less than that of the parent population, providing, of course, that innate tendencies are not obscured by environmental factors. Table II° gives the necessary data. Now the remarkable thing about these standard devia- tions is that in the most cases the variability within the individual ‘‘pure lines” in 1909 is greater than that of a mixture of all the pure lines in 1908. The excess is very striking in several cases. Of the fourteen com- parisons, thirteen show a higher variability within the pure line than in the population. For the ‘‘Viktoria’’ ° For 1908 the population = and C.V. are from Roemer’s Table I. The 1909 population = and C.V. have been calculated by the formule given above. The mean pure line standard deviations have been taken from Roemer’s Tables II-III. None of the constants have been rechecked, since the original data are not available. The mean value of 12.13 for thickness of stem in Table II is obviously a printer’s slip for 2.13. The mean pure line coefficients of variation are from Roemer’s Table X. These were not calculated by dividing the sum of the coefficients of variation of the indi- vidual pure lines by the number of lines, but by dividing the mean standard deviation of the pure lines by the mean average of the pure lines. R 692 THE AMERICAN NATURALIST [Vou. XLV the average M.P.L./Parental ratio is 1.327 while for the ‘*Kapital’’ it is 1.996. The “pure line’’ variability is thus from 30 to 100 per cent. in excess of that of the population. TABLE II COMPARISONS OF MEAN PURE LINE AND POPULATION VARIABILITY 1908-1909. Stand- 1908-15 1909. O 1909-1909. Stand- ard Deviations nt of Variation ard Deviations Character and Comparison | Gelbe Svalifs Gelbe Svalifs Gelbe Svalifs Viktoria- Kapital- Viktoria- | Kapital- | Viktoria-| Kapital- Erbse Er bse Erbse Erbse Erbs Er Weight of Stem POISON sci cic cess 4.250 2.815 32.47 35.23 6.568 TIZ Mean Pure Line.......... 5 .96 6.44 28.53 | 31.40 | 5.96 6.44 L./Pop. Ratio ...... 1.402 | 2.288 .88 89 | .907 | .904 Langit of | > me Population. senii 12.985 | 12.575 E30 15.95 | 16.331 | 20.163 Mean Pubs Line ES TS 13.80 15.98 10.05 10.05 | 13.80 15.98 M.P.L./Pop. Ratio. .....| 1.063 | 1.271 S01 .63| 5i .792 Thickness of Stem laiión sss eiin] TOG | 2081 1: 11.50: 10,16: +. 2.9900 | Bess Mean Pure Line. Eear O18 2.04 8.45 8.46 | 2.18 2.04 L./Pop. pene past .770 .980 13 .83 .891 696 Webes of Pod OU UIALION (05) 6 sscsy deus: 1.364 | 1.805 | 2972 | $2.12 | 1.987 | 4.263 Mean Pure Line..........| 1.80 3.92 27.19 | 34.00 | 1.80 3.92 g Ratio.. 19200) 2173 .91 1.06 .906 .920 Weight of Pod PODOIION, aaa 3.192 | 2.184 | 32.80 | 37.95 | 4.290 | 4.407 ú Pure Lino... 3.80 3.95 33.84 38.50 3.80 3.95 M.P.L. / Pop. Ratio ..... 1.190 | 1809] 1.03 | 101| 385] .896 Number of Seeds Population .. ......-| 6.267 | 8.418 | 31.91 | 35.05 | 9.163 | 18.633 Mean Pure Line. . is 801 | 16.60 33.30 | 36.80 | 8.01 | 16.60 To Pop. Ratio. ..... ~ 1.278 1.972 1.04 1.05 .874 .891 Weight o : i prie bg Bon TE OE EA 2.569 1,833 33.67 40.21 3.511 3.376 Mean Pure Line........... 8.17 3.04 37.05 | 41.90 | 3.17 3. M.P.L./Po p. Ratio..... 1.234 | 1.658 1.10 1.04 .903 .900 The explanation of this anomalous result is first to be sought in the higher means (with the associated higher variability) in the 1909 plants. Basing the com- parison on the coefficients of variation in order to elimi- nate, in so far as possible, the influence of the means, we note that seven of the ratios are greater and seven No. 539] DISTRIBUTION OF PURE LINE MEANS 693 are less than unity, while for all the mean is .936. On an average, therefore, the pure lines have 93.6 per cent. as much variability as the population. The second comparison, that between the variability of the individual pure lines and the population which they form, can be made on the basis of the standard deviations alone since the means are the same. This comparison (the last two columns of Table IT) shows that in both series and for every character the variabil- ity written, the line is less than that for the population. The lowest ratio is .70, the highest is .92 and the mean is .858. This test indicates that they are differentiated. This is, of course, the conclusion which Roemer drew from his selected individual comparisons. Second Test. The Deviation of the Pure Line Means from the Population Mean For characters measurable on a quantitative scale the test for the deviation of the offspring of an individual from its population is given by o° In n(M— m? er) S 2 (m — M) + .67449 Ny rs (1 -7 Where m and M, o and 3, n and N are the means, stand- ard deviations, and numbers of individuals for the family and the population, respectively." = For reasons which will be apparent to the reader later, the data which are given us do not justify calcu- lations to a high degree of refinement.'? We therefore approximate in every point possible. The expressions 1 2n n(M— my ~n’? NN—n) H AMER. NAT., Vol. 45, pp. 471-478. 1911. “ Furthermore, I have serious misgivings that Roemer’s lines comprise 80 few individuals each that the coefficient suggested must be used with caution. It is not needful to consider the point in greater detail here. 694 THE AMERICAN NATURALIST [Vou. XLV may be disregarded, since with such relatively low values of n as those of Roemer’s data the first is practically unity, while the second is generally insignificant and may always be neglected, having a maximum value of circa .010, and in the majority of the cases falling far towards zero. The values of 3?/N are given in Table III. In practically every case the inclusion of 3*/N in TABLE III VALUES OF 2?/N Gelbe Viktoria Svalofs Kapital Char: Erbse * Erbse 1 GION OF PAUL Fics. civ sak ae ee 0233 -0292 2 ngth a oe ee ck 1442 .9204 o. Thokoa OF SOM -o:a 0031 0049 4. Number Of pods i:i csi dans 0021 0105 5. Weight : POA soo ase es ee eS 0100 0112 6 umbe BOOdS sch ae Si a aes 0454 1997 7. Weight of WOE. si Seg ca cs Experiment Department, Division of Horticulture, “Massachusetts Agricultural College. Fic. 3. Type Card. alphabet and in order to secure opportunity for recording a large number of characters they are more or less arbitrarily grouped, each group having if needed the whole twenty-six letters avail- able. Each group is assigned a column in the card. In order to illustrate this a portion of the type designations we have used with beans is here given which with Fig. 3 will clearly show the application of the system. The same blank may of course be used with plants other than beans. Plant type: Fruit types: a= stems green, a= pod flat, b =— stems with purple lines, b= pod oval, c = stems purple tinged, e= pod round, h= pole form, s = pod straight, i= runner form, u= pod curved. j= bush form. Pod types: Flower types: a= light green, a = white b =— medium green, b = light Tok c= dark green, 704 THE AMERICAN NATURALIST c = pink. Leaf types: a = light green, b — medium green, m= no fiber, p = much fiber, s = stringless, v = very stringy. [Vou. XLV e = medium yellow, Seed types: h = broad, a= white, t = medium, ba buff, t c= yellow, d= light red, g = black, m = medium eye, o = dark mottling. The notes shown in Fig. 3 give the following information : The cross in number 272 which is shown by a separate list and by a neighboring guide card to be Prolific Black Wax X Mo- hawk. This cross was one of those made in 1910 (Series 1910) and these plants were grown in 1911. (Crop of 1911.) The color of the card shows at once the generation. Plant 7-34 had green stems and was a bush bean. The leaf was medium green, the flower pink, the pod flat, straight, green, tough and stringy. The seed was black and buff mottled, the black predominating, as shown by o indicating dark mottling. All these observations are very quickly recorded, once they are determined and the record is brief, convenient, definite and easily summarized. It is of course necessary to record clearly and positively just what is signified by each letter. Most characters appear in some established variety and may be fixed by referring to them. Full description or preserved material may supplement such refer- ences when necessary. Additional blank columns are provided to allow for more extended observations or they may be used 1m connection with the space to the right for brief special notes on the individual plant. If more extended notes regarding any plant or group of plants are desired, a description card or a blank card of the appropriate color may be inserted at any point for their accommodation. Photographs or drawings might also be easily filed if desirable. We have used these cards for the past three years for record- ing observations on many thousands of plants, not only beans but several other kinds as well, and the system has given good satisfaction. If it contains any features useful to other observ- ers the purpose of this article will be fulfilled. wae ise J. K. SHAW. THE MASSACHUSETTS AGRICULTURAL COLLEGE The American Journal of Science Established by Benjamin Silliman in 1818 The Leading Scientific Journal in the United States Devoted to the Physical and Natural Sciences, with special reference to Physics, and Chemistry on the one hand, and to Geology and Mineralogy on the other. Editor: EDWARD S.DA Associate Editors: Professor GEORGE L. GOODALE, JOHN TROWBRI IDOR, W. G. FARLOW and ve x DAVIS of Cambr ridge ; Professore A. E. VERRILL, HENRY S. WILLIAMS and L. V. PIRSSON, of N IMES, n; Profes F. BARKER. na Philrdelphia ; arera hea Ha HS. A etnies ee Se S. DILLER, o of Washington Two volumes annually, in monthly numbers of about 80 pages ah This Journal ended its first series of 50 volumes as a quarterly in 1845; its second series of 50 volumes as a two-monthly in 1870; its third series as a monthly ended Dec- ember, 1895. A Fourth Series commenced in 1896. Subscription price, $6 per year or 50 cents a number, postage prepaid in the United States - $6.25 to Canada ; $6,40 to Countries in the Postal Union. Back numbers at reduced prices. taTen-Volume Indexes, Vols. I-X XI-XX, fourth series, price one dollar. Address The American Journal of Science New Haven, Conn. SECOND EDITION, NOVEMBER, 1910 AMERICAN MEN OF SCIENCE A BIOGRAPHICAL DIRECTORY EDITED BY J. McKEEN CATTELL A Bio wographical directory re quires revision if it isto maintain its ulness. Nearly a third of the names in the = cer ites es fe new, and the sketches which a penred $ in the first Saa jae in nearly er k rds revised. The ú ru ed to prepare the revision Nes: bis as great as that given to the first edition. There has been no change in the general rekle of the work. Greater irii has been observed in confining its scope to the natural and “tact sciences ; f j itt omuted. Efforts have reply ae book as complete and accurate as ible. There are of cou omissions, if only Sere Sams Se GEI nol by Xe to repeated requests Jor the information needed. “The hound k leading men of scien ve been again selected rae ot) that were used before, and stars kave been added tothe subjects of research in the fi of 269 new m the “ae places on the list. “The editor’ a in selecting this group of scientific men has been to make a study iy kd aa ions on whic ich scientific research dependa a an Ae may be to conditions There +t the two statistical studies that have bee n made.—From the Preface to the Second Edition. Bi ‘eect econd ene iy the Directory extends to more than 600 pages and contains more than 5500 sketches pa nted on all rag paper and bound in buckram with leather label. ia ere a the work has been in size by fronts: Dg ra cent., it is sold at the same price as the first editi Price: re Dollars, net, ostage paid HE sc SCIENCE PR PRESS GARRISON, N. Y. LANCASTER, PA. SUB-STATION 84, NEW YORK CITY. The American Naturalist 1867, Devoted to the Advancement of the eRe 5 Sciences with Special Reference to the Factors of Organic Evolution and Heredity A Monthly Journal, established in CONTENTS OF THE MAY NUMBER The Inheritance of Polymorphism and ie in Colias Philodice. Professor JoHN H. GEROUL Nucleus and ee in Heredity. Aua MICHAEL F, G A Comparative patie a the oe of the ees ree Organs of Certain beat: oe ee Tide. roltanept? and C G. Cra anore Sco and Discussion: rary Kaikai Paper fs cary cape al Selection 3 in the English Sparro r Notes and tie ture: Biometrics, Yule’s ee colby to the Theory of Statistics. Dr. RAYMOND CONTENTS OF JUNE NUMBE on ae ee in the Domestic oi Dr The Biometric Proof vil the Pure Line Theory. Dr. J. The rahe es ai External Conditions fa essor T: . MONTGOMERY. Shorter Articles and aeg Bieg Ontogeny of & PERE HOBART CLA Notes and Literature: eai Gonia to & Knowledge a the Extinct Amphibia, Dr. Roy L. MOODIE. CONTENTS OF JULY NUMBER marewa Firma and Their Significance, Pro- Further Observations on shs Pose of the Sauropodous Dinosa ered paces P. Hay. A r Articl d Correspond nce: Computati the Coeficient igi Correlation. iege on r H S = ENNINGS Dra V buao atrachoseps atten sare Eich. and Literature: Some Recent Studi es on Vari ation sea ER in Agricultural Plants: Fa Some Recent Books on Fossil Plants Dovelas HOUGHTON CAMPBELL. CONTENTS OF AUGUST NUMBER ka A Comparative Microscopic Study of the Melanin os ntent of Pigmented Skim with special refer- e to the Question of oori kr ritance Amon; Mulattos. Professor H. E. a A Coeficient of ne 1 Prepotency for Studen of Heredity. Dr. J. A So Š The Tee of the pie Professor F. Jean Marchant, an Eighteenth Century Mutationist. Dr. HENRI Hus. Notes and Literature : Notes on Heredity. Dr. W.J- SPILLMAN. CONTENTS OF SEPTEMBER NUMBER eritance of the “Eye” in Vigna. Dr. W. J Heredity of Hair F Form among the Filipinos, Dr. ROBERT BENNETT BEAN. The ang ae pg pe of ne Tomes Indian Archipelago. peA — = aie: Biometrie A Concept. Dr. Rava oo he n a sia at = i qe Nu Correla- ani nungency Tables when the Number c Combination rati 3 Acquired Characters defined. C.L. Teona. a of a ee Sige o n of An € Jseful Ev lution Miiogriehe. V de 5 = x CONTENTS OF OCTOBER gees = E ia in Œnothera. r. R. R tal Observations eine the Behavior of clk Bees in Their Visits to Cotton Blossoms. H. A. ALLARD, III. Reproduction in the Brown Rat. NewToNn MILLER IV. — Articles and Discussion : pa Dialectics d Other Digressions. Dk. J. e v. Topp and Literature: Darwinism agja Hum Life. V.L.E ——— a e Number 35 Cents NATURALIST will be sent to new subscribers for four months for One