®I}e ^. P, pll pbrarg 0H451 r.7A ^^v. NORTH CAROLINA STATE UNIVERSITY LIBRARIES S00626387 W .1 Date Due .arl3'3C*":i3Apr'5?Y ?«V' ])ricf survey of the subject of evolution before 1900 will serve, therefore, to provide a background for the material of the present text, by depicting something of what was going on in the minds of biologists at the time that genetics had its birth. It will also be useful to have before us some of the ideas of evolution as a means of suggesting a wider application of the principles of genetics that are to be taken up. Nothing need be said here of that phase of the evolu- tion enterprise which concerned itself with convincing a doubting public of the mere fact of evolution. The other phase, involving the presentation of ex})lanations of the evolutionary process, will be sketched briefly. The vague ideas of evolution that occupied the minds of men during the earKer history of biology and the fan- tastic speculative explanations that were proposed have Httle more than historical interest for us today. These explanations were based upon meditation rather than investigation, so that they resembled philosophy rather than science. Around the latter part of the eighteenth ccntur\', certain men (notably Erasmus Darwin, Goetiik, and St. Hilaire) developed more accurate notions of evolu- tion, based in good part upon their own observations, I nOFBRTY LIBRARY n. C. State CoUem 2 Outline of Genetics and proposed simple explanations of the process. These explanations called upon the direct influence of the envi- ronment, but, since little effort was made to analyze the process any further than this, these theories have little value for us. The first author to provide any thoroughgoing expla- nation of evolution was Lamarck, and his theory of Use and disuse (1801) still commands the attention of biologists. According to Lamarck, the environment was important, not as a direct cause of evolution, but merely as the occasion for evolutionary change. When an animal came to live under changed environmental conditions, possibly through migration, it encountered certain new needs. These new needs stimulated in the animal the desire to satisfy the needs. Following this, the animal made a conscious effort to satisfy the needs, and in this effort succeeded in exercising certain of its organs more than before. This exercise resulted in the development of the part exercised. At this point Lamarck introduces his basic assumption to the eft'ect that acquired characters are inherited. Whatever gain is made in developing an organ through exercise, is passed on to the progeny. The progeny, living under the same environmental conditions and actuated by the same motives, will make some further gain in the development of the organ in question, and in this cumulative manner the organ will eventually be developed to such an extent that a new species may be said to have originated. The classic example, which seems rather absurd in itself, but serves to illustrate Lamarck's ideas, runs as follows. The horselike ancestors of the giraffe come to live in a new and arid environment, such that the only The Background of Genetics 3 substantial forage is provided by the leaves of occasional trees. These animals need to reach up to the leaves, and therefore desire to do so. Through a conscious ejjorl to stretch up to the leaves, their necks and limbs are exercised in such a way as to lengthen them. The small gain in length, possibly only an inch or less, made during the lifetime of the individual is passed on to the progeny, who are successful in adding another inch. The final result is the giraffe. This evolutionary scheme works also in the reverse direction as the result of degeneration through disuse. Absurd as some of Lamarck's illustrations may seem, he has really provided the elements of a complete and not unlikely explanation of evolution. The major objec- tion lay in his assumption of the inheritance of acquired characters. Practically all of the earlier expcrmients on this problem seemed to demonstrate that inheritance of acquired characters is impossible, and it was for this reason that the majority of biologists discarded Lamarck's theory. Another objection lay in the fact that conscious effort was hardly to be expected among plants. Lamarck of course recognized this obvious difficulty, and revised his theory in the case of plants, where he claimed the changes were brought about through the direct effects of the environment, citing such things as soil, temperature, moisture, and mechanical pressure. The next great explanation was presented in Charles Darwin's Origin of species in 1859. Darwin called attention to the geometric ratio of increase among living organisms, and reiterated the doctrine of ALvlthus to the effect that any population tends to increase more 4 Outline of Genetics rapidly than the means of subsistence. This increase tends to set up a severe struggle for existence or compe- tition, as the result of which an equilibrium of species is established, with approximately the same number of individuals of a given species surviving year after year in any given locality. Darwin next points out the universality of variation among living organisms, such that no two individuals of any species are ever absolutely identical. As for the cause of variation, no explanation is provided, but the nature of variation is rather clearly outlined. Those variations which are important in the evolutionary process are characterized as quantitative, continuous, and fluctuating. By quantitative it is meant that the variations are differences in the degree of development of some part or feature of the organism. When it is said that the variations are continuous, the implication is that further variations will take place in the same direction as the variations that have taken place in the preceding generations. The term ''fluctuating" indicates that reverse variations will take place as freely as do the progressive variations. According to Darwin, variation of this type is going on in all organ- isms. Since this is true, and since a severe struggle for existence is taking place, it is impossible to escape the conclusion that it is the ''fittest" that survive. If a given species is represented in a certain locality by a thousand young individuals, no two of which are abso- lutely alike, and if there is only enough room or only enough subsistence for one hundred of them ever to reach maturity, it must be true that, in general, it will be those that are the best adapted to cope with the conditions of the environment that are the ones to sur- The Background of Genetics 5 vive, while the rest perish in the struggle. The one hundred survivors are ''litter" than were the others of their generation because they happened to have certain useful organs or processes somewhat more fully de\'el- oped. In the following generation, some of the progeny will have the organs in question still more fully de\'eloped than did their parents, while on the other hand there will also be some that have them less fully developed. The former group will again be chosen by nature to sur- vive and perpetuate the species, and thus progress will be made in the direction of better development of useful organs until a degree of development has been attained which may be said to represent a new species. The wa}^ in which nature manipulates these quantitative, con- tinuous variations of Darwin's to bring about this progressive evolution or adaptation can be \'isualized more concretely from this simple diagram. (More poorly adapted forms perish in competition) ^ ^ 6 ^ ^ Environment -> (favors variations in this direction) ^ 8 The numerous objections to Darwin's theory cannot be discussed here; suffice it to say that these objections were directed mainly at the adequacy of the Darwinian variations in accounting for the results of evolution rather than at the idea of natural selection. The ''sur- vival of the fittest" is a rather generally accepted idea. The question whether the Darwinian variations are 6 Outline of Genetics adequate can be considered more critically a little later in the light of the more recent knowledge of inheritance. The third great explanation of evolution was offered by De Vries in 1900. This author was the first to base his conclusions on the results of his own experimental breeding, rather than merely on the extensive observa- tion of plants and animals in nature. Judging from the behavior shown by Oenothera Lamar ckiana (American evening primrose) during the course of the ten or more generations that it grew in his garden, De Vries con- cluded that the real basis of evolution lay in the phenom- enon of mutation. In addition to its ''normal" pro- geny, 0. Lamarckiana produced in small numbers cer- tain distinctly new types, the mutants. The type of variation involved in mutation was distinctly different from the Darwinian, being qualitative, discontinuous, and constant. It was readily seen that the mutants involved qualitative changes from the parent, inasmuch as entirely new characters were shown, rather than merely the quantitatively greater or lesser development of cer- tain of the parental characteristics. It was equally plain that mutation was discontinuous, the direction and nature of mutations being entirely unrelated to any of the mutations that had taken place in the past. And finally, the mutants were strikingly constant, breeding true to their own characteristics rather than reverting in later generations to the original parental type. These mutational changes that De Vries studied intensively in Oenothera were later identified in other species as well. The part that mutation may play in evolution is suggested by a consideration of the characteristics of the mutants. Probably the majority of the mutants differ The Background of Genetics 7 from the parent-form in such ways that thc>' might well be called degenerates; the new characteristics shown serve to adapt the mutant more poorly to the environ- ment than the parent was adapted. Other mutants may show changes only of an unimportant type, so that they are neither better nor more poorly adai)ted than was the parent. A few of the mutants (according to De Vries) may show such characteristics as to be better adapted to the environment than was the parent- form. Upon this miscellaneous mass of mutants natural selection immediately comes to play, quickly eliminating the poorly adapted types and preserving the good. Thus De Vries holds with Darwin in invoking natural selection, but the t^pe of variations involved is dis- tinctly different. According to Darwin, natural selection serves gradually to build up a new species; according to De Vries, numerous new species are born full Hedged, and natural selection merely decides which of them shall survive. Objections to the De Vriesian theory will be mentioned later in this text, in the light of some of the rather recent work in genetics. In addition to these three great explanations of evolu- tion, there are a few others that should be considered briefly. A number of authors (notably David Starr Jordan in this country) have attached primary impor- tance to the principle of "'isolation" in evolution. A few individuals of a species may migrate successfully to a new locality which is, or may subsequently become, sufficiently isolated, by geographic barriers of one t}pe or another, from the original locality that no extensive return migration can take place. The result is that a new colony is established which is sufficiently isolated g Outline of Genetics from the parent realm that free crossing between the colonists and the stay-at-homes does not take place. Any chance variations that may have existed in the few original emigrants will now have an opportunity to pe^etuate themselves instead of being ''swamped out" through free crossing with the other members of the parent species, as would have happened if there had been no isolation. Thus the various populations of a species that we see today somewhat isolated from one another have had a chance to express and later augment chance differences to such a degree that we may now regard them as different varieties. With time the diver- gence of characteristics between the isolated groups will become still greater. The isolation theory, as outlined above, does not by itself provide a ''complete" explanation of evolution. It is best to be regarded as a sort of a coroUary to the Darwinian or DeVriesian schemes. In either event, it is the principle of natural selection that brings about progressive or adaptive evolution, while isolation either serves to multiply species "on the same level," through giving chance variations an opportunity to express themselves and become augmented, or else it serves to enlarge the scope of natural selection by thrusting repre- sentatives of the species into a somewhat new environ- ment or by presenting natural selection with a slightly new population from which to make the choices. Another matter that should be mentioned is ortho- genesis. Orthogenesis may be regarded either as an "explanation" of evolution or merely as the name of a phenomenon. There is considerable evidence support- ing the beUef that the variations (or at least many of The Background of Genetics 9 Ihem) that take place in li\ijig organisms arc determinate^ taking place along predetermined lines, in a predictable direction, rather than indeterminate, with an equal chance of their taking place in any direction, as the other theories would have it. For example, if a variant is slightly dwarfed as compared with the parent type, there will be among the descendants of this variant a further varia- tion involving greater dwarfedness, and subsequently more changes will take place all involving further steps along this same predetermined line. The direction of the variations is not necessarily related to any environ- mental demand. A few authors, regarding orthogenesis as an "explana- tion" of evolution, have visualized an intrinsic ''force" in the organism which guides the variations. Since this view has a vitalistic flavor, it is not popular among scien- tific men. More often it is simply recognized that a certain amount of variation of this sort does take place, and orthogenesis is the name given to the phenomenon, various quite materialistic explanations having been proposed to account for it. "Evolution through hybridization" is a theory that was suggested by Weismann some decades ago, and has recently been developed and championed by Lotsy. It is a fact well known among biologists that crossing two distinct types may result, in the second hybrid generation, in a few new and pure-breeding forms, some- what different from anything that had previously existed. Lotsy has shown by experiment that when such new forms (from Antirrhinum crosses) are returned to grow under natural conditions, nature will select some of the types to survive, but will quickly eliminate lo Outline of Genetics the others. While there is little question that natural hybridization takes place and may be a real factor in producing new varieties, at the same time this theory is not satisfactory as a ''complete" explanation of evo- lution. It seems rather obvious that, although hybridi- zation can multiply variations through crossing forms that are already different from each other, it can never account for the "original" differences. In considering the relative merits of these different explanations, there are three things that it is useful to bear in mind. First, it is certainly not necessary to subscribe to a belief in any one of the theories to the complete exclusion of the others. It is quite possible that every one of them may be a factor in evolution, and altogether probable that no one of them by itself can adequately account for all of the evolutionary change that has taken place. Second, it is not advisable to contrast these explana- tions as though they were coordinate units. The "prob- lem of evolution" is not a single problem, but a complex of numerous ones, and any proposed explanation of evolution is confronted by the necessity of answering several distinct questions. The conspicuous questions to be answered are: 1. What is the cause of variation? 