®i|p ^. p. pm ^tkarg QH5Gb E3 ■an > J ■♦ "JLV (Si.h-^l^^^'^ ■^^ I r 3 ^ *' . ~\ ' ■■■ -•<• NORTH CAROLINA STATE UNIVERSITY LIBRARIES S00614871 R 'O^f ' \ r QH366 ^ -East 10490 ■ihis book may be kept out TWO WEEKS ONLY, and is subject to a fine ^^^0 CENTS a day thereafter. It is due*^ the day indicated below: mm i 15 Feb '50 4 f'FebSSA 268»p92X tznm i3Mar'5i£ | OCT 7 1^4 NOV I 7 J955 44pr'57 * 29Aug'57S ■:'^, 26Apr'60«' UWl^^ T 5M — D47 — V^y. Monographs on Experimental Biology EDITED BY JACQUES LOEB, Rockefeller Institute T. H. MORGAN, Columbia University W. J. V. OSTERHOUT, Harvard University INBREEDING AND OUTBREEDING THEIR GENETIC AND SOCIOLOGICAL SIGNIFICANCE BY EDWARD M. EAST, Ph.D. AND DONALD F. JONES, Sc.D. MONOGRAPHS ON EXPERIMENTAL BIOLOGY PUBLISHED FORCED MOVEMENTS, TROPISMS, AND ANIMAL CONDUCT By JACQUES LOEB, Rockefeller Institute THE ELEMENTARY NERVOUS SYSTEM By G. H. PARKER. Harvard University THE PHYSICAL BASIS OF HEREDITY By T. H. MORGAN, Columbia University INBREEDING AND OUTBREEDING: THEIR GENETIC AND SOCIOLOGICAL SIGNIFICANCE By E. M. EAST and D. F. JONES, Bussey Institution. Harvard Universit THE NATURE OF ANIMAL LIGHT By E. N. HARVEY. Princeton University SMELL, TASTE AND ALLIED SENSES IN THE VERTEBRATES By G. H. PARKER, Harvard University BIOLOGY OF DEATH By R. PEARL, Johns Hopkins University INJURY, RECOVERY, AND DEATH IN RELATION TO CONDUCTIVITY AND PERMEABILITY By W. J. V. OSTERHOUT, Harvard University IN PREPARATION PURE LINE INHERITANCE By H. S. JENNINGS, Johns Hopkins University LOCALIZATION OF MORPHOGENETIC SUBSTANCES IN THE EGG By E. G. CONKLIN, Princeton University TISSUE CULTURE By R. G. HARRISON. Yale University THE EQUILIBRIUM BETWEEN ACIDS AND BASES IN ORGANISM AND ENVIRONMENT By L. J. HENDERSON, Harvard University CHEMICAL BASIS OF GROWTH By T. B. ROBERTSON, University of Toronto COORDINATION IN LOCOMOTION By A. R. MOORE, Rutgers College OTHERS WILL FOLLOW Monographs on Experimental Biology INBREEDING AND OUTBREEDING THEIR GENETIC AND SOCIOLOGICAL SIGNIFICANCE BY ^^ EDWARD M. EAST, Ph.D. HARVARD UNIVERSITY, BUSSEY INSTITUTION AND DONALD F. JONES, Sc.D. CONNECTICUT AGRICULTURAL EXPERIMENT STATION i6 ILLUSTRATIONS PHILADELPHIA AND LONDON J. B. LIPPINCOTT COMPANY i t COPYRIGHT, IQIQt BV J. B. LIPPINCOTT COMPANY Eleclrolyped and printed by J. B. Lippincoti Company The Washington Square Press, Philadelphia, U. S. A. LIBRARY N. C. State College EDITORS' ANNOUNCEMENT The rapidly increasing specialization makes it im- possible for one author to cover satisfactorily the whole field of modern Biology. This situation, which exists in all the sciences, has induced English authors to issue series of monographs in Biochemistry, Physiology, and Physios. A number of American biologists have decided to provide the same opportunity for the study of Experimental Biology. ' Biology, which not long ago was purely descriptive and speculative, has begun to adopt the methods of the exact sciences, recognizing that for permanent progress not only experiments are required but that the experi- ments should be of a quantitative character. It will be the purpose of this series of monographs to emphasize and further as much as possible this development of Biology. Experimental Biology and General Physiology are one and the same science, by method as well as by contents, since both aim at explaining life from the physico-chemical constitution of living matter. The series of monographs on Experimental Biology will therefore include the field of traditional General Physiology. Jacques Loeb, T. H. Morgan, W. J. V. Osteriiout. 6 i ^ ^ Of) PREFACE It is inevitable that each work planned as a member of a series of biological monographs should be somewhat technical. Of necessity each must be concise. In view of the difficulties these limitations involve, one may hardly expect to escape the criticism that the subject matter often tends to be esoteric in its nature, for few can say in Shaw's odd fancy, ^'I tried to do too much — and did if Nevertheless, there has been a serious effort to avoid a mere record of the development of a specific problem in Genetics as an aid to the general biologist. No one could have a professional interest in a subject of this kind with- out the desire that there be some practical application of the results to agriculture and to the many phases of sociology where a knowledge of the laws of heredity is a first requisite. Though such applications of the genetic conclusions are touched but lightly here, there is the hope that the non-biological worker interested in problems of human welfare will find some new thoughts and pertinent suggestions in the compelling logic of the controlled experiments described throughout the pages. At least it was with this idea in mind that the authors prepared the first four chapters. For the zoologist and botanist the well-known facts and elementary principles there dis- cussed would have been unnecessary. 7 8 PREFACE The manuscript has been the product, as it purports to be, of a very intimate collaboration, and the authors join in acknowledging their indebtedness to their fellow biologists for the privilege of copying several illustrations as noted in the legends, to Professor T. H. Morgan for helpful and suggestive criticism, and to Mr. L. C. Dunn and Mr. E. S. Anderson for assistance on the proofs. E. M. E. D. F. J. Boston, September, 1919 CONTENTS CHAPTER PAGE I. Introduction 13 II. Reproduction Among Animals and Plantb 20 III. The Mechanism of Reproduction 36 IV. The Mechanism of Heredity 50 V. Mathematical Considerations of Inbreeding 80 VI. Inbreeding Experiments with Animals and Plants 100 VII. Hybrid Vigor or Heterosis 141 VIII. Conceptions as to the Cause op Hybrid Vigor 164 IX. Sterility and Its Relation to Inbreeding and Cross-breeding 188 X. The R6le of Inbreeding and Outbreeding in Evolution... 195 XI. The Value of Inbreeding and Outbreeding in Plant and Animal Improvement 210 XII. Inbreeding and Outbreeding in Man: Their Effect on THE Individual 226 XIII. The Intermingling of Races and National Stamina 245 Literature 266 ILLUSTRATIONS FIG. PAGE 1. Asexual Reproduction, an Amoeba in Division 21 2. Asexual Reproduction by Means of Runners 22 3. Hermaphroditism in the Tapeworm Proglottid 24 4. Rhopalura, an Example of Extreme Sexual Dimorphism 26 6. Sexual Reproduction in Fucus 26 6. Ulothrix, a Primitive Type of Sexual Reproduction 28 7. An Adaptation for Self-pollination 30 8. An Adaptation for Cross-pollination 34 9. Diagram of Gametogenesis 38 10. Diagram to Illustrate Fertilization 39 11. Formation of Pollen Grains in the Lily 40 12. Fertilization in the Embryo Sac of the Lily 41 13. Entrance of Spermatozoon through Membrane of Egg of Star-fish ... 42 14. Diagram Showing the Distribution of Sex Chromosome in Protenor . . 43 15. Identical Quadruplets in Nine-banded Armadillo 44 16. Diagram to Illustrate Inheritance of Sex-hnked Character 48 17. Diagram Showing Union of like Gametes 52 18. Diagram to Illustrate Mendelism in a Cross between Long-Spiked and Short-Spiked Wheat 54 19. Diagram to Illustrate Gamete Formation in a Dihybrid in Indepen- dent Inheritance 59 20. Diagram to Illustrate Gamete Formation in a Dihybrid in Linked Inheritance 63 21. Diagram to Illustrate Crossing-over 64 22. Curves Showing the Limiting Values of the Coefficients of Inbreeding with Various Systems of Matings 84 23. Graphs Showing the Total Inbreeding and Relationship Curves for the Jersey Bull, King Melia Rioter 14th 86 24. Graphs Showing the Reduction of Heterozygous Individuals and of Heterozygous AUelomorphic Pairs in Successive Generations of Self-fertihzation 90 25. Graphs Showing the Increase in the Body Weight with Age for the Males of Inbred Albino Rats 107 26. Graphs Showing the Increase in Weight of Body with Age for DifTer- ent Series of Male Albino Rats 108 11 12 ILLUSTRATIONS 27. Graph Showing the Average Size of Litters Produced in Successive Generations of Inbreeding Albino Rats by Brother and Sister Matings 109 28. Goliath, an Albino Rat, the Product of Six Generations of the Closest Possible Inbreeding 110 29. Representative Samples of Inbred Strains of Maize after Eleven Generations of Self-fertilization 130 30. Graphs Showing the Reduction of Variability and Segregation of Ear Row Number in Selfed Strains of Maize 132 31. Plants of Maize after Eleven Generations of Self-fertiUzation and Their Fi Hybrid 150 32. Ears of Maize after Six Generations of Self-fertihzation and Their Fi Hybrid 150 33. Graphs Showing Growth Curves of Two Inbred Strains of Maize and Their First and Second Generation Hybrids 152 34. James River Walnut, Hybrid Between Persian Walnut and Butternut 154 35. Growth Curves of Parent Races and Fi and F2 Hybrids of Guinea Pigs 160 36. Diagram to Show How Factors Contributed by Each Parent May Enable the First Generation of a Cross to Obtain a Greater Development than Either Parent 175 37. Cattaloes, the Product of Crossing the Cow and the Bison 180 38. Sterile Hybrid Between Radish and Cabbage 192 39. Tassels of an Almost Sterile Strain Obtained by Inbreeding Maize . . 196 40. Representative Ears of a Cross Between Two Inbred Strains of Maize 202 41. Plants of a Cross Between Two Inbred Strains of Maize 202 42. Diagram Showing a Method of Double Crossing Maize to Secure Maximum Yields, Illustrated by Actual Field Results 203 43. First Generation Cross of Shropshire by Delaine Merino 212 44. First Generation Cross of Hereford by Shorthorn 216 45. "Big Jim" the Product of a Pure Bred Percheron Stallion Mated with a Grade Mare of the Same Breed 220 46. First Generation Cross of Chester White and Poland China 224 INBREEDING AND OUTBREEDING CHAPTER I INTRODUCTION Interest in the effects of inbreeding and of outbreed- ing is not confined to the professional biologist. Histori- cally these are old, old problems, practical problems of considerable significance bound up with man's gravest affairs, his marriage customs and his means of subsist- ence. In these matters, moreover, the passing of time has not diminished the value to be attached to their solu- tion. The questions involved belong to theoretical biol- ogy, it is true, and the professional biologist may lay claim to the first satisfactory analyses ; but relatively his inter- est is that of yesterday, stimulated by the work of Dar\s^in in establishing the doctrine of Evolution. The intimate relation which the effects of various sys- tems of mating bear to these three subjects will be seen more clearly from the following brief explanation. Anthropological investigations have shown that many primitive peoples established rigid customs of exogamy — marriage outside the family or the clan. Such practices, after their identification with totemic systems by Mac- Lennan, became the subject of much notable speculation. In particular may be mentioned the works of Frazer, Lang and Freud. Yet these writers have thrown little light on 13 14 INBEEEDING AND OUTBREEDING the origin of outbreeding as a social habit, and have con- tributed nothing whatever toward the solution of the ques- tions of inbreeding and outbreeding in the sense in which they will be treated here. It is probable, indeed, that these customs usually originated without regard to matters of physical inheritance. The tribes concerned had seldom risen to a state of culture where the welfare of their de- scendants might be expected to cause anxiety, since in few cases had there been that development of animal hus- bandry necessary for the first glimpse into the mysteries of heredity. These observations do not necessarily apply to the marriage folkways which developed in western Asia and Europe and were passed on to the United States. Our laws preventing marriages between certain degrees of kinship have been moulded by the touch of various civil- izations, but in the main they are a legal heritage from the code of Hammurabi through the Hebraic Tahnud. Since they are based largely upon the customs of pastoral na- tions, it may be they had some foundation in experience, half-truths drawn from casual and fragmentary observa- tions of the shepherd and the cattleman. There is no his- torical record of such rational basis, however. Many of the conventionalisms rigidly stabilized by the hand of re- ligious authority have not the slightest biological justifi- cation. Witness the English laws preventing marriage with a deceased wife's sister. On the other hand, if there had not been a dim but real fear of evil consequences aris- ing from inbreeding, there would be something extraordi- nary in the frequencies with which taboos against consanguineous matings have persisted. Among the peoples contributing to European civilization, caste sys- INTRODUCTION 15 terns have been common, and the logical outcome of a caste system is marriage between near relatives. Tnclo of race encourages inbreeding among the ruling class, and power within that ruling class prompts the perpetuation of a serving class in the same manner. Why, then, should exogamy have been continued so commonly throughout epochs marked by rational thought and a high degree of culture? It is true, there are exceptions to this general rule. Rather intense inbreeding was practiced both in Egypt and in Greece when they were at the height of their power and influence. Nevertheless, exogamic customs have prevailed. They exist in Europe and America at the present day, and it is natural to wish to know whether there is any biological justification for them. Let us propose three questions which will show the sociological bearing of the problems under consideration. 1. Do marriages between, near relatives, wholly by rea- son of their consanguinity, regardless of the inheritance received, affect the offspring adversely? 2. Are consanguineous marriages harmful through the operation of the laws of heredity? 3. Are hereditary differences in the human race trans- mitted in such a manner as to make matings between markedly different peoples desirable or undesirable, either from the standpoint of the civic worth of the indi- vidual or of the stamina of the population as a whole? Correct answers to these questions are a matter of more importance than a superficial consideration indi- cates. Settled in accordance with the biological facts, they aid in establishing a concrete scientific basis for marriage, divorce and immigration laws; they give grounds for predicting the changes to be expected in the 16 INBEEEDING AKD OUTBREEDINQ body politic due to differential fecundity, birth control, and other agencies by which the character of the popula- tion is shifted; they even have some relevancy to many problems which one might suppose were wholly of an economic nature, such as minimum wages and mothers' pensions. The second series of phenomena arousing interest in the results of inbreeding and outbreeding comes from ob- servation upon domestic animals and cultivated plants. Plants are included by courtesy, though in reality intelli- gent plant breeding hardly began until the nineteenth cen- tury, and the methods adopted were taken from the pro- cedures in use by animal breeders, with such modifications and improvements as the peculiarities inherent in vegeta- tive propagation made necessary. Animal breeding, on the other hand, is a very ancient occupation, and more or less accurate data on the effects of interbreeding near relatives as compared with the effects of crossing differ- ent strains must have been collected by all of the old agricultural peoples. Since there is no question that under certain circumstances inbreeding does produce un- desirable results — defectives, dwarf -forms, sterile indi- viduals, etc. — it may be that their experience was at the base of some of the antagonism toward close-mating in the human race. Or, it is possible that early breeders ob- served the phenomenon, common to both animals and plants, that when two unrelated stocks are crossed the hy- brids thus produced are often more vigorous than either parent — the phenomenon of hybrid vigor or heterosis, as it is called at the present time. There is no proof of such a sequence of ideas, but it seems to be a logical hypothesis. At any rate, the views of the animal raisers regarding INTRODUCTION 17 inbreeding and the traditions regarding marriage of near kin are very similar. The great majority of breeders have an ineradicable fear of evil consequences if their matings are too close. Only here and there a few fearless ones have used systems of extremely close mating to perpetu- ate their breeds, and by such methods have built up in- valuable races of horses, cattle, swine and poultry. But here a dilemma appears. Inbreeding has deplorable re- sults in certain cases, yet in other instances the returns have been gratifying. What is to be the future practice! To be more than mere trial and error, it must be founded upon a cogent analysis of the whole subject. Finally, interest in the effect of various systems of mating as natural phenomena has been stimulated by the study of organic evolution. The circumstantial data of comparative morphology show that in nature problems similar to those of man have arisen. If these problems are investigated some light may be thrown upon his diffi- culties. Sexual reproduction has been the most success- ful method of providing for the propagation of animals and plants. Does sexual reproduction, therefore, possess an advantage over other methods? Would it otherwise have persisted as it has in both kingdoms? It probably was not the original method of reproduction. Asexual reproduction, reproduction by simple vegetative division, appears to have held the stage when animals and plants were simple and unspecialized. Then, in all probability, came sexual reproduction with separation of the sexes. Secondarily, however, numerous species arose in both kingdoms wherein the sexes are united, male and female cells being produced in the same individual. Thus a sys- tem of mating entailing the greatest possible amount of 18 INBREEDING AND OUTBREEDINa inbreeding was established. But this system appears to have been deticient. {Some evolutionaiy advantage asso- ciated with separation of the sexes was lost. There is reason for assuming that this advantage was connected with cross-fertilization, for tertiary developments in each kingdom brought about numerous mechanisms whereby cross-fertilization was established in hermaphroditic or- ganisms. Still, in spite of the obvious success of bisexual and of cross-fertilized species, as shown by their fre- quency, numerous self -fertilized species, and even species reproducing exclusively by asexual methods have kept their places in the struggle for existence. Both inbreed- ing and outbreeding systems have developed side-by-side under natural conditions. The data of comparative mor- phology, therefore, seem as contradictory as those from anthropology and agriculture. The puzzles presented by these general facts taken from history, husbandry and biology have one common feature. They centre on the problem of inheritance. For- tunately, though less than two decades have passed since the application of quantitative experimental methods to biology became somewhat general, the mechanism of heredity is no longer a riddle; and to-day the effect of inbreeding and outbreeding on plants and animals can be described in considerable detail and interpreted with singular precision. Having this interpretation it may be applied to the three fields of interest we have described. In the ensuing pages the important controlled experi- ments in inbreeding and outbreeding necessary for an orderly and consistent interpretation of the facts are dis- cussed. Uncontrolled experiments, casual observations of stock breeders, data on human marriages between near INTRODUCTION 19 relatives, have been omitted designedly. Numerous data of these types have been available for many years, but they have been of little service in clarifying the situation. This is not altogether due to their fragmentary character in point of time, or even to the fact that they usually lack the precision necessary in data to be used in the analysis of such complex phenomena. Data for a limited number of generations are often useful, and precision is a relative matter. The truth is, the majority of these records was collected without regard to the type of fact required, and without reduction to concrete numerical terms. In other words, in records otherwise accurate, critical data are omitted ; and those given are relatively useless on account of their form. A detailed application of our conclusions to sociology, agriculture and evolutionary theory has not been at- tempted. It is hoped that the suggestions along these various lines are sufficient to show how such application can be made; but human direction of evolution either in man or in the lower organisms is beset with difficulties so numerous and so prodigious that each problem must have its individual solution. CHAPTER II EEPEODUCTION AMONG ANIMALS AND PLANTS In order to obtain a proper orientation of the problem of inbreeding and outbreeding, one must consider first some of the general facts regarding reproduction among animals and plants and their relation to inheritance. The significant changes in both kingdoms have been remarkably similar. The differences are differences in de- tail, and for this reason they are additional arguments in favor of the idea that there are special advantages asso- ciated with the coincidences found in the general processes involved. For example, asexual propagation is more gen- eral in the simpler, sexual reproduction in the higher or- ganisms. But sexual reproduction in animals has largely supplanted the asexual method, in plants sexual reproduc- tion was merely added. Is this not evidence of an im- portance to be attached to the sexual method, apart from a simple provision for multiplication? Again, the diver- sity of sex organs which has arisen among the various groups of animals and plants is highly surprising, yet this dissimilarity may be wholly of a superficial nature. When examined solely with the object of inquiring what systems of mating these variations entail, the parallelisms in each history stand out impressively. If these facts be kept in mind throughout the short discussion of heredity and re- production which follow, their probable evolutionary sig- nificance is not difficult to grasp ; but if the profusion of variation in detail, or even the general mechanism of ac- complishing a particular result is allowed to distract 20 ANIMAL AND PLANT REPEODUCTION 21 attention, the end may be lost to sight through admiration of the ingenuity of the means, (^Tiiere seems to be no question but that sexual repro- duction is a more recent means of propagation than asex- ual reproduction. ' Although asexual reproduction in the narrow sense, that is, by means of simple division or by budding, is common among the protozoa, the sponges, the coelente- rates and the fiat worms, it becomes sporadic in the molluscoids and annelids, and is found in only one or two isolated instances in forms as highly specialized as the arthropods and the chordates. If fragmentation succeeded by regen- eration of the lost parts be conceded to be a true means of reproduction, however, echinoderms and nematode worms are included. Thus of all the great groups of animals only certain worms (TrocJielminthes) FiQ. 1. — Asexual reproduc- tion. An amoeba in division. and the molluscs have no asexual cv. contractile vacuole; ek, coto- , Bare; en, entoaarc; u, nurleua. reproduction m the usual sense or (Kingsiey after sehuize. cour- ^ ^ teay Henry Holt & Co.). the word, and zoologists would hardly feel safe in maintaining its absence in these two phyla since the life history of so many forms is unknowai. But since asexual reproduction is replaced by sexual reproduction to a greater and greater extent as the higher forms are reached one cannot avoid the conclusion that the latter has proved to be the really successful means of propagation. Nevertheless, variations appeared in highly specialized forms which permitted return to an 22 INBEEEDINa AND OUTBREEDING asexual type of reproduction. In the arthropods, as well as in some other forms, mechanisms arose by which the eggs developed without fertilization. This parthenogenetic reproduction has been relatively successful, but only as a stop-gap. Sexual reproduction persists and is used as an occasional means of propagation. It would seem that it possessed advantages too great to be given up entirely. Even as sexual reproduction is a later method of propagation than asexual reproduc- tion, hermaphroditism appears to be a secondary development from forms in which the sexes were separate (gonochorism or dioecism). Omitting the protozoa in which it is difficult to decide such sexual differences, gono- chorism is present in every great animal group but the sponges, and hermaphroditism everywhere except in the Trochelminthes, although in Nemathelminthes, Echinodermata and FiQ. 2.