Dollar THE Z CENCE PRESS n 84: NEW YORK Yearly Subscription, $4-00 VOL. XLV, NO. 540° DECEMBER, 1911 THE AMERICAN NATURALISI A MONTHLY JOURNAL Devoted to the Advancement of the Biological Sciences with Special Reference to the Factors of Evolution CONTENTS ; Page I. The Inheritance of Color in Short-horn Cattle. H. H. LAUGHLIN - - -705 - = = -743 IT. Studies on Melanin. DR. Ross AIKEN GORTNER IN. Shorter Articles and Discussion: A Noteon Certain TOR SEESE Dr. RAYMOND PEARL, LOTTIE E. MCPHETERS ~ z IV. Notes and Literature: The Doctrine of Evolution. V. Index to Volume XLV - - =- + += - Noke o| eo l FEL “ - 763 THE SCIENCE PRESS LANCASTER, PA. GARRISON, N. Y. NEW YORK: SUB-STATION 84 The American Naturalist. . intended for perrea and books, MSS sent to the Editor of THE AME etc., intended for review should be RICAN NATURALIST, Garrison-on-Hudson, New York. rticles containing research work bearing on the tion are especialiy welcome, and will be given preference i prcblems of organic evolu- n. n publicatie drea reprints of contributions are supplied to authors free of charge, ost. One hun Further reprints will be supplied at c dollars a year. e twenty-five cents additional. thirty-five cents. The advertising rates are Four sent to the publishers. The oreign postage is fifty cents and The charge for single copies is Dollars for a page THE SCIENCE PRESS NEW YORK: Sub-Station 84 Lancaster, Pa. Entered as second- l Garrison, N. Y. tter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act ot Congress of March 3, 1879. r E A Naas T E came ae THE BULLETIN—For bargains in Ethnolograph- ical and Pre-historic Specimens. Books on Natural History, Science, Travel, Voyages, ete. See THE BULLETIN post free for 3 cent stamp. 4 Duke St., Adelphi—London—England TENTH EDITION. THE MICROSCOPE; an introduction to ey Methods by SIMON HENRY GAGE, of Cornell University. Over 350 large octavo S and above 250 figures in this new and fall re ition. Price $2.00, postpaid COMSTOCK PUBLISHING CO., Ithaca, N. Y. BIRDS’ EGGS W. F. H. ROSENBERG, 57 Haverstock Hill, London, N. W., England begs to announce the publication of a new Price List (No. 15) of Birds’ Eggs, con- taining over 900 species from all parts of the world. This Catalogue is systematic- ally arranged, with authors’ names, indica- tions of localities, and an index to families. It will be sent post free on application, as will the following lists: No. 11, Birds’ Skins, (5,000 species) ; No. 12, Lepidoptera, (5,000 species) ; No. 13, Coleoptera ; No. 14, Mammals. In course of preparation : New Price List of Reptiles, Amphibians and Fishes. Largest Stock in the world of imens Back or Current Numbers of any American or Foreign technical or trade journal and magazine fur- nished on short notice at moderate rates; all kinds of Government and State Reports in stock. Clippings on special subjects furnished promptly. Large stock of American Naturalist, Science and Popular Science on hand. Magazines, Books and Papers of all kinds bought Special Subscription Price List on request. Information concerning any periodical furnished free of charge. A. W. CASTELLANOS 259 Armstrong Ave., Jersey City, N.J., U.S.A. THE AMERICAN NATURALIST VoL. XLV December, 1911 No. 540 THE INHERITANCE OF COLOR IN SHORT- HORN CATTLE A Srupy 1x Somatic BLENDS ACCOMPANYING GAMETIC SEGREGATION AND InTRA-zyYGOTIC INHIBITION AND REACTION H. H. LAUGHLIN CARNEGIE STATION FOR EXPERIMENTAL EVOLUTION, CoLD Spring HARBOR, N. Y. THE men who made the breed of Shorthorn cattle were in many respects the most skillful breeders of domestic animals. They had many rich and varied inheritance lines to draw upon, and in developing the breed they had high ideals of real excellence, largely ignoring the super- ficial quality of color. A consequence of this neglect of coloris that the great breed of Shorthorn cattle is mongrel in this respect, ranging as follows: Solid red—varying from the richest dark to a light yellowish; spotted red- and-white; red-roan; and white—besides many inter- grades and combinations of these shades and patterns. It is the prevailing experience among Shorthorn breeders that the color of the calf can not be accurately predicted before its birth. Reflecting this experience, Mr. B. O. Cowan, of the American Shorthorn Breeders’ Associa- tion, writes: Owing to the fact that Shorthorns are of mixed colors, you can not with absolute certainty, before birth, tell what will be the color of the 705 Pleasant Valley Bud and Calf. No. 540] INHERITANCE OF COLOR IN CATTLE 707 Courtesy of Thos. Stanton, Wheaton, Ill, Cow—CINDERELLA. Red. CALF—CINDERELLA 2D. Roan. (White star on forehead.) Dam—Clara. Red. Dam—Cinderella. Red. Dam’s Dam—Carrie. Red. Dam’s Dam—Clara. ed. Prince Gloster. Red. Dam ’s Sire—Scottish Minstrel. sag Roan. Sire—Scottish Minstrel. Dark Roan. Sire—Prince Imperial. Light Roa i —Imp. Mistletoe 20th. Roan. Sire ’s Dam—Imp. Helen 21st. Light Roan, Sire’s Sire—Imp. Collynie Mint. Roan. Sire’s Sire—Prince. Red. Courtesy of Geo. M. Rommel, Bureau of Animal Industry, Washington, D. C. Cow—MAvTALINI 17TH. Roan. (A champion Ta T A.) cow.) i 8th. Sire’s Sire—Bapton Conqueror. Roan. CALF. Roan. Dam—Mautalini 17th. Roan. Sir q n. Sire—True Blue. Red and White. Sire’s Dam—Twin Princess 10th. Red and White. Sire’s Sire—Bapton Champion. Roan. Courtesy of F. W. Harding, Waukesha, Wis. Cow—PLEAsANT VALLEY Bup. Roan. Dam—Rosebud llth. Red. Dam’s D CALF. Red. Dam—Pleasant Valley Bud. Roan Sire’s Dam— Valley Gem. Red. Sire’s Sire—Mildred’s Royal. Roan. Fie. 1. 708 THE AMERICAN NATURALIST [Vou. XLV calves. There are a great many instances of red cows bred to white bulls producing red calves, in some instances white calves, and in other instances roans. In some herds in the United States where the breeders have used nothing but red for thirty or forty years it is very rare that they have any calves excepting reds; but even among these occasionally a calf is dropped that is either a roan or a red with some white marks —this is the influence of the blood of ancestors many generations back. Mr. Spangler, of Sullivan County, Mo., reports the following to the Breeders’ Gazette of February 17, 1909: My bull is white, but his sire and dam are both roan. The results are as follows: Since September first there have been fifty-five calves dropped to his service, of these forty-one are roan, nine red, four red- and-white, and one white. Twenty-six are bulls and twenty-nine heif- ers. The cow that dropped the white calf is herself a roan’. . . the rest of the cows are red. Robert Bruce, of County Dublin, Ireland, tabulated the color matings and color progeny of Shorthorns bred by Amos Cruickshank! at Sittyton. This he reports to the Breeders’ Gazette of November 25, 1908, as follows: TABLE I COLOR OF OFFSPRING Color of Matings Red Rd. & Wh. Roan White Total Red. mated with tod ni.i 133 12 3 1 180 Red mated with red and white ....... 31 11 6 0 48 Red mated with roan ... i. Praun] 278 25. 265 0 568 Red mated with white ............... 0 41 4 46 Red and white mated with red and white 0 2 0 1 3 Red and white mated with roan ...... 22 40 2 86 Red and white mated with white ..... 