2. What is the nature of the variations that are important in evolution? 3 . How may variations be perpetuated and multiplied ? 4. How are the variations manipulated to effect progressive evolution? It will be noticed that Lamarck goes farther than any other author in answering question i. Orthogenesis The Background of Genetics ii and the hybridization theory provide suggestions on this point, but the suggestions are hardly satisfactory. For question 2 rather distinct answers are provided by Lamarck, Darwin, De Vries, and the orthogenesis theory. It is in answering question 3 that the isolation theory and the hybridization theory have their chief value. Question 4 is indirectly answered in one way by Lamarck, and indirectly answered in another way in the orthogenesis theory, while all the other theories plainly call upon natural selection to answer this ques- tion. If the theories are to be compared, it can safely be done only after some such analysis as this. Third, discussion of evolutionary theories usually leads to the realization that more exact experimental evidence is needed before much further progress can be made in solving the problems of evolution. Such has been the actual history of the case, for, with the begin- ning of the twentieth century, the study of evolution culminated in, and became diverted into, genetics, the experimental study of inheritance. Of course genetics has not answered all of the questions that have presented themselves in connection with evolution, but many criti- cal and suggestive findings have been made, as will be seen in the following chapters; and unquestionably genetics will contribute a great deal more in the next few decades. CHAPTER II THE INHERITANCE OF ACQUIRED CHARACTERS At the basis of genetics lies the fact that variation occurs in all living organisms. It is possible to classify variations in a number of different ways.^ At the out- set it is important to realize the distinction between non- heritable and heritable variations. As for the former, it is usually evident that these originate as responses on the part of the organism to environmental stimuli. Ac- quired characters of this sort, however, are of little sig- nificance in genetics, inasmuch as they are not passed on from parent to offspring. It is the heritable variations that provide the material of genetics; and the origin of these is a matter of considerable controversy. For the most part, they are ascribed to mutation, meaning that their origin is sudden and spontaneous, seemingly unre- lated to environmental stimuli. There is some evidence, however, which suggests that heritable variations may originate as acquired characters. It will be appropriate at this point to discuss the controversy on inheritance of acquired characters. The idea of inheritance of acquired characters was first clearly developed by Lamarck in connection with his explanation of evolution, the so-called theory of *' appetency," or the effect of use and disuse. Francis Galton, in 1875, was one of the first to express skepti- * A serviceable set of classifications is provided by Babcock and Claussen (i). 12 TJie InJieritance of Acquired Characters 13 cism in regard to such inheritance, but it was Weis- MANN (17) who was most intluential in combating the idea. After Weismann's presentation of the sit- uation, biologists were divided into two camps in reference to the question: (i) neo-Lamarckians, who affirmed behef in inheritance of acquired characters, and (2) neo-Darwinians, who denied it. Until very recently, at least, the bulk of the evidence of genetics has served to refute inheritance of acquired charac- ters. Much of the lack of agreement in this controversy is due to the dehnition of an acquired character. It should be kept in mind that actual characters are not inherited, but only the determiners, which regulate the way in which the organism reacts to its enviromncnt. For example, when it is said that a child inherits its father's nose, the statement is not meant to be literally true; it is meant that just as there was something in the body of the father that was responsible for the develop- ment of a particular type of nose, so there was a similar something in the child's body that developed a similar result. It is merely a matter of convenience to speak of the inheritance of characters. Weismann defined an acquired character as ''any somatic modification that does not have its origin in the germ plasm." This definition is not always easy to apply. Examples of acquired characters in the Weis- mann sense are mutilations, results of function (as in the use or disuse of certain organs), many diseases that affect the bodily mechanism, and, to use a rather vague expression, effects of environment. Weismann gave three reasons for rejecting the belief in inheritance of 14 Outline of Genetics such characters: (i) there is no known mechanism by which somatic characters may be transferred to the germ plasm; (2) the evidence that such a transfer does occur is inconclusive and unsatisfactory; and (3) the theory of the continuity of the germ plasm is sufficient to account for the facts of heredity. When Weismann says that there is no known mech- anism by which somatic characters can be transferred to the germ plasm, to him it is equivalent to saying that it is hard to see how the water that has gone over the dam can return and affect the flow of the water upstream. He assumes, of course, that the genu plasm is isolated from the somatoplasm very early in the development of the fertilized egg into an individual, and that w^hen it is isolated it takes no active part in the history of the body (see fig. i). The somatoplasm is thus merely a carrier of the germ plasm, and is unable to affect the character of it any more than a rubber hot- water bag, although capable of assuming a variety of shapes, can affect the character of the water it contains (Walter 18). This early differentiation of germ plasm and body plasm has been demonstrated rather strikingly in several animals. In Ascaris megacephala, the following cy to- logical situation was demonstrated by Boveri, in 1903 (DoNCASTER 7). Following the first division of the zygote, the two daughter-cells come to differ from each other through the apparent degeneration of some of the cell constituents in one. That daughter which main- tains the full cell equipment of the zygote thereby per- petuates the capacities of the germ plasm, while the other daughter, which has lost certain visible cell constituents, Fig. I. — Diagnim illustrating Weismann's theory of germinal con- tinuity. Three generations are represented, with cells of germ plasm shaded, and those of body plasm unshaded; germ plasm continuous from generation to generation, carried over from parent to ofTsping by z\gote (Z); impossible for body plasm to perpetuate itself into a second generation. 1 6 Outline of Genetics starts a line of purely body plasm cells. A similar differentiation occurs between the granddaughter-cells from the fully equipped daughter, and so on for two more divisions, so that finally there are fifteen body plasm cells and but one germ plasm cell; the germ cells of the adult individual can all be traced to this single initial. A somewhat similar program has been traced in Chry- somelid beetles, where, after numerous segmentation divisions, some of the nuclei associate with certain gran- ules, and it is these nuclei that start the germ plasm. Hegner (Doncaster 7) has succeeded in artificially destroying these granules by means of a hot needle, thus producing embr}'os without germ cells. Equally striking situations have been demonstrated in other animals as well, but nothing of the sort has ever been found in plants. Germ cells in plants are formed from h^-podermal and even epidermal cells, which, during earlier ontogeny, are apparently identical with other somatic tissues. Here there is surely no dis- tinct germ plasm, isolated from body plasm and insulated within it from environmental influences. In fact, there are cases in which ''adventitious" germ cells have been seen to form from tissues which normally are quite as somatic as any plant tissue could be. In this connection, it is worth mentioning that Bateson (2) suspects plants, as genetic machines, differ fundamentally from animals; this idea being suggested to him in good part by the fact that ''in the animal the rudiments of gametes are often visibly separated at an early embryonic stage, whereas in the plant they are given off from persistent growing points." The Inheritance of Acquired Characters 17 A general theoretical objection to Weismanx's view is that every organism is a physiological as well as a morphological unity, and that cells completely insulated in such a unity would be impossible. Cytologists also have come to believe that there are proto})lasmic con- nections between adjacent cells in practically all plant tissues, and, in general, physiolog}' tends to confirm this. Such suggestions voice a growing belief that the body plasm can affect the germ plasm. The reply of the Weismannians is that even though somatoplasm might affect germ plasm in this general physiological way, this is a very different thing from the inheritance of some definite acquired character. To be inherited such a character would have to be exactly redeveloped in the germ plasm, and the intlucnce referred to cannot be so specific as that. This, of course, is a theoretical answer, and the question can only be decided by experimental work. A theoretical rejoinder to this answer may be suggested. It is like the voice in a tele- phone transmitter, which starts vibrations that make the receiver repeat the voice. (Something more delinite on this matter will be considered a little later.) Before arriving at anything like a conclusion on this matter, it will be necessary to consider some of the claimed cases of inheritance of acquired characters. I. Mutilations. — Most of the evidence under this head is in relation to animals. It is a matter of common experience that mutilations are not inherited in man and the domesticated animals. A few quotations from Walter (18) suggest the situation: "It is fortunate that the sons of warriors do not inherit their fathers' honorable scars of battle, else we 1 8 Outline of Genetics would now be a race of cripples The feet of Chinese women of certain classes have for centuries been mutilated into deformity by bandages without the mutilation in any way becoming an inherited char- acter The progressive degeneration or crippling of the little toe in man has been explained as the inherit- ance of the cramping effect of shoes upon generations of shoe-wearers; but Wiedersheim has pointed out that Egyptian mummies show the , same crippling of the little toe, and no ancient Egyptian could be accused of wearing shoes, or of having shoe-wearing ancestors." Sheep and horses with docked tails, as well as dogs with cropped ears, never produce young having the parental deformity. Weismann's early experiments with mice, later verified by other investigators, give additional evidence that mutilations are not inherited. He bred mice whose tails had been cut off short at birth, and continued this performance through twenty-two generations, with absolutely no eft'ect on tail length. Very little serious consideration has been given to the possibility of inheritance of mutilations in plants. Cut- tings for propagation are usually trimmed to prevent excessive transpiration, but no one ever expects to find this mutilation perpetuated, even in the plant developed from the cutting, much less in the next generation devel- oped from seed. In fact, since we have begun to learn of the remarkable powers of regeneration possessed by plants and animals, we would not expect the inheritance of mutilations. There is one bit of work that should be mentioned in this con- nection. Blaeinghem (3) claims to have procured from a single injured individual a Hne of maize plants that show a varying per- The Inlientance of Acquired Characters 19 centage of double and sometimes triple grains. The author calls this a typical case of inheritance of acquired characters, but Ameri- can investigators have hesitated to accept this interpretation of the phenomenon. Characters of practically the same sort have been observed to originate in other cases in corn not known to have been injured in any way. 2. Effects of use and disuse. — Inheritance of the effects of use and disuse lay at the foundation of Lamarck's theory of evolution. Weismann was suc- cessful in discrediting this belief by explaining on some other basis practically all of the supposed examples of this phenomenon that had been advanced. In plants, of course, it would be hard to find anything exactly analogous to the use and disuse of parts in animals; Lamarck himself did not attempt to apply quite the same theory to the plant kingdom. One fact, however, is a common experience of botanists. Functionless