— Asexual repro- Avthropoda it is rare. An extended ductioa by means of runners . . . , ^ • j a t i in the hawkweed. (After experiment ou the subject or hermaph- roditism certainly was made, but that it was an experiment, that hermaphroditism is from the evolutionary standpoint a secondary institution, is clear if one considers the anatomical evidence, as is shown by Caullery.23 Generally, hermaphroditism is a condition associated either with parasitism or with a sedentary life. Furthermore, hermaphroditic organisms do not have a truly simple organization. They have a superficial simplic- ity, due to an adaptation to their mode of life, but if one compares hermaphroditic and gonochoristic species group ANIMAL AND PLANT REPRODUCTION 23 by group, for example unisexual land or fresh-water worms with their bisexual marine cousins, he finds the for- mer to be the more complex, particularly as to their sex organs. The fact that the sponges are hermaphroditic might be considered as weighing against this argument, but it is not without the bounds of probability that the sponges are further along in specialization than is gen- erally admitted, for to find the substance nearest chem- ically to the so-called skeleton of the sponges, one must search among the arthropods — the product of the spin- ning glands of certain spiders and insects. Hermaphroditism, pure and simple, however, was not a success. Only a few degenerate forms retained self- fertilization and persisted. Among them may be men- tioned the tapeworms, certain crustaceans {Sacculina) parasitic on crabs, and the colonial forms, bryozoans and tunicates, the latter being perhaps the most degenerate of all animals since they are wholly unrecognizable as rela- tives of the vertebrates except at one short stage of their life history. In most of the hermaphroditic types new characteristics appeared which enabled them to exercise one of the important functions of bisexuality, cross-fer- tilization, without giving up the obvious energy conserva- tion attainable through the production of both sex cells in a single individual. In nearly all of these forms, this was made possible by the development of the eggs and of the sperm at different times. In a few isolated cases among the turbellarians and the tunicates the eggs develop first and then the sperm; the animal is first a female and later a male (pro- togyny). But in a greater number of species, the indi- vidual is first a male and afterwards a female (pro- 24 INBREEDING AND OUTBREEDING tandry). In the tapeworm (Fig. 3), for example, each segment contains a complete reproductive system, testes, ovaries and accessory glands ; when young the testes func- tion, when older the testes atrophy and the ovaries de- velop. In some of these protandrous species there is even a change in the whole structure of the body, includ- 0000 o. J O do o qOOq O °oOO o ooaoo ,o " '°a.§§ ^°o?,oO.»5^o?o8.'^ O n^A»: ftOoooO*?? Fig. 3. — Hermaphroditism in the tapeworm proglottid. K, genital pore; ov, ovary; re, receptaculum seminalis; t, testes; u, uterus; vd, vas deferens. (Kingsley after Sommer). ing the sexual orifices. The isopods of the family Cynio- thoidcd, a group of crustaceans parasitic on fish, furnish a beautiful illustration. In the male stage the animal is a typical crustacean and would be recognized as such by any layman with a very slight knowledge of zoology ; but when the animal passes over into the female stage it be- comes merely a great ^gg sac many times the previous size. One would hardly suppose the two stages belonged pWPfRTT UBMRT ^ C. State CoUeg* ANIMAL AND PLANT REPRODUCTION 25 to the same order, not to mention a transformation of the same individual. A few other mechanisms which promote cross-fer- tilization have been found in isolated cases. They are not as widespread as the one just described, but are peculiarly interesting nevertheless. Among certain of the cirripedes, the normal individuals are hermaphroditic, but in addition a few tiny degenerate males are developed. They are little more than bags of sperm and are calculated to make some- what amusing any generalization as to the ** stronger sex.'' Darwin, who discovered them, called them com- plemental males. Another means of preventing continued self-fertilization is self-sterility, a condition in which self- fertilization is very difficult or even impossible through some physiological impediment which is not clearly under- stood. It was demonstrated by Castle ^^ for the American race of Ciona mtestinalis. In what appear to be the essential features, the vicissitudes of reproduction have been similar in tlie vegetable kingdom. The problems were solved in differ- ent ways, but the gross results are largely the same. The most striking difference is the varied success of certain mechanisms. In the animal kingdom sexual reproduction wherever instituted practically always displaced asexual reproduction. Only in a few forms which are either fixed or parasitic in their mode of life did the two methods per- sist side by side. In plants, however, where the sessile is the common condition, asexual and sexual reproduction have continued harmoniously side by side clear up through the angiosperms. Again, there is a marked dif- ference in the success of hermaphroditism. In plants hermaphroditic forms became the dominant types in the 26 INBREEDING AND OUTBREEDING highest and most specialized group, the seed plants, while in the highest group of animals, the mammals, only an occasional individual showing rudimentary hermaphro- ditism is found. Just when sexual reproduction first originated in the vegetable kingdom is even more of a question than among animals. Only a few very simple types, the schizophytes (bacteria) and myxomycetes, have passed it by. Perhaps FiQ. 4. FiQ. 6. \ ^\' - T . .V' • T*"-' ' -^ y^-' "»>; I..- ^•- -.■■If I -v. ( Fia. 4. — Rhopalura, an example of extreme sexual dimorphism. (After Caullery.) Fia. 5. — Sexual reproduction in Fucus, giving some idea of the difference in size of egg and sperms. Sperms should be about one-tenth the size shown. (Bergen and Davis after Thuret.) it is for this reason these forms have remained the sub- merged tenth of the plant world. It is tempting, as Coulter ^2 says, to see the origin in the Green Algce, There, in certain species, of which Ulothrix is an example (Fig. 6), spores of different sizes are produced. The large ones having four cilia are formed in pairs in each mother cell, the smaller ones usually having two cilia occur in groups of eight or sixteen in each spore-produc- ing cell. Those largest in size germinate immediately under favorable conditions and produce new individuals. ANIMAL AI^D PLANT REPKODUCTION 27 Those of lesser size also germinate and produce new individuals, but these are small and their growth slow. Only the smallest are incapable of carrying on their vege- tative functions. These come together in pairs and fuse. Two individuals become one as a prerequisite to renewed vigor. Vegetative spores become gametes. Something valuable — speed of multiplication — is given up that some- thing more valuable in the general scheme of evolution may be attained. This is indeed an alluring genesis of sex. It is rather a genesis of sex, however, than the genesis of sex. Various manifestations of sex are present in other widely sepa- rated groups of unicellular or simple filamentous plants, the Peredinew, the Conjugatce and the Diatomece — the Conjugated being indeed the only great group of plants in which there is no long continued asexual reproduction. In these forms one cannot make out such a good case of actual gametic origin, but the circumstantial evidence of sex development in parallel lines is witness of its para- mount importance. After the origin of sex, many changes in reproductive mechanisms occurred in plants, but most of them resulted merely in better protection for the gametes, in increased assurance of fertilization, in provision for better distri- bution, or in greater security for the young plant. First, perhaps, there was physiological differentiation of the gametes. At least such an interpretation may be given to the form of conjugation found in Spirogyra and other Conjugates, where, either by solution of the wall separating them, or by the formation of a tube-like out- growth of one or both cells so that the ends touch, tlie contents of one cell pass over to the other. We may 28 INBREEDING AND OUTBREEDING tliink of the stationary cell as female and the other as male. Another line of development, however, became the dominant one in the plant kingdom just as it did in the animal world. A morphological differentiation of the sex cells occurred. One became a large inactive cell stored FiQ. 6. — Ulothrix, a primitive type of sexual reproduction. A, B, filaments; C, «o- ospores; Z), germination of zoospore; £, gamete formation in filament; F, gametes and their fusion; G, germination of zygospore. (Bergen and Davis after Dodel.) with food, the egg ; the other became small and motile, the sperm. This change is well illustrated in Fucus (Fig. 5), one of the brown algae. It is clear that such a change in- creased the probability of fertilization, since many sperms could be produced without utilizing a great deal of energy, and since the attraction of the egg for the sperm was presumably augmented. A further stage in the evolution of sex was reached when the cells producing the eggs or the sperms were ANIMAL AND PLANT REPKODUCTION 29 differentiated, thus providing for protection of the gametes. Such organs of various types and known by different names have persisted throughout all the higher plants. One may call them ovaries and spermaries and thus keep in mind that in animals the same types of change occurred. The final step in the general development of sexuality is the restriction of the formation of sex organs to a par- ticular phase in the plant's life, which on this account is known as the gametophyte. The remaining stages are known as non-sexual or sporophytic, because they are characterized by the production of non-sexual reproduc- tive cells, the spores. The liverworts, the mosses, the ferns and the seed plants are thus set apart. Since these two phases alternate with each other, pairs of reproductive cells of the gametophyte producing the sporophyte, and the non-sexual spores of the latter giving rise to the gametophyte, the sequence has retained the name of alternation of generations. In the higher liverworts and mosses the gametophyte carries on the greater part of the nutritive work, but in the ferns the sporophyte becomes the dominant structure ; while in the seed plants the gametophyte has degenerated until it consists of but two or three cell divisions. There is no question but that all of these numerous changes are merely insurance against the future, some- thing that may be said of seed production as a whole, since the seed is but the younger generation nourished on the parent stem. And it is interesting to note that just as animals and plants paralleled each other in cramote pro- tection and provisions for assuring fertilization, so also the final step in each, the mammals and the seed plants, 30 INBREEDING^ AND OUTBREEDING was the protection of the young. In certain particulars, however, the higher plants did not simulate the higher animals in their reproductive evolution, and it is not diffi- cult to see the reason for the divergencies. Plants re- tained asexual reproduction as an alternative method of propagation, and made a success of hermaphroditism. The obvious necessity for both was their fixed condition, their slavery to the soil ; but if hermaphroditism with its simplest implication, self-fertilization, had become domi- FiG. 7. — Adaptation for aelf-pollination by means of spiral twietings of stamens and style. (After Kerner.) nant, there would have been little from their life histories upon which to base an argument regarding the respective virtues and defects of inbreeding and outbreeding. This, however, was not the case. Many plants characterized by autogamy persisted and flourished. They even developed numerous devices promoting self-fertilization (Fig. 7), such as pollination before the flower opens, inclination of the anthers toward the pistil or the pistil toward the anthers, rapid elongation of the pistil through a ring of stamens, or various torsions of the accessory floral parts ; but it seems perfectly clear from the exhaustive investi- gations on the fecundation of plants made in recent years ANIMAL AND PLANT REPRODUCTION 31 that only an extremely small percentage of the species of flowering plants which have held their own to the present day in the struggle for existence, have adopted a method of fertilization which permits no crossing. Some of our most vigorous cultivated plants — tobacco, wheat, peas and beans — are naturally and usually self -fertilized, but they each and every one have their flowers so arranged as to pennit an occasional cross. At the same time, one would be too hasty if he con- cluded from these facts that continuous self-fertilization or other means of reproduction which result in a single line of descent is incompatible with inherent racial vigor. At least, there is evidence that various species which seem well able to hold their own seldom resort to crossing as a means of propagation, yet one could hardly use them as examples of degeneration. As illustrations, there is no need to go below the flowering plants, either, although if one desires an example of a long-continued evolution of species and genera without any form of sexual reproduc- tion he is forced to look to the Basidiomycetes. In this large group the fungi are not only asexual themselves, but appear to have been developed in a purely asexual manner from asexual ancestors. But in the flowering plants, many of our most useful types — the potato, the banana, hops and sugar cane — seldom have recourse to sexual re- production. It is true many agriculturists insist that these species sooner or later degenerate for this very reason, but they have never been able to bring forward one atom of critical evidence to uphold their view. Vari- eties of potatoes or of sugar cane do indeed degenerate, but it is probably because of disease which from their method of propagation is difficult to eradicate, and not 32 INBEEEDINa AND OUTBBEEDINa because of the method itself. Again, if one desires further evidence of descent in a single pure hereditary line con- sistent with high specialization and inherent vigor through long periods of time, there is the phenomenon of apomixis to be cited. Apomixis is a general term for certain reproductive anomalies in plants which are really a return to vegetative reproduction. In a broad way it is synonymous with parthenogenesis in animals; but par- thenogenesis in animals includes only reproduction from an unfertilized egg, while apomixis takes in reproduc- tion from certain cells which are not eggs. Some twenty or thirty species of vascular plants have already been found to reproduce in this manner, and unquestionably the list is very incomplete. Examples from Polypodiacece, Ranunculacece and Rosacece are not uncommon, but in particular it is the Composited, the highest group of flow- ering plants, which seem inclined to make this method of reproduction a habit. Of course, one cannot insist that such a return to primitive reproductive methods even by a more modem labor-saving route is wholly for the good of the species concerned. No one in possession of all of the facts could maintain the change to be progressive, or argue that the species adopting it will have a great future as future is measured by the evolutionist. This is not the contention. We merely cite the adoption of apomixis by flourishing genera of the most specialized and highly developed plants as examples of asexual reproduction over long periods without visibly harmful effects. We do this because we believe the emphasis put by Darwin and his followers on supposed ill effects following any type of inbreeding or asexual propagation was misplaced. Certainly the majority, the great majority, of the higher ANIMAL AND PLANT REPRODUCTION 33 plants returned to a type of reproduction which held all the advantages of bisexuality by evolving means for pro- moting cross-fertilization. But it is the advantage of cross-fertilization and not the assumed disadvantage of self-fertilization that should be stressed. The Knight- Darwin Law, ^'Nature abhors perpetual self-fertiliza- tion/' should read Nature discovered a great advantage in an occasional cross- fertilization. The higher plants made a success of hermaphroditism because there was a return to the advantages of gono- chorism through the development of almost innumerable devices tending to promote frequent crossing between plants of the same and nearly related species. Some species did actually return to true structural gonochorism, but in most cases other means of obtaining cross-fertiliza- tion were developed. There was no advantage, consider- ing their sessile mode of life, in relinquishing the possibility of self-fertilization. Some of the various cross-fertilization mechanisms utilized are very reminiscent of those of animals. Monce- cism, the production of male and of female flowers on the same plant, and dichogamy, the maturation of the male and female organs at different times, have their counter- parts in the other kingdom. So also the physiological phe- nomenon self-sterility of which only one instance is known among animals is very common among plants. Some hundred or so species distributed throughout thirty-five or more families have been shown to be self-sterile, al- though the true number is probably many times this figure. Again polygamy, where, in addition to hermaph- roditic flowers, either male or female flowers are devel- 3 34 INBEEEDINQ AND OUTBREEDING oped, has its analogue in the complemental males charac- teristic of the Cirripedes. But by far the most numerous and most iateresting adaptations for cross-pollination are characteristic of plants alone. These are the thousands of structural modi- fications which utilize external agencies. Wind and water have not been despised, but the real servants — ^they are FiQ. 8. — Adaptation for oroBa-pollination, transference of pollen by insects. (After Kerner.) not slaves for they are paid for their services — are the lower animals and in particular the insects. The ideas of Darwin resulting in the tremendous labors of Miiller,Delpino, Kerner, Knuth and others have made it no longer necessary to describe the facts concerning the dispersal of pollen by animals.