'if.■;^■■■ ;' .-vi' GENET 1 1 1 ANlMTRODUCTION'ljjiiP r H E S TV D Y f ^ H E RE D I T Y :|H|i!!l!i| Iji ®Ijp ^. p. pU lltirrarg 0H45I W5 £■ fwC. ST, S00141938 Q T Date Due 2»At.^ 3 j ieMay 58 !« 29Feb'36 i S^'S pt 270cV?f \ I ••<,! ^ gp fa 2 1 1965 JUL 5 f 1967 V968 IWay^S 22mar5|(t 44 3,- t ^' . m,>?o,n 3 01^1^ 17061 GENETICS THE MACMILLAN COMPANY NEW YORK • BOSTON • CHICAGO DALLAS • SAN FRANCISCO MACMILLAN & CO., Limited LONDON • BOMBAY • CALCUTTA MELBOURNE THE MACMILLAN CO. OF CANADA, Ltd. TORONTO GENETICS AN INTRODUCTION TO THE STUDY OF HEREDITY BY HERBERT EUGENE WALTER ASSISTANT PROFESSOR OF BIOLOGY BROWN UNIVERSITY WITH 72 FIGURES AND DIAGRAMS 11 C. COLLEGE OF A. & M. 1. Dept. of Botany THE MACMILLAN COMPANY 1913 All rights reserved COPYKIGHT, 1913, Bt the macmillan company. Set up and electrotyped. Published Februaiy, 1913. J. S. Gushing Co. — Berwick & Smith Co. Norwood, Mass., U.S.A. *u, /ynr.. ^ep^ y *••«».* if.-. - ?? ■•?.*■ THIS VOLUME IS AFFECTIONATELY DEDICATED TO MY MOTHER PREFACE The following pages had their origin in a course of lectures upon Heredity, given at Brown Univer- sity during the winter of 1911-1912, which were amplified and repeated in part the following sum- mer at Cold Spring Harbor, Long Island, before the biological summer school of the Brooklyn Institute of Arts and Sciences. An attempt has been made to summarize for the intelligent, but uninitiated, reader some of the more recent phases of the questions of heredity which are at present agitating the biological world. It is hoped that this summary will not only be of interest to the general reader, but that it will also be of serv- ice in college courses dealing with evolution and heredity. The subject of heredity concerns every one, but many of those who wish to become better informed regarding it are either too busily engaged or lack the opportunity to study the matter out for themselves. The recent literature in this field is already very large, with every indication that much more is about to follow, which is a further discouragement to non- technical readers. It may not be a thankless task, therefore, out of the jargon of many tongues to raise a single voice vii viii PREFACE which shall attempt to tell the tale of heredity. There may be a certain advantage in having as spokesman one who is not at present immersed in the arduous technical investigations that are making the tale worth telling. The difficulties in under- standing this complicated subject may possibly be realized better by one who is himself still struggling with them, than by the seasoned expert who has long since forgotten that such difficulties exist. Among others I am particularly indebted to Dr. C. B. Davenport for many helpful suggestions, to my colleague, Professor A. D. Mead, for reading the manuscript critically, to Dr. S. I. Kornhauser who gave valuable aid in connection with the chapter on the Determination of Sex, and to my wife for assistance in final preparation for the press. I wish to thank Professor H. S. Jennings and Dr. H. H. Goddard, who have given generous permission to copy certain diagrams, as well as The Outlook Company and The Macmillan Company for the use of figures 24 and 66, respectively. The fact that all the suggestions which were at various times offered by my kindly critics have not been incorporated in the text, absolves them from responsibility for whatever remains. H. E. W. Providence, R. I., September, 1912. CONTENTS CHAPTER I. Introduction. 1. The triangle of life ' . 2. A definition of heredity 3. The maintenance of life 4. Somatoplasm and germplasm II. The Carriers of the Heritage. 1. Introduction 2. The cell theory . 3. A typical cell 4. Mitosis .... 5. Amitosis .... 6. Sexual reproduction . 7. Maturation 8. Fertilization 9. Parthenogenesis 10. The hereditary bridge 11. The determiners of heredity 12. The chromosome theory 13. The enzyme theory of heredity 14. Conclusion III. Variation. 1. The most invariable thing in nature 2. The universality of variation 3. The kinds of variation with respect to a. Nature h. Duplication c. Utility d. Direction in evolution e. Source /. Normality g. Degree of continuity . h. Character . i. Relation to an average standard j. Heritability . . . . their PAGE 1 4 5 10 14 14 15 18 20 20 22 24 26 27 28 29 33 35 36 37 38 39 39 39 40 40 40 41 41 41 IX CONTENTS CHAPTER PAGE 4. Methods of studying variation 42 5. Biometry 42 6. Fluctuating variation 43 7. The interpretation of variation curves ... 47 a. Relative variability ...... 47 b. Bimodal curves ...... 47 c. Skew polygons ....... 50 8. Graduated and integral variations .... 52 9. The causes of variation ...... 52 a. Darwin's attitude ...... 52 b. Lamarck's attitude ...... 53 c. Weismann's attitude . .... 55 d. Bateson's attitude ...... 55 IV. Mutation. 1. The mutation theory ....... 56 2. Mutation and fluctuation ...... 57 3. Freaks 58 4. Kinds of mutation ....... 59 5. Species and varieties ....... 60 6. Plant mutations found in nature .... 63 7. Lamarck's evening primrose ..... 64 8. Some mutations among animals .... 67 9. Possible explanations of mutation .... 69 10. A summary of the mutation theory .... 72 V. The Inheritance of Acquired Characters. 1. Summary of preceding chapters ..... 74 2. The bearing of this chapter upon genetics ... 75 3. The importance of the question ..... 75 4. An historical sketch of opinion ..... 76 5. Confusion in definitions ...... 77 6. Weismann's conception of acquired characters . . 78 7. The distinction between germinal and somatic charac- ters 79 8. What variations reappear ? . . . . .80 9. What may cause germplasm to vary or to acquire new characters ? . . . . . .81 10. Weismann's reasons for doubting the inheritance of acquired characters ..... 84 CONTENTS XI CHAPTER 11. No known mechanism for impressing germplasm with somatic characters . . . . . 12. Evidence for the inheritance of acquired characters inconclusive . a. Mutilations b. Environmental effects c. The effects of use or disuse d. Disease transmission 13. The germplasm theory sufficient to facts of heredity . 14. The opposition to Weismann . 15. Conclusion .... account for the PAGE 84 86 87 88 91 92 94 95 96 VI. The Pure Line. 1. The unit character method of attack 2. Galton's law of regression 3. The idea of the pure line 4. Johanssen's nineteen beans 5. Cases similar to Johanssen's pure lines 6. Tower's potato-beetles 7. Jennings' work on Paramecium 8. Phenotypical and genotypical distinctions 9. The distinction between a population and a pure line 10. Pure lines and natural selection . . . . VII. Segregation and Dominance. 1. Methods of studying heredity . 2. The melting-pot of cross-breeding a. Blending inheritance b. Alternative inheritance . c. Particulate inheritance 3. Johann Gregor Mendel 4. Mendel's experiments on garden peas 5. Some further instances of Mendel's law 6. The principle of segregation 7. Homozygotes and heterozygotes 8. The identification of a heterozygote . 9. The presence and absence hypothesis 10. Dihybrids 97 98 102 103 107 108 110 113 115 118 120 120 121 121 121 123 124 128 130 131 132 132 133 Xll CONTENTS CHAPTER PAGE 11. The case of the trihybrid 140 12. Conclusion 143 13. Summary 144 ones new Vin. Reversion to Old Types and the Making of New Ones. 1. The distinction between reversion and atavism 2. False reversion ..... a. Arrested development 6. Vestigial structures .... c. Acquired characters resembling ancestral d. Convergent variation e. Regression ..... 3. Explanation of reversion .... 4. Some methods of improving old and establishing types The method of Hallet The method of Rimpau . . The method of de Vries . The method of Vilmorin . The method of Johanssen The method of Burbank The method of Mendel . 5. The factor hypothesis .... a. Bateson's sweet peas Castle's agouti guinea-pigs Cuenot's spotted mice Miss Durham's intensified mice Castle's brown-eyed, yellow guinea-pigs 6. Rabbit phenotypes ..... 7. The kinds of gray rabbits 8. Conclusion ...... a. b. c. d. e. /. 9- b. c. d. e. 146 149 149 149 150 150 151 151 152 152 153 154 155 155 156 157 159 160 163 164 165 166 169 171 173 IX. Blending Inheritance. 1. The relative value of dominance and segregation . 174 2. Imperfect dominance ...... 175 3. Delayed dominance ...... 177 4. " Reversed " dominance 178 CONTENTS Xlll CHAPTER 5. Potency . a. Total potency b. Partial potency c. Failure of potency 6. Blending inheritance 7. The case of rabbit ears 8. The Nilsson-Ehle discovery 9. The application of Nilsson-Ehle's explanation to the case of rabbit ear-length 10. Human skin color . X. The Determination of Sex. 1. Speculations, ancient and modern 2. The nutrition theory 3. The statistical study of sex 4. Monochorial twins . 5. Selectiv^e fertilization 6. The neo-Mendelian theory of sex a. Microscopical evidence 1. The "x" chromosome 2. Various forms of x chromosomes 3. Sex chromosomes in parthenogenesis b. Castration and regeneration experiments c. Sex-limited inheritance 1. Color-blindness 2. The English currant-moth d. Behavior of hermaphrodites in heredity 7. Conclusion XI. The Application to Man. 1. 2. 3. PAGE 179 179 180 180 182 183 186 193 196 197 198 200 201 202 205 207 207 208 210 210 213 214 216 220 222 5. 6. The application of genetics to man .... 224 Modifying factors in the case of man . . . 225 Experiments in human heredity .... 227 a. The Jukes 227 6. The descendants of Jonathan Edwards . . 228 c. The Kallikak family 229 Moral and mental characters behave like physical ones 230 The character of human traits ..... 231 Hereditarv defects . . . . . . . 232 XIV CONTENTS CHAPTEB 7. The control of defects .... 8. Inbreeding 9. Experiments to test the effects of inbreeding 10. The influence of proximity 11. Inbreeding in the Ught of Mendelism XII. Human Conservation. 1. How mankind may be improved 2. More facts needed ...... 3. More appHcation of what we know necessary . 4. The restriction of undesirable germplasm through a. The control of immigration More discriminating marriage laws An educated sentiment . The segregation of defectives . Drastic measures The conservation of desirable germplasm By subsidizing the fit By enlarging individual opportunity By preventing germinal waste . 1. Preventable death . 2. Social hindrances 6. Who shall sit in judgment ? h. c. d. e. a. b. c. XIII, Bibliography XIV. Index PAGE 235 238 240 241 242 244 245 247 248 250 251 252 254 255 256 258 258 258 259 260 263 265 GENETICS GENETICS CHAPTER I INTRODUCTION 1. The Triangle of Life Within a generation the center of biological inter- est has gradually been swinging from the origin of species to the origin of the individual. The nine- teenth century was Darwin's century. His monu- mental work "On the Origin of Species by Means of Natural Selection," which appeared in 1859, not only dominated the biological sciences but also influenced profoundly many other realms of thought, partic- ularly those of philosophy and theology. Now, at the beginning of the twentieth century, a particular emphasis is being laid upon the study of heredity. The interpretation of investigations along this line of research has been made possible through the cumulative discoveries of many things that were not known in Darwin's day. Trained students have been patiently and persistently bend- ing over improved microscopes, untangling the mysteries of the cell, while an increasing host of in- vestigators, inspired by the Austrian monk Mendel, have been industriously devoting their energies to nOF&RTY UBRARY n, C. State C»Be|i 2 GENETICS breeding animals and plants with an insight denied to breeders of preceding centuries. The study of the origin of the individual, which has grown out of the more general consideration of the origin of species, forms the subject-matter of heredity, or, to use the more definitive word of Bate- son, of genetics. It is not with the individual as a whole that HER! T A G E Fig. 1. — The triangle of life. genetics is chiefly concerned, but rather with char- acteristics of the individual. Three factors determine the characteristics of an individual, namely, environment, training, and heri- tage as expressed diagrammatically in Figure 1. It may indeed be said that an individual is the result of the interaction of these three factors since he may be modified by changing any one of them. Although no one factor can possibly be omitted, the student of genetics places the emphasis upon heritage as the factor of greatest importance. Heritage, or INTRODUCTION 3 *' blood," expresses the innate equipment of the indi- vidual. It is what he actually is even before birth. It is his nature. It is what determines whether he shall be a beast or a man. Consequently in the diagram (Fig. 1), the triangle of life is represented as resting solidly upon the side marked "heritage" for its foundation. Environment and training, although indispensable, are both factors which are subsequent and secondary. Environment is what the individual has, for example, housing, food, friends and enemies, surrounding aids which may help him and obstacles which he must overcome. It is the particular world into which he comes, the measure of opportunity given to his particular heritage. Training, or education, on the other hand, repre- sents what the individual does with his heritage and environment. Lacking a suitable environment a good heritage may come to naught like good seed sown upon stony ground, but it is nevertheless true that the best environment cannot make up for defective heritage or develop wheat from tares. The absence of sufficient training or exercise even when the environment is suitable and the endowment of inheritance is ample will result in an individual who falls short of his possibilities, while no amount of education can develop a man out of the heritage of a beast. Consequently the biologist holds that, although what an individual has and does is un- questionably of great importance, particularly to the individual himself, what he is, is far more important 4 GENETICS in the long run. Improved environment and educa- tion may better the generation already born. Im- proved blood will better every generation to come. What, then, is this "blood" or heritage? Ex- actly what is meant by heredity ? 2. A Definition of Heredity Professor Castle, in his recent book on " Heredity in Relation to Evolution and Animal Breeding," has defined heredity as " organic resemblance based on descent." The son resembles his father because he is a " chip off the old block." It would be still nearer the truth to say that the son resembles his father because they are both chips from the same block, since the actual characters of parents are never trans- mitted to their offspring in the same way that real estate or personal property is passed on from one generation to another. When the son is said to have his father's hair and his mother's complexion it does not mean that paternal baldness and a vanish- ing maternal complexion are the inevitable conse- quences. Biological inheritance is more comparable to the handing dow^n from father to son of some valuable patent right or manufacturing plant by means of which the son, in due course of time, may develop an independent fortune of his own, resembling in charac- ter and extent the parental fortune similarly derived although not identical with it. So it comes about that "organic resemblance" INTRODUCTION 5 between father and son, as well as that which often appears between nephew and uncle or even more remote relatives, is due not to a direct entail of the characteristics in question, but to the fact that the characteristics are "based on descent" from a common source. In other words, an *' hereditary character" of any kind is not an entity or unit which is handed down from generation to generation, but is rather a method of reaction of the organism to the constellation of external environmental factors under which the organism lives. To unravel the golden threads of inheritance which have bound us all together in the past, as well as to learn how to weave upon the loom of the future, not only those old patterns in plants and animals and men which have already proven worth while, but also to create new organic designs of an excellence hitherto impossible or undreamed of, is the inspiring task before the geneticist to-day. 3. The Maintenance of Life So far as we know, every living thing on the earth to-day has arisen from some preceding form of life. How the first spark of life began will probably always be a matter of pure speculation. Whether the beginnings of what is called life came through space from other worlds on meteoric wings, as Lord Kelvin has suggested ; whether it was spontaneously generated on the spot out of lifeless components ; or whether life itself was the original condition of 6 GENETICS matter, and the one thing that must be explained is not the origin of Hfe, but of the non-Hving, no one can say. Leaving aside the first speculation as un- tenable and the third as irrational, since it jars so sadly with what astronomers tell us of the probable evolution of worlds, the theory of spontaneous gener- ation seems to be the last resort to which to turn. In prescientific days this idea of spontaneous generation presented no great difficulties to our imaginative and credulous ancestors. John Milton, with the assurance of an eye-witness, thus described the inorganic origin of a lion : — *' The grassy clods now calved ; now half appears The tawny lion, pawing to get free His hinder parts — then springs as broke from bonds, And rampant shakes his brindled mane." (" Paradise Lost," Book VII, line 543.) Ovid also in his ''Metamorphoses," not to mention a more familiar instance, easily succeeded in creating mankind from the humble stones tossed by the juggling hands of Deucalion and Pyrrha. Although under former conditions on the earth it might have been possible for life to have originated spontaneously, and although it may yet be possible to produce life from inorganic materials in the labora- tory or elsewhere, the exhaustive work of Pasteur, Tyndall and others effectually demonstrated a genera- tion ago that to-day living matter always arises from preceding living matter and this conclusion is gener- ally accepted as an axiom in genetics. INTRODUCTION 7 There are various methods of producing more life, given a nest-egg of living substance with which to start. Any organism, whether plant or animal, is continually transforming inorganic and dead material into living tissue. Through the process of repair, for example, an injury to a form as highly developed even as man is frequently made good, if it is not too extensive, as in the case of a skin wound. When the intake of non-living material is in excess of the outgo, growth results, w4th the consequence that more living substance is built up than existed before. Thus a fragment of a living sponge or a piece of a begonia leaf are each sufficient to restore a duplicate of the original organism. A process similar to the repair of the begonia leaf is that employed so effectively in the great groups of the one-celled animals and plants, the Protozoa and Protophyta, by means of which their numbers are maintained. These one-celled organisms multiply by fission, that is, by equal division into halves, and each half then grows to the size of the parent organism from which it sprang. When two daughter pro- tozoans are thus formed, they are essentially orphans because they have no parents, alive or dead. The parental substance in such a process, along with the regulating power necessary to reorganization, goes over bodily into the next generation in the forma- tion of the daughter-cells, leaving usually no re- mains whatever behind. In primitive forms of this description, continuous life is the natural order, and death, when it does occur, is, as Weismann has 8 GENETICS pointed out, accidental and quite outside the plan of nature. In these cases it is easy to see the reason for "or- ganic resemblance" between successive generations. Parent and offspring are successive manifestations of the same thing, just as the begonia plant, restored from a fragment of a begonia leaf, is simply an ex- tension of the original plant. Many modifications of the process of multiplica- tion by fission occur, all of them, however, agreeing in the fundamental principle that the progeny re- semble the parents because they are pieces of the parents. Thus the greening apple maintains its individuality although coming from thousands of different trees, because all of these trees through the asexual process of grafting are continuations of the one original Rhode Island greening tree grown by Dr. Solomon Drowne in the town of Foster, nearly a century ago. Again, certain fresh-water sponges and bryozoans, quite unlike any of their marine relatives, keep a foothold from year to year within their particular shallow fresh-water habitats by isolating well pro- tected fragments of themselves in the form of gemmules and statohlasts. These structures may drop to the muddy bottom and live in a dormant condition throughout the icy winter when it would not be possible for the entire organism to survive near the surface. In order to meet the conditions imposed by winter, however, these fragments have become so modified INTRODUCTION 9 as temporarily to lose their likeness to the parent generation, although readily regaining that likeness when springtime brings the opportunity. The unity of two succeeding generations, although interrupted by the temporary interposition of something ap- parently different in the form of gemmules or stato- blasts, is thus essentially maintained. The bryozoan colonies of two successive seasons in a fresh-water pond may be regarded as parts of the same identical colony, since they present an "organic resemblance based on descent," although the sole representatives of the parent colony during midwinter may be the sparks of life locked up within the statoblasts buried in the mud. Similarly, the asexual spores of many plants, such as molds, mosses and ferns, may be regarded as gemmules reduced to the lowest terms, namely, to single cells. As in the preceding cases so in this instance the resemblance of the offspring which may arise from these spores, to the parents which pro- duced them, is due to the essential material identity of two generations. These illustrations of heredity in its simplest mani- festations give the key to "organic resemblance" higher up in the scale. Sexual reproduction is no less plainly the direct continuation of life though in this instance tivo sporelike fragments out of one generation contribute to form the new individual of the next generation instead of one fragment. In all cases there is a material continuity between succeeding generations. Offspring become thus an extension of 10 GENETICS a single parent or of two parents, while heredity is simply "organic resemblance based on descent." 4. Somatoplasm and Germplasm In forms that reproduce sexually there theoretically occurs a differentiation of the body substance into what Weismann terms somatoplasm and germplasm. The somatoplasm includes the body tissues, that is, the bulk of the individual, which is fated in the course of events to complete a life-cycle and die. The germplasm, on the contrary, is the immortal fragment freighted with the power to duplicate the whole organism and which, barring accident, is des- tined to live on and give rise to new individuals. The germplasm thus carries potencies for develop- ing both germplasm and somatoplasm, while the somatoplasm, according to this conception, has only the power to reproduce more of its own kind. More- over, the germplasm is not formed afresh in each gen- eration, neither does it arise anew when the individual reaches sexual maturity, but it is a continuous sub- stance present from the beginning. Although this theory of the continuity of the germplasm has been actually demonstrated in comparatively few instances, all the facts we know concerning the behavior of the germinal substance are consistent with it. In many of the Protozoa the entire organism is possibly comparable to germplasm, but in all forms of life that are compounded of several cells the germ- plasm is probably set aside early in the development INTRODUCTION 11 of the individual, and this remains undifferentiated, or in reserve, like a savings-bank account put by for a rainy day, while the somato- plasm is expended in the immediate demands of the tissues that make up the individual. In one instance at least, that of the nematode worm Ascaris, as con- firmed by Boveri, this splitting off or isolation of the germplasm occurs with the very first cleavage of the fertilized egg into the two-celled stage, when one of the two cells forms the future germplasm, while the other differen- tiates by succes- sive divisions into the animal itself. Thus there results a continuous stream of germ- plasm, receiving contributions from other germ- Germplasm V Somatoplasm Fig. 2. — Scheme to illustrate the continuity of the germplasm. Each triangle represents an individual made up of germplasm (dotted) and somatoplasm (undotted) . The beginning of the life cycle of each individual is represented at the apex of the triangle where germplasm and somatoplasm are both present. As the indi- \ddual develops each of these component parts increases. In sexual reproduction the germ- plasms of two individuals unite into a common stream to which the somatoplasm makes no contribution. The continuity of the germ- plasm is shown by the heavy broken line into which run collateral contributions from suc- cessive sexual reproductions. 12 GENETICS plasmal streams at the time of sexual reproduction, as shown diagrammatically in Figure 2, in which individuals are represented by triangles. From this continuous stream of germplasm there split off at successive intervals complexes of somatoplasm, or *' individuals," which go so far on the road of speciali- zation into tissues that the power to be " born again " is lost, and so after a time they die, while the germplasm, held in reserve, lives on. This is what is meant by saying that a father and son owe their mutual resemblance to the fact that they are chips off the same block rather than by saying that the son is a chip off the paternal block. Both somatoplasms are developments at different inter- vals from the same continuous stream of germplasm instead of one somatoplasm being derived from a preceding one. As a matter of fact the germplasm from which the son arises is modified by the addition of a maternal contribution, so that father and son in reality hold the same relation to each other that half- brothers do. From the point of view of genetics, then, the real mission of the somatoplasm, which is so marvelously differentiated into all the various forms that we call animals and plants, is simply to serve as a temporary domicile for the immortal germplasm. Thus the parent becomes as it were the "trustee of the germ- plasm," but not the producer of the offspring. In the light of these preliminary explanations it is plain that the hopeful point of attack in the science of genetics must inevitably be the germplasm which INTRODUCTION 13 is the source, or point of departure, in the formation of each new individual, rather than the somatoplasm, which represents the end stages of the hereditary processes. This has not been the method of the past. The resemblances of the visible father and son have usually been traced instead of the character of their unseen germplasms. By following this old method, investigators have often been misled because the visible or apparent is not always the true index of what lies behind it. A gray and a white rabbit, for example, may produce some offspring that are entirely black just as two white-flowering sweet peas when crossed may sometimes produce purple blos- soms. Consequently it is a great fallacy to affirm that in heredity *'like produces like," since the op- posite is quite often the case. The new heredity, embodied in the science of genetics, attempts to go deeper than the surface appearance of the somatoplasm. It aims to get at the source or origin of organisms, that is, the germ- plasm which is the only connecting thread between succeeding generations of living forms. It is con- cerned not so much with somatoplasm, which repre- sents what the germplasm has done in the past, as with the germplasm and what it can do in the future. CHAPTER n THE CARRIERS OF THE HERITAGE 1. Introduction Heredity, as has been shown in the introductory chapter, is essentially a matter of continuity between succeeding generations of living organisms. This continuity may be direct, as when a mother protozoan divides into two daughters, or it may be indirect, as illustrated by the relationship of a father and son, an uncle and nephew, or any other relatives of varying degrees of kinship which, taken singly or collectively, are somatoplasms derived from a common stream of germplasm. It is the purpose of the present chapter to consider this material continuity between succeeding genera- tions and to discover, if possible, just what are the carriers of the heritage from one generation to another. To this end it will be necessary in the first place to take up what is meant by the "cell theory." 2. The Cell Theory In 1838-1839 the "cell theory" of Schleiden and Schwann, which aflBrms that all organisms, both plant and animal, are made up of cellular units, had its birth. 14 THE CARRIERS OF THE HERITAGE 15 Robert Hooke, as early as 1665, had described *' little boxes or cells distinguished from one another" which he saw in thin slices of cork, and to him is due the rather unfortunate use of the term "cell" which has survived in biological writings to this day. The reason this term is unfortunate is because walls, which are ordinarily the characteristic feature of any cell, such as a prison cell, are usually the least im- portant part of the structure of a living cell, often indeed being entirely absent. 3. A Typical Cell A typical undifferentiated cell is represented diagrammatically in Figure 3. Near the center of the cell the nucleus is shown surrounded by a Cefl wall C«^to plasm Centrosome Nuclear membrane --Nucleus .-Chromatin network Fig. 3. — Diagram of a typical cell. nuclear membrane. The nucleus, in common with the enveloping cytoplasm, is made up of living substance called protoplasm (Hugo von Mohl, 1846), and around the whole there is usually formed a 16 GENETICS wall or membrane which serves to separate one cell from another. Within the protoplasm there may be a considerable amount of non-living substance in the form of salts, pigments, oil-drops, water, and other inclusions of various kinds. The nucleus is to be regarded as the headquarters of the whole cell, since changes which the cell under- goes seem to be initiated in it, while cells deprived of their nuclei cannot long survive. A single instance will serve to show the vital part which the nucleus plays in the life-history of the cell. In 1883, Gruber found that after rocking a thin cover-glass back and forth in a drop of water containing a collection of the protozoan Stentor, which has a long chain-like nucleus, these tiny animals could thus be cut into fragments, which would in some instances recover from the operation and regenerate into complete individuals. Only those pieces, however, which contained a frag- ment of the nucleus regenerated into new Stentors, while pieces of relatively large size which lacked a frag- ment of nuclear substance verv soon disintegrated. The nucleus, it should be said, is made up of more than one substance, a fact that is easily demonstrated by processes of staining, in which certain dyes, through chemical union, stain a part but not the whole of the nuclear substance. The part most easily stained is called chromatin, that is "colored material," and during certain phases of cell life the chromatin masses together within the nucleus into visibly definite structures or bodies termed chromo- somes. THE CARRIERS OF THE HERITAGE 17 Throughout all the various cells that make up the individuals of any one species these chromosomes appear to be practically constant in number with some exceptions to be mentioned later in connection with sex. This law of the constant chromosome number for any species was first stated by Boveri in 1900. The chromosomes of different organisms vary in number from two in the worm Ascaris up to perhaps 1600, according to Haecker ('09), in certain radiolaria. Species which apparently are closely related may differ widely with respect to the number of their chromosomes, while species of unquestionably re- mote relationship may have an identical number of chromosomes in each of their cells. The number of chromosomes characteristic for a species, therefore, is in no way an index to the complexity or degree of differentiation of the species. Besides the nucleus there may often be identified in the cytoplasm of the animal cell a tiny body known as the centrosome. At certain times in the life-cycle of a cell the centrosome becomes the focal point of peculiar radiating lines, which play an important part in the behavior of the cell, particularly during the period of division. Every cell passes through a cycle of life which may be compared with that common to individuals. It is born from another cell ; passes through a vigorous youth characterized by growth and transformation ; attains maturity when the metamorphoses of its earlier life give place to a considerable degree of stability ; and finally, after a more or less extended 18 GENETICS period of normal activity old age ensues, and death completes the cycle. In most instances, however, before this final phase is reached, the cell gives place to daughter-cells through fission, after the manner of most protozoans, and a new cell cycle is begun. Sometimes the road of differentiation has been traveled so far that it is apparently impossible, as in the case of the complicated brain-cells, to retrace these steps of differentiation and begin again. In such instances the outfit of cells provided in the em- bryo determines the numerical limit of the cells available throughout life. When this supply is ex- hausted no more cells appear to replace those w^hich have been worn out. 4. Mitosis The ordinary process by which two cells are made out of one is termed mitosis. It occurs constantly, and particularly during growth, in all cellular organ- isms. A series of diagrams, modified from Boveri, illustrating the typical phases of mitosis is given in Figures 4 to 13. The resting cell (Fig. 4) is characterized by the presence of a nuclear membrane, a single centrosome, and by a chromatin network within the nucleus. In the beginning of the prophase (Fig. 5) the centrosome has divided into two parts, while in the early prophase (Fig. 6) the two centrosomes have moved farther apart and definite separate chromosomes have formed out of the chromatin network. The prophase proper (Fig. 7) is marked by the vanishing of the nuclear THE CARRIERS OF THE HERITAGE 19 membrane and the more compact form of the chromo- somes. At the end of the prophase (Fig. 8) the chro- mosomes have come to lie at the equator of the cell. F»g. 4. TFie resting cell Rg. 5. Be^innin^ Prophase Fl J.6. Earlij Prophase n^.7 Prophase ng.8. End of Prophase nj.9. Metaphase Pig. 10. Beginning Anaphase Rg.ll. Anaphase R^.12 Beginning Telophase n^.i3. End of Telophase Figs. 4-13. — Diagrams illustrating mitosis. After Boveri. being connected by the mantle fibers with the cen- trosomes, each of which has now come to occupy a polar position. In the metaphase (Fig. 9) the chromo- somes split lengthwise, and at the beginning of the 20 GENETICS anaphase (Fig. 10) these half chromosomes commence to separate from each other and to move toward the poles, while the mantle fibers shorten. During the anaphase (Fig. 11) the cell body lengthens and begins to divide, w^iile the migration of the half chromosomes tow^ard the poles is completed. In the beginning of the telophase (Fig. 12) the half chromosomes grow until they attain full size and the division of the cell body into two parts becomes complete. The mantle fibers have disappeared and the nuclear membrane begins to re-form around the chromosomes. Finally, at the end of the telophase (Fig. 13) the nuclear membrane becomes complete, the chromosomes break up into a chromatin network, and two resting cells take the place of the single one with which the process began (Fig. 4). 