organs gradually become aborted, become mere vestiges or even suppressed entirely. For example, a study of the organogeny of flowers shows that when a lloral member is belated in its development it is destined sooner or later not to appear at all. The following theoretical Weismannian (or Darwinian) explanation of this situation is suggested. A given species has a given nutritive capacity; the less it draws upon its nutritive capital for the development of one organ the more it has available to expend on the development of other organs. When an organ becomes functionless it no longer has any survival value; survival is then dependent upon the relative develoi^ment of the other organs. Through "variation" certain individuals develop the functionless organ less than usual and therefore develop the other organs more than usual. Under the new conditions these individ- uals will survive and the others will be eliminated. Gradually abortion of functionless organs would lake [)lace in this waw One would expect that the rate of change would be roughl>- pro- 20 Outline of Genetics pyortional to the size of the organ involved, and that any retrogres- sive evolution of this sort would be slower than progressive evolu- tion. 3. Diseases. — Roughly speaking, diseases are either the results of infection by bacteria or fungi or some inher- ent organic weakness. Since the latter condition is chiefly serious only in inviting attacks by bacteria and fungi, we are concerned chiefly with diseases caused by these pathogenic forms. Realizing this, true inheritance of disease seems to be an impossibility, for if the parasite enters the germ cell it is practically sure to destroy it, and there will be no progeny. It is true that in many cases progeny are born diseased, but this is due to rein- fection of the young embryo from the body of the mother. Many examples of this phenomenon are available in man and other mammals. In plants, also, diseases (e.g., smut) are sometimes passed on by means of spores car- ried upon or even within the seeds. Such a thing, how- ever, can in no sense be spoken of as inheritance, since it always involves a reinfection. In one respect, however, one may speak of disease inheritance. Breeding experiments have shown that predisposition to disease and disease resistance, com- monly called susceptibility and immunity, are inherited. In practically all cases, these characteristics are evidently of germinal origin, having been hereditary in the begin- ning rather than acquired. Such cases, of course, have no bearing on the present problem. There remain a few instances, however, that rather suggest the inherit- ance of acquired characters. GuYER and Smith (10), by inoculating female rabbits either with typhoid vaccine or with the hving bacilli. TJie InJieritauce of Acquired Characters 21 have succeeded not only in building \i\) a high resistance to typhoid in these female rabbits themselves, but in securing from them progeny with a high resistance. In fact, rabbits of the third generation have still shown a high immunity which could have come only from their grandmothers. The likelihood, however, is that this immunity is not passed on through the germ cell itself, but is "reacquired" by offspring while in utero, and nourished by the blood stream of the mother. This, of course, would again be merely a case of ''transmission" rather than true inheritance. The passing on of such an acquired immunity from the male parent to the pro- geny would constitute a convincing demonstration of inheritance of acquired characters, but such a demon- stration has not as yet been made. . It is suspected that a situation similar to the foregoing exists also in man. Racial immunity is believed by some medical men to have been built up not only through a "natural selection" of immune types, but from the passing on from mother to offspring of acquired immunity. There is one fairly well known case of this sort in the plant kingdom. Bolley (4) claims that he can get a resistant strain of flax from almost any known variety. According to him, the resisting ability increases from generation to generation, if the crop is constantly sub- jected to disease attack. He took a ])ure-pedigreed strain of flax which had come original 1>' from a single non-resisting seed. This was planted in slightly ''sick" soil, that is, soil infected with the wilt-i)r()(lucing organ- ism. Most of the individuals died, 1)ut ''a few scrubs" survived. He then planted seeds from these in slightl\- "sicker" soil than before, and thus, by gradually work- 22 Outline of Genetics ing his crop into sicker and sicker soil in the later genera- tions, he finally obtained a fully resistant strain from the pure non-resistant strain with which he started. Such a strain, he says, will not lose its resistance if planted progressively in more infected soils. He gives the fol- lowing theoretical explanation of his results: ''Either (i) the so-called unit character of resistance was present in undeveloped form and becomes stronger from year to year under conditions of disease; or (2) there never was any character present which is entitled to be called a unit character, but it began to develop the first year the parent plant came in contact with the disease, and the protoplasmic nature of the ancestors of the plants which we now have has been such that they accumulated more and more the resisting power from year to year, just as they had opportunity to develop resistance against a constantly acting factor of disease, which, when too powerful, acts as an eliminating factor." BoLLEY inclines to the second alternative. This general conception seems to explain why home-grown seed is regularly more resistant than seed from the same variety which has had a vacation away from home for several years. It has kept in training like a football player. Bolley says that if these conclusions are cor- rect, there are probably no unit characters which are not fluctuating, and there are no fluctuating characters which may not readily be fijced. These results are striking enough, but their signifi- cance depends entirely upon the purity of the strains which were used originally, and also upon the preserva- tion of purity during the experiment. Bolley's phrase "elimination factor," which he uses repeatedly, might The Inherilance of Acquired Characters 23 be taken to suggest selection from an impure strain. If his conception is true, it could be demonstrated by de- veloping a large majority of resistant indiA'iduals among the non-resistant j)lants which were first subjected to disease attack, rather than merely ''a few scrubs." The results as they stand could probably be interpreted as due to the selection of a few resistant individuals from an impure strain. From the foregoing cases, it becomes rather evident that, so far as mutilations, effects of use and disuse, and diseases are concerned, inheritance of acquired characters has not as yet been satisfactorily demonstrated, either in the plant or animal kingdom. One more category of cases, however, remains to be considered. 4. Effects of environment. — This heading is suffi- ciently inclusive to include a number of types of cases. It has now been some years since Castle (6) per- formed his classic experiment on guinea pigs. Animals w^ith white coats will have only white-coated progeny, while a pair with black coats, provided both male and female come from a pure stock, will have only black- coated progeny. Using only animals from pure stock, Castle removed the ovaries from a white-coated female and transplanted them into the body of a black-coated female. The mating between this black-coated '^ foster mother" and a white-coated male resulted in a progeny all of which had white coats. Evidently it was the germ cells alone that were effective in determining the char- acter of the progeny. The decisive results of this experi- ment were very influential in refuting the concept of inheritance of acquired characters. At the same time it must be borne in mind that, whereas such a superficial 24 Outline of Genetics character as coat color might not respond to artificial manipulation of the germ cells, it is still possible that there are other characters, more fundamentally tied up with the metabolism of the organism, that could be affected by such treatment. Numerous experiments confirmed these findings of Castle's, but there was one field of investigation from which rather contradictory results began to be accumu- lated. The numerous studies that have been made dur- ing the last few years on inheritance in the microorgan- isms have been ably summarized by Jennings (13). Here there appear some striking indications of inheritance of acquired characters. ''The germinal or genotypic constitution in most organisms is extremely stable; in many stocks it changes not at all, so far as observation goes. To alter it by physical or chemical agents is usually to kill it. In some of the lowest organisms — rhizopods, bacteria, some infusoria^ — it changes with somewhat greater frequency, though still rarely. The nature of the changes, and whether they may be permanent, or must after genera- tions revert to the original condition, is in some dispute. In these same organisms, environmental agents may pro- duce changes persisting through many generations of uniparental reproduction and even through biparental reproduction, the period of persistence depending partly on the number of generations through w^hich the pro- ducing agent acted. This suggests that inherited char- acters as permanent as any that exist might in time be so produced. In spite of important differences of opinion among investigators, to the reviewer the facts in uni- parental reproduction seem to point more toward the pfOfOnT LBRARf N. C. Stale College The Inlieritance of Acquired Characters 25 production of ev^oluliouary change by the action of the environment on the germ plasm than hy any of the other methods." This behavior on the part of some of the lower organ- isms, difficult to interpret without the assumi)tion of inheritance of acquired characters, fostered the following belief. In higher animals, where germ plasm and body plasm are sharply differentiated, inheritance of acquired characters is an impossibility; in the simpler organisms, however, germ and body plasm are doubtless one and the same thing, with the result that a certain amount of inheritance of acquired characters can and does take place. Such an opinion would not be out of harmony with the views of Weismann, who was early forced to the belief that inheritance of acquired characters must take place in the more primitive organisms. The opinion of the biological world was becoming fairly well settled on this matter when Guyer's startling results (9) were published. It will be seen that Guyer's methods ''strike at the germ plasm" more directly than any that had previously been tried. Grinding up the eyes of white rabbits, Guyer pro- cured a lens-extract. This was injected into the blood stream of fowls. There, since the lens-extract was a foreign and " inhamionious " protein, a reaction took place which resulted in the production in the blood stream of an antibody (following the same principles which apply to the production of antitoxins in medicine). This particular antibody had the peculiar property of "precipitating" or in some way rendering functionless the characteristic protein of rabbit lens. The property is quite specific, so that this antibody may appropriately 26 Outline of Genetics be spoken of as antilens. Serum obtained from the blood of fowls thus ''sensitized," and therefore contain- ing antilens, was injected into the blood stream of nor- mal white rabbits. No noticeable modification was obtained in any case upon adult rabbits that were so treated. When, however, the serum was injected into pregnant mother-rabbits, starthng results were obtained. Some of the resulting progeny had eyes that were clearly defect- ive. Furthermore, the abnormality was readily trans- mitted through the female line for quite a number of generations, without any additional injections being made. The defect did not decrease in degree, but seemed even more pronounced in the later generations. At this point a few questions might be asked. Have such eye defects ever been known to occur among untreated white rabbits; is this the sort of thing that might appear ''spontaneously" through mutation, or a recessive character that might have been segregated out through inbreeding, as is true of so many other functional abnormalities? Careful inquiry has revealed the fact that no such eye defects have been reported elsewhere. Again, is this the sort of thing that might be expected to result from any sort of mutilation, or is it a specific response to a specific stimulus ? This question is clearly answered by the behavior of the controls. Untreated pregnant mothers, mothers treated with serum from unsensitized fowls, and mothers treated with serum from fowls that had been sensitized to rabbit tissues other than lens never gave any defective progeny. An even more critical question is the following: is this another case of transmission rather than true inherit- The Inheritance of Acquired Characters 27 ance; whatever may be the material basis of ilie defect- ive eyes, is it regularly passed from tlie body of the mother to the young in utero rather than through the germ cell proper ? The answer to this question was early suggested by the following facts. Later litters from the mothers that had originally been treated never contained any defective indi\dduals. The influence of the anti- lens seems to die out in the blood stream, suggesting that it is only by being incorporated in the gemi plasm that the character can be perjoetuated. A more con- vincing demonstration of this point appeared in the later experiments. Males with defective eyes when mated with females from a normal line j:)r(3duced only nonnal offspring. When, however, these same males were remated with their own daughters from the fore- going cross, a certain number of defective offspring resulted. Evidently the abnormality can be passed on by a male parent; it behaves in inheritance like a Men- delian recessive character. Inasmuch as the male parent exerts upon the progeny no influence beyond what is contained in a single male gamete, and inasmuch as this gamete is practically nothing more than a nucleus, the foregoing amounts to about as clear a demonstration of inheritance of acquired characters as would be possible. The very definite results of this set of ex[)eriments are sufficient to prove that inheritance of acquired char- acters can take place, even in one of the higher animals where germ plasm and body plasm must be as sharply differentiated as anywhere. It is equally true, however, that only very special conditions can produce the result. A word might be said 011 the theoretical mechanism involved. GuYER proposes that "there is some degree of constitutional 28 Outline of Genetics identity, probably protein homology, between the mature sub- stance of a tissue and its correlative in the germ," and that "basi- cally, inheritance is mainly a question of the perpetuation of spe- cific protein-complexes, and development the result of differential reactions of these same fundamental constituents under different conditions of environment Is it unreasonable to suppose that if changes come to pass which affect certain constituents of tissue cells, this influence, if borne in the circulating fluids of the body, could also affect the hom^ologous constituents of the germ cells?" The same result has been obtained by the direct injection of lens-extract into the blood stream of the rabbits themselves. It is perhaps surprising that rabbits will manufacture antibodies for one of their own tissues. It is evident that they do, however, and this brings us a distinct step closer to something that m'ght occur under natural conditions. If degenerating eyes may them- selves originate antibodies which in turn affect the germ cells, the cardinal principle of Lamarck's theory of evolution through inheritance of acquired characters must be conceded. The fact that the eyes of the pregnant mothers were in no case themselves affected does not necessarily recommend the assumption of "susceptibility" only in the embryonic state, but is doubtless due merely to the fact that there is a relatively much greater blood supply to the lens of embryo rabbits than those of adults. The experiments of Griffith and Detlefsen, final results of which have not yet been published, promise to provide an equally good demonstration of inheritance of acquired characters in mam- mals (see Science 56:676-678. 