^^** ^^'^ The subject has been so fascinating that it is common knowledge how the insects are attracted to flowers by odor and by color ; how they are rewarded for visits by nectar and by pollen ; how ANIMAL AND PLANT REPRODUCTION 35 provisions are made to use them as pollen carriers through numberless modifications of calyx, corolla, sta- mens and pistil; how the animals themselves have devel- oped organs for extraction of food or for attachment to the blossoms (Fig. 8). Perhaps some of these mutual adaptation mechanisms are a little fanciful, but the fact remains that actually an occasional or a frequent cross- poUination is secured by a majority of our 100,000 or more species of flowering plants by means of insects, and the hundreds of mechanisms by which it is obtained are wit- ness of its paramount importance. The thesis of this chapter, then, is simple. The whole trend of evolution in both animals and plants as regards all the mechanisms in any way connected with reproduc- tion, has been such as to provide effectively for continuous descent. In the midst of strenuous competition for place, those organisms which were able to cross with others, at least occasionally, held such an advantage over those which were compelled to continue through one single line of descent, that their descendants have persisted in greater numbers. They have dominated the organic world. Any satisfactory interpretation of the effects of inbreeding and outbreeding must permit a reasonable explanation of this situation. CHAPTER III THE MECHANISM OF EEPEODUCTION Theee is a division of labor in all the higher plants and animals, the result of setting apart definite tissues for producing germ cells. In addition, another important matter is accomplished. The germ cells are insulated from ordinary environmental changes, and are enabled to go through a very exact routine of processes in prepara- tion for the formation of the new organism — the zygote. In general the animal body or the sporophyte of the higher plants can be considered as a double organization. Various parts make up each of the cell units ; but of them all the nucleus, and within the nucleus the chromosomes, seem to be the most important. Each species has a char- acteristic and constant number of these bodies, and it is their distribution which parallels — and probably regu- lates — ^the distribution of the hereditary differences within a species. The double organization of the bodies of the higher organisms is dependent upon the receipt of one set of these chromosomes from each parent. And it is the peculiar method by which these chromosomes are apportioned to the gametes, together with experiments on the actual distribution of characters in the generations succeeding a cross, which have given us a fairly clear idea of heredity as a mechanical process. In ordinary cell division during growth each chromo- some divides longitudinally so that both daughter cells apparently receive an exact half of the chromatin, al- though possibly some sort of a special apportionment is 36 THE MECHANISM OF REPRODUCTION 37 made in the segregation of particular tissues. But when the germ cells are formed, at spermatogenesis and oogen- esis, the chromosomes unite in pairs, a process technically known as synapsis, and at division one member of each pair passes entire to one of the two new daughter cells, thus reducing the number of the chromosomes in the gamete to one-half of those possessed by the body cells. Subsequently there is an equating or halving division similar in appearance to the cell divisions in ordinary growth. Four gametes are thus formed. Leaving out of account the behavior of certain chromosomes believed to control the distribution of sex, there is good evidence that this union of chromosome pairs at synapsis always takes place between two chromosomes, one of which had been received from the father and one from the mother. In other words, it seems clear that each gamete obtains one of each hind of chromosome, although it is a matter of chance whether the cell receives the maternal or paternal representative of any type. Thus, if the chromosomes of the body cells of a particular species are six in number, and we represent them as ABC ahc, regarding A, B, and C as of maternal and a, b, and c as of paternal origin, at synapsis A only pairs with a, B with b and C with c. This procedure, however, will yield eight types of gam- etes, ABC, ABc, AbC, aBC, Abe, aBc, abC, and abc, siuce it is a mere matter of chance which daughter cell receives either member of any pair. In spermatogenesis four sperms are formed from each immature germ cell, but in oogenesis — the maturation of the egg — only one functional gamete is produced, the other three being aborted. Nevertheless, the two processes 38 INBREEDINa A^D OUTBREEDING are similar in all essential features, as may be seen in Fig. 9, the elimination of three out of four of the oocytes taking SPSRMATOOKNBSIS I $pj5.jto. Spermatocyte Secondary — rwato- cytes Spertnato imr Spera- uosoa Multiplication p*riod 056EN£SIS i .^ (^ h. Odgonia •■ Growth period -* Pairing of Chrorsosomes Reducing division J \ / \ Primary oocyte Secondary oocyte (cvoft) and f iret polar body) Mature ovup ^, and poiar bodies Mature ovin WUn* f) 1^^ . mmber cS Fio. 9. — Diagram of gametogenesis showing the parallel between maturation of the aperm cell and maturation of the ovum. (After Guyer.) place in order that their store of nutritive materials may go to make one large egg. Fertilization consists in the fusion of one egg with one sperm, thus bringing back the double number of chromo- THE MECHANISM OP REPKODUCTION 39 somes characteristic of the body cells (Fig 10). and since it is a matter of chance what gametes unite, such gametic differences as we have illustrated would give a possibility of obtaining 8 x 8 or 64 types of zygotes. FiQ. 10. — Diagram to illustrate fertilization; & , male pronucleus; 9 , female pro- nucleus; observe that the chromosomes of maternal and paternal origin, reepectively, do not fuse. (After Guyer.) The mechanism of the process of gametogenesis and fertilization in animals need not concern us further here. We must speak of the process in the seed plants, however, for a rather odd phenomenon occurs there to which there will be occasion to refer later. 40 INBEEEDING AND OUTBREEDING Reduction of the chromosomes takes place in plants just as it does in animals, but the introduction of a gamete generation, the gametophyte, complicates matters. In the seed plants, pollen mother cells are produced in the anthers of the flower which go through precisely the same divisions as in animal spermatogenesis (Fig. 11). But each of the four nuclei thus produced divides during Fig. 11. — Formation of pollen grains in the lily. B, stages in the formation of pollen grains in a group of four (tetrad) within the pollen mother cell; C, mature pollen grain with early stages in the development of the male gametophyte; t, tube nucleus; g, generative nucleus. (After Bergen and Davis.) the formation of the pollen grain, forming a generative and a tube nucleus. The tube nucleus it is that germinates and passes down the style when the pollen grain falls on a ripe stigma. During this period of pollen tube growth the generative nucleus passes down through it toward the ovule, and while so doing divides again, leaving two nuclei each wdth a function to perform. One fuses with the egg and the other with the so-called endosperm nucleus, com- THE MECHANISM OF REPRODUCTION 41 pleting in this manner the peculiar double fertilization characteristic of the angiosperms. In the meantime, the egg and the endosperm yiucleus have been prepared by the accessary cell divisions of the female gametophyte. The reduction division occurs in the usual manner, and as in animals three of the cells are absorbed, leaving a single one to provide for the hereditary succession. Its container enlarges and be- comes the embryo sac, while the cell itself typically goes through three cell divisions resulting in the formation of eight nuclei. Any of these nuclei may become the egg, but generally the egg can be recognized by its position (Fig. 12). Two others from among these nuclei fuse to- gether and become the endo- sperm nucleus, which in turn fuses with the second male nucleus and by succeeding cell divisions forms the en- dosperm of the seeds, the function of which is to furnish food for the young plant, the embryo. Thus, if we represent the chromosome complex of the gametes by x, the embryo is 2x, and the endosperm 3a;. Fertilization in the embryo sac e, egg; fs, first ppcrm; pv, fused polar nuclei ; ss, second sperm. (After Bergen and Davis.) Fia. 12 of the lily 42 INBEEEDING AND OUTBREEDING It is clear from this short description of gametogenesis and fertilization that the processes in plants and in ani- mals are identical in what we deem to be the essential features, the behavior of the chromosomes. If one visual- izes the behavior of hereditary characters in crosses in which the parents differ as the result of the operation of potential factors carried by these bodies, he can correlate i J •V '■iW T Fig. 13. — Entrance of the spermatozoon through the membrane of the egg of a etar- fish giving an idea of the difference in size between the sperm and the egg. (Wilson's "The Cell." Courtesy Macmillan Co.) every fact thus far discovered, with the exception of a few isolated cases found in plants where particular char- acteristics appear to be distributed by the cytoplasm lying outside of the nucleus. Not only can the distribution of ordinary characters be interpreted as functions of the chromosomes, but the distribution of the sexes as well. There is reason to think the behavior of the sex-control- ling chromosomes may perhaps occasionally be influenced by external conditions, but sex itself is determined by the THE MECHANISM OF REPEODUCTION 43 behavior of particular chromosomes of which we have not hitherto spoken (Fig. 14). The evidence in favor of this view of the determination of sex at the time of fertilization through the chromosome complex is from several very different sources. First, there is the phenomenon of multiple births Protevior cf ••• B •••••• Fig. 14. — Diagram showing the distribution of the sex chromosome in Protenor. (After Morgan.) among mammals. In general, animals in which this is the rule, bear both males and females, through all of the individuals must have been under the same environmental conditions. There are multiple births, however, in which the young are invariably of the same sex. Such is the case with those remarkably similar human twins kno^^^l as identical twins. Such is the case with the four young in each litter of the nine-banded armadillo (Fig. 15). Now 44 INBREEDINa AND OUTBREEDING it can be shown that in these two and other similar in- stances, the several young are the product of a single fertilized egg which so develops as to form two or four complete individuals. If sex was determined after fer- tilization, one might expect a random sample of the two sexes here, but this is not the case. I FiQ. 15. — Identical quadruplets in the nine-banded armadillo. Doncaster.) (After Newman from The chief support of this idea of sex determination, however, comes from the microscope and the breeding pen. In an ever increasing number of species, possibly including man himself, it has been found that besides the regular paired chromosomes, the autosomes, there is a single chromosome or possibly a chromosome group com- monly known as the x-chromosome, whose behavior in cell division is somewhat different from the others, and whose THE MECHANISM OF REPRODUCTION 45 distribution absolutely parallels the distribution of sex. There are two types. In the males of animals of Type A, which includes numerous flies, beetles, grasshoppers and bugs from among the insects, as well as representatives from several orders of mammals, a single x-chromosome is present in addition to the regular chromosome pairs, and for this reason two kinds of spermatozoa are pro- duced at spermatogenesis in equal numbers, those pos- sessing the extra element and those without it. In the females, on the other hand, two of these elements are present and the eggs, therefore, always possess it. Thus, on fertilization, half of the resulting young have two x-chromosomes and these become females, while half ha^^e but one and become males. Diagrammatically it is this : Ovum with x fertilized by sperm mth x = female. Ovum with x fertilized by sperm without x - male. In some other cases (Type B), the eggs are di- morphic, while the sperm are all alike, but the result is the same ; the sex distribution follows the chromosome differentiation. In dioecious plants there is some evidence of a similar condition. Strasburger ^^^ found in one of the liverworts, Sphaerocarpus, where the four spores produced by a single spore mother cell hang together and each such tetrad could be planted separately, that invariably two males and two females were produced. More recently Allen ^ has presented evidence of an x-chromosome in this genus. His discovery was made with material of the species Spliwro- carpus Donnellii, but it has been corroborated by one of his students working with Sphccrocarpus texanus. f^ 46 INBREEDING AND OUTBREEDING Again, in the dioecious moss, Funaria, the Marchals;^^^ by a remarkable series of regeneration experiments, have proven the determination of sex at the reduction division. Each spore was found to contain the potentialities of but one sex, but in the sporophyte they demonstrated the po- tentialities of both sexes by inducing direct aposporous development of gametophytes, which proved to have both antheridia and archegonia, the organs of both sexes. The situation in hermaphroditic plants and animals is not so clear. Particularly in plants the peculiar life history with the introduction of alternation of generations, makes experimental work exceedingly difficult. Further- more, there are many species of animals where the sex ratio is nowhere near equality and where both external and internal conditions undoubtedly do have marked in- jfluence, but in such a fundamental phenomenon we can hardly believe these difficulties are insurmountable or will lead to any radically different interpretation of the problem. "Where there is such clear evidence fromvery dif- ferent modes of attack and upon species so unrelated one is constrained to believe the obstacles to a unified theory are only superficial. This is particularly true since there is another line of experimental evidence in favor of the determination of sex by the chromosomes. Our whole evidence on inheritance, in f a-ct, is linked up with chromo- some distribution, so that the easiest way to visualize the process is by supposing that the individual potentialities, the factors, which cooperate in the development of plant and animal characters, are disposed in a definite manner in the chromosomes, as we shall see in the next chapter. The particular discoveries which demand our attention in connection with the phenomenon of sex, however, are THE MECHANISM OF REPKODUCTION 47 those regarding characters commonly known as sex- linked, whose distribution can be accurately predicted if we assume they are definitely coupled with the sex determiner. Such a character is hereditary color-blindness in maji, a condition in which the affected individual cannot dis- tinguish between red and green. It is far commoner in man than in women, and its inheritance is so peculiar that it often seems to skip a generation. A color-blind man married to a normal woman will have only normal children of either sex. The sons will never have color-blind progeny by women with nonnal vision, but the daughters, though married to normal men, will transmit color-blindness to one-half of their sons. If, moreover, a daughter mates with a color-blind man, as might frequently happen in marriage between cousins, on the average one-half of her daughters as well as one- half of her sons will be abnormal. This interesting and apparently complicated inheri- tance is really very simple if we merely assume that the sex chromosomes of the color-blind individuals also carry the determiner for color-blindness. Fig. 16 shows what would be expected. Eepresenting the normal vision by boldface type and color-blindness by outline we see first the result of mating a normal woman with a color-blind man. Since all of her sex-cells, when matured, contain one normal x-element, and since the sex-cells of the male are of two kinds, half containing an abnormal or color-blind determining x-element and half containing no x-element whatever, it is obvious that the sons must re- ceive their x-element only from their mother and the daughters must receive one of their x-elements from their 48 INBEEEDING AND OUTBREEDING father. The sons, therefore, cannot be color-blind and cannot transmit color-blindness, but the daughters, though (Normal) VaU tin* (Color bUDd)| of parenVs o( paresU fiodjr-cellt ot chlldreo Sex-cellt ot cbildroa. Body-cells of ^rand- «bildrea FiQ. 16. — Diagram illustrating the inheritance of a sex-linked character such as color-blindness in man on the assumption that the factor in question is located in the sex chromosome. The normal sex chromosome is indicated by a black X, the one lacking the factor for color perception, by a light X. It is assumed that a normal female is mated with a color-blind male. (After Guyer. Courtesy Bobbs Merrill Co.) they will not be color-blind themselves because one normal x-element is sufficient to determine normal vision, will produce defective x-elements in one-half of their ova, THE MECHANISM OF REPRODUCTION 49 and for this reason will transmit color-blindness to one- half of their sons by a normal man, as will be seen by following out the fourth and fifth columns in the diagram. An egg containing the normal x-element can meet a sper- matozoon carrying an x-element and thus produce a daughter, or it may meet a spermatozoon with no x-ele- ment and thus produce a son; but in either case the chil- dren will have normal vision. On the other hand, an egg containing a defective x-element will by similar fertiliza- tions result either in a normal-visioned daughter, who will carry color-blindness in half of her ova, or in a son who will be color-hlind. Such a scheme of interpretation might seem quite visionary were it not for the fact that similar types of inheritance occur in many of the lower animals. By care- fully controlled experiments with them it has been proven beyond a doubt. CHAPTER IV THE MECHANISM OF HEEEDITY The scientific era in the investigation of heredity be- gan in the latter half of the nineteenth century with the 57ork of Galton and of Mendel. Both enthusiastic and competent investigators, their efforts made with differ- ent material and from diverse points of view, did not fare the same. Galton measured the inheritance of groups of individuals by their resemblance to their progenitors and failed because his method could not take into account the true relationship between the germinal constitution and the body characters of an individual ; Mendel deter- mined the inheritance of a single organism by making the characters of its progeny the criterion and succeeded. Without knowledge of the cell mechanism of gameto- genesis and fertilization, Mendel described the results of his hybridization experiments in terms which agreed pre- cisely with these later discoveries in the field of cytology. Mendelian heredity has proved to be the heredity of sex- ual reproduction: the heredity of sexual reproduction is Mendelian. Progress in the study of heredity through investiga- tions patterned after Mendel's model has been so great that the subject now forms an important sub-division of general biology — Genetics. The details of the subject have outgrown the limits of a single volume, and a knowl- edge of the generalities is no longer confined to the pro- fessional biologist. For such reasons we propose to 60 THE MECHANISM OF HEREDITY 51 discuss here only the broader relationships of Mendelian heredity to the behavior of the chromosomes, since this phase must be emphasized as a basis for correlating the facts from Nature's experiments on inbreeding and out- breeding with the results from the experiments made by man. The Mendelian method of studying heredity consists essentially in crossing forms which differ by well-defined characteristics and in following the distribution of these characteristics separately and quantitatively in the suc- ceeding generations. If a wheat with a long lax head or spike is crossed with one having a short dense spike the F, (first filial) generation bears intermediate spikes. The Fi generation, self-fertilized, however, yields all three types — long, intermediate and short spikes — in the F2 generation ; and in large numbers these types bear a con- stant ratio to each other in the proportion 1 long spike : 2 intermediate spikes : 1 short spike. Nor is this all. The long-spiked plants all breed true to long spikes, the short- spiked plants all breed true to short spikes, while tlv plants bearing intermediate spikes again produce the ratio exhibited by the F^ generation. Diagrammatically the result of the cross is as follows : Pj Long spikes x Short spikes I Fj Intermediate spikes ^-^ II F„ Long spikes Intermediate spikes Short spikes I ^ W ^^ I Fg Long spikes Long spikes Inter- Short spikes Short spikes mediate spikes 52 INBREEDING AND OUTBREEDING If the description of the dual nature of the cells of plants and animals and the result of gametogenesis is recalled, the reason for the production of the ratio of 1 long spike : 2 intermediate spikes : 1 short spike in the F2 generation is plain. Furthermore, it is clear why the types like the grandparents breed true and the type like the hybrid F^ generation does not breed true. The long-spiked wheat has received the factor for long spikes, the something in the germ cell that stands for the production of long spikes, from hoth of its parents ; there- Feviale gantete \likU gamete Fia. 17. — Diagram showing the union of like gametes. fore it breeds true to long spikes. The gametes which it produces all bear the factor for long ears. The diagram illustrating the fusion of the parental gametes, shows why the long-spiked wheat produces gametes, each of which bears the factor for long