0 1 1 1 3 Roan mated with roan .............. 56 10 183 60 309 Roan mated with white .............. 0 0 12 12 24 White mated with white ............. 0 0 1 2 S21 as oss & 120 Professor E. N. Wentworth, of Ames, Ia., supplies the following tabulation from random pedigrees: *Amos Cruickshank, of Sittyton (1808-1895), the most distinguished breeder of Shorthorns, and one of the most skillful breeders of domestic animals, No. 540] INHERITANCE OF COLOR IN CATTLE 709 TABLE II Offspring 43 from white ey white matings. 1 from red by red matings. 83 from roan by roan matings. 127 whites 122 from red by roan matings. 8 from red by red matings. 172 from white by red and white matings. 136 from red by red and white matings. n 207 from roan by roan matings. 645 roans | 439 from red by red matings. 52 from red by red and white matings. 81 from roan by roan matings. 892 reds 320 from red by roan matings. Total 1,664 animals The following table (No. III) records some matings, selected almost at random from the Shorthorn Herd Book, detailing the color of dam, sire and offspring, the last animal of this table, the roan cow Dorothea (Vol. 45, p. 645), herself a roan from two red parents, produced six calves: The first a roan Trout Creek Beauty, by the red- and-white Klondike of Baltimore; the second the red-and- white Lord Strathearn by the red Strathearn Oakland; the third the red Dorothea’s Knight by the red Red Knight; the fourth the white Bapton Favorite by the roan Bapton Ensign; the fifth the roan Dorothea Second by the red March King, and the sixth the red-and-white Dorothea Third by the red March King. It is interesting to note that one cow can produce calves of each color characteristic of the race. In color pattern the red-and-white and the roan-and- white Shorthorns are quite similar to all other breeds of cattle possessing broken patterns—that is to say, there is a tendency toward a white belt at the front flank, a slightly more pronounced one at the rear flank and a white underline. It is known that Angus cattle which are generally black sometimes possess white patches, gener- ally within the line of the rear flank belt. A white Short- horn bred fo a black Angus or Galloway will produce a blue-roan calf, or when bred toa white-faced, roan-bodied [Vou. XLV THE AMERICAN NATURALIST 710 uvoy PIT uvoy INU M PIT uvo pup pay uvo INYM PUD pry INYM PUD pay INYM PUD pay uvoy ONY M PUD pax eoyjo1og, = wojny, euve Apery = BILOJOIA S$, OISSTY 1OJOTA = uByng yoorg uay = 1071A. ISH PIM = uqoy pnorg = yg Weg IMH-FO-8BHOPrA = uByNg MeH OUM puowviq uogdeg SUM, { £yo1e4) Suun eN əuulany pIo Əuuím y Apery TEI0g quy əsodəy 8, 44t S8010 PIN pLogxQ AS IBA snpnəIə, T uI g uueg PIEI IBA ound MEIMN pug əurpəðuðry 3uudsyo PI uvoy PI IUM pay uno I4PVT pee Iy M PUD pay unoy Foy wosurrsy uom PHAS OISSIPY 109TA UvIINS MEH OMMA penp oy} yo dup ulpoane woydeg uvyng woydeg mojsnsny vosy PLA ISIA roronbuog Pueg IB qfy ss010 pN 43ra y S8010 Poy PIOFXO UM 1} wore r sokot PIIM alg Til WTavah KM RR Me MK KS OR a RR KN IMUM PUD PIJ INYM PUD PIJ INYM PUD pay IMUM PUD pay IUM PUD pay uom ouve Apery PIEI 8, orsnog 448 WAV, [[F{-FO-BII0zOT A Wg WAR T TIEE{-FO-BILOpITA, U48 WIV T [H-FO-BMO0pT A Ig WAV T T-FO-BILOqT A, IG WIV [ITH -FO-BIL0qOT A [veg uogdeg No. 540] INHERITANCE OF COLOR IN CATTLE {ii Hereford will produce a white-faced, roan-bodied or red- bodied calf. It is also known that a black Angus bred to a white-faced, red-bodied Hereford will produce a calf with a white face and a black body. A roan Shorthorn Fic. 2. BroopHooKs CHIEF. BROADHOOKS CHIEF 348176. White. Courtesy of F. W. Harding, Waukesha, Wis. Dam—Broadhooks Rose 101234. Roan. Sire—Royal Fancy 93217. Roan. Dam’s Dam—Imp. Roan Rose 75966. Roan. Sire’s Dam—Sensation 7th. Red. Dam’s Sire—Rustic Chief 236800. Roan. Sire’s Sire—Prince of Fashion 64587. Red. mated with a black Angus or Galloway will sometimes produce a black and sometimes a blue-roan calf; as in- stances of the former, Mr. Ralph B. Goodhue, of Don- nelly, Minn., writes: I have had a few animals cross-bred between Angus and Shorthorn and in every instance have the offspring been black, sire and dams pure bred animals. I have bred 31/32 Holstein cows to red Shorthorn bulls and about 65 per cent. have been red-and-white, the rest being black- and-white, more black than white in markings. In breeding grade Shorthorn cows to pure bred Holstein bulls, have got black and white offspring. In the Hereford-Shorthorn, the red Shorthorn bred with the Hereford will most always give a mottled face on the offspring. The roan Shorthorn cows bred to Hereford bulls will give either a ealf looking like a Hereford or a roan calf with clear white face. Professor Wentworth, previously referred to, writes: 12 THE AMERICAN NATURALIST [Vou. XLV In regard to color coats in cattle hybrids, I can give you a few cases from my own experience. We had three Holstein cows at home, two of them carrying a pre- dominance of black, the other a predominance of white. As we had no Holstein bull on three successive years, they were bred to Shorthorn bulls. The first year when bred to a roan one of the calves came a blue gray. This calf was from the cow with the greatest amount of black. The others showed the pattern markings of their mothers. The second year they were bred to a deep red Shorthorn bull (all of the animals mentioned pure breds) and the color pattern showed no trace whatsoever of the Shorthorn parentage. The third year they were bred to a red and white bull. In the ease of the lightest Holstein cow there seemed to be some tinge of red on the ends of the hair in the black pattern; however, at a distance it showed the same color. I have seen Jersey-Holstein crosses usually partaking of the Hol- stein pattern with, perhaps, a slight admixture of dun color on the tips of the hair on the black markings. I have seen Angus crossed on Jersey showing simply the black polled character, although in a few cases the extremities showed a slight tendency towards dun or fawn. I have seen Angus crossed with Holstein and have seen both pure black and black-and-white cows. The instances which I have in mind are about twenty showing pure black and six or seven showing the black-and-white. However, these figures are simply a question of mem- ory and might easily be modified. The case in question is that of a man with a Holstein herd who was forced to breed to an Angus bull one year. Out at the dairy farm we have a Shorthorn cow, roan in color but a grade, which was bred to our Holstein bull, a half brother of Colanta 4th’s Johanna. The calf is roan in color. We also have some Arkansas backwoods cows; they are variegated in color pattern, showing red, dun, yellow, white, brindle and various other markings. A Holstein bull when bred to one of these produced a nearly pure black heifer. The black seemed to be rather tinged with brown at the ends of the hair, but the udder showed a white color. . - . A roan Shorthorn bull bred to a Hereford cow will quite frequently give a roan body with white Hereford markings. A red Shorthorn bull crossed with the Hereford cow is apt to increase the red splotches on the white markings of the Hereford. Mr. P. G. Ross, of