1922). A few examples of the supposed inheritance of the ''effects of environment" in plants should be considered. Zedebaur found that Capsella, which in the course of many years had gradually crept along the roadside up into an alpine habitat and there acquired alpine char- acters, retained these characters when transplanted to the lowlands. This has been accepted by some as an authentic instance of inheritance of acquired characters; but it is possible that this conquest of an alpine habitat TJie InJieritance of Acquired Characters 29 by Capsella can better be explained by the gradual natu- ral selection of just those gemiinal variations that best fitted individuals to cope with alpine conditions. This would result in the gradual establishment of a strain of germ plasm that would produce body structures fitted to alpine conditions. In other words, this is just the way in which natural selection would develop a new elementary species from the original t}pe. If such a type were established, of course its gcnn phism would produce alpine plants, even under lowland conditions. They might not survive long, and natural selection might eliminate them, but their structure would be due, not to the inheritance of somatic structures, but to the inherit- ance of an alpine germ plasm. The objection to Zedebaur's conclusions on the grounds that the result may be attributed to natural selection has been avoided by the famous experiments of Bonnier (5). In 1884, this investigator began mak- ing plantations in the lowlands and at various altitudes in the Alps, so arranged that the two individuals to be compared were produced by dividing one plant. After a lapse of over thirty years he has made the following report. A few of the plants taken from the plains to alpine stations died, but a list is given of fifty-eight species that proved able to maintain themselves at high altitude. These have all undergone changes which make them closely resemble indigenous alpine plants.' In at least ^ The principal chan<::cs arc relatively lari^a- (levelo|)ment of tiie subterranean as compared with aerial parts, shorteninj^ of the leaves and of the internodes of stems, increased hairiness, and relatively larger development of bark and protective tissues. The leaves become thicker in proportion to their surface and are a deeper green, with more higlily developed palisade tissue and a larger numlx'r of ihloroplasts, while the flowers are larger and more highl}'^ colored. 30 Outline of Genetics seventeen species the changes are so great that the plants have apparently been transformed into distinct alpine "species." The reverse experiment, transplanting alpine plants to the lowlands, gave similar but less startling results. These experiments as they stand are really more serv- iceable to the ecologist than the geneticist. The geneti- cist wishes to know whether the transformations will maintain themselves when the plants are returned to their original stations and propagated by seed. Bon- nier has as yet made no clear statement on this latter point. An interesting issue arises in this connection. If the transformed plants, after being returned to their original stations, revert, in the course of a number of generations, are we to conclude that inheritance of acquired char- acters has not taken place ? Should we not rather expect that, if inheritance of acquired characters takes place under a given set of conditions, the reverse conditions will bring the reverse change according to exactly the same principle ?' Such work as that of Bonnier may eventually demonstrate that inheritance of acquired char- acters is a possibiUty in plants, though it may fail to demonstrate that irreversible evolution can be brought about through inheritance of acquired characters. The latter can be fully demonstrated only when an acquired character comes to be represented by a gene or set of genes in the germ plasm, which are as definitely and "permanently" a part of the hereditary complex as any ^ It is of course true that some evolutionary changes are probably irreversible (Herrick 12), but such changes are probably not involved in the Bonnier experiments. TJie Inlicrilance of Acquired Characters 31 of the pre-existing genes. Such a demonstration has as yet not been approached among higher ])lants; it is aj)- proximated for animals by Guyer's white rabbit experi- ments. Attention should be called to another ])hen()menon which can easily be confused with inheritance of acquired characters. If corn is planted in poor soil, weak indi- viduals result. Seed from these weak indi\nduals, when planted in good soil will develop again somewhat weak- ened indi\iduals, suggesting the inheritance of acquired characters. This, however, is merch' the direct effect of environment continuing through the second genera- tion. The weak individuals in the poor soil develop small seeds with low nutritive capacity, and ])lants developed from abnormally small seeds are always weak, whether the individual that produced the seed grew in poor soil or not. There has been fairly good agreement on the point that trees deformed by prevailing winds, like the willows that line the canals in Belgium and Holland, or storm- crippled trees along exposed seacoasts, do not produce progeny showing these characters when the adverse environmental conditions are removed. Mavr (15) has written a notable work on silviculture, in which he claims that only species characters are inherited in trees, and that the effects of climate are not inherited, and there- fore that the source of the seed makes no dilTerence. l\\ other words, seeds of Scotch pine would alwa}s produce Scotch pine progeny, no matter at Avhat latitude or alti- tude the ancestors had been growing. According lo Mayr, therefore, there is no inheritance of acquired characters in trees. 32 Outline of Genetics Dr. Arnold Engler (8) found, however, that in the seedHngs in his nursery growth in height distinctly decreased as the altitude or latitude from which the seed came increased. He also found that seeds from pines which had been crippled by growing in poor soil conditions gave rise to crippled plants when grown in good soil. In many cases, trees of the third generation still showed the habit "acquired" by their grandparents in different habitats. These are striking results, but it is well to bear in mind all of the possibilities. Engler might have been 'deahng with slightly different strains of trees, differing in germinal constitution; or it may have been another case of the "false inheritance of acquired characters" that was explained in connection with corn. Seeds from higher latitudes and altitudes might well have been smaller, so that we should have expected smaller progeny, even when grown in the lowlands.^ There are several examples of what seems to be inheritance of acquired characters in simpler plants, but opinion is not settled on interpretation of results. Jen- nings' statement (see p. 24) includes the bacteria. The work of Hansen (ii) is interesting. This investi- gator took isolated yeast cells, which, when cultivated under ordinary conditions, uniformly gave rise to spore- bearing forms, and subjected them for a time to the highest temperature at which growth could still occur. As a result he procured a race which has been cultivated under ordinary conditions for twelve years without once developing spores. ^ We have reason to believe the size of the seed ma3' affect the size of the resulting plant even in such forms as coniferous trees (Munns i6). The Inheritance of Acquired Characters 33 As an example of the sort of thing that in:iy occur among fungi, the work of Long (14) may be mentioned. Piicciuia Eliis- iana and P. Andropogonis both grow, in one stage, on Aiidro- pogon, where they are to be distinguished by morphological differ- ences in the uredospores. P. Ellisiana has Viola for its alternate host, while P. Andropogonis has Pcntcstemon. P. Ellisiana, how- ever, has been artificially induced to infect Petitestcmon , where it produces spring spores that resemble those of P. Andropogonis. When these spring spores are returned to Andropogon, the result- ing uredospores are morphologically identical with P. Andropo- gonis uredospores. Conversely, P. Andropogonis can be made to infect Viola, where it produces morphologically P. Ellisiana spring spores, and these will bring P. Ellisiana uredospores when returned to Andropogon. This appears to be an actual change in species through a change in the quahty of the nutrition. But can it be demon- strated that the two forms were really distinct "species" in the first place? The findings made by Williams (19) on periodicity of sex cells in the marine alga, Dictyota dichotoma, rather clearly indicate inheritance of acquired characters. In any one locality the male and female organs originate simultaneously and pass through their successive stages of development together; a general liberation of gametes and fertilization takes place on a particular day. This period differs between different localities and varies slightly at any one locality, at all times showing a clear relationship to the tides, and therefore to the amount of light that reaches these submerged plants. So intimate is the foregoing relationship that one can })re(lict the time of liberation and fertilization by consulting the almanac of the locality. On the other hand, there is no evidence of periodicity in seas where there are no tides. Plants transferred to aquaria in the laboratory, and thus 34 Outline of Genetics removed from the influence of tides and varying light, continued to show the characteristic periodicity of the locaHty from which they came. Thus an obvious adjust- ment on the part of the plant to a varying set of environ- mental conditions has evidently become heredity. It would be unwise to attempt any final conclusions on this subject of inheritance of acquired characters; the status of the subject changes as new evidence is gathered. Such evidence as we have considered, how- ever, recommends the following tentative conclusion. Inheritance of acquired characters is possible in many organisms. This possibility is more often realized per- haps in the simpler than in the more complex organisms. In the latter, an unusual set of conditions is required, such that the well-insulated germ plasm will be reached. From the point of view of the geneticist, the small amount of inheritance of acquired characters that might take place is usually negligible. The geneticist deals almost entirely with characters the origin of which is either entirely unknown or 'Spontaneous" (mutation) and not clearly traceable to any specific environmental conditions. For the evolutionist, however, this phenom- enon becomes very significant. The recent appearance of seemingly irrefutable instances of inheritance of acquired characters, taken together with the growing conviction that mutation does not provide the type of change necessary to account for progressive evolution (see chapter on ' 'Mutation"), is leading to a revival, in modified form, of the Lamarckian view. The Inheritance of Acquired Characters 35 LITERATURE CITED 1. Babcock, E. B., and Claussen, R. E., Genetics in relation to agriculture. New York: McGraw Hill Book Co. 1918. 2. Bateson, W., The progress of Mendelism. Xature 104:214- 216. 1919. 3. Blaringhem, L., Production par traumatisme d'unc forme nouvelle de JMais a caryopses multiples, Zca Mays var. poly- sperma. Compt. Rend. Acad. Sci. Paris 170:677-679. 1920. 4. BoLLEY, H. L., The importance of maintaining a constant elimination factor in plant breeding. Ann. Rep. .Vmer. Breeders Assoc. 8:508-514. 191 2. 5. Bonnier, Gaston, Nouvelles observations sur les cultures experimentals a diverses altitudes et cultures par semis. Rev. Gen. Bot. 32:305-326. pis. 2. figs. 4. 1920. 6. Castle, William E., Genetics and eugenics. Cambridge. 1920. 7. DoNCASTER, Leonard, An introduction to the study of cytol- ogy. Cambridge. 1920. 8. Engler, Arnold, Influence of source of seed. Jour. Heredity 5:185-186. 1914. 9. Guyer, M. F., and Smith, E. A., Studies on cytolysins. II. Transmission of induced eye defects. Jour. Exp. Zool. 31: 171-223. pis. 4. figs. 7. 1920. 10. , Experiments with typhoid agglutinins in rabbits. Anat. Rec. 20:214. 1921. 11. Hansen, Emil Chr., Recherches sur la physiologie et la mor- phologic des ferments alcoliqucs. XllL Xouvellcs etudes sur des levures de brasserie a fermentation basse. Compt. Rend. Carlsberg 7:179-217. figs. S. 190S. 12. Herrick, C. J., Irreversible differentiation and orthogenesis. Science 51:621-625. 