spikes. If the letter 8 is substituted for the letter L in the dia- gram, the same illustration holds for the short-spiked wheat. But what happens when the long-spiked variety is crossed with the short-spiked variety? A gamete bear- ing L fuses with a gamete bearing ^S' and a zygote LS is formed. The interaction of the factors L and S produces an F^ plant bearing intermediate ears. "When this hybrid THE MECHANISM OF HEREDITY 53 comes to produce gametes they bear either the one or the other — and never both — of these factors. In other words, the germ cells (both male and female) of the hy- brid are half of them L and half of them S. When the F^ generation is selfed, therefore, it is a matter of chance which of these geiTa cells meet to form zygotes. If a large progeny is produced, there will be a ratio of 1|L|L| : 2|L|/S'| : IIaS'IiS'I, and since the formulas L\L\ and S\S\ are like the zygotic formulae of the long-spiked and short-spiked parents, respectively, the plants that they produce will be long-spiked and short-spiked, as the case may be, and will breed true to that character. The inter- mediates, however, having been produced by zygotes L S lilie the F^ generation, will behave in the same man- ner when selfed. That the ratio will be approximately 1 L\L : 2 L S\ 1 8 S is plain if one thinks for a moment what the result would be if a thousand tickets bearing the letter L and a thousand tickets bearing the letter S were shuffled up in a hat and drawn out in pairs, replacing the pair each time after drawing and recording. Suppose the first member of the pair represents the egg cell; the chances are % that it will be L or S. The second member of the pair represents the male cell and the chances are likewise % that it will be L or ^S'. Therefore, when L is the first mem- ber of the pair, half of the time the zygote formed will be L L and half of the time it will be L S , Likewise, when 8 is the first member of the pair, zygotes \S\L\ and S S| will be formed in equal quantities. Combining these possibilities, the ratio 1|-^|L| : 2|L|5'| : 1| ilar conditions. For example, it having been determined that in the pomace fly; many characters are linked to chro- mosomes whose distribution parallels that of sex, we know it to be much more than a guess to say that the color-blindness of man of which the hereditary distribu- tion was described in Chapter III, is controlled by a factor lying in the sex chromosome and recessive to the normal. Though the whole mechanism in the higher plants and animals can thus be pictured as one of sexual reproduc- tion, in its details the results are still too complex to analyze as concretely as the cases given for illustration. Several thousand concrete differences between plants of the various angiosperm families and between animals in at least three different phyla have been followed through pedigree cultures sufficiently carefully to make possible a definite factorial analysis of their hereditary transmis- sion. This has been possible, first, because variation has taken place in these factors, enabling one to follow the transmission of each member of an allelomorphic pair, and, second, because this variation has been somewhat qualitative in nature. Unfortunately for the peace of mind of the biologist, however, the more numerous differ- ences between animals and between plants are the quanti- tative differences, the variations which make organs a little larger or a little smaller. Now it is a great deal easier to determine the transmission of the factor differ- ences which determine that one flower shall be red and another white than it is to trace the distribution of the factors which determine that one flower shall be one inch and another two inches long. Nevertheless, through the efforts of numerous investigators it has been possible to THE MECHANISM OF HEREDITY 67 show that such hereditary differences behave as should be expected if their inheritance follows the same laws as do the simpler characters. The basis, as one might say, of the Mendelian interpretation of size differences is the proof that practically aU qualitative characters are affected by numerous factors. Sometimes there are two or more factors which produce nearly identical visible results, but more often the character complex is affected in different ways and in various degrees by particular factors. Whether the character develops at all or not seems to be due to the presence or absence of one or more main factors, but given the presence of these factors the degree of development may be influenced by many sub- sidiary factors or modifiers. Now these modifiers being transmitted independently of one another and of the prin- cipal factor or factors, an individual carrying certain modifiers and lacking the principal factor may be crossed with an individual carrying the main factor and lacking the modifiers. The result is a series of recombinations among the germ cells of the F^ generation which produces F2 individuals carrying various groups of modifiers and therefore developing the character complex under con- sideration in different degrees. If one studies carefully such crosses as the one just described, he finds that a number of general conditions are fulfilled. 1. Wlien pure or homozygous races are crossed, the jPi populations are similar to the parental races in uni- formity. This conclusion devolves from observations that if any particular factors AA and an are homozygous in the parental races, they can only form Aa individuals in the Fj generation. 68 INBEEEDING AND OUTBREEDING 2. If the parental races are pure, F2 populations are similar, no matter what F^ individuals produce them, since all variability in the F^ generation is the result of varying external conditions. 3. The variability of the F2 populations produced from such crosses should be much greater than that of the JPi populations, and if a sufficient number of individuals are produced the grandparental types should be recov- ered. The fulfillment of this condition comes about from the general laws of segregation of factors in F^ and their recombination in F2. 4. In certain cases F^ individuals should be produced sho^ving a greater or a less extreme development of the character complex than either grandparent. This is merely the result of recombination of modifiers, as was explained above. 5. Individuals of different types from the F2 genera- tion should produce populations differing in type. The idea on which this statement is based is, of course, that all F2 individuals are not alike in their inherited constitution and therefore must breed differently. 6. Individuals either of the same or of different types chosen from the F2 generation should give F^ populations differing in the amount of their variability. This con- clusion depends on the fact that some individuals in the F2 generation will be heterozygous for many factors and some heterozygous for only a few factors. Such are the conditions which must be fulfilled by crosses exhibiting size differences if we are to visualize their inheritance in the same way as we visualize the in- heritance of qualitative characters such as color. If the size differences are controlled by numerous germ-cell f ac- THE MECHANISM OF HEEEDITY 69 tors, the distribution of the latter cannot be followed with the same ease as one would follow the distribution of cotyledon colors in the garden pea. This is time because the visible effects of certain factors is sure to be very small, and because varying external conditions obscure the effects of inheritance. For example, a plant which through its inheritance should become 6 feet tall under average conditions may become only 4 feet tall if planted in a sterile soil, but a plant which under average condi- tions would become only 4 feet tall might become 5 feet tall if grown in a very fertile soil. Nevertheless, in spite of these drawbacks, one can select size characters for study which are influenced but slightly by external conditions, and by studying large numbers through several generations, and by applying mathematical tests to determine the uniformity or the variability of the resulting populations, he can find out whether quantitative characters satisfy the six require- ments seen to be fulfilled by qualitative characters. This has been done in numerous cases, and the results firmly convince all unprejudiced investigators that the inheri- tance of all types of characters is the same. Table I, from crosses between two varieties of Nico- tiana longiflora ^^ differing in the size of their flowers, illustrates the point. One does not need any refined mathematical methods to see that when the small variety having flowers about 40 mm. in length is crossed ^^dth the large variety having flowers about 94 mm. in length; the result is a uniform F^ population having flowers about 64 mm. in length. The two Fn populations which it produced are much more variable ; and one can easily calculate that if several thousand plants had been gro^^^l instead of 70 INBEEEDING AJ^D OUTBREEDING o o CQ » < > w « QQ O o O o o tf o O o CQ O H QQ Q P » Pi ti 0) ■n s 6 £ fl w 03 i-ii:0(M 05 tH ^ 00 Oi 1 1 Tfl COrH r-H lO O (N 05 1 (NCOl> 1 (N T-l tH (N CO 00 »o (N 00 00 CO(NiO 1-H CO Oi CO Oi lO lO (NCO 00 1-1 CO 1-1 O 05 '-I o CO lO^ (N OO C01> O 1-1 lO CO ■* i-< 1— 1 Tt^ 1> CO 1-1 05 Tti i-< (M O r-l(M O 00 1-1 00 (N "^ CO -^ CO (N CO C0 1>. Oi -rt^ (N 05 O 05i-l CO coco CO iM 1-1 CO lO »0 rH (NCOrJH coo 05 (M rJHCO (N Tt^ (N CO —1 gl 1 COt^OOOS CO CO 1-1 o to CO CO lO CO 1-1 O 1-1 t^i-HCO(N O(Nt-i00 1 T-^ 1 1 1 1 ,-1 CO rt^ ri^ t>. 05T-lr-l(N CO 1 1 1 1 TtH(N (N ^ 00 OCO(M (MTfOrH 00 T— t l-H C^ t>- 1— 1 TJH lO Tt< 00vOt>.i-( lO CO 1-1 CO "^ CM »-( (N TjH Oi uo O to 1— 1 T-H TjH rt^ * 1 Tj<0 CO t^ iO- CO Tt< ^ CO i-H 1 CO CO CO 0) a> rHT-M(NcOOOt^^OO.'>*iOCOt^X00iO00'*'<^0000->*Oi S3g CS i-i(NCO'-<(NCOi-H(Nor» 1—1 »— 1 1 1 '-' r-H ,-< (N i-((N 1 1 1 1 1 1 THC^eoTfi-HcoTj<»ococ^cococococo 1 1 1 1 1 1 i 1 1 1 1 1 1 1 ,-H(NrHTHT-lrHo<: COC ococ rt =3 G. O ft CO fe 1 o (1) 2<-> (2) 4-^— > (3) 8<— >(4) 16-<— > (5) 32<-^ (»)2'» . . ., in which the enclosed numbers represent the ancestral generations (1 = parents, 2 = grandparents, 3 = great- grandparents, etc.), and the other figures the number of ancestors. In the second or earlier generations the an- cestors may not all be different individuals, so that in any generation previous to the parental the actual number of ancestors may be less than the possible number. For ex- ample, in brother and sister mating, any individual in- stead of having four different grandparents, has only two. Expressed symbolically, as above, the representation for this type of mating would be X <-^ (1)2 -^-> (2)4-2/1 -^-> (3)8-7/2 <-> (4) 16-1/3 <-> (5)32-2/4. . . , where 2/1 = 2, 1/0 = 6, y.^ = 14, 2/4 = 30. In this case y has the value of 2" - 2, and this is the highest value it can have in any system of mating where two indi- MATHEMATICAL CONSIDERATIONS 83 viduals are necessary for reproduction. Applj^ng the formula given above, the Coefficients of Inbreeding for each generation in brother and sister mating are : Zo = 100 (2-2)= 2 Z^ = 100 (4-2) =50 4 ^2 = 100 (8-2) =75 8 Z3 = 100 (16-2) = 87.5 16 The figures obtained are the differences between the possible number of ancestors and the actual number ex- pressed as percentages of the former. By plotting these percentages for successive generations on the generation number as a base, a curve of inbreeding is obtained which can be compared to the curves obtained by other systems of matings. This comparison is shown in Fig. 22 for the common types of matings as worked out by Pearl. From these curves it is evident that continued brother by sister and double first-cousin matings have the same effect, although the latter is one generation behind the former. Also the curves for parent by offspring and sin- gle first-cousin matings are similar in type, but show the same differences in position. In any case the concentra- tion of the lines of descent in these systems of inbreeding is rapid, until after fifteen generations no individual can have more than a fraction of one per cent, of the number of ancestors theoretically possible. The Coefficient of Inbreeding alone tells us nothing as to the relation between the different lines of descent. 84 INBEEEDINa AND OUTBEEEDING Two individuals may have the same Coefficients of In- breeding when considered for any given number of gen- erations, but differ greatly in germinal constitution. This is due to the fact that the two lines brought together in the immediate production of any individual may or may not be related. For example, a closely inbred animal of one breed may be mated to another closely inbred animal 100 80 £ 60 a o I 40 20 . /^ 5? ^ 1 / t 1 r f t 1 1 1/ If 1/ 1/ 1 6 8 10 Generations 12 14 FiQ. 22. — Curves of inbreeding showing (o) the limiting case of continued brother X sister breeding, wherein the successive coefficients of inbreeding have the maximum values; (b) continued parent offspring mating; (c) continued first cousin X first cousin mating where the cousinship is double (C^ XC^), and (d) continued first cousin X first cousin mating where the cousinship is single (C* XC^). The continued mating of uncle X niece gives the same curve as C^ X C^ (After Pearl.) of an entirely different breed. The two lines of descent would then be totally unrelated as far as the known pedi- grees are concerned, but the resulting individual would have a high Coefficient of Inbreeding, due to the concen- tration of ancestry separately in the two ancestral lines. To give some measure of the inter-relation of the lines of descent, Pearl has devised the Coefficient of Eelationship, Ky which is essentially the per cent, of the individuals in MATHEMATICAL CONSIDERATIONS 85 each of the descending lines which are also represented in the other line. To give an adequate mathematical esti- mation of the degree of inbreeding, both constants are necessary. There is, generally, some correlation between them, although the Coefficient of Relationship may be zero, and the Coefficient of Inbreeding still be high, as in the illustration just given in which the progeny comes from a pair of individuals from two distinct inbred lines. The application of these methods of determining the amount of inbreeding is illustrated by Pearl from the pedigrees of two Jersey bulls as follows : Inbreeding" " Z" and Relationship '' (K) '' CoeOicients of King Melia Rioter 14th and Blossom's Glorene ^1 Z, {KJ (0) (0) A, Z, (K,) 25 (0) (0) ^3 Z. {K,) 25.00 (50.00) 12.50 (0) ^4 Z\ (K,) 37.50 (G2.50) 12.50 (0) ^5 Z, {K.J 50.00 (75.00) 25.00 (0) ^6 Z, (K,) 71.88 (87.50) 29.C9 (0) ^T Z, (K,) 81.25 (92.19) 35.94 (0) ^8 Z, {K,) 90.63 (92.97) 40.23 (0) The method of making the calculations is explained clearly and concisely by the originator and we shall not undertake to repeat it here. What we are interested in is the genetic meaning of the figures after they have been obtained. The Coefficient of Inbreeding, Z, has to do solely with total relationship, and shows the intensity of inbreeding in the stockman's sense of the word by measuring pre- cisely **the proportionate degree to which the actually existent number of different ancestral individuals fails to 86 INBREEDING AND OUTBREEDING reach the possible number, and by specifying the location in the series of the generation under discussion. * ' King Melia Rioter 14th had less than 10 per cent, of the maxi- mum number of ancestors in the 7th ancestral generation, while in the same generation Blossom's Glorene had nearly 60 per cent. From these figures it is evident that King Meha Rioter 14th is a much more inbred animal than 100 80 2 60 a ® m ' ^^o. 121 rpj^g grgt two investi- gations, together with that of Weismann and von Guaita,^^' ^'^ on mice have been the classic examples of the adverse effects of inbreeding. Crampe^s experiments started with a litter of five 102 INBREEDING AND OUTBREEDING young, obtained by crossing an albino female with a wnite and gray male. These animals were inbred in various degrees for seventeen generations. During the experi- ment many rats showed great susceptibility to disease, divers kinds of abnormalities, diminished fertility, and increased total sterility. Similarly Ritzema-Bos started his investigations with a litter of twelve rats obtained by crossing, this time an albino female with a wild Norway male. This stock was inbred in different ways for six years, during which time he claimed to have obtained about thirty generations. His results did not corroborate those of Crampe in so far as susceptibility to disease or appearance of malformations are concerned, but there was a gradual decrease in size of litter and a gradual increase in percentage of infertile matings, as is shown in the following table : Year of inbreeding 1 2 3 4 5 6 Ave. number ia litter 7.5 7.1 7.1 6.5 4.2 3.2 Per cent, infertile matings 0.0 2.6 5.617.4 50.0 41.2 These investigations, in spite of the habit biologists have of citing them, are not calculated to settle the ques- tion they undertook to answer. Ritzema-Bos himself criticizes those of Crampe, because he believes them to have been started with a weak strain. Miss King, how- ever, thinks the weakness of these rats, as indicated by their susceptibility to disease, the appearance of mal- formations, and their tendency to sterility, was due to the conditions under which they were kept. She had a similar experience during the earlier part of her own experiments, and found that inadequate nourishment was largely the cause. But it is not for this reason that we INBREEDING EXPERIMENTS 103 feel that both of these experiments should be disregarded. Each ivas started with hybrid stock, and such experi- ments with hybrid stock bring in an additioyial compli- cation, Mendelian recombination. The only type of investigation on bisexual animals calculated to offer criti- cal evidence on the effect of inbreeding per se must be carried on with stock which has already been inbred long enough to reduce the genetic constitution of the animals to an approximately homozygous condition. Then, and then only, can the effect of more extended inbreeding be determined without confusion as to the interpretation of the results. Miss King has pointed out a part of the difTiculties involved in starting with a hybrid stock. In one of her experiments the progeny of a cross between a wild Nor- way rat and an albino was inbred for several generations. She found that while the majority of the F-^ females were fertile, at least 25 per cent, of the Fo females were com- pletely sterile and 10 per cent, of those which did breed cast only one or two litters. In the strains extracted from this cross there was variation in degree of fertility, but none was found which exhibited the high degree of fer- tility usually existent in the albino rat. No endeavor to select fertile strains was made and one cannot say whether or not rigid selection would have isolated them, but the researches of Detlefsen^' on hybrids of the genus Cavia to which the common guinea-pig belongs indicate this to be a probability. These investigations as well as those of Easf^^ on the genus Nicotiana show conclusively that various hereditary factors are involved in the partial sterility exhibited in many species crosses, and that tliese factors may be expected to recombine in the usual manner. 104 INBREEDING AND OUTBREEDING It is more than a mere assumption, then, if a great part of the sterility found by Crampe and Ritzema-Bos is attrib- uted to the same cause. The investigations of Weismann and von Guaita are hardly more satisfactory. Weismann inbred a stock of white mice for twenty-nine generations, and found the average number of young for the three 10-gen. periods to be 6.1, 5.6 and 4.2. Where this stock originated, and what method of inbreeding was followed, we are not told. Presumably the gross result was a slight decrease in fer- tility, coincident with the amount of inbreeding, but even this is not certain. As King points out, the average num- ber of litters under observation in the first two genera- tions was twenty-two ; in the last nine generations, three. Clearly there was greater opportunity of selecting healthy breeding stock, as well as a lower probable error, in the earlier part of the experiment, and this might account for the slight difference in fertility found. Von Guaita crossed some of these highly inbred mice with Japanese waltzers and then inbred for six generations. He reports the average number of young in the successive generations as 4.4, 3.0, 3.8, 4.3, 3.2 and 2.3 ; but in view of the vigor almost always expressed in the F^ generation of such crosses one is inclined to doubt the pertinence of these figures to the inbreeding problem. Although, as has been pointed out, there is good rea- son for disallowing the claim of these much cited experi- ments on mammals as proofs of the adverse effects of inbreeding through consanguinity alone, there is no inten- tion of denying the isolation of individuals characterized by undesirable qualities from mixed strains by means of Mendelian recombination. Perhaps it is not wise even to INBREEDING EXPERIMENTS 105 maintain the impossibility of injury to any strain of any species through inbreeding per se, but it is proper to say that the evidence in favor of it is practically nil. Doubtless we could make our case more convincing to the stockman could the enormous number of really well- kept herd records be cited and analyzed. But it is not possible at present to say whether many of these records satisfy the requirements of modern genetic research. This is a task which must be left to the breeding organiza- tions of the future. We can appeal at present to only two investigations on mammals where the effect of Mendelian recombination has been largely eliminated; and these again are on small mammals, the rat and the guinea-pig. The first of these investigations to be reported was that of King.