5. Amitosis Amitosis, or the formation of two cells from one without the machinery of mitosis, is comparatively rare. It occurs in certain rather isolated instances among animals and plants, particularly in old cells late in their life-cycle or in cells that are on the road to degeneration. When amitosis takes the place of the more elaborate process of mitosis it is fre- quently, though not always, a signal of the death- warrant for that particular cell. 6. Sexual Reproduction The mechanism by means of which two cells unite to make one in sexual reproduction is quite as com- THE CARRIERS OF THE HERITAGE 21 plicated as tliat of mitosis by which one cell is trans- formed into two. In sexual reproduction there are two kinds of germ- cells, the egg and the spermatozoan respectively, which take part in producing a new organism. These cells are structurally unlike each other in nearly every particular, but each is a true cell, which von Kolliker made clear as early as 1841, and each has typically the same number of chromosomes in its nucleus, a fact more recently determined by van Beneden in 1883. The egg-cell is often supplied with one or more envelopes of protective or nutritive function, and it is usually distended with stored up yolk, in consequence of which it is comparatively large and stationary. The result is that whatever locomotion is necessary to bring the two cells together for union devolves upon the sperm-cell. Consequently the sperm-cells are practically nuclei with locomotor tails of cyto- plasm, and frequently, in addition, with some struc- tural modification for boring a way into the egg-cell. They are, moreover, much more numerous than the egg-cells, so that although many go astray, never fulfilling their mission, the chances are nevertheless good that some one of them will reach the egg and effect fertilization. Ordinarily only one sperm enters the egg, but when several succeed in penetrating into the cyto- plasm only one proceeds to combine w^ith the egg nucleus, that is, only one sperm nucleus is normally concerned in the essential process of fertilization. 22 GENETICS ^ It was formerly thought by the school of "ovists" that in fertilization the essential process is a stimu- lation of the all important egg by the sperm. The opposing school of "spermists," on the other hand, regarded the egg simply as a nutritive cell the func- tion of which is to harbor the all important sperm. It is now known that both the egg- and the sperm-cell are equally concerned in fertilization, which consists in the union of their respective nuclei within the cytoplasm of the egg. 7. Maturation Certain preliminary changes of a preparatory nature, termed maturation, regularly precede the union of the nuclei of the two sex-cells in fertiliza- tion. These maturing changes result in reducing the outfit of chromosomes in each sex-cell to one half the original number, a proces which is necessary in order to maintain the chromosome count which is characteristic for any particular species and which is known to exist unbroken from generation to genera- tion. If there were no such reduction, then the fertilized egg, formed by the union of egg and sperm nuclei, would contain double the characteristic number of chromosomes, and during the formation of a new individual, the number in all the cells arising by mitosis from such a fertilized egg would like- wise be double. When the germ-cells of such indi- viduals unite in fertilization, the original number of chromosomes would be quadrupled, and so on in THE CARRIERS OF THE HERITAGE 23 geometric progression throughout subsequent genera- tions. In 1883, too late for Darwin to learn of it, van Beneden discovered the important fact that the Primordial Sex Ceils V Manvj similar cell divisions Maturation ^ — — = — =\^ Sperinatocyte 1- OocijU Spermatids 2.-Oocijte V '^bortiva E^* Fig. 14. — Scheme to illustrate maturation of germ-cells. mature germ-cells, as expected, actually contain only half the normal number of chromosomes. The mature egg- or sperm-cell, with half its normal number of chromosomes, is termed a gamete (marry- 24 GENETICS ing cell), while the fertilized egg which is formed by the union of two gametes (mature egg- and sperm- cell), and which consequently has the characteristic number of chromosomes, is called a zygote (yoked cell). A diagrammatic representation of the process of maturation is shown in Figure 14. The number of chromosomes (not shown in the diagram) remains constant in each germ-cell respec- tively until the division of spermatids into sperma- tozoa, and of the second oocytes into mature eggs and second polar cells, when it is reduced to one half the normal number. As spermatozoan and mature egg unite in fertilization, the original number of chromo- somes is restored in the fertilized egg (zygote) . 8. Fertilization The stages concerned in a typical case of fertiliza- tion, according to Boveri, are illustrated in Figures 15 to 23. In Figure 15 the "head" and the "middle piece" of the sperm-cell have penetrated into the egg cyto- plasm, while in Figure 16 the tail of the sperm-cell has become lost and the middle piece, which furnished the centrosome, has rotated 180° so that it lies between the nucleus, or head, of the sperm-cell and that of the egg-cell. Figure 17 shows an increase in the size of the sperm nucleus and a division of the centrosome into two parts which begin to migrate towards the poles. This process of polar migration of the centrosomes is carried further in Figure 18 as f JWPEMT UBRARt THE CARRIERS OF THE HERITAGE 25 fiig.18. Approach of Sperm Nucleus Pig.lS. Entn^ of Sperm Ff^lG. Loss of Sperm Tall R^.IZ Division of Cenlrosome f?g. 19, Increase of Sperm Fi^.ZO. Formation of Nucleus Chromosomes Fi*g.2.I. Splitting of Chromosomes H^. £2. Anaphase n^.SS.Two-celled Sla^e Figs. 15-23. — Diagrams illustrating fertilization. After Boveri. 26 GENETICS well as the increase in the size of the sperm nucleus, until in Figure 19 the process is complete so that the centrosomes have assumed a polar position and the sperm nucleus is equal in size to the egg nucleus and lies in contact with it. In Figure 20 the chro- matin network of the two nuclei has formed into an equal number of chromosomes which in each case is half the number characteristic for the species. Figure 21 shows the complete disappearance of the nuclear membrane, a process that had already begun in the preceding figure, and also the arrangement of the chromosomes, connected with mantle fibers, in the equatorial plane where the former split longitudinally. In Figure 22, when the half chromosomes thus formed pull apart and migrate toward the poles, the segmenta- tion of the fertilized egg has begun, and there finally occurs, as shown in Figure 23, the two-celled stage following fertilization in which each cell contains the normal number of chromosomes, half of which came from the egg and half from the sperm. 9. Parthenogenesis Fertilization is by no means an essential process in the formation of a new individual, even in those ani- mals which produce both eggs and sperms. Many animals and plants reproduce parthenogenetically, that is, the egg-cell may develop without first uniting with a sperm-cell. In these instances the chromo- somes of the egg are not halved during maturation, and the offspring, therefore, have the same number THE CARRIERS OF THE HERITAGE 27 of chromosomes as the parent, since they are simply fragments of the parent. Professor Loeb, by the use of certain chemicals, has succeeded in doing artificially what apparently is never accomplished in nature, namely, making an egg that normally requires fertilization develop par- thenogenetically. 10. The Hereditary Bridge Whatever may ultimately prove to be deter- miners of the hereditary characters which appear in successive generations, it is obvious that, in any event, such determiners must be located in the zygote, that is, in the fertilized egg. This single cell is the actual bridge of continuity between any parental and filial generation. Moreover, it is the 07ily bridge. In the majority of animals the egg develops en- tirely outside of and independent of the mother, thus limiting to the egg-cell itself all possible mater- nal contributions to the offspring. Although there is abundant evidence that half of the filial char- acteristics come from the male parent, the only actual fragment of the paternal organism given over to the new individual is the single sperm-cell, which unites with the egg in fertilization, and the whole of this even is not usually concerned in the process of fertilization. The entire factor of heritage is packed into the two germ-cells derived from the re- spective parents and, in all probability, into the nuclei of these germ-cells, since the nuclei are ap- 28 GENETICS parently the only portions of these cells that in- variably take part in fertilization. To the new individual developing by mitosis from the fertilized egg into an independent organism, the factors of environment and training referred to in Figure 1 are subsequently added. When it is remembered that the human egg- cell is only about 2Vth of an inch in diameter, a gigantic size as compared with that of the human sperm-cell, and, furthermore, when one passes in rapid review the marvelous array of characteristics which make up the sum total of what is obviously inherited in man, the wonder grows that so small a bridge can stand such an enormous traffic. A sharp-eyed patrol of this bridge as the strategic focus of heredity is proving to be one of the most effective points of attack in the entire campaign of genetics. It is not desirable at this time to discuss possible ways in which the determiners of the heritage, what- ever they may be, are originally packed into the germ-cells, for this question can be more conven- iently considered in a later connection. It is im- portant at present, however, to emphasize the ob- vious conclusion that determiners of heredity must inevitably be present in the germ-cells in order to account for the fact of ''organic resemblance based on descent" between parents and their progeny. 11. The Determiners of Heredity What are the determiners of hereditary qualities ? Do they actually exist in the germ-cells as visible THE CARRIERS OF THE HERITAGE 29 entities, and is there such a thing as a mechanical basis for heredity as the German embryologist Wilhelm His suggested years ago when he wrote : "It is a piece of unscientific mysticism to suppose that heredity will build up an organism without mechanical means" ? Can we find these determiners by the aid of microscopes and differential stains, or are they some sort of intangible entities, such as enzymes or hormones or the like, which only the chemist can detect ? Whatever the answer to these questions, it may at least be affirmed that the determiner represents the adult structure without resembling it. It is something which controls the unfolding of the developing or- ganism with respect to both quantity and quality, and which also governs the time and rate of appear- ance of its various characteristics so that certain combinations rather than others shall come about in definite sequence. To use the words of Conklin : "The mechanism of heredity is the mechanism of differentiation." 12. The Chromosome Theory Certain investigators, who seek a morphological basis for heredity, regard the chromosomes as the car- riers of the heritage; in other words, as the source of the determiners of ontogeny or the effective factors in the process of differentiation. A few of the grounds for this theory are briefly indicated below. First: In spite of the great relative difference in 30 GENETICS size between the egg-cell and the sperm-cell, in hered- ity the two are practically equivalent, as has been repeatedly shown by making reciprocal crosses be- tween the two sexes. The only features that are apparently alike in both the germ-cells are the chromosomes. The inference is, therefore, that they contain the determiners which are the causal factors for the equivalence of adult characters in heredity. The existence of an extra chromosome in probable connection with the matter of sex is, as will be pointed out later, an exception to the exact chromosome equivalence of the two sexes, which only goes to strengthen the supposition that the chromosomes are the carriers of hereditary qualities since extra chromo- somes are always associated with the character of sex. Second: The process of maturation, which always results in halving the chromosome material of the germ-cells as a preliminary step to fertilization, is a series of complicated manoeuvers not practised by other cells. During this process no other part of the cells appears to play so consistent and important a role as the chromosomes. Provided they act as hereditary carriers, their peculiar behavior during maturation is just what is needed to bring together an entire complement of hereditary determiners out of partial contributions from tw^o parental sources. Third : Sometimes abnormal fertilization occurs, as in the case when two or more sperm-cells, instead of one, enter the egg cytoplasm and unite with the egg nucleus. This unusual performance has been artifi- cially induced by chemical means in the case of sea- THE CARRIERS OF THE HERITAGE 31 urchins' eggs. The fertilized egg, or zygote, thus formed with an excess of male chromosomes, re- sults in the development of abnormal larvae. It is thought that a causal connection may exist, there- fore, between the additional male chromosomes in the fertilized ovum and the abnormalities of the progeny. Fourth: The fact that chromosomes may retain their individuality throughout the complicated phases of mitosis, as has been proven in some instances, agrees with the corresponding fact that certain characteristics of the somatoplasm maintain their individuality from generation to generation. Moreover, certain chromosomes in the fertilized egg have been identified with particular features in the adult developing from that egg. Tennent sum- marizes his recent work on Echinoderms (1912) by the statement that from a knowledge of the chromosomes in the parental germ-cells, particular characters in the adult hybrids may be predicted, and, conversely, that from the appearance of sexually mature hybrids the character of certain chromosomes in their germ-cells may be predicted. Again, the correlation of a particular chromosome in the germ-cells with a definite adult character, namely sex, has been repeatedly demonstrated in connection with the so-called "extra chromosome" to which reference has already been made. Fifth: Finally, excellent evidence of a definite causal connection between certain chromosomes of the germ-cells and particular somatic characters has 32 GENETICS been furnished by certain critical experiments upon the eggs of sea-urchins. Boveri found that he was able in some instances to shake out the nuclei bodily, chromosomes and all, from the mature eggs of the sea-urchin, Splicer echinus, and when there was added in sea water to such enucleated eggs the sperm-cells of an entirely different genus of sea-urchin, namely. Echinus, the Echinus sperm-cells entered the Sphcer- echinus eggs, which had been robbed of their nuclei, and from this peculiar combination larvae developed which exliibited only Echinus characters! Such cuniulative circumstantial evidence as the foregoing has convinced many that in the chromo- somes we have visibly before us the carriers of heredity. Several biologists, however, raise an objecting voice to this theory, protesting against the mo- nopoly of the heritage by the chromosomes. They point out that there always exists an intimate physiological relationship between the nucleus and the cytoplasm, and that it is unreasonable to expect the isolation of one from the other, since the two must always act together as parts of an organic cell unit. In sexual reproduction, moreover, some small amount at least of spermatic cytoplasm in the form of the so-called "middle piece," which is situated between the head and the tail of the sperm-cell (Fig. 15), may enter the egg about to be fertilized along with the sperm " head " or nucleus, containing the chromosomes. In this way the cytoplasm of the THE CARRIERS OF THE HERITAGE 33 male sperm-cell may not necessarily be entirely excluded from taking part in the formation of the zygote. As a matter of fact, this extra-nuclear part of the sperm-cell sometimes apparently forms the centrosome of the fertilized egg and in consequence may have a hand, as well as the nucleus with the chromosomes, in determining what follows. 13. The Enzyme Theory of Heredity It is not unlikely that the key to this whole prob- lem will be furnished by the biochemists and that the final analysis of the matter of the heritage-carriers will be seen to be chemical rather than morpho- logical in nature. It has been found that the blood of greyhounds and dachshunds is chemically different, although from a morphological point of view it is apparently identical. The idea of "individual albumen" or "protein specificity" for each animal of a species, to say nothing of the animals of different species, has been advanced as not improbable. Miescher has shown that an albumen compound having only forty carbon atoms, a number by no means unusual, would make possible a million com- binations of atoms or isomers. The possibilities in this direction seem to be un- limited if we take into consideration those invisible actuators of chemical processes, the enzymes, which the chemist brings forw^ard with the prodigality of an astronomer dealing in star-dust, to explain dif- ferent chemical reactions. 34 GENETICS Montgomery has suggested that the chromosomes themselves may be masses of enzymes although, ac- cording to the chemist, enzymes are not morpho- logical entities, since they seem to be able to flourish and maintain their identity while bringing about chemical reactions in their neighborhood without being visibly demonstrable. As said before, it is quite likely that in the final analysis heredity will be reduced to a series of chemi- cal reactions dependent upon the manner in which various enzymes initiate, retard, or accelerate suc- cessive chemical combinations occurring in the pro- toplasm. When the same enzymes act upon the same chemical combinations in successive genera- tions, they bring about that "organic resemblance" known as heredity. E. B. Wilson, whose brilliant work in the entire field of cell activity makes it possible for him to speak with authority, has recently said : "The es- sential conclusion that is indicated by cytological study of the nuclear substance is, that it is an ag- gregate of many different chemical components which do not constitute a mere mechanical mixture, but a complex organic system and which undergo perfectly ordered processes of segregation and dis- tribution in the cycle of cell life. That these sub- stances play some definite role in determination is not mere assumption, but a conclusion based upon direct cytological experiment and one that finds support in the results of modern chemical research." THE CARRIERS OF THE HERITAGE 35 14. Conclusion The supposition that the chromosomes, with cer- tain chemical reservations, are the morphological carriers of the heritage forms an excellent working hypothesis, and this chapter may suitably be closed with a second quotation from Professor Wilson. " In my view studies in this field are at the present time most likely to be advanced by adopting the comparatively simple hypothesis that the nuclear substances are actual factors of reaction by virtue of their specific chemical properties ; and I think that it has already helped us to gain a clearer view of some of the most puzzling problems of genetics." CHAPTER III VARIATION 1. The Most Invariable Thing in Nature In the introductory chapter it was shown that "organic resemblance based on descent," by which is meant heredity, is due principally to the fact that offspring are material continuations of their parents and consequently may be expected to be like them. The fact that this is the case in the great majority of instances has given rise to the popular formula, "like produces like," as a rule of heredity. But this formula by no means always fits the facts. Like often produces something apparently unlike. For instance, two brown-eyed parents may produce a blue-eyed child, although brown-eyed children are more usual from such a parentage. It is a common experience, indeed, for breeders of plants and animals ^ to meet with continual difficulties in getting or- ganisms to "breed true." On the other hand, it is exactly these variations which so constantly interfere with breeding true that furnish the sole foothold for improvement. If all organisms did breed strictly true, one generation could not stand on the shoulders of the preceding generation, and there would be no evolutionary advance. 36 VARIATION 37 The most invariable thing in nature is variation. This fact is at once the hope and the despair of the breeder who seeks to hold fast to whatever he has found that is good and at the same time tries to find something better. When the similarities and dissimilarities between succeeding generations are clear, then heredity can be explained. The entire subject of variation is intimately and inevitably bound up with any consideration of genetics. 2. The Universality of Variation Much of the variation in nature is patent to the most casual observer, but it requires a trained eye to see the universal extent of many minor differences. A flock of sheep may all look alike to a passing stran- ger, but not to the man who tends them. A dozen blue violet plants from different localities might easily be identified by the amateur botanist as be- longing to the same species when, to a specialist on the genus Viola, unmistakable differences would doubtless be clearly apparent. The fact that every attempt at an intimate ac- quaintance with any group of organisms whatsoever invariably reveals previously unrecognized varia- tions, indicates that variability is much more wide- spread in nature than is commonly believed. , The key to Japanese art, as pointed out by Dr. Nitobe, consists in being natural and in faithfully copying nature. It is for this reason that the Jap- anese artist makes each object that he produces 38 GENETICS unique, because nature herself, whom he strives to follow, never duplicates anything. The Bertillon system of personal identification is based upon the constancy of minor variations found in each individual. Its importance is shown in Figure 24. The faces of the criminals there pictured would be easily confused by the ordinary observer, but an examination of their thumb prints shows unmis- takable differences between these three individuals. 3. Kinds of Variation A brief enumeration of some of the kinds of varia- tion will reveal their diverse character. a. With respect to their nature variations may be morphological, physiological, or psychological. Under morphological variations are included differences in shape, size, or pattern as well as differences in number and relation of constituent parts. Differences in activity are of a physiological nature. Many animals in captivity are less fertile than when free, while different individuals are well known to vary widely with respect to their susceptibility to disease. Nageli, for example, reports the presence of tubercles in 97 per cent of the cases in 1^ve hundred autopsies, although a majority of the deaths in ques- tion was not due to tuberculosis at all, — a fact which indicates a great diversity in the resistance of differ- ent individuals to the tubercle bacillus. Psychological variations in man, such as those which determine the disposition or mental traits of individuals, are apparent to every one. VARIATION 39 b. With respect to their duplication variations may be single or multiple. A legless lamb ^ is an ex- ample of a single variation or *' sport." Four-leaved clovers, on the contrary, are multiple for the reason that this variation, although not common, neverthe- less occurs frequently. c. With respect to their utility variations may be useful, indifferent, or harmful to the organism possess- ing them. Useful variations are of the kind empha- sized by Darwin as being effectively made use of in natural selection. Indifferent variations, on the other hand, are those which apparently do not play an important part in the welfare of their possessor, as, for example, the color of the eyes or of the hair. Finally, the degree of degeneration in certain organs may be cited as an illustration of harmful variations. The amount of closure of the opening from the in- testine into the vermiform appendix in man is an ex- ample of a harmful variation, since the larger the opening, the greater is the liability to appendicitis. d. With respect to their direction in evolution varia- tions may be either definite (orthogenetic) or indefinite (fortuitous) . Paleontology furnishes numerous instances of the former category, such as the series of variations from a pentadactyl ancestor, all apparently tending in one direction, which have culminated in the one- toed horse. The fact that the paleontologist deals historically with a completed phylogenetic series in which the side lines lack prominence, while the suc- 1 "A Peculiar Legless Lamb." Stockard. Biol. Bull, xiii, p. 288. 40 GENETICS cessful line stands out with distinctness, makes it easy for him to view successive variations as orthogenetic, that is, as definitely directed in one course either through intrinsic (Nageli) or extrinsic (Eimer) causes. Fortuitous or chance variations in all possible directions furnish the repertory of opportunity, according to Darwin, from which natural selection picks out those best adapted to survive in the strug- gle for existence. e. With respect to their source^ variations may be somatic or germinal. Somatic, or body variations, arise as modifications due to environmental factors. They are individual differences which may be quite transitory in nature, while germinal variations may arise without regard to the environment, are deep- seated, and of racial rather than of individual sig- nificance. /. With respect to their normality variations may fall within expected extremes and thus be considered normal, or they may be outside of reasonable expec- tations and consequently be reckoned as abnormal, as in the case of a two-headed calf. g. With respect to the degree of their continuity varia- tions may form a continuous series, grading into each other by intermediate steps, or they may be discon- tinuous in character. An example of continuous variation is the height of any hundred men one might chance to meet, which w ould probably represent all intermediate grades from the highest among the hundred to the lowest. The number of segments in the abdomen of a VARIATION 41 shrimp, on the other hand, which may, for instance, be either eight or nine but cannot be halfway between, illustrates what is meant by discontinuous variation. The widespread occurrence of this later category of variations has been pointed out by Bateson in his encyclopedic volume "On Materials for the Study of Variation." h. With respect to their character variations may be quantitative or qualitative. A six-rayed starfish represents a quantitative variation from the normal number of five rays, whereas a red variety of a flower may differ chemically from a blue variety, or a bitter fruit may differ from a sweet fruit in a qualitative way dependent upon the chemical constitution of the fruit in question. i. With respect to their relation to an average stand- ard variations may have a fluctuating distribution around an arithmetical mean, as when some of the offspring have more and some less of the parental character, or the variations in the progeny may all center about a new average quite distinct from the parental standard and consequently come under the head of mutations, j. Finally, and most important in the present connection, with respect to heritability, variations may possess the power to reappear in subsequent genera- tions, or they may lack that power. It is this aspect of variability which bears most directly upon genetics. Other possible categories might be mentioned, but a suflicient number have been cited to show the great diversity of variations in general. 42 GENETICS 4. Methods of Studying Variations Roughly stated, there are three ways of studying variations : first, Darwin's method of observation and the description of more or less isolated cases ; second, Galton's biometric method of statistical inquir}^ ; and third, Mendel's experimental method. The second of these methods will be considered in this chapter. 5. Biometry The new science of biometry, that is, the applica- tion of statistical methods to biological facts, has been developed within recent years. Sir Francis Galton, Darwin's distinguished cousin, may be re- garded as the pioneer in this field of research, while Karl Pearson and his disciples constitute the modern school of biometricians. Although mathematical analysis of biological data when not sufficiently ballasted by biological analysis of the same facts may sometimes lead the investigator astray, yet often the only way to for- mulate certain truths or to analyze data of some kinds is by resort to statistical methods. Biome- tricians are quite right in insisting that it is frequently necessary to go further than the fact of variation, which may be apparent from the inspection of an individual case, and to deal with cumulative evidence as presented through statistical analysis. In matters of heredity, however, facts as they occur in single cases and definite pedigrees seem to VARIATION 43 offer a more hopeful line of approach than statistical generalizations. It is better to become acquainted with the real parent than to evolve a hypothetical "mid-parent" mathematically. In this connection it is well always to bear in mind the warning of Johanssen, himself a past master in biometry, when he writes : '^ Mit Mathematik niclit als Mathematik treiben wir unsere Studien." 6. Fluctuating Variation With respect to any measurable character there are bound to be deviations from an average con- dition. According to the mathematical laws of chance these deviations sometimes are plus and sometimes minus, and consequently they may be termed fluctuating variations. Pearson gives as a simple illustration of fluctuating variation the number of ribs present in two sets of beech-leaves, as shown below. These sets were taken from two different trees, and each contains twenty- six leaves. Number of Ribs 13 14 16 16 17 18 19 Total First tree . . . Second tree . . 3 4 1 9 4 8 7 2 9 4 1 26 26 Total .... 3 4 10 12 9 9 4 1 It will at once be seen that, while certain leaves might well belong to either tree, as, for example, those with sixteen ribs, the entire group of leaves from 44 GENETICS either tree is unlike that of the other tree. In the first instance the number of ribs fluctuates around Number oi Individuals 35 p. UsTor Constants 30 £5 Arithmetical MeanCA.M)=49 Mode CM) -5 Average Deviation (A.D.')=.5£ Standard Deviation (o-) = .7e4 — CoeKicient oi VariabilitY (C.V.)= 1^1 Formulae zo - 15 W t A.D. = S(x.f) =T cr = rzcx^.f) - c.\/ = n A.M. 2 = sum x= deviation of the class from A.M. f = number in the class n =• total number Number of Rays £ 3 4 5 Fig. 25. — The fluctuating variability of starfish rays. Goldschmidt. 6 7 From data by eighteen as the commonest kind ; in the second case, around fifteen. Such a difference could not easily VARIATION 45 be detected or expressed by any other method than the statistical one. Again, in the case of forty-seven starfishes all of which were collected from one locality the variation in the number of rays proved to be, according to Goldschmidt, an amount indicated graphically in Figure 25, where the data are arranged in the form of a so-called frequency polygon or curve. From such a polygon certain constants may be computed which conveniently express in a single number, for purposes of abstract comparison, dis- tinctions that otherwise could be handled only in the most indefinite way. Thus in this instance the arithmetical mean, ex- pressed by the hypothetical number 4.915, a number which of course does not actually occur in nature, is simply the average number of rays in forty-seven starfishes selected at random. The mode which represents the group containing the largest number of individuals of a kind, namely, thirty out of forty-seven, is five in this particular polygon. The average deviation, which is an index of the amount of variation going on among the starfishes in question, is .52. In other words, .52 is the average amount that each individual starfish deviates from the arithmetical mean of 4.915. Although the one seven-rayed starfish which happens to be in the lot varies from the standard of 4.915 to the extent of 2.085 (7 — 4.915) rays, there are thirty five-rayed starfishes which vary only .085 (5 — 4.915) of a ray, 46 GENETICS and consequently the average of the entire forty- seven amounts to .52 of a ray. In another collec- tion of starfishes where either more seven-rayed or two-rayed specimens might be present, the average deviation would probably be greater. By computing the average deviation, therefore, and using it as the criterion of variation, a compar- ison of the variability of organisms that have been taken from different localities or subjected to differ- ent conditions can be definitely expressed. A measure of variability more commonly in use by biometricians, since for mathematical reasons it is more accurate, is the standard deviation. This is the square root of the sum of all the deviations squared, according to the formula in which x represents the deviation of each class from the arithmetical mean ; /, the number of individuals in each separate class ; 2, the sum of the classes ; and n, the total number of individuals.^ In the present instance the standard deviation is .724, an arbitrary number that has valuable sig- nificance only when brought into comparison with standard deviations similarly derived from other groups of starfishes. Such a variation polygon as the above expresses the law that the farther any single group is from the ^For directions explaining the use of such formulae see Davenport's " Statistical Methods." VARIATION 47 mean of all the groups making up the pol^^gon, the fewer will be the individuals that represent it. 7. The Interpretation of Variation Polygons a. Relative Variability The statistical determination of the relative vari- ability of two lots of organisms with respect to a certain character may be illustrated by the case of the oyster-borer snail, Urosalpinx cinereus, as seen in the accompanying table. Atlantic and Pacific Shells Compared Locality Woods Hole West Shore Penzance Point Nobska Point Nobska Point Nobska Point Barnacle Beach Big Wepecket Mid-Wepecket Average for Mass Cali- f Belmont Beds fornia '[ San Francisco Bay Average for Cal. Difference Number OF Shells 1,001 1,002 1,002 1,001 496 998 1,006 500 1,008 520 A.M. 58.928 61.718 61.737 61.944 66.944 63.932 57.426 57.606 61.066 59.051 60.892 59.664 2.339 2.737 2.152 2.234 2.366 2.604 2.052 2.098 2.335 3.023 3.361 3.138 .803 Prob- able Error ±.0352 ±.0412 ±.0324 ±.0337 ±.0507 ±.0393 ±.0308 ±.0447 ±.0386 ±.0454 ±.0703 ±.0538 The obvious conclusion to be drawn from this table is that the snails which were unintentionally carried from the Atlantic coast to California in the 48 GENETICS transplantation of oysters show more variation in their new habitat than they did in the old one with respect to the particular character measured, namely, the relative size of the mouth aperture compared with the height of the entire shell. ^ b. Bimodal Curves Sometimes two conspicuous modes make their ap- pearance in a frequency polygon, as Jennings found, f|Jonib»r ■ of individuals 25 - t 20 - \ J5 - / \ f\ JO - a\ / b \ 5 _. 