1921. 13. Jennings, H. S., Variation in uniparental reproduction. Amer. Nat. 56:5-15 1Q22. 14. Long, W. H., Influences of the host on the moiphological characters of Puccinia Ellisiatia and P. Amlropogonis. Jour. Agric. Research 2:303-319. 1914. 36 Outline of Genetics 15. Mayr, H., Waldbau auf naturgesetzlicher Grundlage. Berlin. 1909. 16. MuNNS, E. N., Effect of fertilization on the seed of Jeffrey pine. Plant World 22 : 138-144. 1919. 17. PouLTON, Shonland, Shipley, Weisaiann, on heredity. Clar- endon Press. 1 89 1. 18. Walter, H. E., Genetics. New York. 1913. 19. WiLLL\MS, J. Lloyd, Studies in Dictyotaceae III: Periodicity of sex cells in D. dichotyma. Ann. Botany 19:531-560. 1904. CHAPTER III MENDEL'S LAW Mendel's law is the basis of all work in genetics, and should be understood from its original statement to its somewhat complex development. In 1865, Gregor Mendel (3) published in the proceedings of a local scientific society the result of eight years of breeding experiments. The publication was so obscure that scientific men, in general, did not see it, and, in addition to this, Darwinism was at that time absorbing the atten- tion of biologists. For these two reasons, Mendel's work remained unnoticed, and of course unappreciated, until it was discovered in 1900 and became the great classic of genetics. Its influence, therefore, dates from 1900 rather than from the year of its publication. The substance of Mendel's experiments is as follows. Wishing to discover the contributions of each parent to the make-up of their progeny, he chose for his work the simple garden pea, which would breed rai)idly, and ex- hibited well-marked varieties. To magnify his results, he secured hybrids by crossing distinctly different t\pes of peas, and to avoid confusion he considered on\y one character in each experiment. For example, he crossed peas which contrasted in character of lieight, of flower color, and of seeds. In all cases he obtained similar results, so that a single example will sulHce. l^urther- more, he discovered that it made no difference whether the staminate parent was a dwarf and the pistillate 37 T X i I D } I T D > T T T T D T T 1 D f Fig. 2. — Diagram illustrating 38 Outline of Genetics parent tall, or vice versa, and so for all the characters used. In other words, what are called reciprocal crosses gave the same results. The progeny of a tall parent and a dwarf parent were all tall. This generation is known as the first hybrid or the Fi generation. When this generation was inbred, the progeny was made up of tall and dwarf indi\dduals in a ratio of 3:1. This generation is known as the second hybrid generation or the F2 generation. The dwarf forms of the F2 generation subsequently bred true, producing only visible results of Mendel's experi- ^^^^^^^ q^ ^^^ ^^^j ^^ ments. Cross between tall parent (T) and dwarf parent (D) gives one-third bred true and hybrid progeny which are all tall; tWO- thirds split up in juSt hybrid progeny inbred gives 3:1 guch a 3:1 ratio as did ratio in second hybrid generation; ... ,. . inbreeding each of these four indi- their unmediate parents of viduals separately gives for third the Fj generation. This is hybrid generation results indicated expressed diagrammati- in bottom line. ,, . ^ cally m fig. 2. Mendel's theoretical explanation of this behavior involved three distinct theses. I. Independent unit characters. — This means that an organism, although representing a morphologi- cal and physiological unity, from the standpoint of hered- ity is a complex of a large number of independent herit- able units. Thus if one pea plant is tall and another one is dwarf, the behavior of the hybrid produced from them with reference to this character will be the same, no MendeW Law 39 matter what other characters the i)arent-plants may have had. In other words, the characters are independent units, unaffected by other characters or units. The character of tallness from a tall plant with wrinkled seeds or purple flowers will act just the same as from a tall plant with smooth seeds or white flowers. 1\illness is a unit, and its behavior in inheritance is independent of all other units. 2. Dominance. — In the germ plasm there are certain determiners of unit characters which dominate during the development of the body, causing these characters to dominate over others and thus become visible. The characters dominated over and thus not allowed to express themselves are called recessive characters. These recessive characters may be present in the germ plasm, but cannot express themselves and become \'isible as long as the dominant characters are present. When a dominant character is absent, however, its recessive alternate is free to express itself and become \'isible. For example, in the case of tall and dwarf peas, tall- ness is a dominant character and dwarfness is its alter- native recessive. When a dwarf appears, therefore, there is present no dominant tallness to suppress it. In the Fi generation all the individuals were tall because, although they had all received the recessive character of dwarfness from one of the parents, they had received the dominant character of tallness from the other i)arent, and so dwarfness did not appear in an}- of them. Such pairs of alternative characters are now commonly called atlelomorphs. Thus tallness and dwarfness are allelo- morphs in the pea, one dominant over the other, which is therefore recessive. 40 Outline of Genetics 3. Purity of gametes. — A gamete can contain only one of two alternative characters. For example, it may contain the character for tallness or for dwarfness, but not both. In other words, allelomorphs cannot be repre- sented in the same gamete. If the gamete having the character for tallness unites with one having the char- acter for dwarfness, the resulting zygote will have both, but will produce a tall individual because tallness is dominant to dwarfness. When this tall hybrid produces gametes, however, one-half of them will contain the character for tallness and one-half of them the character for dwarfness. Thus the alternative characters are ''segregated" in gamete formation, and no gamete will have both characters. These three theses, independent unit characters, dominance, and purity of gametes (due to segregation), make up the theoretical explanation of Mendel's law. Independent unit characters was of course a necessary conception. It was original with Mendel, and has also been original with other investigators, but this concep- tion does not represent the essential feature of Mendel's law. The idea of dominance had been somewhat vaguely proposed before Mendel's time. In the old literature on animal breeding one meets theories of ''prepotency," which were proposed again and again before the dis- covery of Mendel's work in 1900. In any event, Men- del was the first to formulate definitely the theory of dominance among unit characters. It should be realized also that dominance is not an essential feature of Men- del's theory. Many cases are known in which domi- nance fails, but in other regards the Mendehan inherit- ance is strictly followed. Mender s Law 41 The essential feature of Mi;ndi;i/s theor}- is his con- ception of the purity of gametes, brought about by the segregation of alternative characters. With Mendel this was a purely theoretical scheme, but since his time cytological investigation has discovered an actual physical mechanism which exactly satisfies the require- ments of Mendel's scheme. Every li\ang organism is composed of cells, and these cells are endowed with nuclei. Every nucleus contains a certain number of darkly staining bodies known as chromosomes. The number of chromosomes is always tlie same for a gi\'en species. At the cell divisions which take place in con- nection with the growth of the body, each chromosome is very carefully divided in half, so that the nucleus of each daughter-cell has exactly the same equipment of chromosomes as the mother-nucleus. The exactness of this division in itself suggests that the chromosomes are the bearers of hereditary characters, since none of the other cell constituents seems to be so accurateh' divided at cell division. Even more significant is the behavior of the chromosomes in connection with gamete formation. At that time it becomes evident that the chromosomes exist in pairs; thus there is always an even number of chromosomes in every bcxh' cell of an organism. The two components of each pair of chromo- somes are always morphologically identical. When tlie organism forms gametes, a cell dixision takes place which is fundamentally different from the preceding cell divi- sions. At this division no splitting of the individual chromosomes takes place; instead, the chromosomes Hne up in pairs and the nature of the division is such as to draw apart the components of each ])air. This is 42 Outline of Genetics known as the reduction division, for each of the resulting nuclei has the reduced number of chromosomes, just half of the characteristic number in the body cells. It is important to remember that this reduction is not indis- criminate, but always involves a separation of the two components of each chromosome pair. It is the reduc- tion division that gives rise to the gametes. Gametes, therefore, are characterized by the reduced or haploid number of chromosomes, in contrast with the body cells which have the diploid number. Gametes have just one representative of each chromosome pair that appears in the body cells. When two gametes unite at fertihzation there is, of course, a return to the diploid number in the resulting zygote. This is exactly the mechanism required by Mendel's scheme, on the assumption that the chromosomes are the bearers of hereditary characters. So much data has accumulated to justify this assumption that it will be treated as an established fact in the subsequent descrip- tions. The chromosome mechanism may be applied to the case in hand as follows. For convenience, we will assume that the nuclei of the body cells in Mendel's peas have each four chromosomes (two pairs). This is represented in fig. 3. In the case of a tall plant, two (one pair) of the four chromosomes carry the character for tallness, that is, something that determines the production of the taU character in the somatoplasm. This unknown something is called by various names in the literature of genetics; for the present we shall refer to it as a deter- miner. In our illustration, therefore, two of the four chromosomes carry the determiner for tallness. Mendel's Law 43 Fig. 3 shows a somatic cell with the diploid niuiiber of chromosomes. Tii the fomiation of gametes, this number is reduced to the haploid number, which is in this case two. The diagram shows that the reduction division separates (segregates) the two chromosomes carrying the determiner for tallness, so that each gamete contains one. This occurs for the other characters as well as for that of tallness. From the tall plant, therefore, © © © © Tall Parent © © © © Dwarf Parent Fig. 3. — Diagram illustrating behavior of chromosomes in Mexdel's cross of tall and dwarf peas. Large rectangular figures, nuclei of zygotes or mature individuals; large circles, gametes; snij^l circles within zygotes and gametes, chromosomes; letters on chromosomes, determiners {T, tallness; Z>, dwarfness). all the gametes will contain the detenniner for t;\llness, and from a dwarf plant all of the gametes will contain the determiner for dwarfness. When these two indi- viduals are crossed, the zygote will contain both deter- miners, and these two determiners will be transmitted together in the succeeding cell generations. The indi\id- ual developed from such a zygote will of course be tall, but at the same time it will be carr}ing a recessive deter- miner for dwarfness, and this fact will be shown by its behavior in breeding. The result of inbreeding such hy- 44 Outline of Genetics brids is indicated in the accompanying diagram (fig. 4), which represents the chance matings of the two kinds of gametes. The obvious results are three tall individuals and one dwarf. This is the so-called monohybrid ratio, which means the ratio when a single pair of allelomorphs is considered. Before discussing the further development of Mendel's law, it will be necessary^ to explain some of the terminol- ogy of genetics. When each gamete carries the same kind of determiner the resulting zygote is said to receive a double dose; when a zygote receives only a single such determiner it is said to receive a single dose. In fig. 4, one zygote receives a double dose of tallness and two others a single dose. These phrases are more or less common in the literature of the subject, but the more frequent terminology is as follows. When two similar gametes unite to form a zygote it is called a homozygote; when the two pairing gametes are dift'erent the zygote is called a heterozygote. Using this terminology, it is e\'ident that the 3 : i ratio of the F2 generation is really a 1:2:1 ratio, as follows: i homozygote for the dominant char- acter, 2 heterozygotes, and i homozygote for the recessive character. The 1:2:1 ratio, therefore, is the significant one and appears as a 3 : i ratio only because of dominance. In the experiment represented in fig. 4, three tall indi\dduals appear in the F2 generation. Superficially the individuals look alike, but it is realized that one differs from the other two in germinal constitution, for one will produce only one kind of gamete, while the other two will produce two other kinds. To indicate this situation JOHANNSEN (2) has introduced some appropriate termin- ology. Organisms which appear to be alike, regardless Menders Law 45 of their germinal constitution, arc said to be phciiotypi- cally alike, or to belong to the same phoiolypc. On the other hand, organisms having identical germinal con- stitution are said to be genolypically alike, or to belong to the same genotype. From the standpoint of pheno- types only, Mendel's F2 generation shows the 3: i ratio; but if genotypes are considered, it shows the 1:2:1 ratio. ® ® Fig. 4. — Diagram illustrating behavior ot lirst generation (F,) when inbred. Illustrates meaning of "segregation" and "purity of gametes," and how chance matings of Fi gametes result in 3: i ratio in F2 generation; dwarf individual produced only by zygote in lower right- hand corner. In other words, this group of foniis contains two pheno- types but three genotypes. J- Referring again to fig. 4, several things ma>' bet inferred. It