^^^' ^^o, 121 j-|- ^r^g started with a litter of four slightly undersized but otherwise normal albino Norway rats, two males and two females. From these females two lines, A and J5, were carried on for twenty- five generations by mating brother and sister. In the earlier generations practically all of the females were used for breeding, but in every generation after the sixth about twenty females were selected from approximately a thousand available young. At first the inbred rats exhibited many of the defects reported by Crampe. Numerous females were either ster- ile or produced but one or two small litters. Other ani- mals were characterized by low vitality, dwarfing, and malformations. Stock rats exhibiting the same charac- teristics at this time, however, led to a change in the food, following which the **dire effects of inbreeding'^ prac- tically disappeared. Whether this improvement in the colony was due entirely to the change of diet or may be 106 INBREEDING AND OUTBREEDING attributed partly to selective eUmination of the weaker rats cannot be determined. We are inclined to agree with Miss King in giving greater weight to the first factor, though for a reason which she does not mention. The general success of Miss King^s whole investigation we believe to be due largely to the fact that the experiments were started with stock rats which already must have been very closely inbred and therefore in an approxi- mately homozygous condition. From the seventh generation on, selection was made on the new-bom young with general vigor as the basis, but the two lines were selected differently. In line A only litters having an excess of males were selected to serve as the progenitors of the succeeding generation, while in line B the reverse was the case. The general result was to show that the normal sex ratio in this species, 105 males to 100 females, can be changed. At the end of nineteen generations of selection, line A had produced litters hav- ing a sex ratio of 122.3 males to 100 females, and line B had produced litters having a sex ratio of 81.8 males to 100 females. From these facts there is no doubt but that lines having an hereditary tendency to produce different sex ratios can be isolated, but there is no evidence what- ever in favor of the theory of Diising proposed in 1883 to the effect that inbreeding by lessening the vitality of the mother increases the percentage of male young. The change in the sex ratio was made in two generations. After that the effect of selection ceased. Such a result not only militates against attributing the changed ratios to inbreeding itself, but indicates that a relatively small number of Mendelian factors are involved in the control. The effect of continuous inbreeding on body weight is INBREEDING EXPERIMENTS 107 shown in Fig. 25. This graph is constructed from data collected from the records of males of line A, but graphs constructed from the records of the females of this line and from males and females of line B do not differ from Grow th in bo<^y wci H S«ries A ighl Albtno Rat Males PTn 120 140 160 180 200 220 240 260 260 Age in days Fig. 25. — Graphs showing the increase in the body weight with^age for males of inbred albino rats. (Series A.) A, graph for the males of the seventh to the ninth generati< r.u inclusive; B, graph for the males of the tenth to the twelfth generatinns inclusive; C, graph for the males of the thirteenth to the fifteenth generations iucluhive; D, giaph for the n.ult* of the Grst six inbred generations. (After King.) it in any essential feature. Curve D is further evidence for concluding that the animals of the first six generations suffered from malnutrition, since, as Miss King notes, it is preposterous to suppose that these animals could have given rise to the very large individuals represented by 108 INBREEDING AND OUTBREEDING curve A \1 it really represented their true body weight. How favorably these inbred strains compare with stock animals is shown in Fig. 26. Paralleling the results obtained for body weight were 20 40 eo 320 340 360 380 400 420 440 460 460 Fig. 26. — Graphs showing the increase in the weight of the body with age for different series of male albino rats. A, graph constructed from Donaldson's data for stock albinos; B, graph for males belonging in the seventh to the fifteenth generations of the two series of inbreds combined; C, graph constructed from data for a selected series of stock albinos used as controls for the inbred strain; D, graph for males belonging in the first six generations of the two series combined. (AfterKing.) those upon fertility and constitutional vigor as judged by longevity. Neither was reduced by inbreeding; in fact, there seems to be no doubt but that there was a significant increase in both cases. There was a slight but definite increase in fertility as is evident if one plots the theoreti- cal curve which fits the experimental curve for litter size INBREEDING EXPERIMENTS 109 throughout the course of the experiment (Fig. 27). The whole series of inbreds compares well in this respect with stock albinos for which the average litter size is G.7. Further there was a notable increase in longevity in line A and a marked increase in line B. -■ttt-t: tT-T:!:tT"-- !:::::"":"":: !"V""M :■!;:;;: — "" nu: ' i :"!" !r' -■■"ii" ::_4 _± - _ : : : 1 : " ■ i ;;:;.! 1 It ri i : ; 1 ; ! 1 ~- - — 1- I _ _ ^ ..- _ __^ j'.i.^ 8 Average Size of Litter - - - - ' T A -i. • ^ " ± ■ L ML ^^Bi' ^« i \.^0ir^' ' ^^ \i. I J ^ H L. ■ ■ ^■! 1^ — ^yw 'r^r^ ^^Ni^'*' ^^ ^^^^ 1 ' t M t f ■ ^ * K ^r 1 _ „ . -*«i Im W _L- 1 *:::j| I-- "" :'_ m W 5- J «-*- _ — _ M ^ __ 1 _ _ - _ - _ _ ■ . g_ _„__ -___.. _>_ — _ — ~ __ " _ :: : GENERATIONS :::::::: , L x X 1: - 3- — ^ 4^x|^..|4ffl_ J..., . tr^ii^iMMMi^lMM^II^m.l^N 11 <2 13 14 Ift 16 .17 l« 19 20 flf St fit 2« 25 Fia. 27. — Graph showing the average Bise of litters produced in Bucceseiye generatione of inbreeding albino rats by brother and sieter matings. (After King.) The interpretation of these experiments is wholly in accordance with Mendelian theory. Starting with stock rats which from previous close breeding had already been reduced to a high degree of homozygosity, inbreeding had the tendency to accentuate this purity of type and to segregate slight differences. By selection vigorous uni- form strains were built up, strains somewhat larger, more 110 INBEEEDING AND OUTBREEDING fertile and longer lived than many strains of stock rats. It is clear that this was the result of Mendelian recom- bination for the two lines A and B were in the end some- what different. The rats of line A were slightly more fertile, attained sexnal maturity earlier, and lived longer than those of line B, If this evidence were not sufficient, it is supplemented by the fact that variability gradually be- came reduced during the progress of the inbreeding. The investigations of the effects of inbreeding on the guinea-pig, to which we have referred, were begun in 1906 by G. M. Rommel of the United States Department of Agriculture. In recent years the work has been in charge of SewaU Wright, who has made a very illumi- natiag analysis of the results obtained. This series of experiments was started with thirty- three pairs of stock animals which had been more or less inbred previously. Although maintained exclusively by mating sister with brother, sixteen of these families were in existence at the close of 1917 after some twenty gen- erations of the closest inbreeding. Considered as a whole this inbred race shows distinct evidence of having declined in every character connected with vigor. The litters are smaller and are produced more irregularly. The per cent, of mortality both in utero and between birth and weaning has increased. The birth weights are lower and the rate of growth slower than in control stock. In spite of these facts, however, one is forced to the conclusion that these results are not the effect of inbreeding as a direct cause, but are to be attrib- uted to Mendelian segregation. There are pronounced differences between the various families. Some are still very vigorous, comparing favor- C c a: 3 o O 3 a> •-) p o' 3 3- O o CO CO O m 3 cr 1-1 3 INBREEDING EXPERIMENTS 111 ably with the original stock; others degenerated so rap- idly that they soon became extinct in spite of every etl'ort to prevent such a catastrophe. Among the families still in existence, there is even evidence that vigor as a general term may be divided into various causative factors and that these factors may be combined in various ways, liy grading each family for various characters connected with vigor of growth and reproduction and then classi- fying each family in numerical order for each separate character, Wright has been able to show conclusively that there ar^ many hereditary factors which alfect fer- tility, growth and vitality and that almost any combina- tion of these characters may become fixed in a family through inbreeding. A little later we shall have occasion to speak of sev- eral noteworthy end results obtained by inbreeding the larger domestic mammals, but no further discussion seems advisable in this place because of the lack of quan- titative data. A similar statement holds for birds. The fruit fly, Drosopliila ynelanogaster^ is the only insect which has been used for extended experiments on effects of inbreeding, although there are numerous examples on record where an importation of a rela- tively small number of individuals has resulted in an overwhelming increase — witness the gypsy moth in New England. Castle and his co-workers ^^ bred Drosophila for many generations by continuous brother and sister mat- ings. After fifty-nine generations of this close inbreed- ing the fertility did not appear to be reduced below that shown by the original stock, although it was increased by crosses between certain inbred lines. There was some 112 INBREEDING AND OUTBREEDING indication of reduction in size when inbred flies were compared with random mated stock reared under the same conditions. Far from being exterminated by in- breeding, however, the flies at the end of the experi- ment were apparently fully equal to those with which it was begun,. These experiments showed clearly that inbreeding results in strains of unequal fertility. The less fertile tended to be eliminated by differential productiveness, so that only the more fertile remained. The occurrence of absolute sterility was pronounced in the first part of the experiment, but almost entirely disappeared in the later generations. The figures as calculated from their table are as follows : Per cent. of matings totally sterile Generations 6 to 24 17.80 Generations 25 to 42 18.47 Generations 43 to 59 3.37 Such a result is to be expected when it is remembered that inbreeding produces homozygous individuals, and these, whenever sterile, are, of course, eliminated. Moenkhaus,^^^ Hyde,^^ and likewise Wentworth,^^^ by similar inbreeding experiments with Drosophila found sterility, though increased in the first stages of inbreeding, tended to be eliminated after the process was long continued. The only other experiments on invertebrates which ought to be cited here are those of Whitney ^^^ and A. F. ShuU ^^^ on the rotifer Hydatina senta. Both of these investigators found that inbreeding had a considerable adverse effect on the size of family, number of eggs laid INBREEDING EXPERIMENTS 113 per day, rate of growth, and variability. The proi>er iii- tei-pretation of their resultis is somewhat obscure, unless one postulates the origin of frequent mutations. Thu number of generations bred and the number of families under observation were not sufficient to demonstrate the segregation of differences in these characteristics, thuugh this is to be expected since these qualities are sympto- matic of general, vigor and general vigor was increased by crossing. The difficulty, however, lies in the fact that continued parthenogenetic multiplication which is possible in Hydatina had the same effect as continued inbreeding. Shull introduces the interesting speculation that this sim- ilarity is due to a gradual adjustment of nucleus to cyto- plasm during the asexual propagation — this being as- sumed to bring about the same results as a gradual approach toward homozygosis. We are inclined to at- tribute both changes to environmental causes, believing that if a proper change in diet had been made vigor would have been maintained. While we are not justified in concluding from these experiments that inbreeding accompanied by rigid selec- tion will be beneficial to bisexual animals, they certainly show close mating is not invariably injurious. They in- dicate that the results of inbreeding depend more upon the genetic composition of the individuals subjected to inbreeding rather than upon any pernicious intluenc^e in- herent in the process itself; and, as will be emphasized more strongly later, it is a wholly different matter whether inbreeding results injuriously through the inlieri- tance received, or whether consanguinity itself is respon- sible. Yet such a status for the problem is unsatisfactory. The experiments on animals bring to light no facts which 8 114 INBEEEDING AND OUTBREEDINa may not be interpreted as the result of Mendelian factor recombination ; but if one were to base his judgment on them alone, he could not truthfully make the didactic state- ment that inbreeding per se is not injurious. There would ever be the uncertainty with which the additional variable bisexuality always encumbers a genetic experiment. For- tunately, we may turn to the numerous experiments on hermaphroditic plants for the deciding vote. Many wild species and cultivated varieties of plants are almost invariably self -fertilized, and apparently lack nothing in vigor, productiveness or ability to survive. Amon^ wild plants many species of the family Legumi- nosae, among cultivated plants — ^wheat, rice, barley, oats, tobacco, beans, tomatoes — are types characterized by very nearly continuous self-fertilization, and these plants are in no immediate danger of extinction. On the other hand, the majority of the higher plants is provided with devices which promote natural cross- pollination, and show definite injurious effects when in- bred artificially. Even species which are characteristi- cally self-fertilized are crossed occasionally. This, to- gether with the fact that nearly all plants and animals are benefited by crossing, led Knight as early as the close of the eighteenth century to believe self-fertilization is not a natural process and always produces more or less injurious results. His views were summed up in the state- ment, * * nature intended that a sexual intercourse should take place between neighboring plants of the same spe- cies.'* Darwin, fifty years later, basing his conclusions upon observation of animals and direct experimentation with plants, was even more radical, and concluded that ** nature abhors perpetual self-fertilization." INBREEDING EXPERIMENTS 115 Darwin compared self-fertilizefl plants witli iiit<'r- crossed plants in many different species. In tlie majority of cases the self -fertilized plants were clearly inferior to the crossed plants. These facts led to the belief that the evil effects of inbreedin.2: kept on accnmnlatinsr nntil eventually a plant or animal continuously reproducing,^ in that manner was doomed to extinction. His own ex]>eri- mental results came far short of proving such an assump- tion, however. The two plants with which inhrc'Oflinir was practiced the longest — Ipomea and Mlmnlu^ — showed very little further loss of vigor after the first generation. What the experiments did show, most clearly, was segre- gation of the inbred stock into types differing in tlieir ability to grow as well as in minor, visible, hereditary characters. In both species plants appeared wliich were superior to other plants derived from the same source, some being equal or even superior in vigor to the original cross -pollinated stock. The inbred plants differed from the original material most noticeably in the uniformity of visible characters. Darwin ^s gardener stated it was not necessary to label the plants, as the different lines were so distinct from each other and so uniform among them- selves they could easily be recognized. After several generations of inbreeding, Dar^vin found it made no difference in the resulting vigor whether the plants in an inbred lot were selfed or were crossed among themselves. This he correctly ascribed to the fact that the members of such an inbred strain had become germinally alike. With less justice he attributed this approach to similarity in inherited qualities to tlie fact that the plants were grown for several generations under 116 INBEEEDING AND OUTBREEDING the same conditions, but it is easy to see why he held so tenaciously to this view if one remembers the faith he had in the effect of environment on organisms. Such a view he deemed supported by the fact that crosses of selfed lines with the intercrossed lines (also inbred, but to a less degree) did not give as great increases in growth as crosses of selfed lines with fresh stock from other local- ities. His crosses between inbred lines did give noticeable increases in growth, however, in many cases equaling the original variety. This is well illustrated by Dianthus, in which the selfed line was crossed both with the inter- crossed line and with a fresh stock. The ratio of each crossed population to the selfed population in height, number of seed capsules, and weight of seed produced is as follows : Selfed Selfed X X Intercrossed Fresh Stock Height, comipared to selfed plants 100 :95 100 :81 No. capsules oompared to selfed plants 100:67 100:39 Weight of seed compared to selfed plants 100 :73 100 :33 With Darwin we still attribute the greater increase of vigor in crosses of distinct stocks to a greater germinal diversity, but we differ from him as to the way in which that diversity is brought about. Be that as it may, great credit is due Darwin for being the first to see it was not the mere act of crossing which induced vigor but the union of different germinal complexes. This he states clearly in the following sentences ('* Cross and Self -Fertilization in the Vegetable Kingdom,'' p. 269) : *'A cross between plants that have been self -fertilized during several suc- cessive generations and kept all the time under nearly INBREEDING EXPERIMENTS 117 uniform conditions, does not benefit the offsprin^r in the least or only in a very slight degree. Mimulus and the descendants of Ipomea, named Hero, offer instances of this rule. Again, plants self-fertilized during several generations profit only to a small extent by a cross with intercrossed plants of the same stock (as in the case of Dianthus), in; comparison with the effects of a cross by a fresh stock. Plants of the same stock intercrossed dur- ing several generations (as with Petunia) were inferior in a marked manner in fertility to those derived from the corresponding self-fertilized pknits crossed by a fresh stock. Lastly, certain plants which are regularly inter- crossed by insects in a state of nature, and which were artificially crossed in each succeeding generation in the course of my experiments, so that they can never or most rarely have suffered any evil from self-fertilization (as with Eschscholtzia and Ipomea), nevertheless profited greatly by a cross with a fresh stock. These several cases taken together show us in the clearest manner that it is not the mere crossing of any two individuals which is beneficial to the offspring. The benefit thus derived de- pends on the plants which are united differing in some manner, and there can hardly be a doubt that it is in the constitution or nature of the sexual elements. Anyhow, it is certain that the differences are not of an external nature, for two plants which resemble each other as closely as the individuals of the same species ever do, profit in the plainest manner when intercrossed, if their progenitors have been exposed during several genera- tions to different conditions.'