1 / 1 1 t > 1 1 1 ! 1 ._i^. L ZO 24- £8 32 36 40 44 48 5S 56 60 64 Fig. 26. — The body width of a population of the protozoan Paramecium, showing a polygon with two modes. A, Parajnecium aurelia. B, Paramecium caudatum. After Jennings. for example, in measuring the body width of a popu- lation of the protozoan Paramecium (Fig. 26). 1 " Variation in Urosalpinx." Walter. Amer. Nat. 1910, Vol. XLIV, pp. 577-594. VARIATION 49 It was subsequently found that the two modes in this polygon were due to the fact that the material in question was a mixture of two closely related species, Paramecium aurelia and Paramecium cauda- tum, the individuals of which arranged themselves around their own mean in each instance. Number q\ leaves \z w ;o 9 8 7 6 5 4 3 2 1 Number ot J 3 H 15 16 ribs 17 18 19 10 EI Fig. 27. — The ribs of leaves from two beech trees. When put together they form a polygon which does not reveal its double origin. From data by Pearson. Although such an explanation does not always turn out to be the right one, the biometrician is led to suspect when a two or more moded polygon appears that he is dealing with a mixture of more than one kind of material, each of which fluctuates around its own average. Heterogeneous material, it should be noted, does not always give a bimodal curve. For example, if Pearson's two lots of beech leaves mentioned E 50 GENETICS above are mixed together, they form a regular series from the inspection of which no one could infer their double origin. (See the heavy line in Figure 27.) c. Shew Polygons The direction in which variations are tending may sometimes be determined by the statistical method. As an illustration of this may be cited the number of ray florets on 1000 white daisies {Chrys- anthemum leucanthemum) , 500 of which were col- lected at random by the writer from a small patch in a swampy meadow in northern Vermont, while the other 500 were selected in the same random manner upon the same day from a dry hillside pas- ture hardly more than a stone's throw distant. Among these two lots of daisies the number of ray florets varies from twelve to thirty-eight and their frequency polygons, as shown in Figure 28, form what are termed "skew polygons," because the mode in each case lies considerably to one side of the arith- metical mean. It will be seen that lot A from the swampy meadow, which in spite of the greater fertility of the soil and the unquestionably greater luxuriance of the plants themselves, produced heads with fewer florets, fluctuates around the number 21, while the dry pasture population B, characterized by blossoms which were in general noticeably smaller, fluctuates around the number 34. VARIATION 51 The habitats of the two lots were so near together, however, that there was probably a considerable intermixture of the two types, as shown by the tendency of each polygon to produce a second mode. 90 85 80 15 70 65 60 55 50 45 40 35 30 25 20 ;5 10 5 Fig, Kv!''® 12 13 H 15 16 17 18 19 EO Z1 It 23 24 25 26 27 28 29 30 31 32 33 34 JS 36 37 38 39 . 28. — Variation in the ray florets of the white daisy (Chrysanthe- mum leucanthemum) . A, from a swampy meadow. B, from a hillside pasture near by. Both the polygons are " skew " because in each case there is an admixture of the other type. The distinction between the two types is due to heredity rather than to environment. Thus the A polygon shows that there is an increasing tendency or variability in the twenty-one floret type toward the thirty-four floret type, due probably in this particular instance to invasion resulting from the proximity of the B colony. 52 GENETICS 8. Graduated, and Integral Variations It is comparatively simple to treat statistically integral variations, illustrations of which have been given in the case of beech-leaf ribs, starfish rays, and daisy florets, all of which are characters that can be readily counted. In the same way any measurable character, such as the size of snail shells, may fall into easily limited groups, as, for example, 10 to 11 mm., 11 to 12 mm., 12 to 13 mm., etc. It is somewhat more difficult to classify variations when color or pattern is the character in question, since it then becomes necessary to define certain arbitrary limits for each class of the series within which to group the individual variants. Tower, in his famous researches on potato-beetles, encountered variations in the pigmentation of the pronotum all the way from entire absence of color to complete pigmentation. By cutting up this continuous series of variations into arbitrary groups of equal extent, however, it was quite possible to arrange the data so that they could be statistically treated just as conveniently as the integral variations mentioned above. Groups or classes of this kind are termed graduated variations. 9. The Causes of Variation With respect to the causes of variation authori- tative biologists have taken different points of view. a. Darwin considered variations as axiomatic. An axiom is self-evident, requiring no explanation. VARIATION 53 The absence of variations in organisms rather than the occurrence of variations is, from this point of view, the phenomenon requiring an explanation. Although Darwin himself spent some time in point- ing out the universal occurrence of variability, he accepted it as a primary fact and proceeded from it as a starting point without attempting to seek its causes. b. Lamarck and his followers have regarded the causes of variation either as extrinsic, that is, refer- able to external factors making up the environment of the organism, or as intrinsic or physiological, that is, based upon the efforts which an organism puts forth to fit into its particular environment success- fully. The causes of variation are to be sought ac- t ■oO -L. -L. Af. M, t o/ 50 35 '^0 45 50 55 60 65 70 75 80 85 90 95 100 /o < Ratio of height of head to length of shell > Fig. 29. — Schematic curve of the head height of Hyalodaphnia under various conditions of nourishment. Adapted from Woltereck. cording to the Lamarckian school, in the "environ- ment" and "training" sides of the triangle of life rather than in the "heritage" side (Fig. l)c For example, Woltereck, by controlling the single 54 GENETICS extrinsic factor of food supply, was able to modify the height of the *'head" of the microscopic fresh- water crustacean, Hyalodaphnia, in the remarkable manner indicated in Figure 29. When poor food Number of Flowers Number cf L oUaens lo 6 5 10 9 6 10 9 8 Fig. 30. — Variations in the number of stamens in the flowers of the " live- for-ever" (Sedum spectabile) under various controlled conditions. For detailed description, see text. After Klebs. was supplied, the percentage of the head height to that of the body averaged hardly forty, while with rich food it was increased to over ninety. Similarly Klebs succeeded in changing at will the number of stamens in the common " live-for-ever," Sedum spectabile, by manipulating the environment in which the plants were kept. Some of his results are shown in Figure 30. Polygon A combines the data for 4260 flowers which were raised in well-fer- tilized dry soil under bright light ; polygon B repre- sents 4000 flowers grown in a moist greenhouse under red light ; and polygon C includes 4390 flowers VARIATION 55 from well-fertilized soil in moist hotbed conditions under a weak light. c. Weismann, on the contrary, believes that the causes of variation, at least of heritable variations, are intrinsic or inborn in the germplasm. His con- ception of sexual reproduction is that it is a device for doubling the possible variations in the offspring by the mingling of two strains of germplasm {am- phimixis). By far the greater number of observa- tions recorded go to substantiate this theory. Tower found among his potato-beetles, for exam- ple, that two strains reared in the same environment showed striking differences in variation, — a fact necessarily due to intrinsic rather than to extrinsic factors. Similar cases may be recalled by any one. d. Lastly, Bateson, whose work "On Materials for the Study of Variation" already cited is a classic, takes the agnostic attitude that it is rather futile to guess at the causes of variation before the facts are well in hand. He consequently discourages such attempts by saying : "Inquiry into the causes of variation is, in my judgment, premature." In conclusion, the words of Darwin written half a century ago — "Our ignorance of the laws of variation is profound" — may still be appropriately quoted, notwithstanding the fact that in biometry we have at least an excellent analytical method by means of which considerable insight into variation is being gained. CHAPTER IV MUTATION 1. The Mutation Theory Among the possible kinds of variation already hinted at are so-called mutations which are clearly defined from the fluctuating variations to which reference has just been made. Darwin was fully aware of the existence of muta- tions or "sports" and incidentally gave time to their consideration, but the great task which he accomplished in such a masterly manner was to overthrow the widespread and deep-seated belief of his day in a sudden special creation of distinct species. To this end he marshaled evidence in support of the gradual transition of one species into another, emphasizing fluctuations rather than muta- tions which seemed to him to play a minor role in the origin of species. It remained for the Dutch botanist Hugo de Vries to analyze the character of mutations. There is something distinctly suggestive of Darwin's method in the fact that de Vries worked in silence for twenty years before he gave to the world the *'Mutations- theorie" with which his name will forever be con- nected. 56 MUTATION 51 2. Mutation and Fluctuation A mutation is something qualitatively new that appears abruptly without transitions and which breeds true from the very first. To use the musician's phraseology, it is not a variation elaborated upon an old theme, which would correspond to a fluctuating variation, but it is an entirely new theme. The difference between mutations and fluctuating varia- tions is generally not one of quantity or magnitude, although it sometimes may be so, — since muta- tions are often much smaller than fluctuations. Mutations are discontinuous in the same sense that chemical combinations, such as carbon monoxide (CO) and carbon dioxide (CO2), are discontinuous, but the leap from one to the other may be so small that frequently it is diflicult to ascertain by inspec- tion alone whether the difference is due to a mutation or a fluctuation. The test comes in breeding, for the progeny of a fluctuation will vary around the old average of the parental generation, while the progeny of a mutation will vary around a new average, set by the mutation itself. When a series of mutations is treated statistically, it does not arrange in frequency polygons as readily as a series of fluctuations do. The latter mass around the average standard according to the laws of chance much in the same way that a hundred shots by a good marksman may center around a bull's- eye. Mutations never act in this way. They find no correspondence even with wild shots at the bull's- 58 GENETICS eye. They are shots directed at a different target altogether. To the student of heredity there are two distinc- tions of prime importance with respect to mutations. First, that they usually appear full-fledged without preparatory stages, and second, that they breed true from the start. Fluctuations, on the contrary, ordin- arily "revert" to the parental type in subsequent gen- erations. The great practical importance to the breeder of a knowledge of these heritable mutations is at once apparent. 3. Freaks A further distinction should be made between mutations and so-called freaks or monstrosities, namely, that the former breed true, while the latter do not. A human physical deformity, such as a club-foot, for example, or a humped back, is not a mutation, because it does not reappear as a heritable character. Variations of this kind are not predeter- mined in the germplasm, but are usually instances of something that went wrong during the development of the individual somatoplasm. Thus, among normally "right-handed" snails "left- handed" individuals have occasionally been dis- covered which, when bred, were found to produce all normal "right-handed" progeny. They are therefore not mutations at all, but freaks or mon- strosities due probably to some unusual occurrence early in the cleavage stages of the embryo. MUTATION 59 4. Kinds of Mutation De Vries has classified mutations according to their component units into three categories: pro- gressive, regressive, and degressive. Progressive mutations are signaHzed by the addi- tion of a new character to the sum of complex char- acters making up the individual. If rumor may be believed, iViine Boleyn, the second in the interesting series of wives of Henry VIII, was a progressive mu- tant with respect to at least three characters, for she is said to have been possessed of an extra finger on each hand, supernumerary mammae, and extra teeth. Evidences that each of these three characters occur as heritable mutations is presented in Davenport's *' Heredity in Relation to Eugenics." Regressive mutations are characterized by the dropping out of something. Thus albinism is caused by the absence of pigment or color. Albinic mutants which breed true are well known, particularly among mammals, such as rats, mice, rabbits, cats, guinea- pigs, and even man himself. Degressive mutations include cases of the return of a character which was formerly present in the past history of the race, but which has for generations been absent or latent. Castle's four-toed race of guinea-pigs furnishes an example of this class of mutations. In 1906 Professor Castle discovered a newly born guinea-pig in one of his pens with four toes on each hind foot, from which he has successfully established a four-toed race. The hypothetical an- 60 GENETICS cestor of the rodents probably had five toes on each foot, but the normal number in modern guinea-pigs is four on each of the front feet and three on the hind feet. The individual from which Castle has bred a four-toed race exhibited a degressive muta- tion, tending toward the ancestral type. 5. Species and Varieties The doctors have always disagreed regarding a definition of species. What determines the exclusive boundaries that shall isolate from their fellows any particular group of animals or plants has long been a mooted question, and still remains so. The Linnsean concept of a species was that of an exclusive caste of individuals, inflexibly demarked, over w^hose high barriers no nondescript tramps would dare attempt to climb. When an entomolo- gist of the old Linnsean school encountered an insect which did not conform to the morphological tradi- tions of its fellows, the frequent fate of such a non- conformist was to perish under the boot-heel rather than to find sanctuary in the cabinet of the pre- served. Since it was an exception, and a violator of the divine law of the fixity of species, it deserved to be annihilated! Those were hard days both for heretics and for mutations. The method of the older school of systematists may be described as one which emphasized differences and put up barriers that should keep the unlike apart, at the same time allowing only "birds of a feather to flock together." It was a brave and sue- MUTATION 61 cessful attempt to bring order out of chaos by classi- fying the living world, and it served its purpose well until Darwin's idea of half a century ago, that the origin of all species is from preceding species, put an entirely new face upon the whole matter. Organ- isms of different species were found to be related to one another, and even man could no longer escape acknowledging his poor animal relations. As a consequence, likenesses rather than differences there- after claimed the most attention. During the reconstruction of phylogenetic trees, which seized the imagination and became the prin- cipal business of biologists as soon as the " Origin of Species " was made common property, the crotched sticks in the woodpile of organisms, that had hitherto been largely discarded, were most eagerly sought after. It was just these scraggly sticks, that were neither trunk nor limb-wood but combinations of both, which told the story of continuity and were indis- pensable in building up a reunited whole. As the analysis of the living world gradually came to shift from species to individuals, it was shown that individuals may be regarded simply as ag- gregates of unit characters which may combine so variously that it becomes more and more diffi- cult to maintain constant barriers of any kind be- tween the groups of individuals arbitrarily called species. The old species of the systematist, upon analysis into their respective unit characters, dissolve into numerous "elementary species" and "varieties" dif- 62 GENETICS fering from species perhaps only by the addition or subtraction of a single character, and thus the Fig. 31. — Diagram to illustrate various ideas about "species." Under Species A are represented two groups of individuals which are near enough alike to be placed within a single species, but which are suffi- ciently unlike each other to constitute the " sub-species " or "varie- ties " of Darwin. Under Species B are various groups of individuals distinguished from each other by the addition or loss of one or more characters. These groups represent the "elementary species" and "varieties" of de Vries. 'The "barrier of Linnaeus" attempted to separate species absolutely from each other. Darwin sought to find loopholes in this barrier. To-day attention is directed rather to the relation between individuals than to the boundaries between species. possibilities of analytical classification have become almost limitless. An elementary species, according to de Vries, is a progressive mutation differing from the type species MUTATION 63 by the addition of at least a single character, while varieties are regressive mutations distinguished from the parent type by the loss of at least one character. Both breed true to their respective modifications. These different concepts of what constitutes a species, illustrated diagrammatically in Figure 31, pave the way for a better understanding of muta- tions in connection with heredity. 6. Plant Mutations found in Nature The oldest known authenticated case of a plant mutation is the often cited instance of the *' fringed celandine," Chelidonium laciniatum, which made its appearance in the garden of the Heidelberg apothe- cary Sprengel in 1590 among plants of the "greater celandine," Chelidonium majus. The fringed cel- andine bred true at once and is now a widespread and well-known species. The purple beech has appeared historically as a mutant among ordinary beeches upon at least three occasions in widely separated localities, and it has always given rise to a constant progeny. The "Shirley poppy," notable for its remarkable range of color, originated from a single plant of the small red poppy, Papaver rhoeas, which is commonly found in English cornfields. Instances are known of double flowers among roses, azaleas, stocks, carnations, primroses, petunias, etc., arising from single flowering plants, the seeds of which in turn produce double flowers. 64 GENETICS 7. Lamarck's Evening Primrose The most widely known plant mutations are the progeny of Lamarck's evening primrose, Oenothera lamarckiana, because it was these plants that led de Vries to formulate his mutation theory. It is believed by botanists in general that this plant is a native of the southern United States, al- though it is now, so far as is knowm, extinct as a wild species in America, and native specimens are included in but few American herbaria. It was exported to London as a garden plant about 1860, and from thence it spread to the continent, where, escaping from gardens, it became wild in at least one locality neg.r Hilversum, a few miles from Amsterdam. Here, in an abandoned potato field, it fell under the seeing eye of Hugo de Vries in 1885, and now both botanist and primrose are famous. De Vries found among these escaped plants not only 0. lamarckiana, but also two other kinds or mutants, 0. brevistylis, characterized by short- styled flowers, and 0. Ice vi folia, w^iich has smooth leaves. These tw^o were entirely new species hitherto unknown at the great botanical clearing-houses of Paris, Leyden, and the Kew Gardens. Since the seeds of the (Enothera are produced by self-fertilized flowers, de Vries felt safe in regard- ing these plants as mutants rather than hybrids, and he continued to study them with especial care. Transplanting the mutants along with representa- tives of 0. lamarckiana to his private gardens in MUTATION 65 Amsterdam, where it was possible to maintain them in normal healthy condition, de Vries was able to follow their individual histories with certainty. He found that, out of 54,343 plants of the species 0. lamarckiana grown during eight years, there ap- peared 837 mutants comprising seven different ele- mentary species, all of which, with the exception of 0. scintillans, bred true. See table. Mutants of (Enothera lamarckiana Generation "Born" characters are constitutional, having their origin in the germplasm itself. They are never Weismannian acquired characters and may be illustrated by eye-color, mental disposition, or facial features. Lightning calculators and musical prodi- gies may have their gifts developed and enlarged. 80 GENETICS but the fact that their talent is nevertheless an unmistakable gift, and not an acquisition, remains true. "Achieved" characters are functional and are gained by exercise. Many things are achieved, however, which are not acquired characters, as, for instance, wealth, reputation, or an education. Not any of these are biological characters, and therefore we are not concerned with them in this connection, although in the case of education it should be noticed that the mental exercise necessary to bring about a trained mind, if not the subject matter of the edu- cation itself, is distinctly an acquired character of the "achieved" type. "Thrust" characters are the results of environ- ment. They are acquired w^ithout functional activity on the part of the organism and usually in spite of anything the organism can do to prevent. Some- times these characters are thrust upon the individual before birth, as in the case of blindness caused by parental gonorrhoea or tuberculosis arising from uterine infection, in which case they are termed con- geiiital characters. Congenital or prenatal characters, however, are in no way the same as germinal characters, for they fall just as truly into the category of acquired variations as do those which make their appearance in later life. 8. What Variations reappear ? Returning now to Montgomery's question, — "WTiat kinds of acquired characters are inherited.^" — it INHERITANCE OF ACQUIRED CHARACTERS 81 is apparent that only the "born" ones can be, which have their roots in the germplasm whence tlie new individual arises, and that "achievements" and "thrusts," in order to reappear in the succeeding gen- eration, can do so only by first becoming incorporated in the germplasm. Any character that is not acquired must have been present in the germplasm from which the organism arose, as there is no transfer of characters between organisms except through the germ-cells. Thus it is evident that the only inherited acquisitions are those which, either primarily or secondarily, bring about variation in the germplasm. Such temporary acquisi- tions as a coat of tan or a display of freckles do not impress the germplasm, for when the cause that in- cites their appearance is removed, they soon vanish. 9. What may cause Germplasm to vary or to ACQUIRE New Characters ? The causes which bring about changes in the germ- plasm may be either internal or external. Of possible internal causes may be mentioned first the "amphimixis" of Weismann, that is, the mixture of two nearly related strains of germplasm in sexual reproduction within a species, or secondly, the mixture of two more remotely related strains resulting in hy- bridization. In either case the strain of germplasm undergoes a shake-up that may result at least in new combinations of characters, if not in the production of entirely new characters. This recombination of G 82 GENETICS characters in amphimixis and hybridization will re- ceive further attention in a later chapter. The fact that successive parthenogenetic genera- tions, in which amphimixis does not of course occur, may show a larger degree of variability than sexually produced generations, indicates that amphimixis in itself is by no means sufficient to account for all kinds of variations. The abrupt way, for instance, in which mutations appear in apparent independence of external influences suggests that there may be some internal factor, as yet unknow^n, acting directly through the germplasm, regardless of external causes. The assumption of an unknown factor does not necessarily imply a return to "vitalism," which is so elusive of experimental test and hence so unsatisfac- tory to the scientific mind, nor does it admit, simply because this factor is at present an unknown quantity, that it is consequently doomed to remain so. It is easily conceivable that the external factors acting upon the germplasm may be grouped into two alternative classes : first, external factors that act upon the somatoplasm and through the agency of the somatoplasm affect the germplasm ; and second, those that act directly upon the germplasm without neces- sarily at the same time influencing the somatoplasm. The first category, that of somatic modifications which leave their impress upon the germplasm, in- cludes true acquired characters according to our definition, while the second, which includes cases of the direct influence of external stimuli upon the INHERITANCE OF ACQUIRED CHARACTERS 83 germplasm, regardless of any simultaneous modifica- tion of the somatoplasm, must be excluded as irrele- vant to a discussion of the heritability of acquired characters in the Weismannian sense, since thev are not somatic modifications at all. Many instances of direct influence of external stimuli upon germplasm are known in biological literature, and these have led to some of the misunder- standings concerning the "interminable question" of the inheritance of acquired characters. MacDougall, for example, was able by injecting certain salts into the carpels of plants to stimulate the germplasm of the forming seeds so directly that a progeny of modified character was produced which, in succeeding generations, bred true to the newly in- duced character. Sitkowski, also, fed the caterpillars of the moth Tineola bisellieUa with, an aniline dye (Sudan red III), obtaining therefrom, instead of the normal whitish ones, moths that laid colored eggs, and these in turn hatched into caterpillars still tinged with the color of the red dye. Riddle, with guinea-pigs, and Gage, with poultry, obtained quite similar results. This is an instance of what has been termed *' parallel induc- tion " where somatoplasm and germplasm are affected together by an external factor, as opposed to " somatic induction " or Weismannian acquired characters in which the germplasm is secondarily influenced through, or by the agency of, the somatoplasm. 84 GENETICS 10. Weismann's Reasons for doubting the Inher- itance OF Acquired Characters Weismann's reasons for questioning the popularly accepted view that acquired characters are inherited may be briefly stated as follows : — First, there is no known mechanism whereby somatic characters may be transferred to the germ- cells. Second, the evidence that such a transfer actually does occur is inconclusive and unsatisfactory. Third, the theory of the continuity of the germ- plasm is sufficient to account for the facts of heredity without assuming the inheritance of acquired somatic characters. Let us examine these three statements a little more closely. 11. No Known Mechanism for impressing the Germplasm with Somatic Characters The somatoplasm is something that has traveled out from the original fundamental germplasm along the paths of differentiation and elaboration. The more complex the body cells become, that is, the more successive modifications they undergo, the more difficult it is for these somatic cells to return to their original primitive estate. In many lower forms of life where cell elaboration is not so great, a part lost by amputation is often regenerated, but this process is not possible in higher INHERITANCE OF ACQUIRED CHARACTERS 85 forms where the parts represent cell complexes too hopelessly differentiated to begin anew the unfolding sequences of their elaboration. This difficulty was a very real one in the mind of that famous nocturnal inquirer Nicodemus when he asked: "How can a man be born when he is old ? Can he enter a second time into his mother's womb and be born ?" Not only the development of the race which we call evolution, but also the determination of the individual in heredity, is a chain of omvard-moving sequences like the succession of events in history. It is hard to see how recent events can influence pre- ceding events. It is hard to see how the water that has gone over the dam can return and affect the flow of the river upstream in any direct way. It is like- wise hard to see how differentiated somatoplasm, which represents the end stage of a successive series of modifications, can make any definite impress upon the original germplasmal sources from which it arose. Darwin felt this difficulty and presented with apolo- gies his provisional hypothesis of pangenesis in which he assumed that every bodily part sends contributions to the germ-cells in the form of "gemmules." These gemmules, or hypothetical somatic delegates, then reconstruct in the germ-cells the characters of the en- tire body, including acquired modifications as well as all others, and thus there is no reason why acquired characters cannot readily be transmitted. Unfortu- nately there is no tangible basis in fact for this de- lightfully simple explanation to rest upon. It is a theory assuming that all parental somatic cells take 86 GENETICS part in the formation of the new individual, hence it was called *' pangenesis," or origin from all. Nothing we have subsequently learned of minute cell structure favors this hypothesis, while many facts go quite against it. Moreover, it is directly opposed to the theory of the continuity of germplasm so convincingly set forth later on by Weismann. Darwin indeed advanced it only in the most tenta- tive way, being entirely ready to see it abandoned at any time for something better. It at least per- formed one valuable service to science, namely, that of demonstrating how far investigators were from an adequate conception of any means by which somatic modifications might become incorporated in the germ-cells. We must acknowledge, however, with Lloyd Morgan that the fact that a mechanism for the trans- fer of somatic characters to the germ-cells has not been discovered, is not proof that such a mechanism does not exist. It may simply be beyond our present powers of penetration. 