the famous Maxwalton Farm, Mans- field, O., relates his experience, throwing his observations into approximate percentages, as follows: The color of the offspring of white Shorthorns depends largely on the ancestors, as about 50 per cent. of a bull’s calves will have the color of No. 540] INHERITANCE OF COLOR IN CATTLE dis his dam and her ancestors. . . . We have used white on white and often had roan calves and in one instance had a red calf, but about 75 per cent. are white We have had considerable experience in crossing the Shorthorn on Angus. This we consider the best cross and the offspring is generally better than either of the parents. When crossing red and black the offspring are generally 75 per cent. blacks and even the second cross will not bring 50 per cent. reds; og crossing roan and black, about 50 per cent. will be blue-roans, 10 per cent. red-roans, 10 per cent. reds and 30 per cent. blacks. The Galloway color is much stronger than the Angus, consequently more dark calves will be expected. The Hereford cross is very strong as far as the white face is concerned and about 95 per cent. of white heads ean be expected but the red of the body is easily blended into a roan and about 95 per cent. roan calves can be expected by a white bull, and at least 75 per cent. by roan bull, on Hereford cows. The black of the Holstein seems to be particularly strong and when crossed with red the offspring will be nearly black and will remain very predominant to the third and fourth cross; the broken color shows itself but very little. . . . On the other hand, Holsteins take the roan color very readily and wise crossed on white 95 per cent., and when crossed on roan 75 per cent. of the calves will be ine roan: It is our experience that either the Holstein or the Hereford will take the roan color from a white or roan much oftener than from the red Shorthorn even if part of the red’s ancestry were roans. The Red Polls and Devons seem to be very hard to blend into a roan and when crossed on a white not over 25 per cent. roans can be expected; the balance are red. This we do not consider strange as they have been bred red for so many generations. It would seem that the red color of the Shorthorn is not so strong as the roan when used in erossing, and in our opinion it is the most objectionable. . . . We believe that to maintain the standard we must exert judgment in crossing the best types and colors, as it is evident in both animals and plants that they must have fresh blood to prosper and this is seen much earlier in breeding the short- lived animals such as hogs, dogs, cats and rabbits. We feel that the Shorthorn has given a much better opportunity for crossing than any other breed of cattle as there is very little restriction as to standard color. The different color is, we feel, a very safe rule to go by in cross- ing and we are particularly opposed to using red on red no matter if the ancestors are desirable. We feel that white on red is the proper cross and roan is good to cross on anything. The observed facts fit the following hypothesis so closely that it is presented as a further working basis in solving the problem of the prediction of the color and color pattern in Shorthorn cattle. 714 THE AMERICAN NATURALIST [Vou. XLV Hypothesis—There are two groups of genetically inde- pendent sets of hairs intermingled to make up the Short- horn color coat. One set is alternatively ‘‘ positive white’’ (W) and red (R), in which the white is dominant and the red recessive; the other set is alternatively red (R) or ‘‘albinic white’’ (wr), in which the red is dominant and the white recessive. Dominant white is caused by a Fig. 3. ANOKA ACONITE 2D, ANOKA ACONITE 2p 40311. Roan. Courtesy of F. W. Harding, Waukesha, Wis. Dam—Double Aconite 2d. Vol.53, p.563. Red. Sire—Whitehall ar aea 209776. son Dam’s Dam—Double Aconite. Roan. Sire’s Dam—Imp. Missie 167th. Dam’s Sire—Godoy 115575. Red. Sire’s Sire—Whitehall Suftas vy White. specific antibody existing in the zygote in small quan- tities, retarding or inhibiting the ontogenesis of the determiner for pigmentation. The same body existing in larger quantities reacts with and destroys the deter- miner for pigmentation, causing recessive or albinic white. The dominant white of the Shorthorn is doubtless de- rived from the Romano-British cattle, which it is generally conceded entered into the Shorthorn make-up, which ele- ment is to-day represented by the ‘‘Park Cattle.” They No. 540] INHERITANCE OF COLOR IN CATTLE 715 behave as dominant whites—. e., they themselves are white but sometimes throw red or black (not roan) calves. The recessive white doubtless came in with the Dutch flecked, the colored areas of which took the ‘‘ differential coloring’’ because they lacked the positive graying factor; this recessive white must therefore be attributed to a strain of partial albinism. The spotted color pattern or coarse mosaic doubtless came in with the Dutch bulls of the eighteenth-century importation. The areas com- posing Group One are located about the two flank belts, the underline, the median line and the face and a fine net- work over the remainder of the body; those composing Group Two cover the neck, sides, back, hind quarters and legs in a network exclusive of the areas of Group One. Factors CONSIDERED W = Inhibitor of pigment formation. w = Absence of such inhibitor. R= Determiner for red pigmentation. r= Absence of determiner for red pigmentation. With reference to Set No. 1, or group-unit No. 1, indi- vidual cattle are gametically W.r., WwR, or w.R,. With reference to group-unit No. 2 they are w.R,, WR: or Wr». There are therefore involving these characters nine gametic and three somatic types of individuals, which types are set forth in the following table: TABLE IV GAMETIC COMPOSITION Somatic UNIT PURITY Set 1 Set 2 Aspect Blood Set 1 Set 2 ee a. w.R, W. Red Pure Duplex Duplex r ES w.R, w.Rr Red Mongrel Duplex Simplex APES wR, Wa Roan Pure ulliplex eer WwR, w Roan Mongrel Simplex Duplex Doy WwR, wRr Roan Mongrel Simplex Simplex a WwR, wr, White Mongrel Simplex Nulliplex PAs eek War, w.R, Roan Pure Duplex uplex aa War, wRr Roan Mongrel Duplex Simplex oaa. W.r, Waa White Pure Duplex Nulliplex Roan in this table stands for any animal with red and white hairs interspersed, regardless of the proportion or pattern. 