can be seen what will happen in the F, generation when the F^ individuals arc inbred. The dominant homozygote will ]:)roduce onl}- dominant homozygotes in the F3 generation; the two hctcrozygotcs will split up in the F3 generation in the same 1:2:1 ratio as did their hybrid parents of the Fj generation; and 46 Outline of Genetics the recessive homozygote will produce only recessive homozygotes. It is interesting to consider what would happen if a heterozygote were crossed with a homozygous recessive. It should be obvious that one half of the progeny would be pure recessives, while the other half would be hetero- zygotes, that is, there would be a i : i ratio. Thus far we have considered only what is called the monohybrid ratio, that is, the ratio obtained from one pair of contrasting characters, such as tallness and dwarf- ness. The next step is to consider the dihyhrid ratio. Mendel also used contrasting flower colors, finding, for example, that red flower is dominant to white flower. Introducing this pair of contrasting characters into the situation we have been considering, the dihybrid ratio will be the result. Crossing a tall, rei-floweredjndivid- ual with a dwarf, white-flowered individual, it is evident that all the Fi or first hybrid generation will be tall, red- flowered individuals, since both of these characters are dominant. In the F2 generation, however, the following ratio will appear: 9 tall, red; 3 tall, white; 3 dwarf, red; and I dwarf, white. This 9:3:3:1 is the dihybrid ratio, the explanation of which is indicated in fig. 5. The question may be raised why the characters for tallness and redness are not represented on the same chromo- some. If they were, the result would be a simple mono- hybrid ratio, except that the tall individuals would always be red flowered, and the dwarfs would always be white flowered. The possibility of one chromosome carrying two different determiners will be considered later, but at present we shall assume that these deter- miners are on different chromosomes. MendeVs Law 47 Fig. 5 shows that we are deaUng witli two homozy- gotes, each producing only one kind of gamete, so that all the F, progeny are similar, both phenot>pically and © © © © TaU Red Parent © © ©. © Dwarf White Parent © © © © Gametes ©,© ®© ©3© ®© D) (T R) (R ®© ©,© ® ® ©,o© ®® ® ® ©3© ©,© © © D) D R R ©:© ©.© R^ (w) ©,© ® ® ©,© ©"© ® ® Fig. 5. — Diagram illustrating ciihybrid ratio. Upper part shows how original parents were crossed to give Fi hyi)ri(l; lower i)art shows Fi hybrid producing four kinds of gametes; chance matings among these gametes, when Fi is inbred, results as indicated in the large set of squares and explains the 9:3:3:1 ratio in the Fa generation. 48 Outline of Genetics genotypically, that is, with the same appearance and the same germinal constitution. Each of these Fi indi\dd- uals will produce four kinds of gametes. The possible combinations of these gametes that will occur when the Fi is inbred are expressed by the checkerboard. The resulting F2 involves four pheno types, as follows: nos. I, 2, 3, 4, 5, 7, 9, 10, 13 are tall, red-flowered individuals; 6, 8, 14 are tall, white; 11, 12, 15 are dwarf, red; 16 is dwarf, white. This acounts for the 9:3:3:1 ratio. It will be noticed that nos. 1,6, 11, and 16 are homozy- gotes and therefore will breed true; but the rest are heterozygotes, either for one pair of characters or for both, and these will split into various types upon further breeding. The higher polyhybrid ratios run into quite a string of terms, but involve no further new principles. For example, the F2 phenot^-pic ratio for the trihybrid is 27:9:9:9:3:3:3:1, invoKdng nine pheno types (and 2 7 genotypes), but it can easily be worked out by the same method as was used for the dihybrid. Thus far we have been considering Mendel's law in simple form, and have enlarged but little upon Mendel's original statement. The value of the law is apparent. Upon its republication in 1900, it was taken up by biolo- gists, and numerous breeders set to work to test it. As a consequence, data for and against it began to accumulate. As might be expected, there was much apparent evidence against the law, but as geneticists developed a better conception of the mechanism, the contradictory evidence was explained away. Almost every type of inheritance has now been explained according to Mendel's law. A few of the more important cases will be presentecj, Menders Law 49 PRESENCE AND ABSENCE HYPOTHESIS Tliis may be regarded as a new method of Mendelian thought. It was first suggested by Correxs (i), but later was worked out in detail by other geneticists. It is merely a different way of regarding the MendeHan mechanism. For example, in the case of a hybrid obtained by crossing tall and dwarf parents, the result had been explained by Mendel as due to the fact that one chromosome bears a determiner for tallness and the other one of the pair carries the detemiiner for dwarfness. In other words, each one of a pair of allelomoq)lis is represented by a determiner, two detemiiners thus being present. Dwarfness in this case would be the result of the interaction of that determiner and its enviromuent during the development of the body; and the same for tallness. When both were present, howe\'er, the con- ception of the situation w^as as follows. The determiner for dwarfness, setting up its usual series of reactions, early became paralyzed by the determiner for tallness or its products. This result was called the dominance of the character for tallness. It was as if the detemiiner for tallness completely prevented the acti\'ity of the determiner for dwarfness. This conception was apjxir- ently borne out by the facts and was the ex]3lanation of the mechanism generally accepted. According to the presence and absence h\'])othesis, however, the situation is looked at from a different point of view. Tallness is the result of a detemiiner, but dwarfness is merely the result of the absence of the detemiiner for tallness. The dominant character is produced by an inheritable detemiiner, but the recessive character appears only when the dominant detemiiner 50 Outline of Genetics is lacking. This conception has some evident advan- tages and may modify the previous Mendelian diagram, as shown in fig. 6. This appears to be a simpler mech- anism to account for the phenomenon called dominance. In the case of the dwarf form, there is a "normal" course of development; in the case of the tall parent or hybrid, however, an additional determiner stimulates cell growth, or cell division, or both. This hypothesis introduces some additional termin- ology suggested by Bateson. In our illustration, the Dwarf Parent Gametes Fig. 6. — Diagram showing how the original scheme must be modi- fied to satisfy the presence and absence hypothesis. tall parent has two determiners for tallness and therefore Bateson calls it duplex, having a double dose. For the same reason, the Fi individuals, having only one determiner for tallness, he calls simplex. According to the same terminology, the dwarf parent is nulliplex with respect to its character of tallness. Additional advantages of the presence and absence hypothesis will appear later in connection with a con- sideration of blending inheritance and of cumulative fac- tors in inheritance. Attention, however, should be called to the fact that those who accept the presence and MendeVs Law SI absence hypothesis do not use the fomi of notation thus far used in explaining Mendehan inheritance. Assume that T is used to express the detenniner for tallness, the same letter (/) is used to express its absence. P^or example, instead of using D for dwarfness, / is used for ''lack of tallness" (fig. 7). It is a matter of conven- ience to have a symbol to re])resent the recessive, the absence of something that is present in another indi- vidual. Dwarf Parent Cameti es ■ Fig. 7. — Diagram showing how presence and absence scheme is actually used, with small letter representing "absence." In summary, the essential difference between the presence and absence hypothesis and that of dominant and recessive is that in the former case the recessive determiner has no existence at all, while in the latter case it exists, but is in a latent condition when associated with the dominant. As a matter of fact, some of the later investigations have revealed cases that can hardly be accounted for by the presence and absence hypothesis. In spite of these recogni/.ed e.\ceptions, however, the scheme of notation employed under the presence and absence hypothesis has proved so convenient that it is almost universally employed. 52 Outline of Genetics The checkerboard is an excellent method of depicting the mechanism at play, but the same results can be obtained much more quickly and just as safely by another method. Dihybrid (and other polyhybrid) ratios can be obtained by multiplying together the ratios of the monohybrid components (for this really amounts to the same thing as the checkerboard). In the present case, one monohybrid component gives an F2 ratio of 3 tall:i dwarf, while the other monohybrid component gives 3 red : i white. (3 tall: I dwarf) times (3 red:i white) equals 9 tall, red 13 dwarf, red 13 tall, whiten dwarf, white. The student will find it a very helpful exercise to work out for himself the various phenotypic ratios that will be produced by inbreeding the various genotypes shown in the Fa checkerboard, and by crossing them in various ways. In such problems the method of procedure is to work out separately the results for the different monohybrid components and then to synthesize these results. The synthesis in this case amounts to simple multiplication. First of all, one must familiarize himself with the various possible ratios encountered in connection with monohybrid crosses. These are indicated in the following simple table. (In every case the student must satisfy himself as to how the Mendelian mechan- ism brings about these results.) Phenotypic ratio ^^°^^ Red White i?i?Xanything i : o Rr XRr 3:1 Rr Xrr i : i rr Xrr o : i The solution of dihybrid problems then becomes a simple matter. For example, take the following cases: TtRrXttRR. The monohybrid ratio as regards height is i tall : i dwarf, while the flower color ratio is i red:o white. Multiplying these together gives the dihybrid result, i tall, red: i dwarf, red. TtRrXTtrr. The height ratio is 3 tall:i dwarf, while the color ratio is i red:i white, which results in the dihybrid ratio, 3 tall, red:i dwarf, red: 3 tall, white: i dwarf, white. Menders Law 53 BLENDS This t^pe of inlieritance when first discovered was thought to be in direct conihct w-ith Mendel's law. It is a case in which dominance seems to fail, for the two alternative characters both express themselves and the result is an average between them. It is easy to explain this situation in accordance with the presence and absence h}'pothesis without any violation of Men- del's law. The classic example of blending inheritance was presented by Correns (i) in breeding work upon Mira- hilis Jalapa, the common four-o'clock. Correns crossed red-flowered and white-flowered varieties, and all the hybrid progeny had rose pink flowers. This was a color blend, distinctly intermediate between the colors of the two parents. The Fi generation, therefore, seemed to contradict Mendel's law in that one color character was not completely dominant over the other. The real situation, however, appeared in the F^ generation obtained by inbreeding individuals of the Fi generation which showed the blend. By inbreeding the pink hybrids Correns obtained the perfect 1:2:1 ratio, that is, I red like one grandparent, 2 pink like the hybrid parent, and i white like the other grandparent. Segre- gation was evidently taking place, the only unusual thing being the appearance of the Fj individuals, and that was explained immediately as failure of dominance (see fig. 8). The question this introduces, therefore, is that of a mechanism which would account for such a result. The easiest explanation offered is that the red parent was a homozygote for redness (double dose) and the hybrid 54 Outline of Genetics a heterozygote (single dose) ; the inference is that a single dose produces pink while a double dose pro- duces red. A theoretical explanation of this occasional difference in the result of double and single doses is as follows. Imagine that the body cells of a plant have a certain capacity for expressing heredi- tary characters. In such a case, just as a given quantity of sol- vent can dissolve only a given amount of solute, so the body cells Red Parent Gamete Gamete White Parent (R)p,nkQ ^ggs wm"^ 0P.® R) ( r Pink r ) ( r White Fig 8. — Diagram illustrating blending inheritance, discovered by CoRRENS in Mirabilis Jalapa. can express hereditary characters only to a definite limited extent. In the four-o'clock a single dose of redness may be thought of as half-saturating the body cells, while a double dose completely saturates them. In cases showing a complete dominance, how- ever, a single dose completely saturates the cells and a double dose can do nothing more. This analogy assists in visualizing, on the one hand, the necessary mechanism of blends (apparent failure of dominance), and, on the other hand, that for cases of complete dominance. MendeW Law 55 Problems dealing with determiners for which dominance is lacking differ from those where dominance is present only in so far as the monohybrid ratios must differ. Cross Dominance present Red White Phenotypic ratio Dominance lackinR (Here phenotypic and gcno- typic ratios are the same) Red Pink White RRXRR I RRXRr I RRXrr i Rr Xrr 3 Rr Xrr i rr Xrr o o o o I I I I I o I o o o I I 2 I o o o o I I I. For example, take the following case, where dominance is pres- ent for the height character and absent for the color character: TtRrXttRr. The monohybrid ratio as regards height is i talhi dwarf, while the color ratio is i red 12 pinkii white. Multiplying these together gives the dihybrid result, i tall, red:i dwarf, red: 2 tall, pink: 2 dwarf, pink: i tall, white: i dwarf, w^hite. LITERATURE CITED 1. CoRRENS, C, Die neuen Vererbungsgesetze. Berlin. 191 2. 2. JoHANNSEN, W., Elemente der exakten Erblichkeitslehre. Jena. 1909. 3. Mendel, G., Versuche iibcr Pflanzen-Hybriden. \'erh. Naturf. Vereins in Briinn. 