12. Evidence for Transmission of Acquired Characters Inconclusive The evidence for the inheritance of acquired characters was, for a long time, taken for granted. This theory was the most obvious explanation of many facts and so was accepted without question. An obvious interpretation, however, is not always the correct one. The sun appears to go around the earth, but astronomers assure us that it does not. INHERITANCE OF ACQUIRED CHARACTERS 87 When Weismann began to sift the evidence for the inheritance of acquired characters, he found tliat it was largely based upon opinion rather than fact, much like the popular belief with regard to prenatal influences and birthmarks, or the causation of warts by handling toads. The supposed evidence for the inheritance of ac- quired characters falls chiefly into four categories : — a. Mutilations ; h. Environmental effects; c. The effects of use or disuse ; d. The transmission of disease. a. Mutilations It is fortunate that the sons of warriors do not inherit their fathers' honorable scars of battle, else we would now be a race of cripples. The feet of Chinese women of certain classes have for centuries been mutilated into deformity by band- aging, without the mutilation in any way becoming an inherited character. The same result is also true of circumcision, a mutilation practised from ancient times by the Jews and certain other Eastern peoples. The progressive degeneration or crippling of the little toe in man has been explained as the inheritance of the cramping effect of shoes upon generations of shoe wearers, but, as Wiedersheim has pointed out, the fact that Egyptian mummies show the same crippling of the little toe is unfavorable to this hypothesis, for no ancient Egyptian could 88 GENETICS ever be accused of wearing shoes or of having had shoe- wearing ancestors. Sheep and horses with docked tails as well as dogs with trimmed ears never produce young having the parental deformity. Weismann's classic experiment with mice, an experiment subsequently confirmed by others, is additional negative evidence upon this same point. What Weismann did was to breed mice whose tails had been cut off short at birth. He continued this decaudalization through twenty-two generations with absolutely no effect upon the tail-length of the new-born mice. One may see in the catacombs of the Zoologisches Institut at Freiburg, filed carefully away on shelves, as a " document," long rows of labeled bottles containing the fifteen hundred and ninety-two martyrs to science which made up the twenty-two generations of mice in this famous experiment. Conklin has hit the nail upon the head with respect to mutilations by saying: "Wooden legs are not in- herited, but wooden heads are." b. Environmental Effects Trees deformed by prevailing winds, like the willows that line the canals in Belgium and Holland, or storm-crippled trees along the exposed seacoast are not known to produce a modified progeny when their adverse environmental conditions are removed. Similarly, the persistent sunburn of Eng- lishmen long resident in India does not reappear in their children born in England. INHERITANCE OF ACQUIRED CHARACTERS 89 Sumner kept mice in a constant but abnormally high temperature of 26° C. with the result that the ears, tail, and feet grew noticeably larger than in control animals kept in ordinary lower temperatures, while at the same time the general hairiness of the body decreased. It should be remembered, however, that mice are mammals which pass through an ex- tended uterine existence, so that it is easy to see how the offspring in this case were subjected to the same excessive temperature as the parents for a period sufficient to amply account for their subsequent variation when removed to a normal environ- ment. Lederbaur finds that the wayside weed Capsella, which in the course of many years has gradually crept along the roadside up into an Alpine habitat and there "acquired" Alpine characters, upon being transplanted to the lowlands retains its Alpine modifications. Although this case has been cited as an authentic instance of the inheritance of ac- quired characters, is it not possible that the conquest of the Alps by Capsella has been due, in the course of time, not to the inheritance of acquired characters at all, but to a gradual natural selection of just those germinal variations which best fitted it to cope with Alpine conditions until, finally, a strain of germplasm producing somatoplasm suitable to Alpine conditions has been isolated in the form of an elementary species derived from the original type ? If this is what has happened, of course such germplasm would give rise to Alpine plants whether individual plants grew 90 GENETICS to maturity near the snow-line or in the warm valleys at a lower altitude. Marie von Chauvin was able, by decreasing the amount of water in an aquarium, to transform the gill-breathing salamander Axolotl into the land form, Ainhly stoma, which in its adult form has no gills, but breathes by means of lungs. Both of these forms are sexually mature, reproducing their like, and had long been recognized by systematists as distinct species. More recently Kammerer, by similarly reducing the water supply, succeeded in transforming Sala- mandra maculosa, a salamander that normally pro- duces about seventy eggs which, when hatched in water, become gill-breathing tadpoles, into a sala- mander producing only two to seven young which are born alive without gills and are able to live out of water entirely, in damp situations. These land- adapted offspring, moreover, when supplied with abundant water, produce in turn tadpoles which spend days only, instead of months, in the water under- going their metamorphosis, thus showing an appar- ent inheritance of an acquired character. It should be pointed out, however, that in these cases the gill-breathing forms in each instance rep- resent a case of arrested development. Axolotl is simply a larval form of Amhlystoma that, under nor- mal conditions of an abundant water environment and high temperature, gets no further in its meta- morphosis than the tadpole stage, when it produces eggs and sperms and finishes its life story. A change INHERITANCE OF ACQUIRED CHARACTERS 91 in environment simply permits the life-cycle to go on further. Changing from gill-breathing to lung- breathing is not, therefore, an acquired character, but a purely germinal character that may be either blocked or released by changing conditions in the environment. c. The Effects of Use or Disuse The callosities on the end of a violinist's left-hand fingers are acquired by use, but they are not inherited. There are callosities on the knees of the wart-hog, Phacechcerus, which are also apparently the result of use, for these animals kneel as they root for a living in the African forests, and have done so for untold generations. It has been noticed that young wart- hogs as soon as they are born possess the callosities, so that this instance looks like one of inheritance of a character acquired through use or exercise. The skin on the soles of human feet is thicker than the skin elsewhere, and by use it becomes still thicker. This is apparently another instance of the same sort. The writer has observed, however, that a cross sec- tion through the foot of a "mud puppy," Necturus maculatus, shows a much thickened sole. Necturus, it should be noted, is a very primitive salamander living always under water and never using the soles of its feet in any way to bear its weight, nor is it reasonable to suppose that it ever had any ancestors who did so, for the hands and feet of the Amphibia are the most primitive and ancient hands and feet to be found in the animal kingdom without any known 92 GENETICS ancestral types. The thickening of the skin on the sole of the mud puppy's feet must be due, therefore, to germinal determiners and is in no way an acquisi- tion through use. The same may also be true of the wart-hog's knees and of human soles. The strong arm, the skilled hand, and the trained ear are not inherited. They have always to be reacquired in each succeeding generation just as surely as the ability to walk, or to read and write. Herbert Spencer has defined instinct as "inherited habit." But surely those instincts which determine a single isolated action during the lifetime of the individual, such as the spinning of a peculiar cocoon, cannot be the result of habit, since habits are formed only through repeated action. If, then, some in- stincts require a different explanation from that of "inherited habit," may it not be likely that all in- stincts do ? Dr. Hodge, who succeeded in hatching tame quail chicks out of "wild" eggs, asks the perti- nent question: "How can Si fear hatch out of an eggV The habit of wildness, particularly with precocial chicks like quails, may, under an inciting environment, be very soon established, but it is diffi- cult to see how caution, gained by the experience of the parents, can find its way into the fertilized egg. d. Disease Transmission Many diseases, like tuberculosis, have their im- mediate cause in invading pathogenic bacteria. Bacteria themselves cannot be inherited for the reason that it is not possible for them to become an INHERITANCE OF ACQUIRED CHARACTERS 93 integral part of the fertilized egg and thus cross the "hereditary bridge" which joins two generations. A general predisposition to bacterial disease, that is, a lack of resistance to bacterial invasion due to de- fectiveness in physical or physiological equipment, may be present as a combination of characters in the germplasm, or an individual, as the result of disease, may "acquire" a generally weakened germ- plasm and so produce a progeny exhibiting general liability to disease ; but it is doubtful if such a con- dition can properly be termed the inheritance of an acquired character, since the particular definite disease in question is not demonstrably heritable. When alcoholism "runs in a family," its reappear- ance in the son is probably due to the fact that he is derived from the same weak strain of germplasm as his father. The fact that the father succumbed to the alcohol habit is not the determining cause of drunkenness in the son. The same thing that caused the father to become an alcoholic, namely, weak germplasm, and not the resulting drunkenness in the parent, is the causal factor for alcoholism in the son. At the same time it is entirely probable that hered- itary alcoholism may in some cases arise through "parallel induction," that is to say, acquired alco- holism may end in the simultaneous poisoning and consequent modification of both the somatoplasm and germplasm of the parent, with the result that the germplasm has less resistance to alcoholism in a suc- ceeding generation. The offspring are consequently more likely to succumb to the disease. This, how- 94 GENETICS ever, is not the inheritance of an acquired character or of a definite somatic modification. When a man of the present generation has rheu- matic gout, it is a severe stretch both of patriotism and of the powers of heredity to trace the origin of the affliction back to a revolutionary ancestor who acquired sciatic rheumatism by sleeping on the ground at Valley Forge, yet this is quite as direct as many alleged instances of the inheritance of disease. In the majority of instances, apparent cases of disease inheritance are merely instances of reinfec- tion. This reinfection of the offspring may occur very early in embryonic life, or it may happen after birth, provided the offspring are exposed to the same environment as that in which the parent acquired the disease, but in any case reinfection is not heredity , 13. The Germplasm Theory sufficient to ac- count FOR the Facts of Heredity Weismann holds that the theory of the continuity of the germplasm, already considered in a previous chapter, is sufficient in itself to account for the facts of heredity. Hence it is quite unnecessary to fall back upon the inheritance of acquired characters as an ex- planation, since this theory is at least difficult, if not impossible, of satisfactory proof. To prove the inheritance of acquired characters, according to Weismann three things are necessary : firsts a particular somatic character must be called forth by a known external cause ; second, it must be something new or different from what was already INHERITANCE OF ACQUIRED CHARACTERS 95 exhibited before, and not be simply the reawakening of a latent germinal character; and third, the same particular character must reappear in succeeding generations in the absence of the original external cause which brought the character in question forth. As yet these conditions have not been convincingly met in the evidence which has been brought forward in support of the inheritance of acquired characters. 14. The Opposition to Weismann The opponents of Weismann point out, as a weak place in his argument, the assumption that the germ- plasm is so insulated from the somatoplasm as not to be influenced by it. Weismann assumes, of course, that the germplasm is isolated from the somatoplasm very early in the development of the fertilized egg into an individual, and that when once isolated it there- after takes no active part in, nor is in any way affected by, the vicissitudes through which the somatoplasm, or the body itself, passes. The somatoplasm is thus merely a carrier of the germplasm and unable to affect the character of it any more than a rubber hot- water bag, although capable of assuming a variety of shapes, can affect the character of the water within it. In opposition to this view it is urged that every organism is a physiological as well as a morphological unity, and that cells entirely insulated within such a unity would be a physiological miracle. There is abundant evidence that germ-cells, or rather the sexual organs producing the germ-cells, do 96 GENETICS affect the somatoplasm under particular conditions, as, for example, in cases of castration when those somatic features called "secondary sexual charac- ters" undergo profound modification. If the germplasm thus exercises a constant influ- ence on the somatoplasm, why, it seems legitimate to ask, may not the reverse be true and acquired somatic characters leave their impress upon the germ-cells ? 15. Conclusion But even granting the reverse to be true, that is, that the somatoplasm affects the germ-cells, the in- heritance of acquired characters is by no means thereby established. In order to do this, the precise acquired character in question, which indirectly exercised its influence upon the germ, must be redeveloped, and, although the germplasm might conceivably receive an influ- ence from the somatoplasm and be affected by it in a general way, it is a different matter entirely to develop anew the verisimilitude of the character itself which is supposed to have been acquired. It will be seen in subsequent pages, under the dis- cussion of data furnished by experimental breeding, that the weight of probability is decidedly against the time-honored belief in the inheritance of acquired characters. CHAPTER VI THE PURE LINE 1. The Unit Character Method of Attack In reducing any body of facts to a science, it is first necessary to determine the underlying units out of which the facts are made up. Chemistry was alchemy until the chemical ele- ments were identified and isolated. Histology was terra ohscura until the cell theory brought forward "cells" as the units of tissues. In the same way there could be no science of genetics until the con- ception was developed that the individual is a bundle of unit characters rather than a unit in itself. So it has come about that we now speak of inheritance as applied to unit characters rather than to individuals as a whole. Incidentally the fact that an organism is a com- bination of many units makes it easy to account for the wide diversity of forms found in nature, since the addition of a single unit greatly increases the pos- sible combinations in successive generations. Thus if three unit characters. A, B, and C, are present in each parent, for example, there would be six possible double combinations in the offspring, namely, AA, AB, AC, BB. BC, and CC. If now a H 97 98 GENETICS fourth unit D is added by one parent, there would be not only the original six double combinations, but in addition to these, AD, BD, and CD, that is, as many more as there are unit characters with which the new one may combine. Obviously, when individuals are made up of very many unit characters, as, for instance, a thousand, the addition of one new unit character will increase the possible double combinations a thousand fold. 2. Galton's Law of Regression Galton was one of the first ^ to attempt to express mathematically the relationship between parents and offspring by means of treating statistically a single unit character. According to Galton, a mathe- matical expression of the relationship between two generations should serve as a corner-stone of heredity. What Galton did was to take human stature as a unit character in comparing 204 English parents and their 928 adult offspring, because human stature is not complicated by environmental influences and is, consequently, a purely hereditary matter. Since female height is normally less than male height, the two were reduced, for purposes of comparison, to a common male denominator by multiplying each female height by 1.08, which is the average amount that the male exceeds the female in height. There are always two parents con- cerned in the sexual production of every offspring. 1 " Hereditary Genius," 1869. THE PURE LINE 99 therefore Gallon reckoned a "midparent" in each case, according to the formula 1.00 $ + 1.08 $ in order to represent the double parental generation by a single number for the purpose of easy compari- son with the filial generation. Inches Fig. 33. — Scheme to illustrate Gallon's law of regression. The circles represent graded groups of parental height while the arrowpoints in- dicate the average heights attained by the respective offspring. The offspring of undersized parents are taller, and of oversized parents are shorter than their respective parents. Based on data from Galton. The results of his measurements expressed in inches are shown in the following table, in which the 100 GENETICS offspring are, in each instance, arranged under their respective midparents. Midparental height . . 64.5 65.5 66.5 67.5 68.5 69.5 70.5 71.5 72.5 Average height of offspring . 65.8 66.7 67.2 67.6 68.3 68.9 69.5 69.9 72.2 The mean group for all the midparents, it will be seen, is 68.5, and the offspring of this group average 68.3. The table is expressed graphically in Figure 33 in which the circles connected by the diagonal line represent the graded parental heights, while the arrowpoints indicate the average heights of the off- spring in each group. In order to compare these two series of numbers more readily, they may be reduced to a common basis in which the mean class in each instance is made equal to 100, as follows : — Midparental height . . 94 95.5 97 98.5 100 101.5 103 104.5 106 Average height of offspring . 96 97.5 98.5 99 100 101 101.5 102 105.5 The same series may be expressed in terms of amount of deviation from the mean or middle classes, as shown below. The deviation of each group in the series is marked by the signs + or — according as the heights given are greater or less than 100. THE PURE LINE 101 Midparental height, . . - 6 -4.5 -3 -1.5 + 1.5 + 3 + 4.5 + 6 Average height of offspring . - 4 -2.5 -1.5 -1 + 1 + 1.5 + 2 + 5.5 Finally, the relation between the midparent and the average offspring may be expressed in fractional form by taking the average height of the offspring for the numerato]^ and the height of the midparent for the denominator in each instance. The minus deviations are thus seen to be 6 4.5 2 1.5 which, added together and reduced to a decimal, equal .60. Similarly, the plus deviations are L5 ~3~ 4^ T which reduce to .63. The average of the minus deviations (.60) and the plus deviations (.63) is nearly .62, or about two thirds. That is to say, the fraction f represents the amount of resemblance or "inheritance" between two generations, as determined by the foregoing series, while the remaining | is the measure of "regression" from the general type. This illustrates Galton's Law of Regression or the tendency in successive generations toward medioc- rity. The law may be stated as follows : — 102 GENETICS Average parents tend to produce average children ; minus parents tend to produce minus children ; plus parents tend to produce plus children ; but the progeny of extreme parents, whether plus or minus, inherit the parental peculiarities in a less marked degree than the latter were manifested in the parents themselves. 3. The Idea of the Pure Line It was Galton's law of regression that suggested to the Danish botanist Johanssen a possible means of controlling heredity. In his mind arose the ques- tion whether it would not be possible by continually breeding from plus parents, granting that plus par- ents produce plus offspring and making allowance for some regression to type, to shove over the off- spring more and more into the plus territory and so to establish a plus race. To test this hypothesis, Johanssen selected beans, Phaseolus, with which to experiment, since this group of plants is self-fertilizing, prolific, and easily measurable. Somewhat to his surprise, his beans refused to shove over as much as expected. That is, big beans did not yield principally big offspring, nor little beans little offspring, according to the ex- pectation, although they each produced offspring that varied in the manner of fluctuating variability around an average unlike the parental type. This gave Johanssen the idea that he was using mixed material, so he next isolated the progeny of single beans, which, being self-fertilized, each constituted unmistakably a single hereditary line. In this way THE PURE LINE 103 nineteen beans, now famous, became the known ancestors of Johanssen's original nineteen "pure lines," a further study of w^hich has led the way to some of the most brilliant biological discoveries of recent years. A pure line has been defined by Johanssen as "the descendants from a single homozygous organism exclusively propagating by self-fertilization," and more briefly by Jennings as "all the progeny of a single self -fertilized individual." 4. Johanssen's Nineteen Beans It was found by Johanssen that the progeny of each of these pure lines of beans varied around its own mean, which was different in each of the nine- teen instances. When, however, extremes from any pure line series were selected and bred from, the prog- eny, instead of showing two thirds inheritance and one third regression with respect to the extremeness of a particular character, as Galton found was true in the case of human stature, showed no inheritance and complete regression away from the extreme condition of the parent bean back to the type for the entire pure line in question. That is, selection icithin a pure line is absolutely without effect in modifying a particular character in the offspring of the line in question. This is illustrated in Figure 34 in which the results of selecting for size in the year 1902 is shown for four pure lines only. The average for each pure line is given at the top of its column. When, for example, 104 GENETICS beans weighing 60 eg. were selected from pure lines IT, VII, and XV, the average weights of their progenj^ were 56.5, 48.2, and 45.0 eg. respectively, which in each instance is nearer to the average for the pure line than to the weight of the parental seed. Weight of Cent,- parent «eed 10 20 30 40 » 60 70 10 20 30 10 50 60 70 10 20 SO 40 50 60 70 10 20 }0 40 SO 60 70 ;0 20 50 40 SO SO 70 Pure line number n w w M Fig. 34. — The result of selection in four pure lines of beans. The verti- cal columns, representing the average progeny from different sized parents all derived from the same pure line, contain groups nearer aUke than the horizontal columns, representing progeny from the same sized parents, but different pure lines. All the numbers indicate centigrams. Data from Johanssen. It will be seen at once that the averages in the vertical columns are nearer alike than the averages in the horizontal columns. In other words, the beans bred true to their pure line rather than to their fluctuating parent. As a further example of this law, take the result THE PURE LINE 105 of selection for six years in pure line I as shown in the accompanying table and in Figure 35. TTa'rvitqt' Vitat? Mean Weight of Selected Parent Seed Mean Weight of Offspring Minxis Plus From Minus Parent From Plus Parent 1902 .... 1903 . . . . 1904 .... 1905 .... 1906 .... 1907 .... 60 55 50 43 46 56 70 80 87 73 84 81 63.15. 75.19 54.59 63.55 74.38 69.07 64.85 70.88 56.68 63.64 73.00 67.66 It is evident, for instance, that in 1907 the smallest beans, weighing an average of 56 eg., gave an average progeny weighing 69.07 eg, while the largest ones for the same year, weighing an average of 81 eg., produced nearly the same average in their progeny as did the smallest beans, that is, 67.66 eg. Incidentally all the progeny from both large and small parents averaged notably less in 1904 than all the progeny from large and small parents in 1906, a result due to a "poor year" when certain factors of environment were unfavorable. Such unfavor- able conditions, however, are known to influence in no way the hereditary qualities of the beans. Thus it appears that, although the progeny of a pure line present plenty of variations of the fluctuating type, due probably to environmental differences in nutri- tion, moisture, etc., such variations are quite inef- 106 GENETICS O O. oo \(5>\ c» lO (? ^ 6^ l6> 'O 10 OB V 4^ to / 'o ® TO © 02 a; 03 OS O ^-5 « > o HI a CO -H a^ o o S a M Q, W "^ CI t. •1-1 •!< 05 >J O O >> a •I "a; C3 ^ o: O s ^ o 3 ^ 03 CI a o a a o 'oj ff3 q w 2 a © a; ^^ ^ © 03 _ "5 03 © © c3 O a 03 bO G •pH a 03 o © W) c3 ;h o > © I a 5h CO 7^ o © 03 THE PURE LINE 107 fectual so far as inheritance is concerned, and it makes no difference whether the largest or the small- est beans within a pure line are selected from which to breed, the result will be the same, in that there is a complete return to mediocrity or type with no "in- heritance" of the parental modification. As a matter of fact in 1903, 1906 and 1907 the lighter parents gave a heavier progeny than the heavier parents. It will be seen at once that here is a discovery of far-reaching importance which may require us to reconstruct certain cherished ideas about the part played in the evolution of species, as well as in heredity, by natural selection. 5. Cases similar to Johanssen's Pure Lines Although according to Johanssen pure lines are *'the progeny of a single self -fertilized individual," it is plain that in at least three other possible cases something quite similar to "pure lines" may be obtained. First, when two similar organisms identical in their germinal determiners with regard to a particular character inbreed, their progeny will form a pure line so far as this particular character is concerned just as truly as two parents that are united in a single in- dividual produce a pure line progeny as the result of self-fertilization. Second, in cases of parthenogenesis, the progeny arising from a single female individual without the customary qualitative reduction of chromosomes that 108 GENETICS accompanies sexual reproduction, constitute a pure line or an unmixed strain. Third, in cases of asexual reproduction where the progeny are simply the result of continued fission of the original individual, a pure line may be said to continue from generation to generation. In the second and third categories it should be pointed out that the "pure line" is assured only so long as asexual reproduction continues. It is quite possible for an organism, heterozygotic in composi- tion, to continue to breed true or to produce an ap- parently puje line so long as asexual methods are employed. As soon as such an organism, however, changes to the sexual method of reproduction, seg- regation of characters may occur and different combinations result. 6. Tower's Potato-beetles As an illustration of the effect of selection within pure lines of the first category may be mentioned a case given by Tower in his exhaustive experiments on the Colorado potato-beetle Leptinotarsa decem- lineata. Among the numerous cultures of this beetle which were under control, a considerable variation in color made its appearance. For con- venience in classification these variations wxre graded into arbitrary classes or graduated variants (see p. 52) ranging from dark to light. When a male and a female from the extreme class at the dark end of the series were allowed to breed together, their progeny were not dark, but fluctuated Fig, 36. — Diagram showing the ineffectiveness of selection through twelve generations within a homozygous strain in the case of the Colorado potato-beetle (Leptinotarsa). In each generation extreme dark speci- mens were selected as the parents of the succeeding generation but the progeny always swung back to the type. After Tower. no GENETICS in color around the original average of the entire series. This process of selecting each time an ex- treme pair of dark parents was continued for twelve generations, as shown in Figure 36, without in any way increasing the percentage of brunette potato beetles in the progeny. Thus in a pure line formed by the breeding of two individuals alike with respect to color, the selection of an extreme variant was quite without effect in modifying the color of the progeny. 7. Jennings' Work on Paramecium An instance of the third category of pure lines is furnished by Jennings' remarkable w^ork on the protozoan Paramecium, which was published in 1909. Jennings carried on his experiments quite independ- 206 zoo 194 176 142 US 100 Fig. 37. — Eight pure races of Paramecium. The actual mean length of each race is given in micra below the corresponding outline. Magni- fied about 230 diameters. After Jennings. ently of Johanssen, but he nevertheless arrived at the same general conclusion, namely, that selection within a pure line is without effect. Jennings found that Paramecia differ from each THE PURE LINE 111 other in size, structure, physical character, and rate of niultiphcation as well as in the environmental conditions required for their existence and, further- more, that these differences, in an hereditary sense, are "as rigid as iron." With respect to the character of mean length he was able to isolate eight races, or pure lines, whose average size, drawn to scale, is shown in Figure 37. 256 < MICRA > Fig. 38. — Diagram of a single race (D) showing the variation in the size of the individuals. Magnified about 230 diameters. After Jennings. Each of these pure lines produced a progeny which exhibited a considerable range of fluctuating variation. The offspring of pure line D, for example, varied from ^56 to 80 micra ^ in length with an aver- age of 176 micra, as shown in Figure 38, where samples of the different classes of variants in pure line D are arranged in a series. A single representative of each of the different classes of variants out of all of the eight pure lines bred by Jennings is shown in Figure 39. Each horizontal row represents a single race or ^ A micron is jo^o^th of a millimeter. 112 GENETICS pure line, the average size of which is indicated by the sign + . The mean length of the entire lot, as shown J 55 DOOHa UUUU aO'OHHa 8()898eH08M gyflHH 98 « » » gyjHMM 89 » e + M M « « « « 45 Fig. 39. — Diagram of the species Paramecium as made up of the eight different races shown in Figure 37. Each horizontal row represents a single race. The individual showing the mean size in each race is in- dicated by a cross placed above it. The mean for the entire lot is at the horizontal line. The magnification is about 24 diameters. After Jennings. by the vertical line, is 155 micra. The total number of individuals belonging to each size is not indicated, but in every horizontal line their number is more THE PURE LINE 113 numerous near the average for that line and less numerous at the extremes, thus forming the typical normal frequency polygons of fluctuating variability. The significant fact about these series is this, that extreme individuals selected from any pure line do not reproduce extreme sizes like themselves, but instead, a progeny varying according to the laws of chance around the average standard of the particular line from which it came. 8. Phenotypical and Genotypical Distinctions From the foregoing it will be seen that the be- havior of an organism in heredity cannot always be determined by an inspection of its somatic char- acters alone. For example, six Paramecia, each 155 micra in length and apparently identical, could be selected from the six upper pure lines in Jennings' table given in Figure 39 which would produce six progenies definitely unlike, whereas in the case of pure line D, twenty-four Paramecia, all measurably different from each other in size, would be found to produce twenty-four progenies practically identical. Organisms that appear to be alike, regardless of their germinal constitution, are said by Johanssen to be identical phenotypically , or to belong to the same phenotype. On the other hand, organisms having identical germinal determiners such as those of the varying members of pure line /), are said to be genotypically alike or to belong to the same genotype. 114 GENETICS Organisms belong to the same phenotype with respect to any character when their somatoplasms are alike. They belong to the same genotype when their germplasms are alike. The word "genotype" was suggested by Johanssen in honor of Darwin and his theory of psmgenesis, although there are certain objections to its use in this connection for the reason that systematists have already appropriated it in a different sense. Natural history and common usage deal prin- cipally with phenotypes, that is, with organisms as they appear. The older theories of heredity were likewise concerned wdth phenotypes, but we are now coming to see more clearly than before that heredity must always be a case of similarity in origin, that is, in germinal composition, and that similarity in ap- pearance by no means always indicates similarity in origin or true relationship. The assumption that similarity in appearance does indicate relationship has been made the foundation of many conclusions in comparative anatomy and phylogeny, but to the modern student of genetics who places his faith in things as they are, rather than in things as they seem to be, conclusions based upon phenotypical distinctions alone have in them a large source of error which must be taken into account. In a museum of heredity, should such a collection ever be assembled, the specimens would not be ar- ranged phenotypically as they are in an ordinary museum where things that look alike are placed together as if in bonds of relationship, but they THE PURE LINE 115 would be arranged historically from a genetic point of view to show their true origin one from another. 9. The Distinction between a Population AND A Pure Line A mixture of pure lines has been called a popula- tion. It is not possible to distinguish a pure line from a population by inspection, since both may be pheno- typically alike. Fluctuations about the average occur in both cases with no appreciable difference in character, although such fluctuations, when they occur within a pure line, are simply somatic differ- ences caused in general probably by modifications in nutrition or some other external factor of environ- ment, while fluctuations in a population include not only modifications of this transient nature, but also permanent hereditary differences due to germinal differences in the various pure lines of which the population is composed. Johanssen has made the distinction between pure lines and populations clear by the following figure (Fig. 40), in which five pure lines of beans are com- bined artificially to form a population. The beans which make up the pure lines noted in this figure are represented inclosed within inverted test tubes. The beans in any single tube are all of one size. Tubes vertically superimposed upon each other also contain only beans of one size. Thus it is seen that what may be a rare size of bean in one line, for instance that in the left-hand 116 GENETICS n ^ igji^^ i ji -^ ^ ^ -j^jji tube of jpure line 3, may be identical with the com- monest size in another line, as jpure line 2. The Pure: Line five pure lines represented in Figure 40 are combined in a population at the bottom of the figure^ making a phenotype that marks the five phenotypes above, which are also five geno- types. In the population, how- ever, the five genotypes are hidden within one phenotype. Hence, while selection within a pure line has no hereditary in- fluence, it is evi- dent that selec- tion within a population may shift or move over the type of the progeny Population HI Fig. 40. — Diagrams showing five jyure lines and a population formed by their union. The beans of each pure line are represented as as- sorted into inverted test tubes making a curve of fluctuating variability. Test tubes contain- ing beans of the same weight are placed in the same vertical row. After Johanssen. THE PURE LINE 117 obtained, in the direction of the selection simply by isolating out a pure line of one type. Thus beans chosen from the extreme left-hand test tube in the population cited would belong only to 'pure line 2, while those taken from the extreme right- hand test tube could belong only to pure line 3. Galton's "law of regression," namely, that minus parents give minus offspring and plus parents plus offspring, with a tendency to reversion from genera- tion to generation, depends simply upon a partial but not complete isolation of pure lines out of a population. From this distinction between pure lines and popu- lations it is clear why breeders in selecting for a particular character out of their stock need to keep on selecting continually in order to maintain a cer- tain standard. As soon as they cease this vigilance, there is a "reversion to type" or, as they say, "the strain runs out," which means that the pure lines become lost in the mixed population which inevi- tably results as soon as selective isolation of the pure line ceases. Such reversion must always be the case in dealing with a population made up of a mixture of pure lines, for only by the isolation of pure lines can the constancy of a character be maintained. When, however, a pure line is once isolated, then all the mem- bers of it, large as well as small, are equally efficient in maintaining the pure line in question, regardless of their phenotypical constitutions. 