716 THE AMERICAN NATURALIST [Vou. XLV All of these theoretical types seem to occur except the roan of type 3, which phenomenon will be discussed further on in this paper. With these nine theoretical types of individuals the following forty-five type matings are possible. (The numbers following the color designations refer to the above table describing the individuals somatically and gametically. ) These forty-five cases typify the behavior of two com- panion traits of opposing patency in their dominant phases, thus explaining the behavior of one type of apparent or somatic blend, which is in fact the resultant somatic effect of the lack of synchronism in the behavior of genetically independent units. ‘With these matings it is noted that the cases joined by an arrow (cases 8 and 9,10 and 11,13 and 14,17 and 18, 19 and 20, 22 and 23, 32 and 33, 34 and 35, and 37 and 38) are reciprocal cases wherein the same parental elements enter and the same offspring are expected, but these parental elements are differently combined in each pair of parents—different somatic but identical gametic matings. These principles fit the previously observed facts as follows: As to the attempt to establish a race of Red Shorthorns, the above mating No. 30 (a red by red) expects 25 per cent. roan offspring and amply accounts for the occurrence of roans in such a cross. This phe- nomenon is equally well accounted for by the simple hypothesis that red is dominant; some reds are simplex. It is known that breeders in attempting to eliminate white, spotted and roan from their stock simply destroyed the ‘‘off color’’ calf—the genotypic germ plasm that produced it being continued in the herd. There are, however, reds which will produce only reds, as in matings Nos. 6 and 15. Mr. Spangler’s white bull was produced by two roan parents; such color is expected from such a mating in one fourth of the offspring of matings Nos. 25, 26, 37; in three sixteenths of No. 28 and in one half of No. 34. His whiteness is of either type 6 or type 9 and consequently 717 INHERITANCE OF COLOR IN CATTLE No. 540] ¢ g ULoy L weoy & P me T PA Sap = aym | tym p = kg TMi TM T EM” M T pew p ugoy L uoy l tym p = T a TaT? z ZZ 5 ueo { “AF = EMTT M “UMMA OT waa p Ueo 2 weoy f ymp = Me M z TMI | oe MMM 8 FEM 8 ers g 4 SM Bee L weoy yap = Zat? Be uk oe ZulidsyO Jo uolsodwog dy4¥vulog puv oMeuIey SUOIJB[NI[VH 9}BIPIMIeIUT A Wiavib Cam) [ P Cra) Cam) p woy (a) Cuamm) L weoy Ca) Crm) guəredq pZ Erm) z L Ugoy yoyo orywUI0g quəredq IUO [Vou. XLV THE AMERICAN NATURALIST 718 6 PR T PA z ngë — IJ NYM g + J MYM 8 y M Z + WMS inegau aA > ns Zz pey T pay G uBoy p ueo ( 2 z aye arene == QM p H YR yt F H QE MYM M F + MAM 2 a = pin M í ae o PU T pee ¢ weoy p ugoy É ka INMA M IMM p H MAA p H ay Mp FE YEAH A F ress ee "aM G+ "UMM S Z poy T poy ry a ye g + an yea g H G ueo p weoy g ueo L uoy TUM TMM P HAMM P EMEA g H EM g Ig g + g “aM + MM g HFM II Il f G uroy p ugoy i A å ayia g + g = INYM 8 F INIMA 8 1 oe aM p = c ugoy p ueoy ta enas oe TEMA MM 8 H 'IMYMM 8 ATE od ue pe G ueo p ugoy 8 ugoy L ugoy : ji a à Iya “ag == THEM MM P bE p H EEM p HEM F a. M : : EM g = G ueo p weoy 8 uĽroy L ugoy Wag + gag = TEAM MF EMM MA P H MEM P H EM F mM g +H EEM z ioe SujidsyO Jo UOTTsodu0D OMBUIOg pus OYoUTEH HB[NITV) IVTPA: I (panuyjuoo) A ATAVL Z PN (ay?) (ye) G uvoy (aga) (Cum) Z PN (aga) (ga) G ugoy (igm) Fam) 8 ueo (1g) EEM) CEM) g ueo (iym) (yam) JUIIV PZ guəðreq IUO Sl “BL 719 INHERITANCE OF COLOR IN CATTLE No. 540] 6 Pe g weoy IPM 8 + TMM AA 8 Z pay Gg weoy IJ 8 + IY MAMA 8 ( Z pay G UBoyy 8 uroy IY MYM p + IMMM B+ IYEM ITM F { G ueoy IY MYM M OT G ueo IJN YAM OT G uroy 8g ueoy TPA MM 8 + IMIM 8 G uBoy 8 ueo QM YMM 8 + IMIM 8 Co ey AN g ugoy IJ MIM OT SutidsyO Jo uonsodwop pemos pug oyowery OM g = "2M g + UMM ZS nea >? = "aM g HUM 8 = Ha y = TM HUM HM = wa p = UAM b= nea p = YAM > = nya >? = MMS HEMA = nA p = "IMM 3 HFM g = TE ERS. TPE E desea’ S (panuyuoo) A ATAVL 9 IFM (rm) CaM) g ugoy Cra) Ca) 9 ƏM Cra) Cam) 6 FIM (rm) CEM) 9 HM (Erm) Cam) 6 OTM (ra) CEM) quosed pZ "LE ‘OL [Vou. XLV THE AMERICAN NATURALIST 720 g weoy g pP raya p + IMTM g H YMY p g weoy g P Ermaya g + aytay p + T PU 9 IFM G ueo AA g H EMM 3 H IMAAM P H AMM g G wsoy Ty MM F + g weoy c PON T PA 9 OVUM p weoey 6 OFM 8 ugoy T arm + agim g + tga = tata $ = : zt yë tn = p moog Ta + ag g +y = Ta g + UMM g ,uog | SM tanas t'a = apna Z H EAEE H MEM g H MEM aiaiga + agiagia g H g eM pA g H | Fa EMAG E EA = 9 IUM G uvoy p uvoy zM iym g + ya = "EM MM P + Ty AMM 8 + MMM F she “IMM p = 9 IUM 9 usog FEM ITM M Z + IaM p + atm + aya g + ya = p ugoy 6 NTM 8 uoy L Vaa yam 2+ MZ = MMA g H EMM g H EM p H MM g 6 OTM 8 uvoy L woy ezz aye = Feta p F EM 8g H UMM P a + Ira g eas ts Z pu nea p = Iyya a we p = 3uudsyo Jo up sod oD əpvuog put opowvrg hsb TITOS DN (penuyjuoo) A ATAVL g uvoy (ga) guad p3 gZ Pu (Gya) Caa) ‘08 g ugoy (1ga) yam) "63 g uvoy (ym) (am) "SZ 8 uvo (1m) CEM) "18 8 uoy (1y) Erm) "93 8 deo maG: ev) Crm) "93 T poy (m) (ym) H3 qua eug i 121 INHERITANCE OF COLOR IN CATTLE No. 540] 9 OFT A ‘mg --ay'm g = (ua) ‘ym g++ ym g = (am) g meou 3 pou 9 MUM ¢ ugoy AYM p H TMA p H AM P H M P g ueoy Z Pou 9 FIM c ugoy g usog SEMAIA p TYP Ya p -H EMM M p + MMM F au; ae = a g uroy 3 PA armiya g + yntym g -+ : 9 oun 9 IFM G ugoy 6 OFT A g unoy tag -+ga g = (a) EPMA M F H Typ aay MA M F of Trea g + IMF M z M + TMM ra + FM = (mam) t l tg 9 om g ueog i 6 ONAL l l ( oye — MAMMA 8 + TYTMATA A 8 Bs shops: ‘is Se 9 OVUM g ugoy g uroy aastata |Meat oD 9 80M Q weoy 6 IVM g uroy : - 6 RIM MMM P H IAA p HEMM p H M F ean ved ke a et g = M 9 M G uvoy 6 MTM g uroy 9 ORUM TEMAM P H IMA M P H EME p + M F Ta g+ igmg =, (ta) "YAM 3 HM = (Fam) 6 IM 8 umeog 6 OYM ‘tne taranna | estates | Go SutidsyO JO uopssodmog Ipvwogs pus ojaUIBy SUOPIB[NI[VO AVIAU] quereg pz (panuyuoo) A AIAVL quale IUO "88 "98 [Vou. XLV THE AMERICAN NATURALIST 722 g Weoo : TEM f = FEM YAM OT Ig p = ( g ueoy 9 OIM J Puy ZIM? TM oye FAAS M Ji — TM 8 H OEMTMM 8 amg + UA g = . g UBOT 9 9210 M 6 TM 4} IEM p = EMM p H TaM 8 HITAM F il ara + yam g +M = m -e gus ¢ 9 OMIM 1 ctu p = SIEM ATM 9 TM 6 OTA zm p CAS F Na um g HME = Se 6 OA { utap = z MŽ A My = g ugoy Z Poy bags + mata g oe SEIM g + TPM g ttm p = Buudsyo Jo uoprsoduoo onvwog pur oyaurey SUOTIB[ND[VD IB pomu (panuyuoo) A W'TAVL 9 OTT M (rm) Cunm ) 6 OTM (m) EEM) guaedq pZ 9 UM (ra) (apa ) 6 IUM (Era) EEM) 6 ƏHYM (Erm) CEM) 6 OUM quoieg oud “PP No. 540] INHERITANCE OF COLOR IN CATTLE 723 when mated with red cows only roan calves are expected if the mating be like mating No. 20; 50 per cent. red and 50 per cent. roan if like No. 23; 50 per cent. roan and 50 per cent. white if like No. 35; and 50 per cent. roan, 25 per cent. red and 25 per cent. white if like No. 38. There is ample explanation for throwing a white calf from a roan cow and a white bull. If the mating be like Nos. 31, Fia. 4. FLORINDA SULTANA. FLORINDA SULTANA 70519. Roan. Courtesy of F. W. Harding, Waukesha, Wis. Dam—Gertrude, Vol. 60, p. 1110. Red. oe Hall ApS ap 3. White. Dam’s Dam—Wild Eye Belle 15th. Red Sire’s Dam—Bapton - Roa Dam’s Sire—Judge Wardell 144980. Red. Sire? s Sire—Bapton isni S Roan. 32, 33 or 44 the chance for roan and white colors are equal; if like No. 36 the ratio of white to roan to red expected is 6 to 8 to 2, but if the mating be of type 42 only white calves can be expected. Mr. Bruce’s array of data concerning the Sittyton cattle presents a very telling table of facts. All possible color matings are made, and in most cases the number of offspring is quite large enough to insure a proportional distribution among the expected colors. The facts of this table fit the hypothesis quite significantly. Of special 724 THE AMERICAN NATURALIST [ Vou. XLV interest are the roan by roan matings, which produced 56 reds, 193 mixed color (i. e., roan and red-and-white) and 60 white offspring. This ier not fit well into the now abandoned hypothesis that ‘‘roans are simplex, reds are duplex and whites nulliplex.’’ The number being quite large should approximate more nearly the expected 50 per cent. of roans, or even less than 50 per cent., inasmuch as some reds were later thought to be simplex; however, there are 62.46 per cent. roans. This may mean that some roans are pure and when mated to like animals will pro- duce only roans—as mating No. 1, wherein two roans produce only roans, which in turn are pure and will re- produce themselves. As further explanation in account- ing for an excess of roans—which is common in most herds—note that in matings Nos. 1, 2, 7, 8, 9, 19 and 45 a roan mated with a roan produces roans only. As to the red by red matings, types Nos. 6 and 15 will give only red offspring, while type 30 gives 75 per cent. red and 25 per cent. roan, which fits very well the distribution—133 red, 12 red-and-white, 34 roan and 1 white—with the exception of the one white which will be discussed a little -` further on. As a matter of fact, every possible color mating has been reported to throw every other color characteristic of the breed. The red Shorthorn calf of white parentage is no doubt derived as follows: Save for occasional insignificant red patches in the ears, many Park Cattle are solid dominant white; this element in a few cattle of the Shorthorn breed would in the course of time, by the laws of chance, make the mating Sets 1 and 2 (WwR.), Sets 1 and 2 (WwR,), which would throw 25 per cent. red calves; this appar- ently is exactly what has happened. Moreover, the much more frequently possible white by white mating wR.) (WwR,) Set 1 y Sets 1 and 2 (Wara) will produce 124 per cent. red offspring. t 2 The red-by-red mating that produces a white calf is No. 540] INHERITANCE OF COLOR IN CATTLE 725 either of very rare occurrence or does not occur at all. Besides the cases just referred to, Pearson and Barring- ton? reported two Shorthorn matings reputed to have been red by red that produced white calves. Cases of such rarity and import should be supported by more painstakingly minute evidence than that offered by the es TN Fic. 5. Spicy SULTAN. Spicy SuLTAN 334972. Roan. Courtesy of F. W. Harding, Waukesha, Wis. Dam—Spicy of Edna, Vol. 50, p. 532. Sire—Whitehall Sultan 163573. White. Red, little White. Sire’s Dam—Bapton Pearl. Roan. Dam’s Dam—Spicy of Browndale 9th. Red. Sire’s Sire—Bapton Sultan 163570. Roan, Dam’s Sire—Orange Victor 138562. Red. herd book, which often records an animal as red though it may have white or roan areas of quite noticeable extent, or an animal as ‘‘white’’ that may have, besides the generally characteristic red in and about the ears, small body areas of red or of roan. Mr. E. M. Hall, a promi- nent Shorthorn breeder of Carthage, Missouri, in response to a recent inquiry, wrote: ‘‘I now, March 30, 1911, have one white calf—from red dam and sire, but it is an inbred calf.’ In response to further solicitation he ‘On the Inheritance of Coat-Colour in Cattle,’’ Biometrika, 1905-6, p. 442. 726 THE AMERICAN NATURALIST [ Vou. XLV kindly supplied the data for the following descriptive pedigree: A ft 3 Remus Gerafsine = Choice $ Imp. Efra 554 Reyaf Bug Victoria Riaerss cei ite ja oi r s Res ded Ly VOI p.ms Roa RS GER , Res : i Roda a R Choice Beo$s 15650: Ress Salt Vrolet 94 Ths Con Fuser 219004 Victeria HUS. A beasties riem = Res Dark Rossa (frin) L" dark aleut the heak anb a EE E E H.G3 fp. 074 hick an ramning To R lykter cofor Lack i=B tf body ae —= Choici Goofs Mo Heredity, 508; Mutations in ila, 511 Moths, Influence of Changed Exter- nal Co nditions on Development of Two Species of, THOMAS ‘ ONTGOMERY, JR., 364 Mulattoes, Color Inheritance among, . E. Jo Mutationist, J san Marchant, an Eighteenth Century, HENRI Hus, 493 Natural Selection in the English Ow AR Harris, 314 Newman, H. H, Killifish, 436 Nichols, John ” Treadwell, Pelagic Pipefish, 433; Blennies, 433 Notes and Literatnre, 60, 190, 253, 319, 375, 415, 507, 574, 639, 761 Notropis, Henry W. Fowler , 436 Nucleus and Cytoplasm in Heredity, AEL F. GUYER, 284, 512 (Enothera biennis, some ee of, BRADLEY RE Davis, 193; Mu- tation in, R. R. Ga slong 577 Ogilby, p Pe Fishes of Bris- bane irai a a Genus, AUSTIN HOBART CLARK, 372 i D T. Mio Organice Response, DOUGAL, 5 No. 540.] Origin of Species in Nature, HENRI Hus, 641 PATTERSON, J. p nae A Double Hen ’s Egg, 5 PEARL, RAYMON kad Data on the Rela- tive Conspicuousne ss of Barred and Self-Colored Fowls, 107; an Introduction to the istics : ny antes Plants, 415; Arguments regarding the Genotype Concept, 561; and Lottie E. McPHETERS, Biomet- rical Computations, 756 cha ne: Sa Function in Marin gani F. ALEX. neniet, Pigmented Skins, Study of Melanin Content of, with special reference to Question of Color Inheritance among Mulattoes, H. E. JORD 449 Plants, agar ia hidang and Correlation in, RAYMOND PEARL, 15; Fossil. ecent otis on, UGLAS Hovamtor CAMPBELL, 439 Polymorphism and Sex in Colias groin eo of, JOHN . GEROULD Poss of the ihe Dinosaurs, Further Observations on the, OLIVER P. Hay Primates, Adaptations of the, F. B. 479 Punnett on Mimiery in pete But- terflies, FRANK E. L Pure Line, Theory, Application of Conception of, to Sex-limited In- heritance pte o Sexual Dinorah i : , 65; in the Study of Genetics in er Or- ganisms, H. ENNINGS, 79; ae Proof of, J. ARTHUR , 346; Means, — of, J Ta HUR HARRIS, Quante on Variation and Correla- Tae RAYMOND PEARL, 416 Radcliffe, Lewis, a ug Smith, Butterfly Fishes, 437 Rat, Brown, Reproduetion in, NEw- N MILLER, 623 REDFIELD, = L, Acquired Characters defined, 5 INDEX. 767 Regan, C. Tate, Zeide and Caproide, 432; Flou nders, 432; Systematic Position of the Genus Macristium, ; Lampreys of the Tno siae Pike of Loch 434; iant natoniy an Sussman ‘of the Teleostean Fishes of the Order niomi, 435; Berycide, 435; Fische d dsee, by Albert Gunther, 437; Microcyprini, 437; Salmoperee, 437; Silver Gar and Flying Fish, Fishes related to, 437; Cirrhites, 438 Reprodu ction in the Brown Rat (Mus didi eee NEWTON MILLER, Ribeiro on i OES in Brazil, 435 Rietz, H. L., and L. H. Smith, Varia- tion and "Correlation, RAYMOND PEAR Roberts, a F, Variation and Corre- ati ARL, 42 Russo on Sex- -Chromosomes, 429 Sauropodous Dinosaurs, Further Ob- servations on the Pose of, OLIVER ; Y, 398 Schleip on Sex-Chromosomes, 425 Schoute, J. C., Variation and Corre- Scott, D. H., Studies in Fossil Bot- LAS HOUGHTON CAMP- L, 439 aie Alvin, Fishes of Bantayan Is- nd, 431; Fishes from paws 0, 436 gegen N eno in ise og UR HARRI Bie Chetek ana Boni Fowls, Relative Conspicuousness of, RAY- MOND PEARL, 107 gp AC, Fossil nae o aa HOUGHTON CAMPB d Takeria, fa tion of wow on = Pure Lines . H. Morean, 65; Determina- era ‘Is the Fomle | Frog Hetero- sie in regard to, T. H. MoR- , 253; in Colias - philodice, and Paika sm, pegs tance of, JOHN H. GEROULD, H . K., System sf Recording Mendelian Variations, ria 314, 372, 413, 561, 636, 701, 768 Short-horn Cattle, panera of Color in, H. H. Lau , 705 S HULL, GEORGE Hanmisos, The Genotypes of Maize, 2 Shull, G. H., Variation and Correla- tion, RAYMOND PEARL, 424 . Rietz, Varia- tion ane Correlation, RAYMON PEARL, Society ears Naturalists, 5, 65, 79, 90, 99, 129, 160, 234, 257, ‘984’ 321 , 346 Sparrow, Phen a N a Paper AR- ral Selectio m Ms re UR ea