4:1865. CHAPTER IV THE FACTOR H\TOTHESIS Mendel concluded that each plant character depends upon a single detenniner. Inheritance, however, has proved to be a much more complex phenomenon than was indicated by Mendel's peas. Ratios have appeared that were puzzling, and geneticists have been forced to the conclusion that there may be a complex of deter- miners for a single character. This conception is known as the factor h\^pothesis, and much of the growing com- plexity of genetics has developed around this hypothesis. Previously we have used the word ''determiner," imply- ing Mendel's idea that a single determiner is responsible for the development of a plant character, and this has been true of the examples of inheritance pre\iously considered. It is understood, now, however, that a character is frequently determined by the interaction of two or more separately heritable factors, and hence the factor hypothesis. The distinction between factors and determiners should be clear. In case only one heredi- tary unit is involved in the production of a character, this unit should be referred to as a determiner; in case two or more units interact in the production of a char- acter, these 3i,xe factors .^ ^ This distinction of terms has pedagogical value, but is frequently violated in the literature, where "factor" is frequently used in the sense of "determiner." A less restricted term, gene, refers to the heredi- tary unit without implying whether it acts as a factor or simple determiner. 56 The Factor Hypothesis 57 I. Complementary factors. — This is the simplest expression of the factor hypothesis; it may be illustrated by some of East's work (3). Crossing red-grained and white-grained corn, this investigator obtained an Fi progeny which was all red. This would suggest that the F2 generation would show 3 red to i white; but instead it showed 9 reds to 7 whites, which might seem to violate the Mendelian method of inheritance. It is quite in accord with Mendel's law, however, if we consider that two complementary factors are necessary to produce the red character, and that each of these factors is inher- ited separately. Such a situation would give a dihybrid ratio, as indicated in fig. 9. It will be seen that, out of the 16 individuals in the F2 checkerboard, 9 will be red, for they alone contain both complementary factors; the other 7 will be white. The situation is thus explained by the dihybrid ratio; but, although only one character is involved, that character depends upon two comple- mentary factors. Another situation is worth noting. No. 6 of the checkerboard is white because it contains only one of the necessary factors; no. 11 is white for the same rea- son, but its germinal constitution is just the opposite. What would happen if these two were crossed ? There is only one possibiUty, since each is a homozygote produ- cing only one kind of gamete. The result would be red, and thus a cross between two whites would produce only reds. What would be the result if nos. 6 and 15 were crossed, the former being a homozygote and the latter a heterozygote ? It is obvious that the resulting progeny would be one-half red and one-half white. The same result would be secured in crossing nos. 11 and 14. A S8 Outline of Genetics cross between nos. 14 and 15, both of which are het- erozygotes, would produce 3 white to i red. These Fig. g. — Diagram illustrating behavior of complementary factors in cross between red-grained and white-grained corn. R and C must both be present to produce red-grained corn. illustrations show how differently the same phenotype may behave in inheritance. In each case 2 whites were crossed, that is, the same phenotypes, but 3 different The Factor Hypothesis 59 ratios were obtained because the genotypes were dif- ferent. The striking feature of this situation is that one can cross two whites and get a red. This gives an insight into the so-called phenomenon of reversion. For ex- ample, in the course of numerous breeding experiments Bateson (i) obtained two strains of white sweet peas, each of which when normally *'selfed" bred true to the white color; but when these two were artilicialh' crossed all of the Fi progeny had purple flowers, like the wild Sicilian ancestors of all cultivated varieties of the sweet pea. This appeared to be a typical case of reversion. Further breeding, however, showed that this was just such a case of complementary factors as we have been considering. One of Bateson's white strains had one of the factors for purple flower color and the other strain had the other factor. It is interesting to note that if an investigator should cross homozygote no. i with homozygote no. 11, the Fx and F2 results would lead him to conclude that the red character was due to a simple Mendelian determiner. R would remain a ''determiner" until a strain of corn was discovered which lacked the C factor; crosses with such a strain would reveal the real mechanism of the situation, and thereafter R would be known as a "factor." Complementary factors have been defined and the method of their inheritance described, but is there any mechanism to explain the situation ? A suggestion may be obtained from plant chemis- try (2). The most prominent group of pigments in plants is the group of anthocyanins, which are produced as follows. Plants contain compounds called chromogens, which are colorless them- selves, but which produce pigment when acted upon by certain oxidizing enzymes or oxidases. This would provide a mechanism to account for the behavior of complementary factors. If one of East's white strains of corn contained a chromogcn capable of 6o Outline of Genetics producing red, but lacked the necessary oxidase, it would remain colorless. If the other white strain contained the oxidase but no chromogen, it too would be colorless. In crossing them, however, chromogen and oxidase would be brought together and a red- grained hybrid would be the result. In breeding, such red-grained individuals would of course give red and white progeny in a ratio of 9:7, as explained in connection with East's corn. It should be realized that the foregoing is no more than a suggestion. So far as the genetics of the situation is concerned, complementary factors may be regarded as an established fact; but what either one of these factors actually amounts to, in physiological terms, has as yet only been guessed at. In fact, it would be safe to state that there is no known case where the exact physico-chemical nature of any factor or determiner has been demonstrated. In the foregoing instance, neither chromogen nor oxidase may be the effective units involved. There are, how- ever, plenty of possibilities within the field of chemistry where the bringing together of two inert substances initiates a reaction sufficient to result in a new plant "character." Within limits, hybridizing is very much like mixing chemicals in a test tube. The origin of complementary factors is an interesting field of speculation. Did they originate together or separately? A natural inference would be that they originated together, for neither would be of any use without the other. It should be remembered, however, that the Darwinian idea of usefulness as explaining the occurrence of everything in a plant is frequently inadequate. One must think rather of a plant as a complex physi- co-chemical laboratory. No one claims that all chemical reactions are useful; they are simply inevitable; and plant characters are the result of chemical reactions and physical necessities. The other alternative is to suppose that these factors origi- nated independently in the plant's history. In this case, of course, the first to be produced would remain functionless until its complement came into existence. This might be an explana- tion of what have been called "latent" characters. Also, not only might they have originated at different times but in different varieties or species. In this case, if natural hybridizing should The Factor Hypothesis 6i bring them together the result would be the appearance of a new character; a few authors (notably Lotsy 5) believe that this is a very important factor in the origin of species. The behavior of the red-grained and white-grained corn serves as an introduction to the factor hypothesis and as an illustration of one of the important t}'pes of factor interactions. Complementary factors are always to be recognized by the following behavior. A comple- mentary factor interacts with a dissimilar factor to produce a particular character. 2. Supplementary factors. — A supplementary fac- tor interacts with a determiner (or factor complex) to modify the character produced by the latter. An excel- lent illustration of this factor type can also be provided by the inheritance of color in grains of corn (East 3). Interacting with the factors R and C is a third factor P} This P factor is inherited quite independently of the other two, but whenever it is present the red color becomes modified to purple. The behavior of this factor is revealed when we consider the phenotypes to which the following homozygotes belong. Corn of the formula PPRRCC has purple grains, ppRRCC is red grained, while PPRRcc and PPrrCC are both white grained. From these facts, one can draw the following conclu- sions: (i) when P, R, and C are all present a pur])le grain results; (2) red color can appear only when P is absent; (3) P itself is entirely ineffective in the absence of either R or C. * In the literature on inheritance in corn, this factor is referred to by the symbol Pr, its absence being denoted b>' pr (see work of Emerson and others). It is to avoid possible confusion in the mind of the student that it is referred to by the single letter in the present text. 62 Outline of Genetics These principles reveal the nature of supplementary factors. Unlike complementary factors, they never produce characters, but merely modify characters already present. Otherwise they follow the same principles of independent Mendelian inheritance with which we are now familiar. The student will find a very useful exercise in the solution of various problems which involve simultaneously several types of factors. In solving such problems, it is neither necessary nor advisable to employ any "rule-of-thumb" method. As usual the checkerboard can be depended on to provide an accurate solu- tion, but this is too slow and cumbersome. Instead, the problems can always be solved rather easily by keeping in mind, at all times, the fundamental mechanism of inheritance that is at play and proceeding with the solution in a logical, orderly manner. (If the student simply remembers the Mendelian mechanism he can really work out his own methods of solving the problems.) In attacking polyhybrid problems, the first principle to remember is to work out separately the solution for each pair of allelomorphs, and then to put together these monohybrid solu- tions. This ''putting together," in the case of simple determiners, amounts merely to multiplication, as was described before, but where factors are at play other methods of putting together are necessary. In every case, the method of putting together the monohybrid solutions is a rather obvious one, and is clearly indi- cated by the definition of the factor type with which one is dealing. A few examples involving the P, R, and C factors will serve to illustrate. In these cases, one must consider the P factor last, since it is effective only when both R and C are present. PPRyccXPpRRcc. Considering first the C factor, one finds that a o: i ratio will result; that is, all of the progeny will lack the C factor. Since this is the case, one need go no farther with the solution, since absence of C is sufiicient in itself to insure that all of the progeny will be color- less, or white. PpRrCcXppRrCC. The Factor Hypothesis 63 Considering first the C factor, one finds that a i :o ratio will result; that is, all of the progeny will have the C factor. So far as C is concerned, then, all of the progeny may be colored. Taking up now the R factor, it is evident that this will bring a 3 : 1 ratio; that is, f of the progeny have both C and R and will therefore be colored (whether red or purple will be decided later) ; I have C but lack R and will therefore be colorless. The final fate of this I is thus decided, and one need give no further atten- tion to this group. The f , however, is eligible to be carried on for further consideration under the P factor. The ratio produced by the P factor is 1:1; that is, ^ have P and I lack it. So far as phenotypes are concerned, however, the present fractionation applies only to the f that was carried over, making | with and | without P. Summarizing, } have C but lack R and are therefore colorless; I have C and R but lack P and are therefore red; | have C, R, and P and are purple. The resulting phenotypic ratio is, therefore, 3 purple: 3 red: 2 white. PpRrCc ''selfed" (equivalent to PpRrCcXPpRrCc). Considering the C factor, the result is a 3 : i ratio, or | with and \ without C. The latter group, \ is hereby dropped, since it must be colorless, while the f is carried on for further con- sideration. The R factor also gives a 3 : i ratio, or f with and \ without R. But this fractionation applies only to the f that was carried on. Therefore we have f Xf or A which has both C and R, and f Xj or -i\- which has C but lacks R. This x\ is hereby dropped, since it also must be colorless, while the 1^0 is carried on for further consideration. The P factor gives 3:1, or f with and \ without P. This fractionation applies to the iV which was carried on, so that we have iVX 4 or H which has C, R, and P and is purple, and ^i Xi or i;^ w^hich has C and R, but lacks P, and is therefore red. Summarizing, \ plus i'',; or (jv is colorless; ,m is red; and l\ is purple. The phenotypic ratio, then, is 27 purple:9 red: 28 white. 