118 GENETICS 10. Pure Lines and Natural Selection From the foregoing statements it appears that by means of selection within a population, such as occurs normally in nature, it is not possible to get anything out that was not already there to begin with. If this is so, the origin of species cannot have come about, as Darwin thought, through natural selection by a gradual accumulation of slight favorable varia- tions. The best that selection can do is to isolate pure lines. Within pure lines it is quite powerless to change the genet ypical characters. In other words, natural selection can only maintain and strengthen the frontier posts that are already es- tablished. It cannot break into the wilderness and create new centers. Since the extreme members of a pure line, having the same genotypical constitution, always tend to backslide to mediocrity within the limits of the line in question, the crucial question is : How can the critical step from one genotype to another, a step indispensable in the evolutionary derivation of species, ever occur ? That it has repeatedly oc- curred in the course of time is amply proven by the fact that somehow or other we have gone from Ameba to man. At present the only loophole of escape seems to lie either in the unlikely inheritance of acquired char- acters, or in mutations which make the leap from one character to another, and so eventually" from one type to another, without the aid of selection. THE PURE LINE 119 It is interesting to note that Johanssen himself, who has been so prominently concerned in erecting this barrier in the way of the evolutionary derivation of species by natural selection, has recently reported mutations arising within his pure lines of beans. It must be admitted that to the skeptical there is a vicious circle here, for when a variation fails to re- appear in a subsequent generation, it may be ex- plained as the failure of natural selection to act within a pure line, but when a variation does reap- pear it is hailed as a mutation! In any event the way of experiment lies open, and the evidence of investigators in this critical field will be awaited with keen interest. CHAPTER VII SEGREGATION AND DOMINANCE 1. Methods of Studying Heredity Modern studies in heredity have been pursued principally in three directions : first, by microscopical examination of the germ-cells ; second, by statistical consideration of data hearing upon heredity ; and third, by experimental breeding of animals and plants. The first two of these methods of approach have already been touched upon as well as experimental breeding with reference to "pure lines." In the present chapter attention will be directed to a con- sideration of experimental breeding with reference to hybridization, that is, breeding from unlike par- ents, a process which Jennings characterizes by the expressive phrase, "the melting-pot of cross-breed- mg. 2. The Melting-pot of Cross-breeding Hybridization, or cross-breeding, as formulated by Galton (1888), results in one of three kinds of inheritance, namely, blending, alternative, or par- ticulate. 120 SEGREGATION AND DOMINANCE 121 Of these, blending inheritance may be called the typical "melting-pot" in which contributions from the two parents fuse into something intermediate and different from that which was present in either parent. Galton illustrated this process by the inheritance of human stature in which a tall and a short parent produce offspring intermediate in height. A more thorough consideration of this type of inheritance will be presented in Chapter IX. By the method of alternative inheritance the pa- rental contributions do not melt upon union, but retain their individuality, reappearing intact in the offspring. In inheritance of human eye-color, for ex- ample, the offspring usually have eyes colored like those of one of the parents when the parental eye- color is unlike in the two cases, rather than eyes intermediate in color between those of both parents. According to Galton particulate inheritance results when the offspring present a mosaic of the parental characters, that is, when parts of both the maternal and paternal characters reappear in the offspring without losing their identities by blending or without excluding one another. Piebald races of mice arising from parents with solid but different colors have been cited as illustrations of this sort of inheritance, although it will be seen later in connection with the ''factor hypothesis" that another interpretation of this phenomenon is not only possible but probable. The distinctions between these three categories of inheritance are diagrammatically represented in Figure 41. 122 GENETICS In blending inheritance the offspring are seen to be unhke either parent, because the parental deter- miners fuse into a new thing. In alternative in- heritance, on the contrary, the offspring may be like either parent, since the characters in question do not lose their individuality upon union, as shown in the diagram. Only one or the other of the two Blending Alternative Particulate Characteristics of parental jermplasm as Shovun in the Somaplasm Double germplasm termed from Contributions from both parents Possible kind* of apparent dfsprinj this jermplasm maif produce Fig. 41. — Three kinds of inheritance described by Galton. mutually exclusive characters thus becomes effective in determining the nature of each offspring. Finally, in particulate inheritance the double germplasm which determines a new individual may be imagined to undergo a diagonal rather than a vertical cleavage upon maturation, thereby causing unblended fragments of both parental characters to become effective at once, in this manner producing a mosaic offspring. SEGREGATION AND DOMINANCE 123 3. JOHANN GrEGOR MeNDEL Our understanding of the working of inheritance in hybridization we owe largely to the unpretentious studies of an Austrian monk, Johann Gregor Mendel, who, although a contemporary of Darwin, was prob- ably unknown to him. For eight years Mendel carried on original experiments by breeding peas in the privacy of his cloister garden at BrUnn and then sent the results of his work to a former teacher, the celebrated Karl Nageli, of the University of Vienna. At the time Nageli's head was full of other matters, so that he failed to see the significance of his old pupil's efforts. However, in 188[6 Mendel's results appeared in the Transactions of the Natural History Society of Brunn,^ an obscure publication that reached hardly more than a local public. Here Mendel's investigations were buried, so to speak, because the time was not ripe for a general apprecia- tion or evaluation of his work. At that time neither the chromosome theory nor the germplasm theory had been formulated. More- over, much of our present knowledge of cell structure and behavior was not even in existence. Weismann had not yet led out the biological children of Israel through the wilderness upon that notable pilgrimage of fruitful controversy which occupied the last two decades of the nineteenth century, and the attention of the entire thinking world was being monopolized 1 Verhandlungen naturf. Verein in Briinn. Abhandl. IV, 1865 (which appeared in 18G6). 124 GENETICS by the newly published epoch-making work of Charles Darwin. Mendel died in 1884, and his work slumbered on until it was independently discovered almost simul- taneously by three botanists whose researches had been leading up to conclusions very much like his own. These three men were de Vries of Holland, von Tschermak of Austria, and Correns of Germany. Their papers were published only a few months apart in 1900 and were closely followed by important papers from Bateson in England and Davenport and Castle in America, with a rapidly increasing number from other biologists the world over. To- day the literature upon this subject has grown to be very large, and the end is by no means yet in sight. Concerning Mendel, Castle has well said: "Mendel had an analytical mind of the first order which en- abled him to plan and carry through successfully the most original and instructive series of studies in heredity ever executed." 4. Mendel's Experiments on Garden Peas What Mendel did was to hybridize certain varie- ties of garden peas and keep an exact record of all the progeny, in itself a simple process but one that had never been faithfully carried out by any one. Before examining Mendel's results it may be well to state the difference between normal and artificial self-fertilization. Self-fertilization occurs when from the pollen and ovule of the same flower are derived SEGREGATION AND DOMINANCE 125 the two gametes which uniting produce a zygote that develops into the seed and subsequently into the adult plant of the next generation. In artifi- cially crossing normally self-fertilized flowers it is necessary to carefully remove the stamens from one flower while its pollen is still immature, and later, at the proper time, to transfer to it ripe pollen from another flower. Mendel's cross-breeding experiments on peas showed certain numerical relations which gave rise to what has come to be rather indefinitely known as "Mendel's law." This law may be temporarily formulated as follows : — When parents that are unlike with respect to any character are crossed, the progeny of the first gen- eration will apparently be like one of the parents w4th respect to the character in question. The parent which impresses its character upon the off- spring in this manner is called the dominant. When, however, the hybrid offspring of this first generation are in turn crossed with each other, they will produce a mixed progeny, 25 per cent of which will be like the dominant grandparent, 25 per cent like the other grandparent, and 50 per cent like the parents resem- bling the dominant grandparent. An illustration will serve to make plain the man- ner in which this law works out. Mendel found that when peas of a tall variety were artificially crossed with those of a dwarf variety, all the resulting offspring were tall like the first parent. It made no difference which parent was 126 GENETICS selected as the tall one. The result was the same in either case, showing that the character of tallness is independent of the character for sex. When these tall cross-bred offspring were subse- quently crossed with each other, or allowed to pro- duce offspring by self-fertilization which amounts to the same thing, 787 plants of the tall variety and 277 of the dwarf kind were obtained, making approx- imately the proportion of 3 to 1. On further breeding the dwarf peas thus derived proved to be pure, producing only dwarf peas, while the tall ones were of two kinds, one third of them "pure," breeding true like their tall grandparent, and two thirds of them "hybrid," giving in turn the proportion of three tall to one dwarf like their parents. These crosses may be expressed as follows : — Tall, r, X dwarf, U = tall, T(t). That is, tallness crossed with dwarfness equals tallness with the dwarf character present but latent. Mendel termed the character, which became ap- parent in such a hybrid, in this case tallness, the dominant, and the latent character which receded from view, in this instance dwarfness, the recessive. When now the hybrids, T(t), were crossed to- gether, the result algebraically expressed was as follows : — T -{- 1 (all possible egg characters) T-\-t (a ll possible sperm characters) TT+ Tt Tt + tt TT+2T(t)+tt SEGREGATION AND DOMINANCE 127 That is, one out of four possible cases was dwarf, t, in character and the other three were apparently tall, although only one out of the three was pure tall, T, while the remaining two were tall with the dwarf character latent, T (t). The same thing may be expressed more graphically by the checkerboard plan. Male Gametes T I to ^.T which Punnett suggested (Fig. 42). Each square of the checkerboard rep- resents a zygote which, having received a gamete from each of the two par- ents, may develop into a possible offspring. The character of the gametes of the parents is shown outside of these squares, while the arrows repre- sent the parental source from which the offspring have received their heredi- tary composition. The essential feature of Mendel's law is briefly this: hereditary characters are usually independent units which segregate out upon crossing, regardless of temporary dominance. Mendel carried on further experiments with garden peas, using other characters. He obtained practically the same result as in the instance already given, for the actual progeny in the second generation of the cross-bred offspring figured up, as seen in the table z < la < Z u Fig. 42. — Diagram to illustrate theoretically the formation of the four possible zygotes in the second filial generation of a monohybrid. 128 GENETICS below, very nearly to the expected theoretical ratio of 3 to 1. Character Number of Dominants Number of Recessives Ratio Form of seed .... 5474 smooth 1850 wrinkled 2.96 to 1 Color of seed coat 6022 yellow 2001 green 3.01 to 1 Color of flowers 705 colored 224 white 3.15 to 1 Form of pods . . 882 inflated 299 constricted 2.95 to 1 Color of unripe pods 428 green 152 yellow 2.82 to 1 Position of flowers 651 axial 207 terminal 3.14 to 1 Length of vine 787 tall 277 dwarf 2.84 to 1 Total .... 2.98 to 1 Darbyshire repeated the yellow-green cross with garden peas, obtaining in the second generation the large total of 139,837 individuals of which 105,045 were yellow and 34,792 green, wliich is very close to 3 to 1. o. Some Further Instances of "Mendel's Law " Since the rediscovery of Mendel's law the ratio of 3 to 1 in the second generation has been found by a number of different investigators to be constant in a large array of characters observed both in animals and plants of diverse kinds wdien these are cross-bred with reference to the characters in question. Botanists have the advantage perhaps in this matter, as they deal with forms which usually produce a large number of offspring from a single cross, — a very desirable thing in estimating ratios. On the SEGREGATION AND DOMINANCE 1^29 other hand, they are handicapped by being unable usually to obtain more than one generation in a year, while zoologists may secure from many animals like rabbits and mice several generations in a year, al- though ordinarily the number of progeny is much Organism Author Dominant Recessive P Nettles Correns Serrated leaves Smooth-margined leaves '03 Sunflower ShuU Branched habit Unbranched habit •08 Cotton Balls Colored lint White lint '07 Snapdragon Baur Red flowers Non-red flowers '10 Wheat Biff-en Susceptibility to rust Immunity to rust '05 Tomato Price and Drinkard Two-celled fruit Many-celled fruit '08 Maize de Vries Round, starchy kernel Wrinkled, sugary kernel '00 Silkworm Toyama Yellow cocoon White cocoon '06 Cattle Spillman Hornlessness Horns '06 Pomace fly Alorgan Red eyes White eyes •10 Horses Bateson Trotting habit Pacing habit '07 Land snail Lang Unbanded shell Banded shell '09 Mice Darbyshire Normal habit Waltzing habit '02 Guinea-pig Castle Short hair Angora hair '03 Canaries Bateson and Saunders Crest Plain head '02 Poultry- Davenport Rumplessness Long tail •06 Man Farrabee Brachydactyly Normal joints '05 Barley von Tsehermak Beardlessness Beardedness '01 Salamander (Amblystoma) Haecker Dark color Light color •08 smaller and the ratios obtained have a larger chance of error than is the case with the more numerous plant offspring. Semi-microscopic animals, as, for example, the pomace fly, Drosojphila, which produces a large progeny every two weeks or so, may combine the general advantages mentioned for the two groups of organisms 130 GENETICS indicated above, but they have the disadvantage of being so small that the detection of their distinctive phenotypic characters is attended with considerable technical difficulty. What the modern experimenter in genetics desires is an organism, first, that possesses conspicuous distinc- tive somatic characters, and, second, which w411 come to sexual maturity early and breed either in captivity or under cultivation both numerously and frequently. The preceding table, compiled chiefly from Bateson ^ and Baur,- might easily be much extended. It shows from what diverse sources confirmatory evidence of the truth of Mendel's law has been derived within the past few years. 6. The Principle of Segregation The essential thing which Mendel demonstrated was the fact that, in certain cases at least, the deter- miners for heredity derived from diverse parental sources may unite in a common stream of germplasm from which, in subsequent generations, they may segregate out apparently unmodified by having been intimately associated with each other. This "law of segregation" depends upon the conception that the individual is made up of a bundle of unit characters. It may be illustrated by the separate flowers picked ^ from a garden which, after being made into a nose- gay, may be taken apart and rearranged without in 1 "Mendel's Principles of Heredity," 1909. '^ " Einf iihrung in die experimentelle Vererbungslehre," 1911. SEGREGATION AND DOMINANCE 131 any way disturbing the identity of the separate blossoms. The general formula of segregation that covers all cases of organisms cross-bred \\ath respect to a single character, that is, monohybrids, is given in Figure 43. U (Dominant) K (Recesiln) D(R) 1 DD 2 D(R) RR i OD DD 2D(R) RR RR ^ 1 DD uf ' Fig. 43. — General Mendelian formula for a monohybrid. 7. HOMOZYGOTES AND HeTEROZYGOTES A character which is present in the offspring in double quantity because it was present in both parents is said by Bateson to be homozygous, while an or- ganism which is homozygous with respect to any character is called a hoinozygote so far as that particu- lar character is concerned. In contrast to the homozygous condition, an organ- ism is said to be heterozygous when it derives the determiner of a character from one parent only. Such an organism is described as a heterozygote with respect to the character in question. A homozygous and a heterozygous dominant may appear alike. 132 GENETICS although not necessarily so, that is, they may have the same phenotypical constitution, but their geno- typical composition is always different. 8. The Identification of a Heterozygote "Homozygote" and "heterozygote" are terms then descriptive solely of the genotypical constitution of organisms, and, as has been said, it is not always possible to distinguish one from the other by inspec- tion, although it may frequently be done, as will be pointed out later. The only sure way to identify a heterozygote is by breeding to a recessive and observing the kind of offspring produced. Peas of the formulae TT and T{t), for example, both look alike, since a single determiner for the tall character, T, is sufficient to produce complete tallness. When, however, these two kinds of tall peas are each bred to a recessive dwarf pea, of the formula tt, the progeny will differ distinctly in the two cases as follows : — Case I. r + r X < + / = 100 per cent ^(0. Case II. T + tXt + t = oOper cent 1(1) + 50 per cent ti. That is, if the dominant to be tested is homozygous (Case I), the entire progeny will exhibit the dominant character, but if the dominant to be tested is heterozy- gous (Case II), then only one half of the progeny will show the character in question. 9. The Presence and Absence Hypothesis Mendel's conception that every dominant character is paired with a recessive alternative is now being SEGREGATION AND DOMINANCE 133 largely replaced by the ^presence and absence hypothesis which was first suggested by Correns but later logi- cally worked out by others, particularly by Hurst, Bateson, and Shull. According to this latter inter- pretation, a determiner for any character either is, or is not, present. When it is present in two parents, then the offspring receive a double, or duplex, *'dose," to use Bateson's word, of the determiner. When it is present in one parent only, then the offspring have a single, or simplex, dose of the character. When it is present in neither parent, it follows that it will not appear in the offspring. In this case the offspring are said to be nulliplex wdth respect to the char- acter in question. Take the case of tall and dwarf peas, the determiner for tallness when present pro- duces tall peas, even if it comes from one parent only, but if this determiner for tallness is absent from both parents, the offspring are nulliplex, that is, the absence of tallness results and only dwarf peas are produced. The difference between the presence and absence theory and the dominant and recessive theory is that in the former case the ''recessive" character has no existence at all, while in the latter instance it is present, but in a latent condition. 10. DiHYBRIDS So far reference has been made exclusively to mono- hybrids, any two of which are supposed to be similar except with respect to a single unit character. ]\Iono- hybrids are comparatively simple, but when two 134 GENETICS organisms are crossed which differ from each other with respect to two different unit characters, the situa- tion becomes more comphcated. Mendel solved the problem of dihybrids by crossing wrinkled-green peas with smooth-yellow peas. He found that smoothness S is dominant over wrinhled- ness W and that yellow color Y is dominant over green G, or, as it would be stated according to the presence and absence theory, smoothness is a positive character which fills out the seed-coat to plumpness while its absence leaves a wrinkled coat, and yellow- ness is a positive character due to a fading of the green which causes the yellow to be apparent. In the absence of this green fading factor or determiner the green, of course, appears. If smooth-yellow SY and wrinkled-green WG are crossed, all the offspring are smooth-yellow, but they carry concealed the recessive determiners for wrinkledness and greenness according to the formula S{W)Y(G). When the determiners of these cross- breds segregate out during the maturation of the germ-cells, they may recombine so as to form four possible double gametes, namely, smooth-yellow SY and wrinkled-green WG, which are exactly like the grandparental determiners from which they arose, and in addition, two entirely new combinations, smooth-green SG and wrinkled-yellow TT^F. Since the male and the female cross-breds are each furnished with these four possible gametic combina- tions, the possible number of zygotes formed by their union will be sixteen (4x4 = 16). That is, the SEGREGATION AND DOMINANCE 135 monohybrid proportion of 3 to 1 in tlihybrid com- binations is squared, (3 + 1)- = 16. It of course does not follow that the offspring in dihybrid crosses will always be sixteen in nunil:>er, or that they will always conform strictly to the theoreti- cal expectation of (3 + 1)-. The ofispring ol)tained undoubtedly obey the laws of chance, but the greater the number of offspring, the nearer they come to fall- ing into the expected grouping. The sixteen possible zygotes resulting from a diliybrid cross will give rise to sixteen possible kinds of indi\dduals which in turn, as will be demonstrated directly, present four kinds of phenotypic and nine kinds of genotypic constitutions. A dihybrid mating, using the same symbols em- ployed in the case just described, would be expressed algebraically as follows : — SG+ WY+ SY+ WG = all the possible egg gametes SG+ WY+ SY+ WG = all the possible sperm gametes SGSG+ SGWY+ SGSY+ SGWG SGWY +WYWY+ WYSY+ WYWG SGSY + WYSY +SYSY+ SYWG SGWG + WYWG + SYWG+WGWG SGSG+2 SGWY+2 SGSY+2 SGWG+WYWY+2 WYSY+ 2 WYWG+SYSY+ 2SYWG+ WGIVG The second and the ninth items in this result are alike ; by combining them the revised result reads : — SGSG+4 SGWY+2 SGSY+2 SGWG+WYWY+2 WYSY+2 WYWG+SYSY+WGWG There are then these nine different combinations of germinal characters or nine difTerent genotypes in any dihybrid cross. By placing the recessive char- acters in parentheses, whenever the corresponding dominant is present to indicate that the dominant 136 GENETICS causes the former to recede from view, these nine genotypes may be combined into four phenotypes as follows : — Phenotypes . . 9SY 3SG 3 WY \WG Genotypes . . ^S(G)(W)Y SGSG WYWY WGWG 2S(G)Sr 2SG{W)G 2 WYW{G) 2SY{W)Y SYSY From this analysis it may be said that the Mendelian ratio for a typical dihybrid is phenotypically 9:3:3: 1, while that for a monohybrid, as we have already seen, Cf- SG WY SY WG Q i i 4. I s&-» WY^ SY WG SG SG © WY SG © SY SG ® WG SG ® SG WY ® WY WY ® SY WY © WG WY ® 5G SY © WY SY ® SY SY ® WG SY ® SG WG @ WY WG @ SY WG @ WG WG @ Fig. 44. — Diagram to illustrate the possible combinations arising in the second filial generation (F2) following a cross between yellow-smooth YS and green-wrinkled GW peas. is 3:1. This expected ratio corresponds essentially with the actual results Mendel obtained in crossing smooth-yellow and wrinkled-green peas. SEGREGATION AND DOMINANCE 137 Figure 44 presents a graphic representation of the different combinations resulting from a dihybrid cross follomng the checkerboard plan used in Figure 42 to illustrate monohybrids. The nine genotypes and four phenotypes which result from a dihybrid cross are shown in the following squares. Number in Each Class Genotype Number of Squares 1 ^ in Fig. 44 , Phenotype Number in Each Class 1 SYSY 11 SY 2 {W)YSY ' 7-10 9 2 S(G)SY 3-9 4 S(G)(W)Y 2-5- 12- 15 1 SGSG 1 SG 3 2 SG{W)G 13. 4 1 WYWY 6 WY 3 2 WYW{G) 8-14 1 WGWG 16 WG 1 16 ! 16 Another illustration of dihybridism is shown in Figures 45 and 46 which is based upon data fur- nished by the Davenports.^ In the matings given here, dark or pigmented hair, represented by the solid black circles, is dominant over light-colored, that is, unpigmented or slightly pigmented hair, symbolized by the open circles, while curly hair is dominant 1 "Heredity of Eye-color in Man," Mcnce. X. S. 26, p. 589, 1907; " Heredity of Hair Form in Man," Amer. Nat. 42, p. 341, 1908. Daven- port, C. B. and G. C. 138 GENETICS over straight, represented by crooked and straight lines respectively in the diagram. In other words, the presence of pigment is dominant over the ab- sence of pigment, while the factor that causes curli- ness is dominant over the absence of this factor, with respect to human hair. HA/f^ ' ^ — — r;^ T^' — . ^/ KEY TO Symbols • = Dark O = Li^ht J = CurlLj — = Straight '°^s DARK Fig. 45. — The heredity of human hair according to data by C B. and G. C. Davenport. The arcs represent the somatoplasms of four indi- viduals. Within the arcs are the gametes formed by these individuals. The dominant character is placed on the outside of the arc where it will be visible. SEGREGATION AND DOMINANCE 139 When a homozygous individual with dark curly hair crosses with a homozygous individual with light straight hair, all the offspring have dark curly hair. The dark curly-haired individuals of this second generation, however, are heterozygous with respect to each of these two hair characters. When any two individuals having this particular genotypic compo- sition mate, therefore, they may produce any Nu.-nber in 6AcK GENOTrPE 4 2 DdrA cur/ij © Ddrk strdight Light cur/ij one of four possible phenotypes — dark curly, dark straight, light curly or light straight haired individ- uals. These four phe- notypes in turn will present nine different genotypic combina- tions out of sixteen pos- sible cases, as shown in Figure 46. Figure 45 further- more serves to make clear, first, the distinc- tion between somato- plasm and germplasm ; second, the maturation of germ-cells ; third, the segregation of gametes; and fourth, the formation of zygotes in sexual reproduction. The cells of the somatoplasm are represented as le © Phenotype: Light strdight Nuii9 Grandmother grandfather .t*oMOr, %lu}\^ Duplex HorflOZYGOTE HOMOZYGOTE Duplex'j / I Heterozygote \ ^ ' Simplex ^ Mother B. 1 \ y / / Meter o'zYGOTE 1 \^v^ ^' Sim'plexX I ■ » ~1 I > "^ T |/\f A V 1 f"!^ ^- ■ -^ * ^02Y&° '^t^/liplei^ HETfROZYGOTE HOMOZYGOTE HETEROZYGOTE Simplex Duplex 5/mplex Fig. 48. — Three generations of a Mendeliau monohj^brid. The outlines represent the somatoplasms with the phenotypio character on the out- side. The black symbols inclosed within the somatoplasm stand for the germplasm in the form of gametes. The short dotted arrows indi- cate the relation between germplasm and somatoplasm. The long dotted arrows indicate possible recombinations of germplasms. will have blue eyes unlike their own, but like the two blue-eyed grandparents. Such a blue-eyed child would be an instance of atavism. The explanation 148 GENETICS of this apparently inconsistent hereditary behavior is perfectly simple in the light of the Mendelian ratios, as shown diagrammatically in Figure 48, in which the circles represent the blue-eyed and the squares the brown-eyed character. This figure also illustrates what typically occurs in the formation of Mendelian monohybrids of the first and second filial generations. The squares are symbols for the dominant characters, while the circles are symbols for the recessive characters. When the two are superimposed, the circle recedes from view, The large outside figures indicate the somatoplasm, therefore the phenotype. The small inclosed figures indicate the germplasm, therefore the genotype. The short dotted arrows indicate what it is that deter- mines the somatoplasm in each case, while the long dotted arrows show what possible recombinations of germplasms can be made. Child No. 4 is an " ex- tracted recessive" derived from dominant parents, but with one recessive grandparent on each side. It is a case of "atavism," or taking after the grand- parent. Notice that atavism can occur only by alternative inheritance. To quote Davenport: *'In the majority of cases atavism is a simple reappearance in one fourth of the offspring of the absence of a character due to the simplex nature of the character in both parents." An illustration of reversion would be the reappear- ance of the ancestral jungle-fowl pattern in domestic poultry or of the slaty blue color of the ancestral rock-pigeon among buff and white domestic pigeons, OLD TYPES AND NEW 149 for the ancestral character or characters in this type of hereditary behavior, as said before, reappear only after a lapse of many generations. 2. False Reversion "Around the term * reversion,' " Bateson observes, "a singular set of false ideas have gathered them- selves." In proof of this statement there may be cited at least five categories of apparent reversion which properly ought not to be classed as true reversion. a. Arrested Developjnent Feeble-mindedness is not reversion to ancestral forms of less intelligence, but an instance of arrested development when, for some reason, the individual fails to accomplish his normal cycle of development. Likewise harelip in man is not a case of reversion to rabbit-like ancestors in which harelip is the nor- mal condition, but it is ordinarily due to an arrest or failure of certain embryonic steps that are essential to the development of the usual form of human lip. b. Vestigial Structures These are the vanishing remains of characters that were formerly of significance. They do not represent something latent that is now /•^'appearing, for they have never yet disappeared phylogenetically, and con- sequently they cannot be regarded as true reversions. The muscles under the scalp which enable those persons possessing them to wiggle the ears ; the palatine ridges in the roof of the mouth of many 150 GENETICS babies and some adults which resemble the ridges in the roof of a cat's mouth ; the vermiform appen- dix, a necessary part of the digestive apparatus of many animals but fraught so often with evil conse- quences to man ; these and scores of similar charac- ters, which, taken together, make man in the eyes of the comparative anatomist a veritable old curi- osity shop of ancestral relics, are the last traces of characters which formerly had a significance in some of man's forbears. Having lost their usefulness, these structures still hang on to the anatomical household as pensioners. They have not been re- called from the past, but have always been with us, although of diminishing importance. In no sense, therefore, can they be called reversions. c. Acquired Characters resembling Ancestral Ones Sometimes the drunken descendant of a drunken great-grandparent has acquired this characteristic through his own initiative quite aside from any an- cestral contribution to his germplasm. This is not reversion. It is a reacquisition which resembles the ancestral condition. Again, tame animals that run wild acquire habits resembling those of their wild ancestors, but this is not necessarily reversion. It is the natural response of feral animals to the conditions of wild life. d. Convergent Variation The European hedgehog, Erinaceus, an insecti- vore, the American porcupine, Erithizon, a rodent, and the Australian spiny anteater. Echidna, sl OLD TYPES AND NEW 1.51 monotreme, are all mammals which have developed in a similar manner the very peculiar device of der- mal spines. There is no reason, however, for regard- ing this character as due to descent from a common spiny ancestor. It is not reversion to an ancestral type, but rather a case of convergent variation. Similarity does not always indicate genetic continuity. In the case of birds albinism, melanism and fla- vism are modifications of ordinary pigmentation which appear irregularly among many different species as pathological " sports," but no one of these conditions can be regarded as reversions to ancestral white, black, or yellow types. e. Regression Galton's "law of regression" refers to the w^ide- spread phenomenon already explained of a constant swinging back to mediocrity which the breeder must oppose with continual selection in order to maintain the standard of any particular strain. We have seen that within a "pure line," regression is complete and that in populations made up of a mixture of pure lines it is a factor always to be reckoned with. Regression, however, has to do with fluctuating varia- tions and does not bring about a permanent change of type. It should, therefore, not be confused with reversion. 3. Explanation of Reversion Darwin, who did not always differentiate between reversion and atavism, suggested that reversion was 152 GENETICS due sometimes to the action of a more natural en- vironment, as in the case of animals set free after having been in captivity, and sometimes to hybridi- zation, since there seems to be a general tendency of hybridized organisms to "revert" to ancestral types. It is now known that reversion, like atavism, is simply a case of latent characters becoming apparent according to the Mendelian principle of segregation. To quote Davenport: "There is nothing more mys- terious about reversion, from the modern standpoint, than about forming a word from the proper com- bination of letters." 