^ Unfortunately, in Darwin's time the key to the sola 118 INBREEDING AND OUTBREEDING tion of the problem of inbreeding was lacking. Mendel's work was yet unrecognized ; the principles of inheritance of separate characters, of segregation, of chance recom- bination, Darwin was not permitted to know. Had he realized the way in which recessive characters can be con- cealed for many generations without making their ap- pearance until homozygosity was brought about by in- breeding, doubtless his views on the subject would have been materially changed. As we have just indicated, and as we shall have occa- sion to emphasize again, the greatest advance in our knowledge of the significance of inbreeding has come through linking its effects with Mendelian phenomena. The first experiments on the subject made in the light of this discovery were those of G. H. ShuU and of East, undertaken independently in 1905 with maize, an ideal cross-fertilized species, as the subject. Shuirs investigations were not begun with the object of studying the effects of self-fertilization, but the studies having involved parallel cultures of cross-pollinated and seff-pollinated lines, it was impossible not to have noticed the smaller stalks and ears and the greater susceptibility to attacks of the corn-smut (Ustilago maydis) shown by the latter. Interest thus aroused, data were collected bearing on the subject of inbreeding, and in 1908 his first conclusions on the subject were published. His observation that the progeny of every self-fer- tilized maize plant is inferior in size, vigor and productive- ness to the progeny of a normal cross-bred plant derived from the same source, corroborated preceding investi- gations made by Morrow and Gardner ^^2' ^^^ and INBREEDING EXPERIMENTS 1 1 9 Sliamel ^^^ ; but the conclusion whicli lie drew was now. The universality of this decrease in vigor was to Shall a proof that the injurious effect of inbreeding could not be due to an accumulation of deficiencies possessed by the parents since superior and inferior parents yielded sim- ilar results. Further, Shull noted that this decrease in size and vigor accompanying self-fertilization, instead of proceeding at a steady or even at an increasinir rate as might be expected from this older view, actually became less and less in succeeding generations — ])resiiinably in- dicating an approach to stability. The neatness with which these observations fit a ^Nfendelian interpretation of inbreeding did not escape notice. It was pointed out how one might consider a com field to be a collection of com- plex hybrids whose elementary components may lie separated by self-fertilization through the operation of the fundamental Mendelian laws of segregation and recombination. "With this working hypothesis the investigations were continued for several years, papers on the subject ap- pearing in 1909, 1910 and 1911. Evidence of the hybrid nature of ordinary commercial maize plants and their de- pendence upon hybridity for their vigor was found in the decided differences in definite, hereditary, niorpholoia- cal characters exhibited by self-fertilized families having a common origin, but a further proof of the validity of the hypothesis came in testing the conclusions to which the view leads. Obviouslv crosses between ])lants of a single family, which by long-continued self-fertilization has be- come homozygous in nearly all its characters, should show little increase in vigor over self-fertilization; but crosses 120 INBEEEDING AND OUTBREEDING between distinct self -fertilized lines should often result in high-yielding F^ generations possessing great vigor and showing a high degree of uniformity. Again, crosses between different near-homozygous strains, though uni- form and vigorous in the Fj generations, should become much more variable and much less vigorous in the F2 generation. These general propositions Shull tested in a limited way in 1910 after his families had been self-fer- tilized for ^ve generations. The variability of two such strains and the crosses between them for a definite and easily determined character — number of rows per ear — is shown in the following table : Strain Mean Coefficient of variation A B AXB (FO BXA (Fi) AXB (FO B X A (FO 8.30±.06 14.10±.15 12.71±.15 11.77±.07 11.84±.ll 13.79±.ll 8 . 50 per cent. ± . 47 per cent. 9.66 per cent. ± .74 per cent. 10.00 per cent. ± .87 per cent. 8. 13 per cent. ± .42 per cent. 14.64 per cent, i .67 per cent. 10.62 per cent. ± .56 per cent. Clearly the F^ generation, made with either type as the mother, is as uniform as the parent strains, but the F2 generations are both more variable. To test the other corollaries, nine different self-fer- tilized families of the fifth generation were compared with families obtained by crossing two plants belonging to each family; seven families were raised as first-genera- tion hybrids between these different selfed strains; ten crosses between F^ individuals were compared with ten self-fertilizations in the same families ; and ten families were grown in which self-fertilization had been precluded for five years. The average height in decimeters, number INBREEDING EXPERIMENTS 121 of rows per ear, and yield in bushels per acre of these fifty-five families are g-iven in the following table: Selfed Fclfod X Kibs Fi F. F, Holfed Fi Sib CroMes Cro»- breds Average height . . Average No. rows Average yield 19.28 12.28 29.04 20.00 13.20 30.17 25.00 14.41 68.07 23.42 13.07 44 . 02 23.55 13.02 41.77 23.30 13.73 47.40 22 . 05 15.13 01.52 The sister-brother (sib) crosses give a slightly greater height, number of rows per ear and yield per acre than the corresponding self-fertilized families, an indication, as Shull states, of some heterozygosis still remaining in the selfed families ; in other particulars Mendclian expec- tation is wholly confirmed. The experiments of Shull on the effect of inbreeding in maize were continued only from 1905 to 1911. We may be pardoned, therefore, if we describe the experiments be- gun in 1905 by East at the Connecticut Agricultural Ex- periment Station in somewhat greater detail, for they are still being carried on by Jones. In fact, in point of num- bers and scope they are the most extensive experiments on the problem of inbreeding. The general method of procedure has been merely to self-pollinate individual plants from different varieties of all the principal types of maize. The seed from such self-fertilized plants has been grown and some plants again self-fertilized. Thus a selfed plant has been the parent of each population. Over thirty different varieties, with several lines in each variety, have been inbred in this way. The old- est strains have now been self-fertilized for twelve consecutive generations. In every case there has been a reduction in size of plant and yield of grain. Besides this result, to which 122 INBREEDING AND OUTBREEDING there has been no exception, the several inbred lines originating from the same variety have become more or less strikingly differentiated in morphological characters. Some of the differences which characterize the several inbred strains in various combinations are as follows : Colored and colorless pericarps, cobs, silks and glumes. Profusely branched tassels and scantily branched or unbranched tassels. Long ears and short ears. Round cobs and flat cob®. Narrow silks and broad silks. Ears with various numbers of rows. Ears with straight rows and ears with irregular rows. Ears with large seeds and ears with small seeds. Ears high on the stalk and ears low on the stalk. Stalks with many tillers and stalks with few tillers. Leaves with straight margin and leaves with wavy margin. Many other character differences governed by definite inherited factors have been observed, but these may serve as illustrations. Along with these normal differences a number of characters have appeared which might well be called mon- strosities, using the term not because of any abnormality in the method of their inheritance, but because they are not fitted to struggle for place either in agriculture or in nature. A common occurrence is the isolation of dwarf plants which are rarely capable of producing seed from their own pollen. Plants manifesting various degrees of chlorophyll deficiency are also frequently found. This may show in the form of an entire lack of chlorophyll, as seen in pure albino plants which live only until the supply of food in the seed is exhausted; or, it may appear as a yellowish green, the plants struggling through to seed INBREEDING EXPERBIENTS 123 production— though with some diniciilty. Some phuita are obtained with ear malformations and thus produce but a minimum amount of seed. Other phmts hick brace roots and are unable to stand upright. Still others show vaiious grades of pollen and ovule abortion, and suscepti- bility to disease. The variability of the inbred lines in respect to the above characters decreased as inbreeding was continued. After four generations they were practically constant for the grosser characters. From the eighth generation on they have been remarkably uniforai in all characters. Inbreeding the naturally cross-pollinated maize plant, then, has these results: 1. There is a reduction in size of plant and in produc- tiveness which continues only to a certain point and is in no sense an actual degeneration. 2. There is an isolation of subvarieties differing in morphological characters accompanying the reduction in growth. 3. As these subvarieties become more constant in their characters the reduction in growth ceases to be noticc^able. 4. Individuals are obtained with such characters that they cannot be reproduced or, if so, only with extreme difficulty. A large amount of data has been obtained upon which to base these statements, but since most of them have been published it seems desirable to include only a few illus- trations here. The strains which have been the longest inbred will serve to show something as to the effect which inbreeding has had upon yield of grain, lieight of ])lant and other maize characters. The original experiment began ^nth four individual 124 INBEEEDING AND OUTBEEEDING plants obtained from seed of a commercial variety grown in Illinois known as Learning Dent. This variety was given the number 1, and four plants which were self- pollinated and selected for continuation of the inbreeding experiment were numbered 1-6, 1-7, 1-9, and 1-12. These four lines were perpetuated each year by self-pollination and will be referred to hereafter as the Leaming strains. In the second inbred generation two self -pollinated plants in the 1-7 line were saved for seed and from them two inbred lines were split off which consequently came origi- nally from one line inbred two generations. These were numbered 1-7-1-1 and 1-7-1-2. Many other inbred strains coming from different material have been started from time to time and several of them are still being continued. There is no need to mention them specifically, except as they bring out special features. TABLE III The Effect of Inbreeding on the Yield and Height of Maize Year grown 1916 1905 1906 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 No. of genera- tions selfed Four inbred strains derived from a variety of Leaming dent corn 1 2 3 4 5 6 7 8 9 10 11 1-6-1-3-etc. Yield bu. per acre 74.7 88.0 59.1 95.2 57.9 80.0 27.7 41.8 78.8 25.5 32.8 46.2 Height inches 117.3 l-7-l-l-etc. Yield bu. per acre 86.7 96.0 " 97.7 103.7 74.7 88.0 60.9 1908460 63.2 25.4 Height inches 1-7-1-2-etc. 1-9-1-2-etc. 117.3 39.4 47.2 24.8 32.7 42.3 81.1 83.5 84.9 78.6 Yield bu. per acre 74.7 88.0 60.9 190759.3 1908597 68.1 41.3 Height inches 117.3 90.5 58.5 19.2 37.6 88.0 86.9 83.8 Yield 1 bu. per Height inches acre 74.7 117.3 88.0 42.3 51.7 35.4 47.7 26.0 76.5 191338.9 1914454 85.0 191621.6 1 "630.6 78.7 191^31.8 82.4 INBREEDING EXPERIMENTS 125 In Table III the yield of grain and heiglit of plant of the four inbred Learning strains are given in the suc- cessive generations of self-fertilization. In 191G seed of the original variety, which had been gro\vn in the mean- time in the locality in Illinois from whence it was originally secured, was obtained and grown for comparison with the inbred strains. This variety in Illinois in 1905 yielded at the rate of 88 bushels of shelled grain per acre and in Connecticut in 1916 at the rate of 75 bushels. There is no reason for supposing that the variety had changed to any great extent in the intervening years. Coming from Illi- nois, it was placed at a disadvantage as compared to the inbred strains, because it was not adapted to the local conditions, while the inbred strains, grown for several years, had been selected more or less unconsciously to meet the prevailing conditions. Even with this in favor of the inbred strains they yielded only from one-third to one-half as much as the original variety grown under the same conditions. With regard to the rate of reduction in yield or the constancy of the varieties during the later generations, it is difficult to draw conclusions from these figures, owing to the fluctuation in yield from year to year due to sea- sonal conditions and to the difficulty of accurate testing in field plot work, a fact recognized by all who have made such tests. No yields for any of the strains were taken in 1912. The yields for 1909 and 1915 were too low on account of poor seasons. The yields in 1914 were too high for the opposite reason. In 1915 the. yields were unreliable because only a few plants were available for calculation, most of the plants having been used for hand pollinations. 126 INBREEDING AND OUTBREEDING In 1916 and 1917 the inbred strains were grown in some- what larger plots and the yields are fairly reliable. With these points in mind, an examination of the table shows that from the beginning of the experiment to the ninth generation there has been a tremendous drop in productiveness, so that in that generation the strains were approximately only one-third as productive as the variety before inbreeding. From the ninth to the eleventh gen- eration there has been no reduction in yield and prac- tically no change in visible characters. Height of plant, as far as the available figures show, followed the same course. The reduction which has taken place occurred in the first eight generations ; after that there has been no appreciable change. All along the several Learning strains have shown considerable difiFerences in productiveness and in height. Strain No. 1-6 has given the largest yields and the tallest plants. It gave nearly 50 per cent, larger yields than the poorest yielding strain in the eleventh year, and was about 30 per cent, higher than the shortest strain. One of the strains, No. 1-12, was lost in the sixth gen- eration. Previous to this time it had been the poorest of the five. It was partially sterile, never produced seed at the tip of the ear and was perpetuated only with care. Since the difficulty of carrying along any inbred strain is great, owing to failure to pollinate at the correct time, to attacks of fungus on the ears enclosed in paper bags, and to poor germination in the cold, wet weather common in New England at com-planting time, the loss of this strain might be easily accounted for without assuming continu- ous deterioration. The strain probably could have been retained if sufficient effort had been put forth; but in INBREEDING EXPERIMENTS 127 view of the further reduction in other strains, it would have been extremely diliicult. {Since plants are frequently produced which cannot be perpetuated, however, it is to be expected that some strains will also be found which cannot survive. This is good evidence that strains, differing markedly in their ability to grow, are isolated by inbreeding. Plants of the surviving strains, while smaller in size and lower in productiveness, are perfectly healthy and functionally normal in every way except that in many of them there is an extreme reduction in the amount of pollen produced. These infertile types are dependent on other plants for pollen in order to make the yields they show in open field culture ; when grown by themselves the yield is less due to an inadequate supply of pollen. On the other hand, this extreme reduction in pollen production is not shown by all the strains, some inbred strains producing pollen abundantly. Prom the data given in Table III there is considerable evidence that these plants have reached about the limit of their reduction in size and productiveness and tluit whatever changes have taken place in the last three years have been slight. Further inbreeding is necessary for one to be positive on this point. But as the crosses within these inbred strains have given no significant increases over the selfed lines, and as there has been no visible change in morphological characters, in the past three years at least, it seems apparent tliat the reduction in vegetative vigor and productiveness is ver>' nearly, if not quite, at an end. Reduction is shown by inbred maize plants in other characters. Length of ear, as well as height of plant and 128 INBREEDING AND OUTBREEDING yield of grain, is smaller. There is also a slight reduction in number of nodes and in rows of grain, but in contrast to the other three characters the change is almost negli- gible. The last two are only slightly affected by environ- mental factors as compared with the others. A plant may be reduced to one-half its normal height by being grown in a poor situation, but the number of nodes will be nearly the same in the two cases. Hence, we see that inbreeding affects plants much in the same way as poor environmen- tal conditions. In all of the characters mentioned there is a reduction in variability and change in mean differing in the several lines. This is illustrated in Table IV, in which are given the data for number of rows of grain on the ear of four different plots of the origiual non-inbred variety and four strains derived from this variety after ten genera- tions of self-fertilization. The marked reduction in vari- ability is apparent both in the restricted range of the distribution of the inbred lines compared to the variety, and in the coefficients of variability. This reduction in variability applies only to each in- bred line separately. If all the different lines were com- bined together into one population the variation would be greater than that shown by the original material. This is readily apparent from the table ; it also follows from the fact that many characters are produced by inbreed- ing which are seldom seen in the regularly cross- pollinated stock. Inbreeding reduces variability within separate lines, but increases variability in the descend- ants as a whole. From the curves on inbreeding given in the preceding chapter (Fig. 24), it was seen that the production of com- INBREEDING EXPERIMENTS 129 I «0' I I to to ^^ I I I I I I CO 4^ •-' to CC o >*^ CO oo CtiOi CO Cii H-^f I I I I I I I I I I 1 I a a d o I I I I I I to to I I to to t I CO CO I I to to I I I I I I I I I t-t>-i t~A^^ OOCOCOM ^-.f? II \ T I I I I t-* 1-^ (-* »— • 4^ »i^ ik. »^ I I I I J J ' I d C;i Cn en 4k. 4^ »L 4ik 4^ 4^ 4k. 4k. 4» *k. 4k- I I CO H-' to to I 4k. I I I I to to to to I I Cn -^ I— i Ci ■ I I I to to »— ' »-• III CO a« >— ' 4k. I I I t— toco I i- to to to to oto 00 OGO to • t-- 05 Cn COO) 4^Cn t-* to •-'CO •--CO 4^0 H-* »-' CO to t-'COtO OiCn coco ^ • »— ' to Cn >— » t— « I— ' H- cn -vi en 4' -^ •—•(-» to ►-* Ol -^ CO ►-' 00 Oi I— » t-» t— ' toco -J 00 4k- to t-» Oi M CO C;i toco to 4i. to to 4^4»' toco COOi Cn Oi Oi tOOi to Oi 4^ Cn Oi Ci Oi 00 CO to to Cn CT> Cn Cn •M CD O t-" -vi (-« H-* »-» H-» to to to to t-" t-i »-» l-» Cncn CnOi Ot--»tOt— 4^.4^.0105 coco H-K- ►"^ 00000 K-H-H-H- CO 4k -^CO H-ht-H-H- .oo.cKja.iais 4*. t0 4i. tO^ hfH-K-tt-H- OOCO OiCn tO»— 'to*— ' t— 'I— >Ot— ' OO^OOi OOC04*. H- to to to to t0 4* CO ^ W OiCO coco 00-4000 MMOiCO 4kOi4^t04^ Oto 00 H- H-H- OiO H-K- CO 4i- I-' 4k- 4'OOOCO hf-H-tt-H- co ro 4^-00 Oi CO 00 CO tO H- 4^ <-• to Oi -^ to to K-H-hf H-hf COO^ .OiO> ->44^0iCn 00 CO Oi 00 i-» CO 4*- H- Oi Oi 4»- Ci O I— ' O I— ' 4k. CO CO 00 CO 01 C7i CO 4- ^- c I o Vj ro o to ts) to c a c 3) CO o D <-► or n n p •^ < o n H CO o O l^ S3 N W > q "A K o ta M a" d o 50 "i ^^ o wa t o o^ M « S3 ^ D H "^ 2 » M ^ t-H o 55 £2 n :^ o 9 130 INBREEDINa A^D OUTBREEDING pletely homozygous types by self-fertilization is greatest in the generations from the third to the sixth if a large number of factor differences are involved at the start. The experimental results obtained from these inbred strains of maize fit this theory well. It is not until after about three generations of self-fertilization that extreme types begin to appear. While there has been a reduction in size and productiveness before this, it is at this time, or during the next two or three generations, that the greatest diversity of types occurs. It is here that most of the monstrosities and plants which are unable to re- produce themselves appear. From Table IV we see that equally striking changes in the mean row number also take place. The averages have been shifted both up and down from the original conditions. The greatest segregation has taken place between the first and the eighth generations. In the eighth generation the lines were again split up, but show no marked change after this point. Differences ia the ears of these inbred strains of com are shown in Fig. 29. The rate of reduction in variability and rate of change of mean are shown by the data for row number of two of the inbred strains for successive years in Table V and Fig. 30. These two Hues are descended from the same plant in the second year of self-fertilization. The figures previous to the third year are not available, and in that year only for one of the strains, but since then a marked change in average row number, and a reduction in vari- ability have taken place without conscious selection one way or the other. Though the number of plants grown in the generations from the 7th to the 10th are too few to be a basis for accurate conclusions, the sharp increase in ";AS ^v :-y^ y . ."^^ m K "N • Fio. 29. — Representative samples of inbred strains of maize after 11 tifncratmn.s OiO^ rf^ 4k. CO CO I I I I I I I I I I I I I I I I •vIM -4^ I I I i II II I I tot-- I I tot-- I I CO 4^ I I tOM I I I I t— ' 4». I I I I CO to I I to>-» I I tot-* I I CO *«. I I to^ I I 1—1 en t-i )U' I I 4>-Ot I I tOH-i I I tot-- I CO I to 4»- I — I I tOH- '1 I I to^ I I tOH- I I tot- CO CO • to- I--- CO- to • o> 4»> to t-* . CnCO bOH-* Oi' 4k- to to I-* to too 00 o — to 00 o to to c o 00 CO WW cn to H-H- »-to COCi to to Cn to 00 O M to coco 4^ to to o O ^ WW WW 4k 00 CO Cn 00 ^J 05 00 cn CO H- co o H- m C « r "^ ft H •» • ■ It •T3 5 B C B cr CO to CO 4*> 4>> 3> C5 OC to •4 O ^ K5 >^ M ■^ »-* to g 5? c 3 cr 2. 