SPILLMAN, W. J., Heredity, 60, 507; Inheritance of the c Bye? GE Vigna, 513 Starks, E. thi and W. F. Thompson, Stopes, Mari D rie nt Plants, DOUGLAS ionann CAMPBELL, 439 s B., Some Effects owing a Subsequent Genera- F. M., Variation and Cor- only used, H. S. Jwunxes, 123 ae on Effects upon or Mice, and Pers — or su t Generation, 2, 9 Extinct Am- 3 and W. Stock- berger, r, Variation snd Correlation, RayM 15 Oda and ar 7 **Gold-eye of Lake Winnipeg,’’ 431; Lobotide and Lutianide, 431; Scienide, 432 Thompson, W, F., and E. C. Starks, , 4 5 inism and Hu- Tower, Wm. Da ' Heredity, Ww. J. SPILLMAN , 60 THE AMERICAN NATURALIST [ Vou. XLV Tracy, Henry C., Fishes of Rhode Island, 435 Variation and Correlation in Agri- cultural Plants, RAYMOND PEARL, 415 Variations, System of recordin 701 be as Inheritance of the “Eye” . J. SPILLMAN, 513 Mendelian, Je 3 SHAW Wagner, George, Cisco from Lake Michigan, 434; Stickleback of Lake Superior 4 Waite, Ed ., Fishes from Ker madee and Norfolk Talands, 434; New Zealand Fishes, Waldron, L. R., Va enin Pe Sid Corre- lation, RAYMOND PEARL, 415 Waugh A, K. Shaw, ugh, le Variation and “Commi tet: Ray- MOND PEARL, 421 i Max, a of New Guinea, Weed, Alfred C., and Barton A. Bea , Anteliochimera, 432; Morel omous Toad-fishes of South ‘Am ica, 432; Habits of Electric er been ange. J. M., Variation, RAY- D PEARL , 424 Wheldale, M., Formation of Antho- 507 Whiting, P. W., 2 G Wieland, G. R. ao dene Fossils Cycads, DoUGL AS HOUGHTON CAMPBELL, Williston, S. W., Extinct Amphibia, Roy L. ’ MOODIE, 375 Wilson, E. B., Chromosome Be- havior, 507 Winkler, H., Oun i DOUGLAS HOUGHTON CAMPBELL, 43 Woodward, A. major, 375 Smith, Bothriceps T G. aes Introduction to the Theory Statistics, RAYMOND PEARL, 319 Zaleskiego, E. I., Variation and Correlation, RAYMOND PEARL, 416 oogeography of the East Indian Archipelago, P. N. VAN KAMPEN, “ 537 The American Journal of Science Established by Benjamin Silliman in 1818 The Leading Scientific Journal in the United States Devoted to the Physical and Natural Sciences, with special reference to Physics, and Chemistry on the one hand, and to Geology and Mineralogy on the other. : EDWARD 8. DANA Associate peat nk Professor GEORGE L. GOODALE. JOHN TROWBR RIDGE, W. G. FARLOW and WM. M. DAVIS of G hoy e; Professors z x VERRILL, HENRY S. WILLI AMS and L. V. P of New ; Professor 0. ARKER, of geet egypt: Professor JOSEPH 8. AIMES, o more; MR. J. § . DILLER, of Washington. Two volumes annually, in monthly numbers of about 80 pages each. This Journal ended its first series of 50 volumes as a quarterly in 1845; its second series of 50 volumes as a two-monthly in 1870; its third series as a monthly ended Dec- ember, 1895. A Fourth Series commenced in 1896. Subscription price, $6 per year or 50 cents a number, postage prepaid in the United States ; $6.25 to Canada ; $6,40 to Countries in the Postal Union. Back numbers at reduced prices. sa Ten-Volume Indexes, Vols. I-X XI-XX, fourth series, price one dollar. Address The American Journal of Science New Haven, Conn. SECOND EDITION, NOVEMBER, 1910 AMERICAN MEN OF SCIENCE A BIOGRAPHICAL DIRECTORY EDITED BY J. MeKEEN CATTELL A Biographical direc ires revision if it is to maintain its- mess. Nearly a third of the names in the Present edition are new, “ane od the elas — a peared $ in Ags sheng division have in ae every en been revised. — amount of work required to prepare the rev as that given to the first edition. change +n the general plan of the work. "Cran predia A ayia in y tis a ms gia natural and exact sciences, and for this reason a few names included in as fal edition have been o make the book as complete and accurate as possible. There are of course suse aed T only a Se men will oe ge Sige to Sones: requests for the information needed. The thousand leading men of science have been again selected by hods that were before, and stars have added to the is of research in gid case of 269 new men who have obtained places on the list editor’ s object in selecting this group o” —— en to make a study a the -o = pog scientific pasta pega ara 80 = as may be to im - iron the nate the “Seon 2 i The second ag of the D 600 FER 5500 sketches. rec is well Bge on all rag paper rand bound in rakor with leather mhara Although the work has been n size by more than 50 per cent., it is sold at the same price as the first editio Price: Five Dollars, net, Postage paid THE SCIENCE PRESS GARRISON, N. Y. LANCASTER, PA. SUB-STATION 84, NEW YORK CITY. The > American Naturalist stablished in 1867, Devoted to the Advancement of the Biological Sciences Factors oe ‘Spaced Reference to the of Organic Evolution and Heredity CONTENTS OF JUNE NUMBER Inheritance of Fecundity in the Domestic Fowl. Dr RAYMON The seine wowed of the Pure Line Theory. Dr, J. ARTHUR HAR The Influence a ante nged External Conditions on e Development of fades = rae of Moss, Pro- ale THomas H. M r Articles and sonata ang ge Ontogeny of a Genus, AUSTIN HOBART CLARK. Notes and Literature: Recent Contributions to s Knowledge of the Extinct Amphibia, Dr. Roy L. Mooprz. CONTENTS OF JULY NUM Germ-cell Bet harman and Their emacs Pro- fessor Further eg ga on the re of the Sauropodous Dinosaurs, si us a pri d Tien Notes and asx nai Some Recent Studies ca evil ation an z Correlation n Agricultural Plan RAYMOND PE x-chromosomes n Yee saephrodithie - Professor W. E. CASTLE. Notes on Ichthyology: President DAVID STARB Joani Some Recent his on Fossil Plants: Projessor Dovelas HOUGHTON CAMPBELL. posi OF AUGUST NUMBER A Comparative aacupte Study of the Melanin ae of Nera: Skins, with special ae ce to the Question of Color In a BIRGA Mi essor H. E, Jor A Coefficient of Individoat E oe k for Stud of Heredity. gore Prepotency e The Adaptations o = Peat tes. Professor F. B. Pag o Marchant, an Eighteenth Century Mutationist. Notes and Literature : Notes on Heredity. Dr. W, J- SPILLMAN, ad CONTENTS OF SEPTEMBER NUMBER eg ome of the “Eye” in Vigna. Dr. W. J PILLMAN. Heredity « of Hair Form among the Filipinos, Dr. ROBERT BENNETT BEAN. = weography of th of n — Indian Archipelago. Ey rter Articles and nae ssion: Biometrie Argu- ments regarding the Genorita Con Eai. Dr. RAYMOND PEARL. On the s when te of Borres 3 ; e 3. defined. C. L. REDFIELD. Notes and Literature: The Present Day Peotone — ote e of Animal Psychology; A Useful Evo tion Bibliography. V. L. K. CONTENTS OF OCTOBER NUMBER I. Mutation in Œnotbera. Dr. R. R. Ga ATES. II. Some Experimental Observations concerning = Behavior of Various Bees in Their Visits to Cotto Blossoms. H. A. ALLARD. Tir. es P Newro: IV. Shorter Articles and Discussion and Other Disian Dk. J. V. Notes avd Literature: adeila nin pea Life V.L. K. N MILLER — kinken ARR CONTENTS OF NOVEMBER NUMBER The Origin of Species in Nature. Dr. HENRI HUS. Some Experimental Observations concerning the Be- havior of Various Bees in their Visits to Cotton Blossoms, H. A. ALLARD. The Distribution of Pure Line Means. Dz. J. ARTHUR Harris. Shorter Articles and Discussion: A System sa a ing Mendelian Observations. Dr. J. K.8 ee tent NATURALIST will be sent to new subscribers for four JHE SCIENCE PRESS —— 84: NEW YORK Yearly Sabios $4.00 four months for One Lancaster, Pa.