4. Some Methods of improving Old and estab- lishing New Types a. The Method of Hallet This method, which was formulated by the English wheat-grower Hallet in 1869, has been in common use for a long time. It consists in placing the organ- isms to be bred in the very best possible environment and then choosing those individuals which make the best showing as the stock from which to breed further, a procedure based upon the deep-seated belief that acquired characters are inherited. For example, in a field of wheat, plants near the edge of the field which, from lack of crowding or by reason of proximity to an extra local supply of fer- tilizer or any other favorable environmental factor, make a more vigorous growth than their neighbors. OLD TYPES AND NEW 153 are selected in the hope that the gains made by them will be maintained in their offspring. We have seen that it is very questionable whether acquired characters which are due to environmental conditions play any role whatever in heredity. The phenotypic character does not always indicate what the germplasm will subsequently do, and when the true genotypic constitution of the germplasm is still further masked by the temporary fluctuations caused by a modified environment, it is increasingly difficult to select wisely from the display of variants those which will produce the best ancestors for the future stock. That this common procedure of selecting the best- appearing animal in the flock and the biggest ear of corn in the bin, has met with a large degree of success in the past is due entirely to the fact that in many instances the phenotypic character is an actual ex- pression of the genotypic constitution. This is not always the case, however, and we cannot now fail to see that the method is blind and full of error. Its successes are due to the indirect results of chance rather than to a direct control of the factors of hered- ity. The great proportion of failures resulting from this procedure now find a reasonable explanation from the standpoint of Mendelism. b. The Method of Rim pan Contrasted with the Hallet method of augmenting acquired characters and then selecting the best display of them, is the method of Rimpau, who experimented 154 GENETICS for two decades with various grains and, finally, among other results, produced the famous Schand- stedt barley. Rimpau's method is to sow grain under ordinary conditions with a minimum rather than a maximum amount of fertilizer and then to select individuals, neither from the rich spots nor from the edges of the field where there is little crowding, but from situa- tions where the environmental conditions are ordi- nary or even unfavorable. Individuals making a good showing under such usual, or even adverse, conditions are worthy by nature rather than by nur- ture and are consequently most desirable as progeni- tors of future stock. By this method the attempt is not to keep the progeny of single individuals sepa- rate, but to mass together the best as they appear under ordinary normal environment. This again is an indirect method of procedure, although the character of the germplasm is more nearly hit upon in this way than by Hallet's method, since the mask of temporary accessory modifications is stripped so far as possible from the somatoplasm, and the phenotype made to approximate the geno- typical constitution. c. The Method of de Vries The method of de Vries has already been in part described in Chapter IV. It depends upon the pres- ervation and exploitation of the mutations occurring in nature. It recognizes clearly the fact that change of type is dependent upon a germplasmal variation OLD TYPES AND NEW 155 which is largely, if not entirely, independent of environ- mental factors. Accordingly, the work of the successful })reeder consists in simply taking what nature spontaneously furnishes to him rather than in attempting to force nature into producing something new. These muta- tions, when isolated, may become the progenitors of desirable new lines. d. The Method of Vilmorin This is an isolation method which has been success- fully applied to the sugar-beet industry. The seeds from each plant to be tested are sown in separate beds from which upon maturity samples are taken and tested for sugar content. The plants from the bed furnishing the sample which contains the highest per- centage of sugar are then used as the seed producers for the next generation. In this way by continual selection an improved strain may be maintained. e. The Method of Johanssen The method of isolating pure lines or homozygotes out of a mixed population has been considered in Chapter VT. As in the method of de Vries of isolating mutations so, too, in the pure line method it is recog- nized that the germplasm is the source of initiatory changes and that the technique of establishing new types consists in sorting out homozygous strains of this germplasm. The method of Johanssen is quite different from those of Hallet and Rimpau in that the ideal organi- 156 GENETICS zation is not sought for among phenotypes, but among genotypes. It is not the somatoplasm, but the germplasm that is selected. /. The Method of Burhank This is a method of greatly increasing the number of variants by promiscuous hybridization and then of eliminating all except those of a desired phenotypic combination. Indirectly it depends upon the princi- ple of the segregation of unit characters which makes possible rearrangements of these characters according to the laws of chance. The characters themselves remain unchanged, since nothing new is produced by hybridization except new arrangements of existing characters. The spectacular success of Luther Burbank in "creating" new plant forms is due largely to his very extensive hybridizations, his skill in detecting among the varying progeny the winning phenotype and his ruthless elimination of the great majority of varia- tions that do not quite fill his requirement. The successful combinations must be propagated in most instances asexually by grafting, cuttings, bulbs, etc., rather than sexually through the medium of seeds, because new genotypes which will breed true are not necessarily isolated by this procedure. The consequence is that Burbank's method cannot be utilized in animal breeding to any great extent where the maintenance of a desirable strain by asexual prop- agation is out of the question. It will be seen that this method, like the first two, is OLD TYPES AND NEW 157 fortuitous and to a certain extent unscientific in that no one can repeat the exact conditions of the experi- ment and arrive at the same results. It depends upon the chance mixing up of a large number of possibilities and then in not being distracted or blinded by the good while selecting the best. In the hands of a skilful plant breeder with unlimited resources at his com- mand it may result in much practical achievement, but it does not particularly illuminate the path of other breeders who wish to repeat the experiment. It is after all a selection of phenotypes and, therefore, forever open to error, since phenotypes do not always indicate what the behavior of their constituent geno- types will be in heredity. g. The Method of Mendel The method of Mendel, like the foregoing, depends upon hybridization with the difference that the desired combination is sought directly by definite predetermined crosses, according to the expectations of the Mendelian ratios, rather than through the random result of fortuitous combinations. This method has been rendered possible by the determina- tion of Mendel's laws of dominance, and of the inde- pendence and segregation of unit characters which give to the experimental breeder definite expectations and a method of procedure. If, upon hybridization, the desired character be- haves like a recessive, then all that is necessary to establish a pure stock exhibiting the character in question, is to breed two rccessives together, because 158 GENETICS recessives are always homozygous and, regardless of their ancestry, breed true. On the other hand, if the desired character proves to be a dominant, then it is necessary to determine whether it is present in a duplex or a simplex condi- tion; in other words, whether it is homozygous or heterozygous, for only homozygous organisms breed true. Establishing a strain consists, consequently, in making an organism homozygous. The test to determine whether a dominant character is homozygous or heterozygous, that is, whether it will breed true or not, can be made by a single cross according to the procedure outlined in para- graph 8 of Chapter VII. If, upon crossing the individual to be tested with a recessive, it produces an entirely dominant progeny, then its germplasm is duplex for this character, and it will always reproduce the character in either duplex or simplex condition according to what it may be crossed with. When crossed, for instance, with another duplex dominant like itself, a pure homozygous strain of the character in question will be perpetuated. If, on the contrary, the dominant character to be tested proves to be simplex or heterozygous, as de- termined by the fact that, when crossed with a re- cessive, 50 per cent of the progeny are recessive, then it requires more than a single generation to establish a homozygous dominant strain. In random inbreeding of diverse strains if the re- cessives are constantly eliminated as they appear, a population is gradually obtained which is composed OLD TYPES AND NEW 159 of an increasing number of dominants so that after only a few generations the chances are much reduced that recessives will appear, which means the practical purity of the strain. 5. The Factor Hypothesis It has been ascertained within the last decade that some characters require more than a single de- terminer to bring them to expression. The idea of compound determiners for a single character may be termed the factor hypothesis of heredity. The con- verse is also true, that certain single determiners may control more than one character. For instance, the determiner for gray hair in rats also produces a lighter color on the belly. Mendel, whose experiments led him to believe that each character depends upon only a single deter- miner for the reason that he worked on characters severally belonging to different parts of the plant, was apparently unaware of the existence, in certain cases at least, of compound determiners. These compound factors may be arranged in va- rious categories. For example, there may be, — (1) A complementary factor which is added to a dissimilar factor in order that a particular character may appear ; (2) A supplementary factor which is added to a dissimilar factor with the result that a character is modified in some way ; (3) A cumulative factor which, when added to 160 GENETICS another similar factor, affects the degree of expression that a character is given ; (4) An inhibitory factor, which prevents the action of some other factor, and so on. It will be profitable to consider the factor hypothe- sis in some detail, since it helps to explain both rever- sion and the formation of new types. a. BatesoYis Sweet Peas In the course of numerous breeding experiments Bateson obtained two strains of white sweet peas. Lathy r us, which, when normally self -fertilized, each bred true to the white color. When these two strains were artificially crossed, however, the progeny all had purple flowers like the wild ancestral Sicilian type of all cultivated varieties of sweet peas. Here was apparently a typical instance of reversion, but according to the factor hypothesis the explanation is this. The character of purple color is dependent upon two independent factors w^hich, though sepa- rately heritable, are both required to produce it. Each of these white strains of sweet peas possesses one of these factors which can produce colored flowers only when united with its complement, a proof of which appeared upon interbreeding hybrid purples from such a cross. In short, the color purple depends upon the action of two complementary factors which follow the behavior of a dihybrid. (See Chap. VII, par. 10.) The gametic formulse for the two strains of white sweet peas used in this experiment are Cp and cP, respectively. C stands for a color factor without OLD TYPES AND NEW 161 which no color can appear, even lliougli pigment for color may be present, and c is the absence of this factor, while P represents a purple pigment factor which only finds expression in the somatoplasm when Fig. 49. — Diagram to illustrate the possible progeny from two hetero- zygous purple sweet peas according to data from Bateson. C, color factor (large circles) ; c, absence of C (small circles) ; P, pigment fac- tor (large crosses) ; p, absence of P (small crosses). In the zygotes within the checkerboard squares the gametic symbols are superimposed. taken together with the color factor C. The small letter p stands for the absence of the purple pigment factor. It will be seen that each of the white sweet peas whose formukie are given above lack one of the two essential factors for purple color. When the M 162 GENETICS two are crossed, however, all the progeny are purple with the formula CcPp. These hybrid sweet peas upon gametic segregation theoretically produce four kinds of gametes, CP, Cp, cP, and cp which may combine as any other dihybrid in sixteen different ways. In this case, however, these combinations group themselves into only two phenotypes, purple and white, as indicated in the accompanying diagram (Fig. 49) in which C and c are represented by large and small circles respec- tively, while P and p are correspondingly indicated by large and small crosses. The gametic symbols are superimposed to form the zygotes. The theoretical expectation here shown was closely approximated in the actual results. It may be noted in passing that the seven kinds of white sweet peas resulting from the above cross, while phenotypically alike, that is, in the zygotic symbols of Figure 49, lacking either the large circle (color) or the large cross (pigment), belong to three distinct genotypes as follows : — 1 2 3 Without the pigment factor (large cross) Without the color factor (large circle) Without either pigment (large cross) or color (large circle) Number of Zygote in Figure 49 6- 8-14 11 -12 15 16 Among the purple peas are the following four geno- types : — OLD TYPES AND NEW 163 1 2 3 4 Duplex for both color (large circle) and pigment (large cross) Duplex for color (large circle) but simplex for pigment (large cross) Simplex for color (large circle) but duplex for pigment (large cross) Simplex for boLh color (large circle) and pigment (large cross) Number op Zygote in Figure 49 2-5 3-9 4-7-1013 b. Castle's Agouti Guinea-pigs An illustration of a supplementary factor that acts only in conjunction with some other to bring about a modification, is the pattern factor demonstrated by Castle in his guinea-pigs. The wild gray, or "agouti, " color of the hair of cer- tain guinea-pigs is due to the fact that pigment is distributed along the length of each hair in a definite pattern. The tip of a single hair is black followed by a band of yellow, while most of the proximal part which is more or less concealed by overlapping hairs is a leaden color. The distribution of pigment in such a pattern gives the characteristic gray, or agouti color to the coat when taken as a whole. Castle demonstrated the separate nature and be- havior of such a pattern factor when he discovered that it is transmitted independently of pigment, which is necessary to bring it to expression. He showed that upon crossing a solid bhick guinea-pig, imquestionably possessing pigment but no "pattern,'' with a wliite 164 GENETICS albino guinea-pig having no pigment, some of the offspring "reverted" to the ancestral agouti, or "pattern" type, thus proving that the pattern must be carried in this case by the white or albino guinea- pig as a factor independent of the color which is necessary for its expression. c. Cuenofs Spotted Mice Another instance of the interaction of supple- mentary factors is seen in the spotting of piebald mice. Cuenot discovered that such spotting is due to the absence of a uniformity factor which if present causes color to be uniformly distributed over the entire coat. Both of these independent factors, spotting and uniformity, are real and not imaginary, since they may be separately transmitted through albino animals in the same way as the pattern factor mentioned above, notwithstanding that in albinos both are hidden through the absence of pigment, upon the presence of which their visibility depends. Whenever piebald or spotted animals appear in a progeny derived originally from self-colored stock, it is evidently due to the absence of such a "uni- formity" factor as has just been described. Gal ton's theory of "particulate inheritance" (page 121) is now satisfactorily explained as true al- ternative inheritance in which the mosaic appearance is caused by a Mendelian determiner, in this instance a spotting factor or, in other words, the absence of a factor for uniformity. OLD TYPES AND NEW 105 d. Miss Durham s Intensified Mice Miss Durham, in her work with mice, has demon- strated an intensifying factor, the absence of wliich she calls a diluting factor. The action of the former produces, as its name implies, intensity of color, while that of the latter serves to lessen the degree of intensity in which color appears. These factors of intensity and diluteness, it should be observed, do not in any way correspond to the duplex and simplex condition of a dominant color character, either of which would straightway appear if crossed with an albino. The factors of intensity and dilution of color are of an entirely different nature, as they have been proven to be indepen- dently transmissible through albinos where a color character could not appear because of the absence of pigment. The following illustration of this kind of sup- plementary factors taken from Miss Durham's experiments will serve to make the case clear. The symbols employed are : — B = black pigment which masks brown, or chocolate. h = the absence of B, consequently chocolate. I = intensity factor. i = dilution factor or absence of intensity. C = a complementary color factor acting with P. P = a complementary pigment factor acting with C. BICP = black. BiCP = blue or maltese (dilute black). hICP = chocolate. hiCP = silver-fawn (dilute chocolate). 166 GENETICS The crosses which were made are represented in the table below, in which the expectation according to the Mendelian dihybrid ratios is given in paren- theses after the actual results of each cross. Black {BICP) X Silver-fawn {hiCP) Blue {BiCP) X Chocolate {hICP) Blue {BiCP) X Silver-fawn {hiCP) Black {BICP) Blue [BiCP) Choco- late {bICP) 9(9) 42(45) 0(0) 4(3) 16(15) 33(36) 3(3) 14(15) 0(0) Silver- fawn {biCP) 2(1) 8(5) 12(12) It will be seen that the actual results, even when such small totals are concerned, approximate very closely the expectation and are entirely consistent. e. Castle s Brown-eyed Yellow Guinea-pigs Recently Castle has shown that in guinea-pigs there is an independent factor for extension of pig- ment distinct from the uniformity factor already mentioned. The absence of this extension factor ("restriction ") is manifested by a lack of black or brown pigment everywhere except in the eyes and to a slight extent in the skin of the extremities, while the distribution of yellow is wholly unaffected by it. That such "extension" and "restriction" factors really exist, is proven in the following way : — When a brown (chocolate) guinea-pig is crossed with an ordinary black-eyed yellow one, the young are all black pigmented, but by cross-breeding OLD TYPES AND NEW 167 these hybrid young four varieties are obtained in the next generation, viz., black, brown, bhick-eyed yellow, and brown-eyed yellow, the latter a variety unknown before Castle's experiment in breeding was made. For the sake of clearness the formation of the brown-eyed yellow is shown below in Figure 50. cT Y^ V Be V bE be S^ BE BE Black 5 Be|]BE Black bE| BE Black 13 BE b G Black V Be ^ BE Be Black 6 Be Be Bfack-eijed Yellow JO bE B© Black 14 be B Black-eHed Yellow \y be 3 BE bE Black 7 Be bE Black I] bEllbE Chocolate 4 BE be Black 8 B e be Black -eHedYellovj 12 bE be Chocolate 15 16 be bE be be Chocolate Brown-e\jed Yellow Fig. 50. — Diagram to illustrate the origin of a brown-eyed yellow guinea- pig from two heterozygous black parents based upon Castle's experi- ments. The factor for yellow (F) is present in every gamete and is consequently duplex in every zygote but is hidden whenever the fac- tor B is present. B, black pigment hiding brown or chocolate ; h, chocolate (absence of B) ; E, extension of B over the entire body hiding Y ; e, restriction of B to eyes alone thus exposing Y over the entire body. 168 GENETICS Symbols B = black pigment, hiding brown or chocolate. b = absence of B, or chocolate. Y = yellow pigment, hidden by B. E = extension of B over entire body, hiding Y. e = restriction of B to eyes alone, thus exposing Y over the entire body. C = complementary color factor acting with P to produce color. P = complementary pigment factor acting with C to produce color. (The factors C and P may be omitted for the sake of simplicity, since they are present in each instance.) First Cross "Extended" chocolate (bEY) X black-eyed yellow (BeY) = black {BbEeYY). Second Cross When these cross-breds are mated with each other, they each form four kinds of gametes, BEY, BeY, bEY, and beY, which unite into sixteen theoretical genotypic possibilities, shown in Figure 50. These fall into four phenotypes, nine black (BEY), three black-eyed yellow (BeY), three chocolate (bEY), and one brown-eyed yellow (beY). The actual results in Castle's experiments gave all four kinds in close numerical agreement with this expectation. The action of extension and restriction factors is, therefore, plainly a case of Mendelian dihybridism in which two independent pairs of alternative char- acters are concerned. OLD TYPES AND NEW 1G9 6. Rabbit Phenotypes Perhaps no better application of the factor hy- pothesis may be found than the case of the color of rabbits. There are many varieties of rabbits so far as color is concerned, particularly among domesticated races. These varieties are now quite explainable by the factor hypothesis, as indicated in the table below. The sixteen kinds of rabbits there catalogued have The Factor Hypothesis applied to Colors of Rabbits Constant Factors Alternative Factors Gametic Formula Phenotypic Character when Crossitd with the Same Kind of G\METir CoMniN\Tinv 1 2 3 4 C 5 E 6 I i 7 8 B Y U u A AUIEClYBBr] Gray a A a aUIEC [YBBr] Black AuIEC [YBBr] Gray spotted auIEC [YBBr] Black spotted U A AUiEC [YBBr] Blue-gray a aUiEC [YBBr] Blue (Maltese) u U A a A AuiEC [YBBr] Blue-gray spotted Br auiEC [YBBr] Blue spotted e I i AUIeC [YBBr] ( Yellow (with white belly ( and tail) a aUIeC [YBBr] ( Sooty yellow (with yellow i belly and tail) u A a AuIeC [YBBr] Yellow spotted auIeC [YBBr] Sooty yellow spotted U A AUieC [YBBr] Cream a aUieC [YBBr] \ Pale sooty yellow u A a AuieC [YBBr] Cream spotted auieC [YBBr] Pale sooty yellow spotted 170 GENETICS been obtained by Castle and other experimental breeders as well as many of the albino types that would double this list if c, or the factor for absence of color, should be substituted for C, the presence of color, in column 4 of the table on page 169. Explanation of Symbols in the Foregoing Table Br = a factor acting on C to produce hroivn pigmentation. B = a factor acting on C to produce black pigmentation. F = a factor acting on C to produce yellow pigmentation. The three factors, F, B, Br, are present in every rabbit gamete and up to date have not been sepa- rable as independent unit characters, although they have been separated out in guinea-pigs and mice. There are no brown rabbits, because black always goes linked with brown covering the brown factor. Yellow rabbits result, as explained below, through the action of factor e. C = a common color factor necessary for the production of any pigment. It was discovered in 1903 by Cuenot. c = the absence of C which results in albinos, regardless of whatever pigment factors may be present. By changing C to c, sixteen kinds of albinos would be added to this catalogue, an addition of one phenotype and sixteen genotypes, all looking alike but breeding differently. E = Si factor governing the extension of black and brown pigment, but not of yellow, e = the absence of extension or restriction of black and brown pigment to the eyes and the skin of the extremities only, while yellow remains extended and visible. Demonstrated by Castle in 1909. OLD TYPES AND NEW 171 / = an intensity factor which determines the degree of pigmentation. It can be transmitted indepen- dently of C through an albino. Discovered by Bateson and Durham in 1906. i = the absence of intensity or dilution. Dilute black = blue. Dilute yellow = cream. Dilute gray = blue-gray. U = SL factor for uniformity of pigmentation or " self- color" discovered by Cuenot in 1904. u = the absence of uniformity which results in spotting with white. A = Si pattern factor for agouti, or wild gray color, which causes the brown and black pigments to be ex- cluded from certain portions of each hair, resulting in the gray coat. When present in the rabbit, it is also associated with white or lighter color on the under surfaces of the tail and belly. It was demonstrated bv Castle in 1907. a = the absence of the agouti or pattern factor. 7. The Kinds of Gray Rabbits Each of the apparent kinds of gray rabbits indicated in the foregoing table may be made up of various genotypes. For instance, there are thirty-two differ- ent genotypes, each of which is phenotypically a gray rabbit. The zygotic formula for each of these thirty- two possibilities is displayed in the next table, and it will be seen that these range all the way from rabbits homozygous in all their variable characters (No. 1) to those homozygous in none (No. 3'2). The progeny of these various types of gray rabbits when inbred wall consequently vary from the pure 17^ GENETICS The Kinds of Gray Rabbits (Color only) Num- ber of Genotype Zygotic Formula Het- EROZY- GOTIC Phenotypes When inbred, these kinds are produced Fac- tors 1 AAUUIIEECCl YBBr][YBBr] None X 2 AAUUIIEECc [ YBBr][YBBr] One X X 3 AAUUIIEeCC [ YBBr][YBBr] One X X 4 AAUUIiEECC [ YBBr][YBBr] One X X 5 AAUuIIEECC ! YBBr][YBBr] One X X 6 AaUUIIEECC [ YBBr][YBBr] One X X 7 AAUUIIEeCc YBBr][YBBr] Two X X X 8 AAUUIiEECc YBBr][YBBr] Two X X X 9 10 AAUuIIEECc AaUUIIEECc YBBr] [YBBr] YBBr][YBBr] Two Two X X X X X X 11 12 13 14 AAUUIiEeCC AAUuIIEeCC AaUUIIEeCC AAUuIiEECC YBBr] [YBBr] YBBr] [YBBr] YBBr] [YBBr] YBBr] [YBBr] Two Two Two Two X X X X X X X X X X X X X X X X 15 16 AaUUIiEECC AaUuIIEECC YBBr] [YBBr] YBBr] [YBBr] Two Two X X X X X X X X 17 AaUuIiEECC YBBr] [YBBr] Three X X X X X X X X 18 AaUuIIEeCC [YBBr] [YBBr] Three X X X X X X X X 19 AaUuITEECc YBBr] [YBBr] Three X X X X X 20 AaUUIiEeCC YBBr] [YBBr] Three X X X X X X X X 21 AaUUIiEECc YBBr] [YBBr] Three X X X X X 22 AaUUIIEeCc YBBr] [YBBr] Three X X X X X 23 AAUuIiEeCC YBBr] [YBBr] Three X X X X X X X X 24 AAUuIIEeCc YBBr] [YBBr] Three X X X X X 25 AAUuIiEECc YBBr] [YBBr] Three X X X X X 26 AAUUHEeCc YBBr] [YBBr] Three X X X X X 27 AAUuIiEeCc [YBBr] [YBBr] Four X X X X X X X X X 28 AaUUIiEeCc [YBBr] [YBBr] Four X X X X X X X X X 29 AaUuIIEeCc YBBr] [YBBr] Four X X X X X X X X X 30 AaUuIiEECc [YBBr] [YBBr] Four X X X X X X X X X 31 AaUuHEeCC [YBBr] [YBBr] Four X X X X X X X X X X X X X X X X 32 AaUuIiEeCc [YBBr] [YBBr] Five X X X X X X X X X X X k; 2- CO t3 o X o v; CO o e+ a> X o p 3 X 71 X o ►1 p 3 w ■o o & X 'i > o •-< o p '< CO o m p" a m •o o w c CD IK 1-1 p o c a- a> CO o o OLD TYPES AND NEW 173 gray, as in No. 1, to a gray from which sixteen pos- sible types of young may be expected as in No. 32. Up to the time when Castle's paper upon the factor hypothesis ^ was published in 1909, nine genotypic kinds of gray rabbits had been obtained in his ex- periments, whose genotypic formulae correspond to the following numbers in the list : 1, 3, 6, 10, 13, 20, 22, 28, 29. 8. Conclusion That a relatively small number of factors may pro- duce an extensive array of combinations is evident from this data. The analysis of germplasm by the factor hypothesis is now being generally applied by geneticists to the particular organisms with which they are concerned. It has been carried out notably in detail by both Bateson and Davenport for poultry and byBaur for the snapdragon. Antirrhinum. Finally, the elucidation of the factor hypothesis makes any further explanation of reversion super- fluous. It is now easy to see how a particular char- acter may remain latent for generations and at last come to expression only when the missing factor necessary to its activity is supplied by some cross. It is also clear how hybridization, in w^hich many characters are concerned, is bound to furnish far more new combinations than would, at first thought, be expected. ^ "Studies of Inheritance in Rabbits." Carnegie Institution Publica- tions, No. 114, 1909. W. E. Castle in collaboration \vith Walter, Mul- lenix and Cobb. CHAPTER IX BLENDING INHERITANCE 1. Relative Value of Dominance and Segre- gation Of the three fundamental principles which underlie "Mendel's law," namely, segregation, independence of unit characters, and dominance, the principle of dominance has been found to hold true in a surpris- ing number of cases and in relation to very diverse organisms, notwithstanding the fact that the time spent in the investigation of dominance, as that term is now understood, has been comparatively short. Doubtless future experimentation will demonstrate the existence of dominance to a far greater extent than has at present been discovered. Its universal application is by no means assured, however, since the mathematical precision with which it works, that following its discovery in 1900 has so captivated the biological world, is beginning to give way in the face of many exceptions which have been steadily accumulating. Even Mendel himself noted certain exceptions to the law of dominance, and his followers have pointed out with increasing emphasis that it is subject to many modifications. It is now understood, indeed, 174 BLENDING INHERITANCE 175 that segregation, not dominance, is the most essential factor in the Mendehan scheme. 2. Imperfect Dominance It frequently occurs that dominance is so imperfect that a heterozygous, or simplex, dominant may be distinguished at once by simple inspection from a homozygous, or duplex, dominant, whereas the test of crossing with a recessive is necessary wlienever dominance is complete, as has been previously ex- plained. The single dose of the determiner in such a case has plainly, then, less phenotypic effect than a double dose. There are many cases of imperfect dominance among flowering plants. Cor reus has shown that when plants of a white-flowering race of the "four-o'clock," Mirahilis jalapa, are crossed with those of a red- flowering race, all the offspring in the first filial genera- tion, unlike either parent, exhibit rose-colored flowers. When, however, these rose-colored flowers are crossed with each other, they produce red, rose, and white in the Mendelian ratio of 1 : 2 : 1 ; that is, three colored to one white. The red-flowering race thus proves to be homozygous and the rose-flowering race hetero- zygous. Here color dominates the absence of color, or white, but the degree of the color depends upon whether the dose of pigment is duplex or simplex. A classic illustration of imperfect dominance among animals is the "blue Andalusian fowl," the hereditary behavior of which is illustrated below (Fig. 51). It will be seen that when two blue Andalusian fo\Als, 176 GENETICS characterized by a mottled plumage, are bred together, they produce three kinds of offspring in the ratio of 1:2:1. Twenty-five per cent are clear black, 50 per cent are blue Andalusian, and 25 per cent are white "splashed" with black. Both the black and the splashed white fow^ls from this cross prove, upon further breeding, to be homozygous, while the blue Andalusian itself is heterozygous and can, therefore. Andalusian Andalusian i~" Black T T 1 Andalusian Anda T T T usian Splashed White T BlacK Black Andalusian Andalusian Sp<- White SpL White Andalusian Fig. 51. — The heredity of the blue Andalusian fowl, an illustration of "imperfect dominance." never be made to breed true. In order to produce 100 per cent of blue Andalusian chicks, it is necessary simply to cross a splashed white with a black Anda- lusian. There is nothing in this case to indicate whether the black or the splashed white should be regarded as the homozygous dominant, since dominance is im- perfect. In either case the heterozygous blue Anda- lusian is at once evident in the first filial generation without further crossing. A similar case of imperfect dominance is furnished by the roan color of cattle w^hich results when red and white are crossed. If two roans are mated, they BLENDING INHERITANCE 177 produce red, roan, and white offspring in the propor- tion of 1:2:1, thus showing that roan is a heterozy- gous character in which the dominance of red is imperfect. Even in cases of apparently perfect dominance it is sometimes possible by close inspection to detect differences between a pure dominant {DD), Figure 43, and a heterozygous dominant [DR] when a superficial examination is not sufficient to distinguish them. For instance, in the cross between smooth and wrinkled peas, a microscopic examination of the starch-grains in the cotyledons of the hybrid peas shows that they are of two kinds. Darbyshire calls attention to the fact that, in the power of absorption, hybrid smooth peas {DR) are intermediate between their pure dominant smooth {DD) and pure recessive wrinkled {RR) parents. 3. Delayed Dominance A character which is really dominant is sometimes so late in manifesting itself in the individual growth of the offspring that it may properly be termed a delayed dominant. Dark-haired individuals often do not acquire their definitive hair color until adult life, and it is common knowledge that the eyes of an infant for a consider- able period provoke no little speculation among ador- ing relatives as to " whose eyes " they are. According to Davenport, when a white Leghorn fowl is crossed with a black Leghorn, white being dominant in this case, chicks are produced that are N 178 GENETICS white with black flecks in their plumage. These black flecks, however, disappear at the time of the first molt. The complete dominance of white is, there- fore, simply delayed. 4. "Reversed" Dominance In certain instances there seems to be a reversal of dominance, as may be illustrated by Lang's results w^ith snails (Helix) . He has proven in his experiments that red snails are generally dominant over yellow snails, although in certain cases there is apparently an exception to the rule, for snails with yellow shells dominate those with red shells. Davenport also has shown that although extra toes are usually dominant over the normal number in poultry, yet, in something like 20 per cent of the cases, the normal number is dominant. To speak of these cases as instances of "reversed dominance," is open to serious objection, since such an explanation does not agree with the generally accepted "presence and absence" idea of heritable characters. It is difficult to see how the presence of a certain determiner can dominate in a part of the offspring of any cross and the absence of the same determiner be able to dominate the remainder. It is perhaps nearer the truth to conceive that in cases of apparent "reversal" of dominance there is an insufficient amount of a particular determiner available to bring the character concerned into expression. In other words, although a dominant character may be present in two cases, yet in one BLENDING INHERITANCE 179 it fails, for some reason, to become effective. This interpretation agrees with the facts brought out by subsequent breeding in cases of this sort. It sometimes occurs that a character wliich is dominant in one species may be recessive in anotlier. Horns are dominant in sheep, but recessive in cattle. White color is recessive in rodents and sheep, but dominant in most poultry and in pigs. 5. Potency Davenport seeks to explain modifications in typical dominance as variations in the potency of determiners. He defines potency as follows: "The potency of a character may be defined as the capacity of its germi- nal determiner to complete its entire ontogeny." That is, if the potency of a determiner, for some reason, is insufficient, there may be either an incom- plete or delayed manifestation of the character in question, or it may fail entirely to develop. The variations of potency may be grouped into three general categories according to the degree of their manifestation ; namely, total potency, partial potency, and failure of potency. A further word of explanation for each of these three kinds of potency seems desirable at this point. a. Total Potency This is complete Mendelian dominance in which even the heterozygotes produced by a sini])lex dose of a character are indistinguishable phenotypically, that is, by inspection, from the homozygotes produced 180 GENETICS by a duplex dose of the same character. It is as if a single bottle of black ink poured into a jar of water was just as effective as two bottles of ink, in forming an opaque fluid. h. Partial Potency Partial potency covers all cases of incomplete dominance, such as those of the four-o'clock {Mirahilis) and blue Andalusian fowls, where a simplex dose of a determiner does not produce the same visible effect as a double dose. The dominant prickly Jamestown weed {Datura) ^ when crossed with a recessive glabrous variety of the same plant, produces cross-breds in the first generation which show only a few prickles (Bateson) (Baur), following the law of partial potency. Banded and uniformly colored snails also, when crossed together, produce snails with shells showing only a pale banding (Lang). Numerous further instances of incomplete domi- nance could be cited. c. Failure of Potency If for any reason a determiner fails to accom- plish its possibilities in whole or in part, then the character in question may never become evident, and the result, so far as appearances go, is the same as if it was a recessive lacking the determiner entirely. That the failure of potency is not identical with the absence of a determiner can usually be demon- strated by further breeding, because dominants failing BLENDING INHERITANCE 181 in potency, which are either of the formula DD or /)/?, may, if bred inter se, give a various prof:^eny among which the dominant character D is hkely to again become manifest, while recessives, of the formula RR, on the contrary, will always give offspring which all agree in the entire absence of the character in question. Davenport cites an extreme case of failure of potency in one of two rumpless cocks from the same blood. The character of rumplessness is due to an inhibitor of tail development. That these two cocks both possessed this character was demonstrated by the entire absence of any tail in either case. The in- hibiting determiner for tail growth was so weak in cock No. 117, however, that, to quote Davenport's exact words : "In the heterozygote the development of the tail is not interfered with at all, and even in ex- tracted dominants it interfered little w^ith tail develop- ment, so that it makes itself felt only in the reduced size of the uropygium and in-bent or shortened back. But in No. 116 the inhibiting determiner is strong. It develops fully in about 47 per cent of all the heterozygotes and in extracted dominants may pro- duce a family in all of which the tail's development is inhibited." Here were two birds of the same blood, pheno- typically alike and presumably genotypically alike, which because of an individual difference in the potency of the determiner for rumplessness produced quite different results in their offspring although bred to precisely the same array of hens. 182 GENETICS 6. Blending Inheritance In the instances of imperfect dominance given above, where the progeny of unhke parents present an intermediate condition, it is found that, upon cross-breeding these offspring, segregation into the grandparental types occurs just as truly as in instances of complete dominance. In poultry, for example, when Cochins, which are "booted," and Leghorns, which are clean-shanked, are crossed, booting of an intermediate grade of four results, on a scale in which ten represents complete booting, and zero no booting or clean shank (Daven- port). The character of booting and its alternative absence, however, segregate out in true Mendelian fashion when these hybrids are subsequently crossed together. It is evident that dominance plays only a secondary role in such cases, and that the all-im- portant factor is segregation. Are there, then, any cases where true fusion of hered- itary parental traits occurs, in other words, where segregation in the second filial generation does not appear? Does the "melting-pot of cross-breeding" ever "melt" the characters thrown into it .