5' o a > (9 P < 5 H > o n 9 93 2 CD o t M 132 INBREEDINa AND OUTBREEDING average row number and decrease in variability in the eighth generation are probably due to the favorable grow- ing conditions of that year — ^witness the high yields for the inbred strains in that year as given in Table III. The apparent rise in variability after the eighth generation is in part due to the fact that the ears had become some- 4» C V ■H O t %4 tt) n 20 , 15 . U o I o O rt C o Ave. No. RoVS of 1-7-1-1 Ave . No . Rows of 1-7-1-2 Ave.C.V. — T— z 4 S 6 7 8 9 10 U Generations Intred Fia. 30. — Graphs showing reduction of variability and segregation of ear row number in selfed strains of maize. what more irregular in row number, so that accurate de- termination of number of rows has been more difficult in the later generations. However, this rise is more appar- ent than real as the values for the coefficients of variabil- ity in the intermediate generations are probably lower than they would have been if an adequate number of plants had been grown. The effect of inbreeding upon variability is even more INBREEDING EXPERIMENTS 133 apparent in details of plant and oar structure which are difficult of statistical expression. The beautiful uni- foimity of these plants in all characteristics at the pres- ent time is one of their most striking features. This can be seen fairly well for the ear chanicters in accompany- ing illustrations (Eig. 29). In the minutiie of the tassels, leaves and stalks they show the same striking uniformity. These minor details which characterize each of these groups of plants are difficult to describe adequat(^ly, but are perhaps the most noticeable feature about them. The tassels or the ears of all these four Learning strains, if mixed together, could be separated without the slightest difficulty. Some characters appear so rarely in plants they have been generally considered to be due to what might be called physiological accidents rather than to inheritance. An illustration of this kind is furnished in maize by the occurrence of doubled or connate seeds. Instead of one embryo enclosed in a pericarp, separate embryos and endosperms are present, with the seeds arranged back to back and the embryos facing in opposite directions. A few seeds of this kind have been described from time to time, but never more than one or two on an occasional ear. From twelve inbred strains of a variety of maize other than the ones previously described, two linos have been obtained which produce these peculiar seeds as a common feature. One of the strains show^s from one to six or more on practically every ear. The second strain shows them more rarely and the other ten strains derived from the same variety have never been obsor\'ed to boar them. Here, then, is a character which does not appear except at rare intervals when the plants are crossed 134 INBREEDING AND OUTBREEDING and in full vigor. When the plants are brought to homozygosity and the vigor of the plants is reduced, the doubled seeds appear in abundance in some lines, but not in all. A character, then, may be governed in its ex- pression by other characters and modified by the vigor of the plant, but in the final analysis it is dependent upon definitely inherited factors. In the same way such indefinite and complex charac- ters as susceptibility and resistance to disease are shown to be capable of segregation. In 1917 one of the inbred Learning strains had not a siugle plant affected by the smut fungus, although 1000 plants were grown in differ- ent places. Other strains derived from the same variety and grown side by side with the susceptible race showed from 5 to 10 per cent, of plants infected. Susceptibility of maize to smut thus seems to be dependent upon in- herited factors. As the result of inbreeding, these fac- tors may be segregated tato some lines and not into others. Although there has been a striking reduction in size of plant, general vegetative vigor and productiveness, and in comparison with non-iribred varieties the inbred plants are more dififtcult to grow, emphasis must be put upon the fact that they are normal and healthy. No actual degen- eration has occurred. The monstrosities which are com- mon in every field of maize, such as the occurrence of seeds in the tassels, anthers in the ears, dwarf plants, completely sterile plants, and other similar anomalies, now no longer appear in these inbred strains. These facts, taken together, should be sufficient to demonstrate beyond doubt that by far the greatest amount of the general vari- ability found among ordinary cross-fertilized plants is due to the segregation and recombination of definite and INBREEDING EXPERIMENTS 135 constant hereditary factors. Some of the characters which appear after long-continued inbreeding are seldom seen in continually cross-pollinated plants, and never are so many seen in combination. This is because they are recessive in nature and complex in mode of inheritance The most significant feature about the cliaracters wliirh make their appearance in inbred plants is that none of them can be attributed directly to a loss of a physiological stimulation, although undoubtedly many of them may be modified by the vigor of the plants upon which they are borne. There is no one specific feature common to all inbred strains, but simply a general loss of vigor, a gen- eral reduction in size and productiveness accompanied by specific characters more or less unfavorable to the plant's best development. But these unfavorable char- acters are never all found in one inbred strain, nor is any one of them found in all inbred strains. Although no systematic selection has been practiccnl throughout these inbreeding experiments, a great deal of selection upon many characters has been unavoidable as is the case in any inbreeding experiment. In maize, the difficulties of hand pollination result in the selection of plants whose staminate and pistillate parts are matured synchronously. Any great difference in this respect, par- ticularly towards protandry, renders self-fertilization difficult or impossible as the pollen is viable but a short time. Of course, all plants wiiich are weak, steril«\ dis- eased or in any way abnormal, tend to become eliminated wherever these causes reduce the chance of obtainincr seed. This unconscious selection becomes more ricrid in the lat«^r generations of inbreedin.s: as reduction in vicror and pro- ductiveness becomes more pronounced. Acrain, the small 136 INBREEDING AND OUTBREEDING amount of seed produced by hand pollination under the most favorable circumstances, necessitates the using of the best ears obtained for planting in order to have enough plants upon which to make any fair observations. These factors tend to prevent the attainment of com- plete homozygosity. Nevertheless, all the evidence at hand indicates that the four strains of Leaming com which have been continuously self-fertilized for twelve generations are now very nearly, if not completely, homozygous in all inherited characters. As stated before, this evidence com- prises cessation of reduction in size and productiveness, of reduction in variability, and of change of average row number and other characters. But there are still other ways of testing the proposition. On the theory that in- crease in growth results from crossing when the individu- als united differ in respect to some inherited qualities, if no increase results, then the parents have no differences. These strains have been tested in this way by crossing different plants within a strain and comparing the crossed plants with self ed plants. While some increases in growth resulted from such crossing they were balanced by de- creases in other cases, so that the inconsistencies are most likely due to difficulty in securing an accurate test. At the same time one should not shut his eyes to the possi- bility that some of the strains have reached complete homozygosity, while others, as yet, have not ; although no sure evidence of such a state of affairs has been obtained. Most of the direct experimentation to determine the effects of inbreeding has been with cultivated plants and domestic animals. The question will undoubtedly be asked, therefore, as to whether the results would have been the same had wild species been investigated. It would be INBREEDING EXPERIMENTS 137 futile to maintain that there is every reason to supjwse wild species should behave exactly as their domestic cousins. Wild types, in general, nii.irht not present such an appearance of injur>^ under inbreeding as is often sho^vn by cultivated species. This would not be due to differences in their method of inheritance, however, but because wild species are usually exposed to a more rigor- ous struggle for existence and tlic indivi(Uials are, there- fore, less likely to differ by a large number of hereditary factors. For such reason one should expect experiments on different wild species to give rather varied results, and in the comparatively small number which have ])e('n made this is the case. Castle's experiments on the fniit fly pave no markedly unfavorable results. Collins states that self- fertilizing teosinte, a semi-wild relative of maize, causes no loss of vigor. Yet Darwin compared self-fertilizod and intercrossed plants of several species which are largely cross-fertilized in the wild with great disadvantage to the former. This discussion of the effects of artificial inbreeding in certain plants and animals has been given in some de- tail in order to bring out the many important considera- tions involved. There has even been repetition in order to emphasize the most important points. Details are merely by way of parenthesis, however. Let us now get out of the parenthesis and into the main ar^irument. From the preceding ohservatioyi it can he said that inbreeding has hut one demonstrable effect on organisms subjected to its actioii—the isolation of homozygnu^ types. The diversity of the resultmg types depends di- rectly upon number of heterozygous hereditary factors present in the individuals with which the process is he- 138 INBEEEDING AND OUTBEEEDING gun; it is likely, therefore, to vary directly with the amount of cross-breeding experienced by their immediate ancestors. The rapidity of the isolation of homozygous types is a function of the intensity of the inbreeding. Take the case of maize as an example. Maize is one of the most variable of cultivated plants, and is usually cross-pollinated under natural conditions. In other words, the individuals making up any commercial variety of maize are each and every one heterozygous for a large number of hereditary factors — a heterozygosis that is kept up by continual crossing and recrossing. "When such a variety is inbred there is automatic isolation of homo- zygous combinations, following simple mathematical laws as we have already seen. If self-fertilization is practiced, stabilization through an approximately complete homo- zygosis occurs after a relatively small number of genera- tions ; if a less intense system of inbreeding is followed, the result is the same, but it is obtained more slowly. Dur- ing thisf process, before stabilization is reached, there is reduction in size, vigor and productiveness following somewhat roughly the reduction in per cent, of hetero- zygousness. We can think of this reduction in vigor as a change correlated with approaching homozygosis if we wish, although as we shall see there is reason to believe it to be a result of linked inheritance. What does occur is a reduction in vigor of the population as a whole in each generation associated with the isolation of individuals more homozygous than their parents. Any particular in- dividual may be vigorous or weak, fertile or sterile, nor- mal or monstrous, good, bad or indifferent, depending wholly upon the combination of characters received, many of the characters which become homozygous will be INBREEDING EXPERIMENTS 139 recessives or combinations of recessives which seldom are seen under ordinary circmnstances, because they are hid- den by their dominant allelomorphs. These recessives are the ''corrupt i'mit" which give the bad name to inbreeding, for they are often— very often— undesir- able characteristics. The homozygous inbred strains after stability has been reached are quite comparable to naturally self -fer- tilizing species provided they have passed as rigorous selection as the latter have had to undergo by reason of natural competition. And Darwin, as well as others, found that artificial self-pollination causes no reduction in such genera as Nicotiana, Pisum and Phaseolus where self-fertilization is the general rule. Are then the immediate results of inbreeding some- times injurious? In naturally cross-fertilized organisms they most emphatically are — nay, more, even disastrous — when we recall the reduction to over half or one-third in production in grain and a corresponding decrease in size of plant and rate of growth in maize. But maize is prob- ably an extreme case. With other organisms the results are not so bad, and in some cases, especially when selec- tion has been made, no evil effects are apj)arent. In fact, there may be an actual improvement. But the truth is, we did not set out to answer that question. It had already received a correct answer. What ive undertook to inquire was whether inbreeding is inj^irious merely by reason of the consangmnity. We answer, No/ The only injury proceeding from inbreeding comes from the inlieritance received. The constitution of the individuals result ins< from a process of inbreedine: depends upon tlio chance allotment of characters preexisting in the stock before in- 140 INBREEDING AND OUTBREEDING breeding was commenced. If undesirable cbaracters are shown after inbreeding, it is only because they already existed in the stock and were able to persist for genera- tions under the protection of more favorable characters which dominated them and kept them from sight. The powerful hand of natural selection was thus stayed until inbreeding tore aside the mask and the unfavorable char- acters were shown up in all their weakness, to stand or fall on their own merits. If evil is brought to light, inbreeding is no more to be blamed than the detective who unearths a crime. Instead of being condemned it should be commended. After con- tinued inbreeding a cross-bred stock has been purified and rid of abnormalities, monstrosities, and serious weak- nesses of all kinds. Only those characters can remain which either are favorable or at least are not definitely harmful to the organism. Those characters which have survived this **day of judgment'' can now be estimated according to their true worth. As we shall see later vigor can be immediately regained by crossing. Not only is the full vigor of the original steck restored, but it may even be increased, due to the elimination of many unfav- orable characters. If this increased vigor can be utilized in the first generation, or if it can be fixed so that it is not lost in succeeding generations, then inbreeding is not only not injurious but is highly beneficial. As an actual means of plant and animal improvement, therefore, it should be given its rightful valuation. CHAPTER VII HYBKID VIGOR OR HETEROSIS Whether or not inbreeding in a race of plants or ani- mals results injuriously depends primarily, as we have attempted to show, upon the hereditary constitution of the organism. The beneficial effect of crossing, heterosis, is a more widespread phenomenon. It may be expected when almost all somewhat nearly rehited foniis are crossed together. Even plants or animals which show no harmful results of inbreeding are frequently improved thus in a remarkable way. Moreover, this stimulating effect is immediately apparent in the individuals result- ing from the cross. It is then at its maximum. It is natural, therefore, that the early writers on the subject should have noticed and emphasized the good to be derived from crossing rather than the bad which some- times results from inbreeding. Almost without exception the great horticultural writers of the late eighteenth and early nineteenth centuries noted the occurrence of hybrid vigor, and many of them described it in great detail. Among them may be mentioned Kolreuter (1763), Knight (1799), Mauz (1825), Sageret (1826), Borthollot (1827), Wiegmann (1828), Herbert (1837), Lecoq (1845), Giirt- ner (1849). In fact, in Focke^s compilation of this early work, **Die Pflanzen-Mischlinge*' (1881), cases of heter- osis worthy of special mention w^ere found in fifty-nine families of the flowering plants as well as in the conifers and the ferns. Animal husbandmen were somewhat loss 141 142 INBREEDING AND OUTBREEDING inclined to acknowledge and discuss tlie matter, although they had an excellent example before them in the mule — an animal known and appreciated for over four thousand years. But the necessity of their following the custom of maintaining breeds true to certain fixed standards prob- ably accounts for their conservatism in estimating the importance of the phenomenon. K61reuter,i25 the first botanist to study artificial plant hybrids, made many interspecific crosses in the genera Nicotiana, Dianthus, Verbascum, Mirabilis, Da- tura and others, which astonished their producer by their greater size, increased number of flowers and general vegetative vigor, as compared with the parental species entering into the cross. He -gives many exact measure- ments of his hybrids and speaks with some awe of their ** statura portentosa'* and '^ambitus vastissimus ac alti- tudo valde conspicua," Later, after some observations on certain structural adaptations for cross-pollination which he interpreted correctly, he made a passing re- mark which plainly showed he thought Nature had intended plants to be cross-fertilized and that benefit ensued therefrom. Some forty years after, Thomas Andrew Knight,^22 ^ horticulturist who was a very keen observer, noticed sim- ilar instances of high vigor in his crosses : in the descrip- tion of these experiments we note the following remarks concerning a cross between two varieties of peas : By introducing the farina of the largest and most luxuriant kinds into the blossoms of the most diminutive and by reversing the process I found that the powers of the male and female in their effects on the offspring are exactly equal. The vigor of the growth, the size of the seeds produced, and the season of maturity, were the same, though the one was a very early, and the other a very late variety. I had, in this HYBRID VIGOR OR HETEROSIS 143 experiment, a striking instance of the stimulating effects of cro6tiing the breeds; for the smallest variety, whose height rarely exceeded two feet, was increased to six feet, whilst the height of the large and liLxunaul kind was very little diminLshed. It is evideut that iu tiiLii particular cajse Krnglil was dealing with dwarf and standard peas, and dominance of the tali standard habit of growth is to be expected. This is not the correct interpretation of the majority of his ob- servations on hybrid vigor, however; a sulhcient number of really striking manifestations of the phenomenon were found to give adequate foundation for his anti-inbreeding principle, elaborated by Darwin lifty years later. Probably the most extensive series of early experi- ments on hybridization were those of Gartner."^^ This enthusiastic worker crossed, or attempted to cross, every- thing available to him. According to Lindley,^-** he made 10,000 pollinations between 700 species, and produced 1250 different hybrids. Many of his attempted crosses either failed to produce seed, or if seed was produced, gave feeble plants; but a great number of the hybrids, where the crosses were made between plants not too distantly re- lated, showed distinct evidence of hybrid vigor mani- fested in many different ways. Gartner speaks especially of their general vegetative luxuriance, increase in root development, height, number of flowers, the facility of their vegetative propagation, their hardiness and early and prolonged blooming. He says : One of the most conspicnons and common eliaracteristics of plant hybrids is the luxuriance of all their parts, a luxuriance that is shown in the rankness of their growth and a prodisral development of root shoota, branches, leaves, and blossoms that could not be induced in the parent stocks by the most careful cultivation. The hybrids usually reach the full development of their parts only when plante-} ' relation to the phenomenon of hybrid vigor, for such individuals may be alike in constitution. Danvin's repeated emphasis of the good derived from crossing plants whose ancestors were exposed to different conditions was because he thought such differences in environment brought about germinal changes. This attitude, therefore, does not detract from his general position tliat it is differences in germinal construction which bring about liybrid vigor; and this is the principal point at issue. 150 INBREEDING AND OUTBREEDING Tlie manifestations of heterosis are most noticeable as increases in size. This gain in size in plants which are more or less determinate in their number of parts is made up of an increase in the size of parts rather than in the number of parts. In maize the number of nodes is in- creased much less in comparison to length of internodes. For example, in a large series of crosses between inbred strains of maize height of plant on the average advanced 27 per cent., whereas the number of nodes rose only 6 per cent. Corresponding to the increase in intemode length there is an extension in diameter of stalk, length and breadth of leaves. Root development is proportion- ally augmented. Both the tassels and ears are larger, and frequently two ears develop on crossed plants where either parent produces one, the color of the foliage tes- tifying to the greater vigor. The greatly enhanced growth of a plant may be made up by increase in the size of cells, as well as by a multi- plication in the number of ceUs. However, in a cross between different species of Catalpa no differences could be seen in tracheid length, although the cross was con- siderably taller and larger in diameter. The principal effect of crossing maize is shown by an additional production of seed. A number of crosses have given 180 per cent, increases in yield of grain over their inbred parents. Examples of what can be done are seen in the accompanying illustrations (Figs. 31 and 32). Im- provement in yield is shown by crosses between inbred strains derived originally from the same variety, as well as between crosses of strains derived from different vari- eties or even from quite distinct types. The results have been very wonderful as a whole, giving at the very least o 3 P L^.r; ^>W^^': 2.T O £3 00 to 01 N — y r2 ►t-t; - o 3 7) O) o re HYBRID VIGOR OR HETEROSIS 151 a return to the condition of the original stock before in- breeding was commenced. Some combinations regularly give greater increases than others, but in every case such differences are small as compared with those between the crosses and the inbred parents. Although, in the main, reciprocal crosses give about the same result, some variation in this respect is habitu- ally shown. In general, there is a correlation between the yield of the better parent strain and the yield of the cross. The crosses in which strain No. l-f) has been used as the female parent have regularly given the highest yields, and this strain is the most vigorous and productive of the four inbred Leaming strains used in our illustrations. In a comparison of crosses between inbred strains of maize with ordinary outcrossed varieties the inbred hy- brids are handicapped because they have to start from small, poorly developed seeds. This handicap is brought out clearly by a comparison of second generation plants grown from self -fertilized seed produced on vigorous hybrid plants, with hybrid plants grown from seed pro- duced on inbred plants. The first generation starts off poorly, as shown in the accompanying illustration (Fiir. 33), but soon catches up and passes the second generation. At maturity the second generation is shorter and loss productive, although it has a much greater variability. The third generation from selfed plants of this particular cross has been grown, and there is still further loss of the stimulation which is at its maximum in the first gonora- tion. On continued inbreeding these families presumably would exhibit a continuation of the same course of reduc- tion in size, vigor and variability shown in the original inbreeding experiment, until homoz^-gosify was acrain 152 INBREEDING AND OUTBREEDING 100 Growth Curves of Two In"bred Strains of Maize and Their Hybrids. 