^ It was formerly believed that diverse parents generally produce intermediate offspring, and that this intermediate condition continues without any segregation at all in the form of "blending inheritance," but within the last decade apparent cases of blending inheritance have been thrown out of court one after the other by the Mendelians. Bateson, in an inaugural BLENDING INHERITANCE 183 address at Cambridge University in 1908, stated that what was once beheved to be the rule has now be- come the exception. He goes on to say: *'One clear exception I may mention. Castle finds that in a cross between the long-eared lop rabbit and a short-eared breed, ears of intermediate length are produced ; and that these intermediates breed approximately true." Let us examine this "one clear exception" a little more closely. 7. The Case of Rabbit Ears As a typical example of blending inheritance in rabbit ears may be cited the following case : — A female Belgian hareAvith an ear-length of 118 mm. was crossed with a male lop-eared rabbit with an ear-length of 210 mm. The average of these ear- lengths is 164 mm. Five offspring from this pair had ear-lengths, when adult, approximating this aver- age as follows: 170, 170, 166, 156, 170, of which two were females and three were males. When from this litter one of the females measuring 170 mm. in ear-length was subsequently crossed with her brother having an ear-length of 166 mm., two litters were produced in which the individuals when adult at- tained ear-lengths of 170, 166, 168, 160, 172, and 168 mm. These results are represented diagram- matically in Figure 52. This illustration is typical of many other breed- ing experiments made by the same investigators ^ * Castle, in collaboration with Walter, Mullenix and Cobb. "Studies of Inheritance in Rabbits." Carnegie Institution Publications, Wash- ington, No. 114, 1909. 184 GENETICS ;t ....b' / 2 34 56 Offspring ofZsnd 5 •»- r- t- c~- (B lo CM ' — '- e e oo 6669 dd E E E- e E e e e s e E E o «X) to «o p w (S) ^Q iO iO t*» r* Offsprin^of land? Fig. 52. — A case of three generations of ear-length in rabbits, a-b, average ear-length of the first filial generation (Fi). a'-b', average ear-length of the F2 generation derived from 1 and 7. Data from Castle, in collaboration with Walter, Mullenix and Cobb. upon the ear-length of rabbits which included 70 different litters of rabbits containing 341 individuals. In none of these experiments could the blend in the BLENDING INHERITANCE 185 second filial generation be called perfect, but it may at least be said that evidence of segregation, that is, a return to one or the other of the parental types, was much less apparent than evidence of blending. Furthermore, crosses were made in which lop ears of various fractional lengths were obtained as desired, including |, |, f , J, f , J, and | lengths. Not one of these fractional lengths apparently segregated in subsequent generations after the Mendelian fashion, but all bred approximately true. Moreover, ears of one half lop length, for instance, were obtained in three ways : first, by crossing full- length lops with short-eared rabbits as indicated in the first cross of the case cited above; second, by crossing one half lop lengths together, demonstrated by the second cross in the illustrative case given, and third, by mating J and f lop lengths. Theoretically, I and I as well as f and f lop lengths would also pro- duce I lop lengths, for in all of the crosses that were made the length of ear behaved in a blending fashion. These results were based, not upon a single measure- ment of each specimen, which might be open to considerable error, but upon daily measurements from the time the rabbits were two weeks old until their ears ceased to grow at about twenty weeks. The growth curves drawn from these daily measurements showed continually an intermediate or blending condi- tion in progeny derived from diverse parents. A Mendelian explanation of this apparently excep- tional case of blending inheritance has been suggested by Lang based upon the result of Nilsson-Ehle's 186 GENETICS discoveries while breeding wheats at the Agricultural Experiment Station of Svalof in Sweden. 8. The Nilsson-Ehle Discovery Nilsson-Ehle found in breeding together different strains of wheat that a certain wheat with brown chaff crossed with a white-chaffed strain yielded only brown-chaffed wheat in the first generation. These heterozygous or hybrid brown-chaffed w^heats when crossed with each other produced, not the expected proportion of three brown to one white, but fifteen brown to one white. This was not explainable as the chance result of a single cross, but was the conclusion drawn from fifteen different crosses all of the same strains that yielded a total progeny of 1410 brown- chaffed to 94 white-chaffed plants, which happens to be exactly the proportion of fifteen to one. In other experiments it was discovered that although dominant red-kerneled strains of wheat crossed with white-kerneled varieties usually gave the three- to-one proportion upon segregation in the second filial generation, yet one particular strain of red- kerneled Swedish wheat in the second generation gave approximately sixty-three red to one white- kerneled strain. The explanation of these two unexpected results is this. In the case of brown-chaffed wheat there are two independent determiners for the character of brown color, and these simply follow the Mendelian laws for a dihybrid, w^hile in the case of the red- kerneled wheat there are three independent deter- BLENDING INHERITANCE 187 miners for the character of red color, each of which is able to give red color to the wheat. Taken together, these three determiners behave cumulatively, follow- ing the law of a trihybrid. For example, if a brown-chaffed wheat with the for- mula BB\ in which B and B' each represent a brown- chaffed factor, is crossed w^ith a white-chaffed wheat of the formula bb\ in which b and b' each represent the absence of B and B^ respec- tively, then all the progeny of this cross will be brown-chaffed, having the zy- gotic formula BBW. When upon matura- tion the gametes form out of the germ-cells from such hybrids, the following four combina- tions are pos- sible, and no others : BB\ Bb\ bB\ bb\ These represent, there- fore, the possible gametes present in each sex of the first filial generation, and upon intercrossing they can combine into sixteen possible zygotes to form the second filial generation, as shown in Figure 53. bb' Bb' bB' b b bb' bb; bb @ Bb', BB (D bB' BB' © b b' BB' © Bb BB' Bb' ® B b' Bb' ® b B' Bb' ® b b' Bb' ® bB' BE' b B' © B b' b b' © b B' b B' ® b h' b B' ® BB' bb' b b' ® B b' b b' ® b B' b b' ® b b' b b' ® Fig. 53. — Diagram of the possible combinations in the F2 generation of brown-chaffed wheat according to experiments of Nilsson-Ehle. B and B' are cumulative factors for the brown- chaff character, b and b' denote the absence of B and B' respectively. 188 GENETICS The numbers in the squares indicate how many times a brown determiner is present in each zygote. It will be seen that only one out of the sixteen possi- bilities lacks a brown-chaff factor, and this one will n n n m u n n + + + f 3 ^ / Number of doses of the brown determiner Fig. 54. — The distribution of the sixteen possibilities resulting when two similar determiners (brown-chaff) act together as a dihybrid. consequently produce only white chaff, while the re- maining fifteen possibilities, each of which has at least a single determiner for brown, will all yield brown chaff. The brown-chaff factor, moreover, is present in BLENDING INHERITANCE 189 varying doses among these fifteen possil)ilities, as indi- cated by the numbers in the squares. It is evident, therefore, that several shades of brown will be rep- d< ©00 ©© ©60 (D ©0 © e ©0 4 ©60 1) ©0 3 © o ©0 3 ©60 a> ©0 3 © 6 ^ 00 2 6 ©00 e 4 ©0 © G 3 ©90 © 6 3 <» ©0 © © 3 ©60 ©00 z © 3 o ©0 e ©00 e 3 ©0 6 2 © © Q e 2 <" © 0) 6 2 ©60 (D e 1 a> © © 6 1 ©60 © 6 1 © 6 © 6 Fig. 55. — Diagram to illustrate Nilsson-Ehle's case of trihybrid red ' wheat. The large screwheads each represent a single determiner for the red character. The small screwheads symbolize the absence of the red character, or white. The number in each square indicates how many doses of the "red" determiner is present. For further explanation see text. resented depending upon the number of doses of the brown determiner in each instance. Figure 54 shows how these different shades of brown arrange themselves in the manner of a fre- quency polygon of fluctuating variation with the greatest number in the halfway class and the least 190 GENETICS numbers at the two extremes. In this instance six out of sixteen individuals of the second gen- eration theoretically present a perfect "blend" be- tween the original brown- and white-chaffed grand- parents, although complete segregation has actually occurred. The same explanation holds true as displayed in Figure 55 for the trihybrid case of red- and white-kerneled wheats in which only one white- kerneled to sixty-three red-kerneled individuals ap- pear in the second filial generation. The number of red determiners in each zygote is indicated by the figure at the bottom of each square. The large screw- head symbols with vertical, horizontal and diagonal slots each represent an independent determiner for red kernel, while the small screw heads symbolize the absence of each of these determiners, or white kernel. When the pure strain of red-kerneled wheat is crossed with a pure strain of white-kerneled wheat, the first generation is all a heterozygous red of a Pure red -h cuhite = Hybrid red Fig. 56. — The result of crossing white wheat with trihybrid red wheat. somewhat lighter shade than the original pure red strain. When plants of this heterozygous sort are crossed together, they yield plants producing red-kerneled and white-kerneled wheats in the proportion of sixty- three to one. The sixty-three kinds of red wheats are BLENDING INHERITANCE 191 self-crossing plants of the second generation. It was to be expected that, if these hybrid wheats of the second generation carried one, two, three, or more determiners for a red kernel as the theoretical tables in Figures 55 and 57 demand, their progeny would be distributed with reference to the number of red- and white-kerneled individuals, in the following ratios : — m m m m m of varying degrees of redness and may be classified after the manner of fluctuating variations with the greatest number of kinds at the intermediate degree between pure red and pure white. (See Figure 57.) In order to test whether the sixty-four kinds of wheats produced in the second filial generation, as theoretically displayed in Figure 55, really contain separable, though indistin- guishable, determiners for red-kernel, Nilsson-Ehle produced families of the third filial generation by m « » a m m m m m m » m m # # # # # # # # # # # # # # # + + + + + + 1 6 ' 5 r[G. 57. — Thedistributionof the sixty-four possibilitifs in tin- F-i generation when three similar determiners act together as a trihybrid. 192 GENI^TICS 3 red to 1 white when 1 determiner for red is present. 15 red to 1 white when 2 determiners for red are present. 63 red to 1 white when 3 determiners for red are present. All red to no white when 4 or more determiners for red are present. Among seventy-eight sample families of the third generation inbred to test this theoretical conclusion, the actual results were : — 8 families giving the ratio of 3 red to 1 white. 15 families giving the ratio of 15 red to 1 white. 5 families giving the ratio of 63 red to 1 white. 50 families giving the ratio of all red to no white. . It has been actually demonstrated therefore, in the case of this particular strain of wheat: (1) that the factors producing red kernel are several in number; (2) that they act independently of each other in heredity; (3) that these several independent factors segregate; and (4) that any one red factor acting alone produces a "red" result. The Nilsson-Ehle principle of cumulative determin- ers has been confirmed in America by East in a mas- terly series of breeding experiments upon maize. In connection with the Nilsson-Ehle principle, it will ,be seen that the possible number of intergrades between the two extremes increases rapidly as the number of duplicate determiners increases. Thus with six duplicate determiners for the same character present, the ratio of possible dominants to recessives in the second filial generation would be 4095 to 1. The reappearance of this single recessive among 4095 BLENDING INHERITANCE 193 dominants would be extremely unlikely, and it might easily be mistaken for a mutation or a freak. Appar- ent blends of all intermediate degrees, however, would be sure to appear. Yet these are not blends in the *' melting-pot" sense at all, but strictly cases of Men- delian dominance and segregation. 9. The Application of the Nilsson-Ehle Ex- planation TO the Case of Rabbit Ear- length The so-called blending rabbit ears, along with other similar cases, can now be made to fall into line, as pointed out by Lang, with the Mendelian law of segregation. If we assume that the long ear of the lop rabbit has only three independent but equal determiners for excess length, the case becomes one of Mendelian trihybridism with cumulative factors, which works out like Nilsson-Ehle's red-kerneled wheat in the following manner: — In general the average for full lop ear-length may be placed at 220 mm. and for the ordinary short- eared rabbit ^ at 100 mm. The difference, or the ex- cess length of the lop ear, is 120 mm., which, according to the trihybrid formula, corresponds to the six doses of the character symbolized in the U])per left-hand square in Figure 55 by six large screw heads, three ^ Not the Belgian hare, as cited in the illustration given in Figure 52. The Belgian hare has typically a somewhat longer ear than the ordinary short-eared rabbit, o 194 GENETICS coming from each parent respectively. If all of these independent determiners are equal as regards excess ear-length, each factor would represent an excess of 20 mm. above the normal ear-length found in short- eared rabbits, that is, — 220 mm. - 100 mm. 6 = 20 mm. When according to this computation a lop (20 mm. X 6 + 100 mm. = 220 mm.) and a pure short- eared rabbit (20 mm. X + 100 mm. = 100 mm.) are crossed, if imperfect dominance occurs, which is a very common phenomenon, it is true that the offspring might present a "blended" appear- ance. If now these cross-breds of the first gen- eration prove to be trihybrids with respect to excess ear-length, there w^ould be sixty-four possibilities in their progeny segregating out just as in the red- kerneled wheat. These possibilities would be arranged in the fol- lowing f reqviencies : — Number of Excess Ear- Number of Cases occur- Total Length in Milli- length Determiners ring OUT OF 64 meters OF Ears resulting 6 1 220 5 6 200 4 15 180 3 20 160 2 15 140 1 6 120 1 100 BLENDING INHERIJANCE 195 Since the average litter among rabbits is about five, the chances that these five rabbits will breed true to their hybrid parents and form a perfect blend between their grandparents is 20 out of 64, while the chance of their being like either grand- parent is only one out of 64. It should be noted further that 50 out of 64, or 77 per cent, of these hybrids of the second filial gen- eration would have an ear-length between 140 and 180, thus approximating a "blend" closely enough to be so classified upon a casual inspection. Moreover, if it should be found that excessive ear-length in rabbits is due to more than three dupli- cate determiners, the possibilities of getting anything but an apparent blend would be much decreased. The fact, furthermore, that the fractional ear- lengths of the hybrid rabbits in Castle's experiments bred approximately true in the second and subse- quent filial generations, may also be explained by the Nilsson-Ehle hypothesis. For example, half lop lengths, according to this explanation, are those with three doses of the deter- miner for excess ear-length. It follows that the progeny of two rabbits each carrying three doses of a determiner will likewise, after the reduction during the maturation of the germ-cells, have three doses of the determiner {—^ — = ^V It would be interesting to breed rabbits having ears of one eighth lop length in w^hich, according to the foregoing hypothesis, there would presumably be 196 GENETICS present only a single determiner for excess ear-length, with ordinary short-eared rabbits having no excess ear-length, in order to see if the expected Mendelian three-to-one proportion for a monohybrid would ap- pear in the progeny. 10. Human Skin Color Finally, although accurate published data is wanting, it is probably true that skin color in all kinds of hybrids resulting from crosses between negroes and whites is not a case of blending inherit- ance, as commonly supposed, but rather of true Mendelian segregation. In fact, there is frequently visible evidence that segregation does occur, as shown by many authentic instances where the offspring of diversely colored parents produce children with skin color of different shades. If human families included hundreds of offspring in a single generation instead of the usual number, the problem of skin color in man could doubtless be quickly solved since ratios could then be obtained large enough to reveal the underlying laws of inher- itance. CHAPTER X THE DETERMINATION OF SEX 1. Speculations, Ancient and Modern From the earliest times the desirabihty of con- trolhng the sex of an unborn child, in particular instances at least, has seemed very great. Likewise the wish to be able to predetermine sex among do- mesticated animals has made breeders quick to grasp at every clue that promised success. There has been no want of speculations concerning the determination of sex. J. Arthur Thomson, who with Professor Geddes wrote "The Evolution of Sex" in 1889, says : "The number of speculations as to the nature of sex has been well-nigh doubled since Dreylincourt, in the eighteenth century, brought together 262 'groundless hypotheses' and since Blumenbach caustically remarked that nothing was more certain than that Dreylincourt's own theory formed the 263rd. Subsequent investigators have long ago added Blumenbach's theory of 'Bildungs- trieb' or formative impulse, to the list." It maybe added in passing that the hypothesis of the deter- minative action of external factors upon developing germ-cells which Geddes and Thomson elaborated in the book just referred to, has, in its turn, accord- ing to most biologists, joined the long roll. 197 198 GENETICS Hippocrates thought that sex of the offspring depends upon the relative "vigor" of the parents, while Sadler (1830) concluded that the relative ages of the two parents is the determining factor. Other writers, on the contrary, have thought that the age of the mother at the time of childbirth deter- mines the sex of the offspring, and Thury (1863), in the days before the facts of maturation were known, ascribed the determinative factor to the relative degree of "ripeness" of the egg when fertilized. It was once assumed also that the right ovary or the right testicle is the seat of one sex and the left ovary or left testicle of the other. Galen, who did the biological thinking for several centuries of mankind, asserted that the right side of the body, "being warmer" than the left, consequently produces males. Schenk cites a most amazing bit of folk-lore to the effect that: "In Servia if a man has a stye on his eyelid he comes to the conclusion that his aunt is pregnant. If the stye is on the upper eyelid, the child will be a male; if on the lower, a female." Modern theories of sex determination, like the earlier speculations, may be resolved into two groups, namely, those which depend upon controllable ex- ternal or environmental factors such as food, climate, chemical dosage and will power, and those which de- pend upon internal factors at present beyond control. 2. The Nutrition Theory Of external factors which may exert a moulding influence upon the sex of the offspring, nutrition is THE DETERMINATION OF SEX 190 possibly the most potent. This factor may be conceived to act either upon the parent previous to the maturation of the germ-cells, upon the germ- cells themselves, or upon some susceptible embryonic stage of the life cycle subsequent to that of the fer- tilized egg. It has been suggested that since the egg is char- acterized by possibly a more advanced metaboHc condition than the sperm due to the presence of the nutritive yolk, consequently the more yolk or nutri- tion there is, the more femaleness will characterize the egg. In other words, femaleness is a nutritive condi- tion associated in the egg with the presence of yolk. A generation ago Professor Schenk of Vienna, by controlling the nitrogenous diet of certain royal prospective mothers, gained a soothsayer's reputa- tion as a prophet of sex which was based upon several correct predictions. Of course, any prediction of sex is bound to turn out correct in 50 per cent of the cases, regardless of what it is based upon, since in man the two sexes are approximately equal in numbers. Adherents of all sorts of theories, therefore, have always been able to produce considerable ''evidence" to sub- stantiate their speculations, however crude the latter have been. Statisticians have pointed out that in times of unusual hardship, like famine or war, when tlie amount of available nutrition for pregnant mothers is presumably reduced, there seems to be a prepon- derance of males born. 200 GENETICS A series of nutrition experiments upon frogs per- formed by Born ('81), Pfluger ('8^2), and Yung ('85) showed that the percentage of female offspring, which normally is slightly over fifty, could be changed to over 90 per cent by regulating the food supplied to the mother before the egg-laying period. Cuenot and King, however, working independently, repeated these experiments with great care, taking into account all the eggs that were laid and not simply the ones that developed, and both obtained negative results. They concluded, therefore, that the high percentage of female tadpoles appearing in the initial experi- ments was due to a greater mortality among the males and not to the transformation of possible males into females. There seems to be no doubt that nutrition may affect the percentage of those which reach maturity. If one sex requires a greater amount of nutrition than the other to carry out successfully the more strenuous metabolic changes in its life-cycle, then unequal percentages between the sexes of the sur- vivors resulting- from modified nutrition do not in any way help to solve the problem of determining the sex of the individual. In other words, the elim- ination of one sex through modified nutrition does not "determine" the other sex. 3. The Statistical Study of Sex From statistical sources it has been ascertained that ordinarily there is produced a practical equality in the numbers of the two sexes. THE DETERMLXATIOX OF SEX 201 Oesterlehen in Europe summarized the daUi for nearly sixty million human births and found that an average of 106 males are born to every 100 fe- males. According to various authorities, the relative num- ber of males per 100 females is given for horses as 99, for cattle 94, and poultry 95, while in pigs, rabbits, pigeons, and greyhounds the corresponding number of males is slightly over 100. This practical equality of the sexes in all sorts of natural environments indicates the improbability of the assumption that external conditions determine sex. 4. MONOCHORIAL TwiNS There are two kinds of twins, namely, ordinary twins, which come from two separately fertilized eggs each inclosed in its own chorion, and "identical twins," that have their origin in one egg which is in- closed in one chorion. Of the former, something like 30 per cent in man are reported as being of two sexes, thus showing that it is neither nutrition nor envi- ronment which determines sex. Usually when twins are of the same sex, they exhibit as great a range of difference in mental and physical traits as do ordinary children of the same fraternity born at different times, but occasionally "identical twins" are born, and such monochorial twins are always of ihe same sex. This is evidence that sex, like other somatic characters, is determined in the germplasm at the time of fertilization. 202 GENETICS Similarly, in the clialcid fly, Ageniapsis, sl chain of embryos is formed from a single egg, and these, ac- cording to Marschal, are all of the same sex. Newman and Patterson also have shown that in the armadillo, Tatusia, there are customarily produced four young within a single chorion, all of which are of the same sex. These facts point toward the conclusion that the determination of sex takes place at the time of fer- tilization. 5. Selective Fertilization Within the last ten years considerable evidence has been collected in support of the supposition that sex is a Mendelian character. Mendel himself, without elaborating this idea into a definite hypothe- sis, suggested the probability that sex is a heritable character behaving in the same way as other herit- able characters. In 1903 Castle published a paper ^ in which a tentative explanation, since abandoned, of the phe- nomenon of sex determination was advanced, based upon three assumptions : first, that all germ-cells are heterozygous for sex and, therefore, upon maturation there are formed both male and female eggs as well as male and female sperms ; second, that in fertili- zation the gametes always unite with their opposites so far as sex is concerned ajid never with their like, with the result that each fertilized egg must carry 1 Castle, W. E., "The Heredity of Sex." Bull. Mus. Comp. Zool., Harvard, Vol. XL, No. 4, 1903. THE DETERMINATION OF SEX 203 determiners for both sexes and be lieterozygous, as indicated in Figure 58; and third, tliat the character of sex follows the law of alternative dominance, ac- cording to which in the male offspring the nude determiner dominates M(F), while in the female the female dominates {M)F. This hypothesis is simply an attempt to explain the numerical equality of the sexes, and also the fact that the determiner for the opposite sex may be car- MALC Gametes FEMALE. M(r) M(M) (DF r(M) MALE - DO NOT OCCUR - FEMALE ZYGOTES Fig. 58. — Diagram to show Castle's 1903 theory of the heredity of sex. ried by either parent, but it leaves unanswered the question of what causes ''selective fertilization" and "alternative dominance." There appears to be some evidence that selective fertilization, which was assumed in Castle's 1903 theory, may actually occur under certain circum- stances. For example, homozygous or pure yellow mice, that is, mice with a duplex determiner for yellow color, are not known. In breeding, all kinds of yellow mice behave as if heterozygous or simplex with respect to yellow color, for when any two yellow mice are bred together, they produce a certain per- centage of recessives which would not happen if they 204 GENETICS were pure yellow. In a Mendelian monohybrid cross, as has been previously pointed out, the expectation is that in the second generation one fourth of the offspring will be recessives {DR X DR = DD -\-2DR-\- RR) , but when yellow mice are bred together, the percentage of recessives approximates one third instead of one fourth. This apparent exception to the Mendelian ratio finds an explanation, however, when it is assumed that selec- tive fertilization takes place in such a cross, and thus, since a D gamete never unites with another D gamete, but always with its opposite, R, pure yellow mice are unknown. This supposition is further supported by the fact that the litters of young from yellow mice are, on an average, only three fourths as large as normal litters of mice, which is exactly what would be expected if one fourth of the possible gametic combinations {DD) fail to produce offspring. Castle's tentative explanation of the determina- tion of sex at least breaks away from the old concep- tion that the sperm-cell produces male offspring and the egg-cell, females. It agrees, too, with Darwin's idea that both sexes are present in each individual with one sex latent. In certain parthenogenetic rotifers, aphids and daphnids, both sexes are plainly present in the female, since two kinds of easily dis- tinguishable eggs are produced, one of which develops into males and the other into females without fer- tilization or any kind of a union with a sperm-cell. THE DETERMINATION OF SEX 205 6. The Neo-Mendelian Theory of Sex Correns (1906) avoids the difficulties of alternative dominance, which Castle's hypothesis offers, by sup- posing that one par- Type I ZYGOTES (Female) QcT (Male) Cfcf n (Female! (Male) 99 9cf GAMETES 9 (F. 9 :) ( -h Male) W (Female) QcT (Male) Fig. 59. — Diagram to show tho neo-Men- delian theory of the heredity of sex, using sex symbols. ent only is hetero- zygous with respect to sex, and this sup- position is becom- ing more and more probable as evidence accumulates. Ac- cording to this idea, there are two types of cases, one when the female is the heterozygous parent and the other when the male is the hetero- zygous parent, as represented in Figure 59. The formulse for these types may be expressed in the nomenclature of the presence and absence theory, as follows (Fig. 60), in which the sym- bol X represents the female determiner in the heterozygous case of type I, and XX the female de- terminer when the male is the hetero- zygous parent. The formulae may be still further modified, accord- ing to Morgan, for the satisfaction of those who ob- Type Zygotes Gametes (Female) X o X -H O I (Male) o o o xo (Female H- O oo ) ( Male ) n (Female) X X x + X (Male) X o X XX ( Female + O XO ) (Male/ Fig. 60. — Diagram to show the neo-Men- delian theory of the heredity of sex accord- ing to the presence and absence hypothesis. 206 GENETICS ject to regarding the male factor as nothing positive, but simply the absence of femaleness, by assuming that a universal factor of maleness (m) is present in all cases, as shown in Figure 61. Thus in type I of this scheme it is only when the dominant female factor F is entirely absent that male- ness becomes expressed in the somatoplasm, while in type II it is neces- sary to have a double dose of the factor F in order to produce a female, since a single dose results in a male. All of these three theoretical schemes agree in assuming that one sex is hetero- zygous, while the other is homozygous and that femaleness is the result of an added factor in excess of maleness. The evidence for these conclusions has been ob- tained chiefly from four sources : first, from a mi- croscopical examination of the germ-cells ; second, from castration and regeneration experiments ; third, from the results of hybridization in "sex-limited inheritance"; and fourth, from the behavior of hermaphrodites in heredity. Type Zygotes GrAMETES I (Female) Fmfm Fm -1- fm (Male) fmfm fm -\- fm Fmfm fmfm (Fema\e) (Ma\e) n (Female) FrnFm Fm -\- Fm (Mole) Fmfm Fm -f fn\ FrnFm Fmfm (Fema\e) (Male) Fig. 61. — Diagram to show the neo-Men- delian theory of the heredity of sex with Morgan's modification, making male- ness (m) present as a positive character in every gamete. THE DETERMINATION OF SEX 207 a. Microscopical Evidence A, 1. The "X" Chromosome In 1891 Henking called attention to the presence of two kinds of spermatozoa in the firefly, Pyrrho- corisy and later McClung (1901), in studying the spermatogenesis of the grasshopper, discovered a similar phenomenon with respect to the chromosomes of its spermatozoa. Soon after, Stevens and Wilson Fig. 62. — Diagram to show how numerical equality of the sexes results when one parent is homozygous (the female in this instance) and the other is heterozygous for the sex character. working independently on various species of insects, and Boveri, on sea-urchins, found that when the male is characterized by two kinds of sperm-cells, one of which has an "extra" chromosome (the so- called "accessory" or "a:" chromosome), while the other does not, the female of the same species, upon maturation of the eggs, produces mature eggs, all of which possess one "a:" chromosome. The result of this heterozygous condition of the male and homo- zygous condition of the female with respect to the X chromosome is the theoretical equality of the sexes among the individuals formed by their union, as 208 . GENETICS shown in Figure 62 or in type II of Figures 59, 60 and 61. It will be seen that when a male gamete bearing an X chromosome unites with a female gamete also bearing an x chromosome, the outcome is a fertilized egg containing xx chromosomes. Such an egg is con- sequently homozygous for sex, and will develop into a female individual. In the same way when a male gamete lacking an x chromosome, as half the gametes derived from a heterozygote do, unites with a female gamete bearing an x chromosome, as all gametes from a homozygote must do, then the fertilized egg will be heterozygous, carrying only one x chromosome, and will develop into a male indi\adual. The chromatin difference between the two sexes may be qualitative, as Wilson holds, or quantitative, as Morgan assumes, but in either case it seems cer- tain that, with difference in sex, there is invariably associated a definite difference in the character of the chromosomes present in the germ-cells. These conclusions have been abundantly con- firmed in various species by a large number of inde- pendent w^orkers, and are now well established as a part of biological science. In fact, it is not at all unusual to find the technical confirmation of the X chromosome theory given as a part of the routine class work in university courses. A, 2. Various Forms of X Chromosomes The extra chromosome in different species may assume various forms or degrees of complexity. It THE DETERMINATION OF SEX 209 may be either single or multiple. It may be paired before maturation with its absence, or with an unlike ("i/") chromosome. It may be linked inseparably with some one of the ordinary chromosomes (auto- somes), or resemble the autosomes so closely that its presence can only be assumed from analogy with other cases, and not definitely determined at all. In all of these cases, however, there is one point of likeness, and that is that there always seems to be additional chromatin material associated with the female sex. The reason for this may lie in the more highly metabolic requirements of the female, who must produce yolk or provide in some way for the main- tenance of the young in addition to furnishing half of the germinal heritage. In the microscopical evidence on this point there is one apparent exception to the rule that females are homozygous and males heterozygous with respect to sex. Baltzer (1910) found that in one of the sea- urchins an extra sex chromosome is associated with the female sex, so that two kinds of mature eggs are produced upon maturation and only one kind of sperm-cells. In other words, in this case the female is heterozygous for sex and the male homozygous, instead of the reverse which is true for all other forms thus far microscopically investigated. Such cases as this of the sea-urchin are theoreti- cally provided for in the formulae under type I given above in Figures 59, 60 and 61. 210 GENETICS Ay 3. Sex Chromosomes in Parthenogenesis The behavior of the chromosomes in cases of parthenogenesis, where the union of an egg-cell and a sperm-cell are not necessary for the production of a new individual, throws additional light upon the relation between chromosomes and sex determi- nation. For instance, among the social Hymenoptera, bees, ants, wasps, etc., the "queen" produces eggs which upon maturation, if unfertilized, develop into males or drones, all of whose cells contain a reduced amount of chromatin (Fig. 63). It is only when sexual repro- duction occurs through the union of a mature egg- cell and a mature sperm-cell or spermatozoan, that the full complement of chromatin is restored to the fertilized egg and females are again produced. Castle says : "In all known cases of parthenogenesis the female is in the duplex (2 n) condition, and the male is in the simplex (n), or partially duplex (2 n — 1 condition. The female in all cases has the greater chromatin content." b. Castration and Regeneration Experiments Certain characters which are known as "secondary sexual characters," such as the ornamental plumage in male birds, the beard in man or the sting in worker bees, are often associated with a definite sex. When an individual is castrated, it is quite common not only for these peculiar secondary sexual characters to dis- appear, but also for the secondary sexual characters of THE DETERMINATION OF SEX 211 9 SOMA (Queen) LAR BODY OCYTC — Polar body ^— PcLAR BODY Mature cgo 5perm cells ©fERTlUZEO EGG Fig. 63. — Diagram of the heredity of sex in bees, ants and wasps. The outline chromosomes represent sample somatic chromosomes. The solid black chromosomes stand for sex. The female has two sex chromosomes while the male has but one. the opposite sex to develop to a certain degree in their stead. This indicates that the determiners for sex are intimately associated with those for the sec- 212 GENETICS ondary sexual characters, and also that the determin- ers for the opposite sex are often present in a latent condition, or, in other words, that the organism, either male or female, is heterozygous with respect to sex. If a female of the annelid worm Ophiotrocha, for example, is cut in half, it is effectually castrated, be- Male PardSiliwllij castrated male Female Fig. 64. — The crab, Inachus, parasitized by the cirripede, Sacculina. Evidence that a Mendelian sex determiner is correlated with "sec- ondary sexual characters " and that the male is heterozygous for sex while the female is homozygous. After Smith. cause the ovaries are in the posterior part of the body. It has the power of regeneration, however, but when a new posterior part is formed, it contains, not female, but male reproductive organs. The worm is, there- fore, now a male, as shown by the presence of testes instead of ovaries, proving that it was originally heterozygous with respect to sex, carrying one sex latent. THE DETERMINATION OF SEX 213 According to Smith, parasitic castration is performed on the crab Inachus, which is found in the Bay of Naples, by a cirripede, Sacculina. The male crab of this species has one large claw and a narrow abdo- men, while the female has no large claw, but a broad abdomen. When Sacculina parasitizes the female, the secondary sexual characters of the female are stunted, but not materially changed. When, on the contrary, the male is parasitized, it not only loses its distinctive large claw in subsequent molts, but it also takes on the broad abdomen of the female (Fig. 64). This apparent anomaly is quite explainable upon the as- sumption that the female is homozygous for sex and the accompanying secondary sexual characters, while the male is heterozygous. When maleness is destroyed in the male by the castrating parasite, therefore, the femaleness that is latent in this sex becomes manifest through the appearance of female secondary sexual characters ; but when the female is castrated, no other secondary sexual characters than those already present make their appearance, since only femaleness is present in the homozygous female sex. c. Sex-limited Inheritance Additional evidence that sex is a character depend- ing upon determiners which behave in Mendelian fashion is furnished by what is called sex-limited inheritance. There are certain characters known as sex-limited characters that are in no sense to be con- fused with secondary sexual characters which appear to be always linked with the determiner for either one 214 GENETICS sex or the other. They are, therefore, well described by the term *' sex-limited." (1) Color-blindness This phenomenon may be illustrated by the in- heritance of human color-blindness, a character which appears to be linked with the determiner for sex. It requires a duplex, or homozygous, dose of the deter- miner for color-blindness to produce a color-blind female, while only a simplex, or heterozygous, dose is Gametes (x f: 90 9@ Fig. 65. — General diagram for sex-limited inheritance. The underscored symbol (21) represents a sex determiner with some other character (as color-blindness) linked with it. needed to produce a color-blind male. These facts agree perfectly with the idea that the female is homo- zygous and the male heterozygous with respect to THE DETERMINATION OF SEX 215 sex, and that the factor for color-blindness is Hnked with the determiner for sex. Sex-Hmited inheritance, as shown in this case, may be ilhistrated l>y the (ha- gram on the opposite page (Fig. 65) in which, for the sake of simphcity, only sex chromosomes and the de- terminers for color-blindness are represented. Under- scored 2< represents a color-blind determiner linked to a sex chromosome. From this diagram, which agrees substantially with the facts, it is apparent that a color-blind male mated to a normal female will produce no color-blind offspring, although the females will be "carriers" of color-blindness, that is, will possess the factor in simplex form and will, therefore, carry it for the fe in ale in a latent condition. The sons of such a mating having a normal mother and a color-blind father will be absolutely free from the defect and cannot produce color-blindness in any of their offspring when mated with a normal strain. If, however, the "carrier" daughters from such a parentage, who are genotypically heterozygous for color-blindness but phenotypically normal, mate with normal individuals, the expectation is that one half of the sons, and none of the daughters will be color- blind, but that one half of these daughters will carry the color-blind determiner in simplex form, that is, in a condition ineffective for producing color-blindness in female individuals. All of the various possibilities in the inheritance of color-blindness according to the sex-limited interpre- tation are indicated in the following table : — 216 GENETICS Parents Expected Offspring Normal 9 Color-blind Color-blind 9 Carrier Normal Carrier ^ color-blind 1 normal ^ carrier ^ normal Color-blind Normal Normal Carrier Color-blind Color-blind Color-blind Color-blind Color-blind Carrier ^ color-blind ^ normal ^ color-blind ^ carrier (2) r/^^ English Currant-worm A famous case of sex-limited inheritance is that of the EngKsh currant- worm, Abraxas, which occurs in two varieties, viz.. Abraxas grossulariata and Abraxas Fig. 66. — Abraxas grossulariata, the English currant-moth, and (on the right) its paler lacticolor variety. From Punnett's " Mendelism." lacticolor (Fig. 66). The Hghter-colored lacticolor is recessive to the darker-colored grossulariata variety and has been found in nature associated only with the female sex. THE DETERMINATION OF SEX 217 Doncaster and Raynor, in 1908, published the results of various crosses between these two varieties which demonstrate clearly that sex is a Mendelian character and that, in this instance, maleness is homozygous and femaleness heterozygous with the determiner Key to Symbols Phenotype Gross.c? Gross, c? Lact.c? GR0SS.9 Lact.^ Conslilulion with respect to ihe GROSSULARIATA factor Dupl ex Simplex fSlulliplex Simplex Nulliplex Genotype (joi Ooi Iqi Ooi ()0I Jfll Oof flo Ooi Do Gametes 01 Fig. 67. — Key to the symbols employed in Figures 68-71. The outline symbols represent samples of the autosomes or somatic chromosomes. The black symbols stand for the "extra" or sex chromosomes. G above a black symbol indicates the grossulariata factor linked with a sex chromosome. The variety lacticolor occurs whenever the grossu- lariata factor is absent. for maleness linked with the factor producing the variety grossulariata. A study of Figures 67-71 will make this case clear. Outline symbols represent ordinary chromosomes or autosomes, several of which are omitted for sake of clearness. The black sym- bols represent sex chromosomes. The letter G placed 218 GENETICS above a black symbol represents the grossidariata factor linked with a sex chromosome. The variety lacticolor occurs whenever the factor for grossulariata is absent. In this case two sex determiners are neces- sary to produce a male, and only one to produce a female. In the following theoretical diagrams the actual number of offspring obtained by Doncaster and Raynor in each cross is indicated outside the circles that represent the zygotes, and the parenthetical numbers refer to the five kinds of individuals cata- logued in Figure 67. In the first cross (Fig. 68) where a lacticolor female (5) and a grossulariata male (1) were bred together, the entire progeny was grossidariata in character with an approximate equality between the sexes, that is, 45 males (2) to 50 females (4). When these hybrid grossulariata individuals, (2) and (4), were mated with each other in Cross 2 (Fig. 69), the character of grossulariata appeared again in both sexes, (1), (2), and (4), while the character lacticolor was confined as usual to females alone (5). It was only when grossulariata hybrid males (2) were crossed back to lacticolor recessive females (5) in Cross 3 (Fig. 70) that individuals of both varieties and both sexes appeared, (2), (3), (5), (4), in practically the expected equal numbers, namely, 63, 65, 70, 62. The lacticolor male (3) obtained by bringing together the two sex determiners necessary for maleness, each of which had been dissociated through the foregoing crosses from the sex-limited grossulariata factor, was en- tirely new to science, never having been found in nature. THE DETERMINATION OF SEX 219 Cro ss 1 GAMETE5 Zygotes 1-5 Gross, d* (2) (0 Gross cf Gross. ^ (4) Fig. 68. — The formation of heterozygous grossulariata individuals, both male and female, by crossing pure grossulariata males with lacticolor females. Cro 5s £ (2) Gross. cf GRoss.d GRoss-d* Lact. 9 Gross, (f (/) (Z) (5) (4) Fig. 69. — The cross-breeding of heterozygous grossulariata individuals. 220 GENETICS Finally, when these newly made lacticolor males (3) were crossed with heterozygous grossulariata females (4) (Fig. 71), the proportion of sexes was approxi- mately equal, as expected, that is, 145 males to 130 females, but all of the males were of the heterozygous grossulariata type (2) and all of the females of the re- cessive lacticolor type (5), showing a return to the sex- limited condition. All of these curious results find a satisfactory and complete explanation in the assump- tion, first, that sex is a Mendelian character carrying tw^o determiners for maleness and one for femaleness ; and, second, that the determiner for the character of grossulariata when present is always linked to the sex determiner. This case is of particular interest, since it agrees with the microscopical evidence already referred to in connection with the chromosomes of Baltzer's sea- urchins, in which the male was likewise homozygous and the female heterozygous with respect to sex. The chromosomes of Abraxas present certain techni- cal difficulties which at present have not been over- come, so that we do not yet know whether the evi- dence of the heterozygous character of one sex and the homozygous character of the other, obtained from the breeding experiments of Doncaster and Raynor, will be confirmed upon a microscopic examination of the chromosomes in the germ-cells. (3) The Behavior of Hermaphrodites in Heredity Certain plants occur in monoecious form, that is, as hermaphrodites, and also in dioecious form, that is, with THE DETERMINATION OF SEX 221 Cross 3 (2) Gross d Gross, d* Lact. c? (2) (3) (Ooioi) 7(X ^ 6Z-' Lact. 5 Gross ^ (5) m Fig. 70. — Heterozygous grossulariata male crossed with lacticolor female. One fourth of the progeny are lacticolor male, not known to occur in nature. Cross 4 GROSS d" tACT. ^ (2.^ (5) Fig. 71. — Back cross of lacticolor male with grossulariata female produc- ing the original sex-limited condition in which all the females are of the lacticolor type. Data for Figures 68-71 from Doncaster and Raynor. £22 GENETICS the sexes on separate plants. Among such dimorphic plants, Bryonia in particular has been investigated by Correns and Lychnis by Shull. Without describing the crosses made in their experiments in detail, it may be stated that when dioecious types are recipro- cally crossed with hermaphroditic forms, the result- ing progeny indicate plainly that one sex is homozy- gous while the other is heterozygous with respect to the sex character. This confirmatory evidence is quite in line with that already brought forward that sex is a Mendelian character the determiners of which are carried in the germplasm. 7. Conclusion The evidence thus far obtainable from all sources points to the conclusion that sex is unalterably fixed at the time the egg is fertilized, by definite deter- miners which act in the same way as other Mendelian determiners. Dr. Shull, whose exhaustive studies in sex determination place him in the front rank as an authority on the subject, makes this conservative statement: "Nearly all the recent investigations indicate that sex is at least predominantly dependent upon the genotypic nature of the individual." If this is so, while it furnishes the best of confirma- tory evidence in support of Mendel's law, it shows that it is not possible for man to predetermine the sex of his offspring, which he has long hoped to be able to do. The following quotation from Castle may suitably close this chapter: "Negative as are the results of our study of sex control, they are perhaps THE DETERMINATION OF SEX 223 not wholly without practical value. It is something to know our limitations. We may thus save time from useless attempts at controlling what is un- controllable and devote it to more profitable employ- ments." CHAPTER XI THE APPLICATION TO MAN 1. The Application of Genetics to Man Human civilization goes hand in hand with the degree of successful interference which man exerts upon the natural forces surrounding him. Primitive man was overwhelmed and outmastered by his environment, but civilized man harnesses nature to do his will. Savages are not proficient in the arts of cultivating plants and domesticating animals, while these are the very things upon which human prog- ress fundamentally depends. The degree of civiliza- tion of any people is closely correlated with the degree of their success in exercising a conquering control over plants and animals. Any knowledge of the laws of heredity, therefore, as applied by man, either directly to himself or indirectly to animals and plants, is a distinct contribution to human progress. In 1900 the National Association of British and Irish Millers, as Kellicott points out, being dissatis- fied with the quality and quantity of the annual wheat yield, engaged Professor Biffen to apply his knowledge of heredity to the practical problem of improving their wheat crop. The characters desired were a short full head, beardlessness, high gluten 224 THE APPLICATION TO MAN 225 content, immunity to rust, strong supporting straw, and a high yield per acre. In the short time that has elapsed. Professor Biffen has succeeded in pro- ducing strains of wheat that combine all these de- sirable characters to a remarkable degree. Such an immediate result would not have been pos- sible before 1900, when the rediscovery of Mendel's law revolutionized man's know^ledge of the action of heredity in nature. This same knowledge which has made possible the improvement of wheat may be applied to the breed- ing of man, for there is no reasonable doubt that man belongs in the same evolutionary series with all other animals, as Darwin showed, and is consequently subject to the same natural laws to a considerable degree. It must be admitted that thus far in the progress of civilization more attention has been directed to the scientific breeding of animals and plants, little as that has been, than to the scientific breeding of man. Let us hope that the future will have a dif- ferent story to tell ! 2. Modifying Factors in the Case of Man There are certain qualifying factors which make the problems of genetics somewhat different in the case of man than of other organisms. For example, mankind has come to be partially exempt from some of the natural laws that affect other organisms. Thus with respect to the w^orkings of natural selection man is partially under "grace" Q 226 GENETICS rather than "law." Nature no longer "selects" good eyes in man by long, patient, and devious processes when poor eyes are made good almost in- stantly by a visit to the oculist. She has long since given up providing natural weapons of defense for those who have the w^its to supply themselves more efficiently with artificial means of self-preservation, and she no longer attempts to improve the natural powers of locomotion of those who are able to tame a horse to ride upon, or who build steamships, rail- roads, automobiles and aeroplanes, thus accom- plishing at once what would require ages at least to evolve. Neither does the law of the survival of the fittest in its original sense apply equally to man and to other organisms. Human society to-day protects its unfit in hospitals, asylums, and through various philan- thropies, while physicians devote themselves to the art of prolonging life beyond the period of usefulness. We do not desire these results of our modern civili- zation to be otherwise, but the fact remains that some of the most inflexible and universal "natural laws" are ineffective in the case of man, and it is profitable to bear this in mind when applying the laws of ge- netics to man. The laboratory for human heredity is the wide world, but it is obvious that the experimental method which has proven so effective in studying the heredity of animals and plants is impracticable in the case of man. The consideration of human heredity, there- fore, must always be largely from the statistical side, THE APPLICATION TO MAN 227 consisting in an analysis of experiments already per- formed rather than in initiating new experiments. Such institutions as insane asylums, prisons, sanitariums, and homes for the unfortunate are excellent foci for studying certain phases of human heredity, because they are simply convenient places where the results of similar experiments in genetics have been brought together. 3. Experiments in Human Heredity a. The Jukes A classic example of an experiment in human hered- ity which has been partially analyzed by the statisti- cal method is that furnished by Dugdale in 1877 in the case of "Max Jukes" and his descendants. It includes over one thousand individuals, the origin of all of whom has been traced back to a shiftless, illit- erate, and intemperate backwoodsman who started his experiment in heredity in western New York when it was yet an unsettled wilderness. In 1877 the histories of 540 of this man's progeny were known, and that of most of the others was partly known. About one third of this degenerate strain died in infancy, 310 individuals were paupers who all together spent a total of 2300 years in alms- houses, while 440 were physical wrecks. In addition to this, over one half of the female descendants were prostitutes, and 130 individuals were convicted crim- inals, including 7 murderers. Not one of the entire family had a common school education, although 228 GENETICS the children of other families in the same region found a way to educational advantages. Only 20 individuals learned a trade and 10 of these did so in state's prison. It is estimated that up to 1877 this experiment in human breeding had cost the state of New York over a million and a quarter dollars, and the end is by no means yet in sight. b. The descendants of Jonathan Edwards In striking contrast to the case of Max Jukes is that of Jonathan Edwards, the eminent divine, whose famous progeny Winship describes as follows: "1394 of his descendants were identified in 1900, of whom 295 were college graduates ; 13 presidents of our greatest colleges, besides many principals of other important educational institutions ; 60 physi- cians, many of whom were eminent; 100 and more clergymen, missionaries, or theological professors; 75 were officers in the army and navy ; 60 were prominent authors and writers, by whom 135 books of merit were written and published and 18 impor- tant periodicals edited; 33 American States and several foreign countries and 92 American cities and many foreign cities have profited by the beneficent influence of their eminent activity ; 100 and more were lawyers, of whom one was our most eminent professor of law; 30 were judges; 80 held public office, of whom one was vice-president of the United States ; 3 were United States senators ; several were governors, Members of Congress, framers of state THE APPLICATION TO MAN 229 constitutions, mayors of cities, and ministers to for- eign courts; one was president of the Pacific Mail Steamship Company ; 15 railroads, many banks, in- surance companies, and large industrial enterprises have been indebted to their management. Almost if not every department of social progress and of public weal has felt the impulse of this healthy, long-lived family. It is not known that any one of them was ever convicted of crime." c. The Kallikak Family A more convincing experiment in human heredity than the foregoing, since it concerns the descendants of two mothers and the same father, is furnished by the recently published history of the " Kallikak " family.^ During Revolutionary days, the first Martin Kalli- kak, — the name is fictitious, — who was descended from a long line of good English ancestry, took advantage of a feeble-minded girl. The result of their indulgence was a feeble-minded son who be- came the progenitor of 480 known descendants of whom 143 were distinctly feeble-minded, while most of the others fell below mediocrity without a single instance of exceptional ability. " After the Revolutionary war, Martin married a Quaker girl of good ancestry and settled down to live a respectable life after the traditions of his forefathers. From this legal union with a normal woman there have been 496 descendants. All of 1 " The Kallikak Family." H. H. Goddard. The Macmillan Co. 230 GENETICS these except two have been of normal mentality and these two were not feeble-minded. . . . The fact that the descendants of both the normal and the feeble-minded mother have been traced and studied in every conceivable environment, and that the re- spective strains have always been true to type, tends to confirm the belief that heredity has been the determining factor in the formation of their respec- tive characters." 4. Moral and Mental Characters behave LIKE Physical Ones These instances of human breeding show unmis- takably that "blood counts" in human inheritance, even though the hereditary unit characters that lead to these general results have not yet been analyzed with the clearness that is possible in dealing with the characters of some animals and plants. There is of course no question of moral and mental traits in plants, and the role that these play in animals is not easy to determine ; but in man the case is undoubtedly much more important and complex, since mental and moral characteristics have a large share in making man what he is. There is, however, no fundamental scientific distinction which can be drawn between moral, mental, and physical traits, and they are undoubtedly all equally subject to the laws of heredity. For instance, as an illustration of the heritability THE APPLICATION TO MAN 231 of non-physical traits, in the Jukes pedigree three of the daughters of Max impressed their pecuh'ar moral and mental characteristics in a distinctive way upon their offspring. To quote Davenport : "Thus in the same environment, the descendants of the illegitimate son of Ada are prevailingly criminal ; the progeny of Belle are sexually immoral; and the offspring of Effie are paupers. The difference in the germplasm determines the difference in the prevailing trait." 5. The Character of Human Traits Of the mental, moral, and physical traits which are heritable in man, some must be regarded as generally desirable, some as indifferent, and others as defects to be avoided if possible. In general the majority of human traits, those which together make up man as distinguished from other animals, do not particularly claim the attention because they are so universal. Some which stand out from the mass, such as the physical traits of eye-color and the color and character of hair, may be regarded as indifferent so far as the welfare of the individual is concerned, while others like skin color and certain racial features that characterize particular strains of *' blood" may, under certain circumstances, work a social handicap upon their possessors according to the traditions of the community in which they appear. A long list of desirable mental traits miglit be enumerated that seem in a general way to be subject to the laws of inheritance,, although they have not 232 GENETICS yet undergone the careful analysis demanded by modern genetics which deals in unit characters rather than in lump inheritance. Musical, literary, or artistic ability, for example, mathematical aptitude and inventive genius, as well as a cheerful disposition or a strong moral sense are probably all gifts that come in the germplasm. They may each be developed by exercise or re- pressed by want of opportunity, nevertheless they are fundamentally germinal gifts. A genius must be born of potential germplasm. No amount of faithful plodding application can com- pensate for a lack of the divine hereditary spark at the start. 6. Hereditary Defects Undesirable hereditary traits are usually defects due to the absence of some character. For instance, albinism, which occurs in several kinds of animals and also in man in one out of every 20,000 individuals (according to Elderton), is due to the absence of pig- ment in the skin, hair and eyes. Albinic individuals have poor eyesight because they are unable to stand strong light, being without protective pigment in the eyes. This peculiarity of albinism behaves as a recessive character both in man and in other animals. An albinic individual may, therefore, marry a normal individual without fear of producing albino children, although the children of such a mating would carry heterozygous germplasm with respect to albinism, THE APPLICATION TO MAN 233 and in cousin marriages might subsequently produce some albino children. Davenport, in his recent work on "Heredity in Re- lation to Eugenics," brings together a long catalogue of human hereditary defects, although in most instances they are extremely difficult of accurate analysis. This is the case, first, because these defects so often probably depend upon a combination of determiners rather than upon a single one, and, sec- ond, because the available data are usually scattered and incomplete. Deafness, for example, is a defect which is heredi- tary though exactly to what degree, it is at present impossible to state. The following table taken from the extensive work of Fay (1898) upon "Marriage of the Deaf in America" gives some idea of the results of different matings lumped together statistically. Condition op Parents Percentage of Deaf Offspring Both born deaf 25.9 One born deaf, one wnth acquired deafness 6.3 One bom deaf, one normal 11.9 Both with acquired deafness 2.3 One with acquired deafness, one normal . . 2.2 That two parents born deaf do not produce more than 26 per cent of deaf children is probably due to the fact, first, that each parent is in all likelihood heter- ozygous for deafness and that, second, the same com- 234 GENETICS bination of factors which is the cause of the parental defect on either side of the pedigree does not happen to recombine after segregation to form the new individ- ual. Deafness will be produced in the offspring only when matings occur in which the proper factors are combined. Such an undesirable result is much more likely to happen if both parents come from the same, or related, hereditary strains than if they are derived from families in no way connected by blood. Herein lies the biological objection to cousin marriage which tends to bring together, and thus to perpetuate, like defects. Outcrossing, on the contrary, through the law of dominance, tends to conceal defects and to prevent their expression. Many other cases of human defects, such as im- becility or insanity, are extremely difficult of analysis from the standpoint of heredity because, in the first place, the defective conditions descriptively included under these vague terms are made up of a multitude of diverse conditions each of which must have a different array of determiners and, in the second plaice, because any one definite sort of insanity or imbecility may be conditioned by a variety of factors. However, the difficulty of the problem is no reason for abandoning the attempt to reach its solu- tion and to learn, if possible, "whence come our 300,000 insane and feeble-minded, our 160,000 blind or deaf, the 2,000,000 that are annually cared for by our hospitals and Homes, our 80,000 prisoners and the thousands of criminals that are not in prison, THE APPLICATION TO MAN ^285 and our 100,000 paupers in almshouses and out " (Davenport) . 7. The Control of Defects The method of possible control of human defects depends upon whether they are positive or negative, that is, dominant or recessive. In those cases where a given defect is due to a single determiner the Mendelian expectation for the possible offspring arising from various matings is indicated in the fol- lowing table in which D stands for the defect and d for its absence : — The Mendelian Expectation for Defects 1 2 3 4 5 6 7 8 If the Defect is Positive (dominant) If the Defect is Negative (recessive) When both DD XDD = all DD parents show DD XDd = hDD +hDd dd Xdd = all dd the defect DdXDd = \DD + hDd+\dd When one parent only- DD Xdd = all Dd dd XDD = all Dd shows the defect DdXdd = hDd + hdd dd XDd =- \Dd + \dd When neither dd Xdd = all dd DD XDD = all DD parent shows Dd XDD = iDD + iDd the defect DdXDd = i DD +i Dd + i dd If the defect is positive and in a duplex or homo- zygous condition in one parent, as in 1, 2, and 4 above, all the offspring will possess it regardless of the germinal constitution of the other parent. In 236 GENETICS two cases only, namely, in 3 and 5, where the de- fective parent is heterozygous, is there any chance of unaffected offspring, and even in these cases the defect is quite as likely to appear as not. It is ob- vious that the only way to rid germplasm of a dominant defect is by continued mating with recessive individuals. By this method it is possible in time to shake off the defect. When it once dis- appears in any individual, it will never return unless crossed back to a similar defective dominant strain. In other words, such a recessive extracted from a heterozygous ancestry will breed just as true as a recessive which was pure from the start. In both instances there is an entire absence of the character in question, and it is clear that this character can thereafter never again reappear, since something cannot be derived from nothing. On the other hand, if a defect is negative, depending upon the absence of a normal dominant determiner, as is usually the case with defects, it behaves as a Mendelian recessive, that is, it is always apparent in individuals developing from the homozygously de- fective germplasm. It is certain, for example, that an imbecile which has arisen from homozygous defective germplasm carries only the determiner for imbecility in his own germplasm, and when two such recessives mate, noth- ing but imbecile offspring can result, for recessives breed true. Nothing plus nothing equals nothing. An illustration of this principle is given in the fol- lowing pedigree (Fig. 72) furnished by Goddard, 1910. THE APPLICATION TO MAN 237 The result is quite different, however, when one parent only shows the defect. If the other parent is a normal homozygote, as in case 4 of the accompanying table, all the offspring will be normal in appearance, but with the bar sinister of defectiveness in their germplasm, while if the other parent is heterozygous (Case 5), one half of the progeny will be defective. Finally, when neither parent shows defectiveness EkO N c N]-r{N)@[r]®LN c [Nh