75 CO « o C CO (D 50 25 P (1-9) (1-7) 30 40 T 60 I 70 80 50 Numljer of Days from Planting 90 100 33. — Graphs showing growth curves of two inbred strains. of maize and thier first and eeoond generation hybrids. HYBRID VIGOR OR HETEROSIS 153 reached. The resulting inbred strains would have about the same amount of development as the original inbred strains, but would probably differ from them in appear ance through the possession of diflerent combinations of characters. The principal point is that the vigor and size lost by inbreeding are immediately restored by crossing, but lost again on further inbreeding. It is a transit^)ry effect, for the most part, impossil)le of fixation. Increases in yield of grain are also fre^juently ob- tained when ordinary commercial varieties of maize are crossed. Rarely are the increases greater than 10 per cent., however, and even this is more commonly to be expected when varieties of somewhat different iypa are used; for example, flint and dent. Most varieties of corn are now so widely crossed and furthermore are so near the limit of production that great advances are not to be expected. Collins ^^ has obtained especially large incre- ments in >deld by hybridizing typos of com from different geographical regions. Three different varieties of corn from the southwest — Hopi, Brownsville and Hairy Mexi- can — each gave an increase of 100 per cent, or more when crossed with a variety from China having seeds with a different type of endosperm. Even before the plants are obtained there is a striking effect of crossing in an immediate increase in the size of seed. This was noted by Roberts, ^^''^ and established very clearly by Collins and Kempton ^ through pollinating ears of maize with a mixture of the plants own pollen and of a different sort. By taking advantage of the phenom- enon of double or ** endosperm** fertilization, the experiment was so designed that the ontrTossed seeds could be distinguished by differences in endosponn color. 154 INBEEEDING AND OUTBREEDING Advances in average weight of seed ranging from 3 to 21 per cent, were obtained. With inbred strains as parents, the increases are even greater, ranging from 5 to 35 per cent. The seeds have a heavier embryo as well as a heavier endosperm, yet curiously enough they mature faster than the selfed seeds on the same ears. It is a point of some interest, perhaps, that there is no selective action favoring the foreign pollen when these pollen mixtures are applied. This matter has been deter- mined very carefully on account of its bearing on Men- delian theory, but it also answers in the negative the question of whether there is an effect of heterosis manifested by a selective chemotropism before the zygote is formed. Darwin, in ^' Cross and Self -Fertilization in the Vege- table Kingdom, '' compares the time of flowering of 28 crosses between different types of plants which had shown distinct evidence of hybrid vigor. Of them, 81 per cent, flowered before the parents. In other cases, where no heterosis was shown in other characters there was no ac- celeration of the blooming period. These results have been corroborated in crosses between garden varieties of tomatoes and of sweet corn, where a tendency to put for- ward the time of both flowering and maturing has been shown to accompany increases in size. Shortening the time of growth thus seems to be one of the many ex- pressions of an increased metabolic efSciency on the part of the hybrid plant. Increased longevity, viability, endurance against un- favorable climatic conditions, and resistance to disease have also been frequently noted as properties of hybrids. Kolreuter ^^^ and Wiegmann "^^ both mention these points. bt Fig. 34. — James River ^\'alIlUt, a faiijnii.s tree cuiisidcriMl to iu- a n tween the Persian walnut and tlic coninion liuttcmnt. .\c• •r HYBRID VIGOR OR HETEROSIS 155 and Gartner "^^ gives them bis especial attention. Under the heading, "Ausdauer und Lebenstouacitiit der P^as- tardpflanzen, ' * he makes the following statements: There is certainly no essential difference between annual and biennial plants and between these and perennials in regard to their longevity, for frequently different individuals oi' the same species have a longer life at times as, for example, Draba venia which has bolh annual and biennial forms. The longevity of a plant thei-eby f umi^hea no specific difference but at most only signitictj a variabihty. ILowever, in hybrids this difference deserves special consideration. In mo^t hybrids an increased longevity and greater endurance can bo obeerved as compared to their parental races even if they come into bloom a year earlier. The union of an annual, herbaceous female plant uith a perennial, shi-ubby species does not shorten the life cycle of the forth- coming hybrid, as the union of Hyoscyanms agrestis with niger, Nicotiunn rustica "with perennis, Calceolaria plantaginea with rugosa shows. So also in reciprocal crosses w^hen the perennial species furnishes the seed and the annual species supplies the pollen, as Nicotiana glauca v^ilh Langsdorfjiij Dianthus caryophyllus with chitiensis, Malva sylcestns with mauritiana or biennials with perennials and reciprocally, as Digitalis purpurea with ochroleuca or lutea, and lutea with purpurea, or ochroleuca with purpurea. From, the union of two races of different longevity a hybrid usually results into Avhich the longer life of one or the other of its parent races is carried whether it comes from the male or female parent species. Many more instances are given by Gartner supporting the conclusion previously reached by Kolr enter that the longer life of hybrid plants is to be counted among their usual properties. Gartner also gives several examples of endurance to unfavorable weather conditions by hybrids. Many of his tobacco hybrids actually survived the winters in the open field in south Germany when the parents were killed. The hardiness of hybrids is frequently shown by a great resistance to parasitism. Gornert'^'^ states that teo- sinte and the first generation cross of teosinte and maize 156 INBREEDING AND OUTBREEDING are not attacked by the apMds which damage maize. The cross between the inbred strain of maize most susceptible to smut, previously mentioned, and the strain not affected, gives a hybrid which is only slightly parasitized. The same thing has been noted in crosses between other strains of maize, some of which are quite badly damaged by an unidentified leaf blight organism. Radish seedlings which were naturally cross-fertilized were much less dam- aged by damping-off fungus than uncrossed seedlings from the same plants. Resistance is not shown by all first generation hybrids when the parents diif er in suscep- tibility. Some cases are known in which the hybrids are fully as susceptible as the less immune parent. In the majority of crosses reported, however, in which re- sistance to parasitism is a factor the hybrids tend to show resistance. Among the diverse manifestations accompanying heterozygosity may be mentioned viability of seed. In maize, crossed and selfed seeds from the same ears have shown a difference of 16 per cent, germination in favor of the crossed seeds. The crossed seedlings appeared earlier and grew faster from the first. Increased facility of vegetative propagation of hy- brids was frequently noted by the early hybridizers. Sageret ^^^ makes particular note of a hybrid tobacco which easily propagated itself vegetatively. Many of our cultivated fruits which are propagated by buds, grafts, cuttings, etc., owe part of their excellence at least to the fact that they are in a heterozygous condition. Moreover, there is no evidence to prove that plants lose any of their hybrid vigor in long continued vegetative multiplication through innumerable generations. HYBRID VIGOR OR HETEROSIS 157 In general, as noted before, there is similarity between the effect of heterozygosis and that of a good environ- ment. Those characters which are quickest to be modi- fied by external factors also show the greatest change on crossing. A good illustration of this is a Nicotiana cross which was ahove the average of the parents in botli height and leaf size. The length of the corolhis, on tlio other hand, a character very slightly affected by the en- vironment, was not increased. There is at least one difference between the two, however; in time of maturity, environment and heterosis have somewhat opposite ef- fects. Generally speaking, favorable growing con<]iti()ns tend to delay flowering and maturing, whereas conditions which stunt the plants tend, like heterosis, to hasten them. Each of these effects is by no means always present when, a cross is made. The usual and of course the most noticeable effect is the increase in size. But crossing may have a stimulating effect upon certain part^ and a de- pressing effect on others. This is shown in many species crosses in which reproductive ability is greatly roducod or even totally eliminated, while at the same time vegeta- tive growth is enormously increased. Freeman and Sax have independently obtained seeds from crosses between common bread wheats and macaroni wdieats which were shrunken in appearance and small in size, owning to a poor development of the endosperm. The embryos were well developed, however, and the plants produced gave dis- tinct evidence of hybrid vigor. In this discussion there has been a noticeable omission of the effect of crossing on animals. Illustrations are not lacking that crossing frequently is highly ])onori('ial to animals; but animals do not funiish as desirable research 158 INBEEEDING AND OUTBREEDING material for this particular problem as plants, on account of their bisexuality, as was explained earlier, and for this reason but few quantitative data are available. There is no question but that animals behave the same as plants in heredity ; therefore, one might transfer the conclusions reached in the one kingdom to the other without apology, for the effects of inbreeding and cross-breeding are wholly and solely the working out of the laws of heredity. At the same time, it will not be amiss to present some of the results obtained by zoologists, for they strengthen the case immensely. In the two cultivated species of insects which form our sole instances of domestication, bees and silkworms, there seems to be evidence of increased vigor on crossing only in silkworms (Toyama ^^^), In the fruit fly, however, upon which the greatest amount of genetic work has been done, Castle,2i Moenkhaus,^*^ Hyde^^ ^nd Muller^^* all found size, fecundity and general constitutional vigor increased remarkably, particularly when the strains crossed had been inbred previously. In the rotifer, Hydatina, Whit- ney 216 and A. F. Shull ^®® obtained similar results. Fur- ther, Gerschler^® describes and figures first generation crosses between different genera of fishes which show very marked increases in size. In birds also there is such an increase in vigor that poultry fanciers often cross two distinct strains and sell the progeny because of their rapid growth and large size. No attempt is made to breed from the hybrids ; they are simply produced because of their vigor. When very great differences in size exist, there is not, of course, an increase in size sufficient to throw the individual of the first hybrid generation above the larger parent, as is shown by the HYBRID VIGOR OR HETEROSIS 159 work of Pliillips ^•*^ on crosses between the large Freiicli Rouen and the small domestic Mallard duck, and by the work of Punnett ^^^ on crosses betwen the iSilver Sebright bantam and Gold-pencilled Hamburgh breeds of poultry. There is an increase over the average of the two parents, but the i*'i's do not reach the size of the larger parent race. Part of the reason for the comparatively small sizes of the i^\'s in these crosses, however, is due to the fact that the crosses were always made on the i>mall hens allowing the hybrid birds to get their start in life with only the nutriment stored in the smaller eggs. The greatest amount of data on this subject, just aa there is the greatest amount of interest, has been obtained from the mammals. In the meat breeds of cattle, swine and sheep, as in poultry, it is a common practice to cross distinct races and sell the progeny. The increase in size and the rapidity with which this size is obUiined are so general a phenomenon that it bids fair partially to replace the older method of pure line breeding. Not only are varietal crosses thus characterized, but speciiic crosses. We have already mentioned the mule. With the disad- vantage attached to sterility, the mule certainly would not have held its own throughout the past forty centuries were it not for its tremendous capacity for work and its remarkable resistance to disease. Crosses between the ass and the zebra, and between the cow jmd the zebu also give animals of considerable merit, and one can hardly refrain from thinking that within a few years some con- siderable use ^\dll be made of them. For precise data on the effect of cro.^sing ditTeront II races, however, we must turn to the small maninmls used so constantly in experimental work, the mouse, the rat, the 160 INBREEDING AND OUTBREEDING guinea-pig and the rabbit. One need go no further than to cite the work of Castle and his students at the Bussey Institution of Harvard University, the work of Miss King at the Wistar Institute of Anatomy and that of Wright at the Bureau of Animal Industry of the United States Department of Agriculture. The painstaking researches of these investigators show mthout question that the effect of crossing on animals is the same as upon plants. 1 Weight ini (f F,. Cut "Band C Oran fififi 8 d^aeeB X dF,.Ci tVBandC 600 / ^ dCullen 400 200 ^ y [^ ^ Age in Daya 40 do 120 100. 200 240 sso 320 3G0 40O FiQ. 35. — Growth curves of males of race B guinea-pigs and Cavia cutleri and their Fi and Ft hybrids. (After Castle.) The results from one genus is typical of them all. Castle ^^ made a cross between a domestic guinea-pig and a wild cavy, Cavia cutleri. The first generation hybrid males weighed about 85 grams at birth, which is slightly more than the young of either pure race, and retained this lead throughout their subsequent life as is shown by the growth curve in Fig. 35. At maturity they weighed about 890 grams, as compared with 800 grams for the guinea- pig ancestor and 420 grams for the cutleri ancestor. The second generation hybrids of both sexes were HYBRID VIGOR OR HETEROSIS ICl smaller than the first generation hy])ri<]s from birth on, showing that some of the growtli impetus produwHi by the hybridization had not beun retained. 13ut the growth curve of the second generation hybrids rises rapidly at first, showing the healthy start in life they obtained from their vigorous F^ mothers. Perhaps no such increase in vigor as that shown in the species cross just described is usually found when dif- ferent sub-races are crossed. It would not be expecti'd, for ordinary races of mammals are contiinially iMMng crossed within the variety and, therefore, hybriilization would not be expected to increase heterozygosis to any marked degree. But results similar to those obtaineil in plants may be expected if the genetic conditions are sim- ilar. This is proved by the data Wright obtained when ho crossed guinea-pigs bom of unrelated inbred mothers and fathers. The cross-breds were distinctly superior to their inbred relatives in nearly all characters coimected with vigor. In spite of the fact that their inbred mothers were small and somewhat deficient in vigor, a slightly larger per cent, of cross-breds than of inbreds were bom alive, and a distinctly larger per cent, of those born alive were raised. They were somewhat heavier at l)irtli in a given size of litter and gained in weight much more rapidly between birth and weaning. They matured earlier and produced larger litters and produced them more regularly than the inbreds. Thus the results with animals are comparable to those obtained with plants in all essential features. Brietly, in crosses which are fertile the effects aro sueh as to con- tribute to a greatly increased reproductive ability, making 11 ry.' 162 INBREEDING AND OUTBREEDINa possible a larger number of offspring. The degree to which heterosis is expressed is correlated, within limits, with the differences in the uniting gametes. When homo- zygous forms are crossed, it is at its maximum in the first hybrid generation, and diminishes in subsequent gen- erations of inbreeding as segregation occurs and homo- zygosity is again attained. It is a widespread phenom- enon and accompanies heterogeneity of germinal consti- tution whether the organisms crossed are from the same or diverse stocks, whether they have been produced under similar or under different environmental conditions; although it is not apparent until the zygote is formed, from that time on it is expressed in many ways through- out the lifetime of the individual and is undiminished by asexual propagation. These are the effects of cross-breeding upon develop- ment in which we have been particularly interested, those in which the organizations of the combining gametes are sufficiently compatible to permit continued propagation. But it must not be forgotten that we have dealt with only one part of the problem. As the differences between the forms increase limits are reached beyond which the organ- isms neither reproduce nor flourish. One can arrange a series in plants in which (1) the parents are so diverse the cross cannot be made; (2) the seed obtained fails to germinate under any set of conditions; (3) the hybrids are so weak they are unable to reach maturity; (4) the hybrids are extremely vigorous, but sterile except pos- sibly in back-crosses ; (5) the hybrids are fully fertile and more vigorous than either parent; or (6) the parents may be so closely related no effects whatever are to be noted. HYBRID VIGOR OR II1:TKKoSIS IG.'i A somewhat similar series can be arranged with animals, although usuiilly in wide crosses if the hybrids can Ix' obtained at all they are as large or larger than the average between parents. A satisfactory interpretation of the vigor of hybridization must take all these facts into con- sideration, even though tliey may not be the result of the operation of one single law. CHAPTER VIII CONCEPTIONS AS TO THE CAUSE OF HYBEID viaoR The early plant hybridizers, although they frequently discussed the increased size and vigor of their crosses, seldom commented on the effect of inbreeding, and made no speculations as to the cause of either. The animal breeders of the period were more imaginative. Ac- quainted with both phenomena, but more familiar with the results of inbreeding, they unhesitatingly linked the two — the first as an antidote for the second. They attributed most of the injurious effects which appeared in their herds to the concentration of undesirable traits. If unfavor- able characters and tendencies to disease were present, mating similar animals brought out these undesirables more pronouncedly; whereas, if healthy animals from un- related herds were brought in, such tendencies were checked, the defects disappeared, and the health and vigor of the herds returned. Darwin, however, refused to ascribe any large part of the effects of inbreeding to this cause. He knew of many cases in which weakened animals from different in- bred herds had been mated together, and gave progeny of full health and vigor and of increased size. The unde- sirable features induced in both herds by inbreeding dis- appeared when animals of the different herds were mated. Instead of a concentration of the less favorable traits of the two parental lines the reverse seemed to have oc- curred. Similar cases in plants were familiar to him, and 164 CAUSE OF HYBRID VIGOR 1G5 proved beyond question tlie ^a'eat advantage to be gainLxi by crossing even when the individuals themselves were weak. These facts, taken together with the many mar- velous and intricate contrivances of phuits to insure cross-pollination, led him to believe that selt'-fertilizatioii was inherently harmful and something to be avoided it" possible. The benefits accming Trom crossing he ascrilied, as we have seen, to the meeting of sexual elements liaving diverse constitutions. After Darwin's contribution to the problem of inbree