64 Outline of Genetics (Any such mass of fractions is readily handled on the black- board or scratch pad by considering the C, R, and P ratios under separate columns, and scratching out fractions as they are dis- posed of by being carried on. At the end, if the procedure has been orderly, the proper fractions will be found under the proper columns.) Other methods of solving these problems could be worked out. All that is necessary in any method is clear vision of the Mendelian mechanism, analysis into monohybrid components and separate solution of each, followed by an orderly synthesis, the method of synthesis always being indicated by the nature of the factors with which one deals. As before, the physiological mechanism accounting for the behavior of the P factor has been only guessed at. It has been found that the purple pigment is produced by the same substance as the red, but represents a higher state of oxidation, which sug- gests the following possibility. C is oxidized by 7? up to a certain point, where red is produced; P, an additional enzyme, is capable of oxidizing the red pigment still further to purple. P is incapable of attacking the original chromogen, but when R carries the attack to a certain point, P can function and carry the oxidization further. As a consequence, P without R gives white grains, while R gives red grains only in the absence of P. 3. Inhibitory factors. — ^An inhibitory factor pre- vents the action of some other determiner or factor. This factor type also can be illustrated in connection with inheritance of grain color in corn (East 3). If one were to cross a purple-grained race of corn, having the formula PFRRCC, with any white-grained race, he would expect all of the Fi generation to be purple. In this case, every gamete produced by the purple parent would have the formula PRC. As a consequence, no matter what might be the formula of the white parent, every Fi individual would have to have at least one dose of P, R, and C, and this in itself should be sufficient to The Factor Hypothesis 65 insure the production of the purple color. As a matter of fact, just this result has been actually realized in practically all such crosses. In one case, however, startlingly different results were obtained. Crossing just such a purple parent as was mentioned with a white- grained race resulted in an Fi generation which was all white. Inbreeding this Fi gave an F2 generation which con- tained some colored grains, but a decided majority of whites. Evidently the colorless condition is dominating over the colored. This would be surj^rising in any case of color inheritance, for we always expect colored to be due to the presence of something that is absent in col- orless. Analysis of the present case revealed the fact that the white-grained race that had been used was homozy- gous for the presence of an inhibitory factor, /; whenever this factor is present no color of any kind can be produced. This readily explains the foregoing results. If the purple- grained race that was used had the fonnula iiPPRRCC and the white-grained race was IIPPRRCC, the result- ing Fi would be IiPPRRCC, which would be pheno- t}^pically white owing to the presence of the / factor. Inbreeding would then result in the following F2: i IIPPRRCC, phenot>TDically white; 2 IiPPRRCC, white; I iiPPRRCC, purple. It is clear that the colorless con- dition is actually dominating, but it is dominating on account of the presence of the / factor. It is evident that the purple and red types with which we had been dealing before must all have been homozy- gous for the absence of /. Corn grains can be colorless for any of three reasons, absence of R, absence of C, or 66 Outline of Genetics presence of /; whereas the colored condition occurs only when three conditions are simultaneously satisfied, presence of R, presence of C, absence of /. Tetrahybrid problems, involving /, P, R, and C can be worked out on the same principles as were previously outlined. It is probably most convenient to consider the / factor first. UPpRrCc X iippRRCc. Considering the / factor, there will be | of the progeny with and ^ without /. This first | is now dropped because it is bound to be colorless on account of the presence of /, no matter what may be the rest of the germinal composition; and in this case it is the latter ^ that is eligible to be carried on for further consideration. The C factor gives f with and \ without C. Applied to the \ that was carried on, this becomes ^Xf or f without / and with C, and |Xi or | without / and without C. This last \ is hereby dropped, since it lacks C and must be colorless; while the f is carried on for further consideration. The R factor gives a ratio of i with:o without R. Conse- quently no further fraction is dropped into the white phenotype at this point, and the whole f is carried on. The P factor gives \ with and \ without P. Applied to the I, this becomes f X| or A which lacks / and has C, R, and P, and is therefore purple; and f Xf which lacks /, has C and R, and lacks P, and is therefore red. Summarizing, | plus | or if is colorless, A is red, and A is purple. The phenotypic ratio is, 3 purple 13 red: 10 white. liPpRrCc ''selfed." / gives f with I:\ without /. Drop the f and carry on the \. C gives f with C:\ without C. Applied to the \ that was carried on, this becomes |Xf or A which lacks / and has C; and |Xi or j^,i which lacks / and C both. Drop the Vo and carry on the A. R gives f with R : \ without R. Applied to the iV that was carried on, this becomes i^Xf or if which lacks / and has both C and R; and AXi or ^^ which lacks /, has C, and lacks R. Drop the 6^4 and carry on the o^. The Factor Hypothesis 67 P gives f with P\\ without P. AppUed to the A that was carried on, this becomes o^ Xf or /5V which lacks / and has C, R, and P, and is therefore purple; and ^aX\ or of.n which lacks /, has C and R, and lacks P, and is therefore red. Summarizing, f plus iV plus ^t^ or .7^,1 is colorless, 250 is red, and 2^5^ is purple. The phenotypic ratio is, 27 purple 19 red: 220 white. Thus three factor types are at play in the inheritance of color in grains of corn. Four distinct factors are interacting, but all are inherited independently and quite in accordance with Men- delian principles. Again, the exact physiological nature of the / factor is not understood, but can be only guessed at. We have assumed that color is produced when an enzyme is present to oxidize a chromo- gen. Enzymes are sensitive; their activities may be affected or completely checked by various agents. Assume that / is such an agent, and the necessary mechanism is provided. When / is present R is paralyzed, so that it cannot oxidize C. 4. Cumulative factors. — These are considerably different from the other types and will be considered in another chapter under the caption "Inheritance of quantitative characters." These four great factor types are really the only ones encountered in genetics, each representing a distinct type of interaction. In the literature of the subject many other descriptive titles are given to factors, but no fundamentally new mechanisms are introduced. A few more words might be said on the three factor types that have already been considered. Of these three, the rarest type is the inhibitory factor, complementary and supplementary factors being quite common. It is not surprising to lind that true doniiiKint inhibitory factors are rather rare, as is suggested by the following reasoning. In nature, there has been for countless generations a struggle for existence among the individuals of a species, with a survival of the 68 Outline of Genetics fittest. It follows from this that there has been a struggle for existence among unit characters. Those dominant unit char- acters which are "fit," which serve better to adapt the organism to environmental conditions, are the ones to survive; while "unfit" dominant unit characters are eliminated with the elimi- nation of the organisms that contain them. (The same reasoning would not apply to the recessives, which are "protected" from natural selection when they occur in heterozygotes.) These dominant unit characters which exist today are, for the most part, "survival" characters, being important in the economy of the organism since they are serving to adapt it to the environment. In addition, there must of course be a number of "indifferent" characters, which cannot be construed as adaptations; these have persisted simply because, since they have neither positive nor negative survival value, there has been no reason to eliminate them. Since most of the dominant unit characters have positive survival value, it follows that anything which prevents the expres- sion of these dominant unit characters must have negative sur- vival value. This is exactly what the dominant inhibitory factor amounts to; it is something which prevents the expression of a dominant unit character. One is forced to the following conclu- sion. Although dominant inhibitory factors may have come into existence just as frequently and numerously as the other types of factors and determiners, most of these inhibitors would have been ehminated through natural selection on account of their negative survival value. The few dominant inhibitory factors which per- sisted would be those which inhibited "indifferent" characters. Color in grains of corn is doubtless an "indifferent" character. (There are two types of things which are fairly common and which might be confused with inhibitory factors: (i) simple determiners which produce such characters as to "mask" other characters without really inhibiting other determiners; (2) "lethel factors" which will be explained on p. 69.) On the other hand, it is not surprising to find that supple- mentary factors are rather common. The supplementary factor carries further a reactioh which has been brought up to a certain point by some other force. Exactly this sort of mechanism must play a large part in the ontogeny of most organisms. The litera- The Factor Hypothesis 69 ture of genetics is full of such things as "intensifying" factors, ''diluting" factors, and "(listril)ution" factors, all of which follow the supplementary factor mechanism. Complementary factors are also common, quite as we should expect. Any complex machine contains numerous parts, capable of applying numerous dififerent forces, each one of which may be quite functionless in itself, but, in interaction with some of the others, will produce a visible result. The living organism is just such a complex machine. It is not necessary that complementary factors exist only in pairs. In corn itself there is an additional complementary factor A for color of grain (Emerson 4). R, C, and A must all three be present for the grain to have color. (A cross between the red type AARRCC and the white type aarrcc would give a red Fi, AaRrCc, and an F2 which would show a ratio of 27 reds 137 whites.) Many sizable sets of complementary factors doubtless exist in the organism without our knowing it. One can of course never iden- tify a factor wdthout discovering a race in which this factor is lacking. Where the character involved is an "indifferent" one, races lacking one of the complementary factors are frequently discoverable. Where, however, the character involved is vital to the existence of the organism, it is impossible to discover a race lacking one of the necessary complementary factors, since such a race cpuld not live. In this connection a word should be said of "lethal" factors. Lethal factors, of which a surprisingly large number have been identified during the last few years of investigation, are by defini- tion factors which bring death to the organism. One might conclude from such a definition that these are inhibitory factors which are inhibiting some vital function of the organism, but this is practically never the case. In almost all cases, it is the homozy- gous recessive condition only which brings the lethal eft'ect, so that it is really the absence of the factor that is lethal rather than the factor itself. How are we to interpret this behavior in terms of the mechanisms which we have already described ? One might assume that merely a simple determiner is involved, a determiner for some vital function, so that its absence brings the lethal eft'ect. On the other hand, it is altogether likely that more than one gene yo Outline of Genetics is necessary to the success of this vital function. On such a basis, the lethal would be regarded as a factor rather than a determiner, and would of course be thrown into the complementary factor class. The present writer does not believe that genes should be called factors simply because there theoretically may exist other genes necessary to the production of the character in question; but holds the view that the term "determiner" should be main- tained so long as only one of the effective genes in the set has been identified. In the case of the lethals, however, there have been discovered in the same organism (the fruit fly) a number of genes, the absence of any one of which will bring the lethal effect. It is reasonable, therefore, to regard these as composing one or more complementary sets governing the performance of certain vital functions. (The student may wonder how it is possible to identify lethal factors when their absence simply brings death, an unrecordable phenotype in the population. This will be understood later when the subject of "linkage" is discussed.) It should be realized that genes, be they factors or deter- miners, may at times have more than one role. The A factor in corn, mentioned above, interacts with R and C in a complementary set for the production of aleurone color in the grain. A also has an effect in producing pigment in the vegetative parts of the plant. The R and C factors in stocks (Saunders 6) are a complementary pair for the production of colored flowers. R, C, and a third factor, H, must all be present for there to be hairs on the leaves. Such phenomena support the belief that the gene is not a "vital- istic" unit endowed with a specific function in connection with a single plant character, but rather is of the nature of many chemi- cals, the presence of which will inevitably affect the course of more than one type of reaction. LITERATURE CITED 1. Bateson, W., Mendel's laws of heredity. Cambridge. 1909. 2. CzAPEK, P., and M. E., Biochemie der Pflanzen. Jena. 1913. 3. East, E. M,, and Hayes, H. K., Inheritance in maize. Conn. Agric. Exper. Sta. Bull. no. 167. pp. 142. ph. 25. 191 1. The Factor IlypotJiesis 71 4. Emerson, R. A., A fifth pair of factors, Aa, for alcuronc color in maize, and its relation to the Cc and Rr pairs. Cornell Univ. Agric. Exp, Sta. iMem. 16. pp. 231-289. 1918. 5. LoTSY, J. P., La theorie du croisement. Arch. Xeerland Sci. Exact, et. Nat. Ill B 2: 1-61. 1914. 6. Saunders, E. R., Further contribution to the study of the inheritance of hoariness in stocks {Matthiola). Proc. Roy. Soc. B 85. 1912. CIL\PTER V INHERITANCE OF QUANTITATIVE CHARACTERS This phase of the factor hypothesis, if true, is of fundamental importance, not only to genetics but to general biology. It is based upon the conception of cumulative factors, and as it is presented it will be real- ized that it throws light not only upon numerous breed- ing problems, but also upon variation in general, which means evolution also. A cumulative factor may be de- fined as one which, when added to another similar factor, affects the degree of development of a character. It will be recalled that Correns crossed red and white strains of Mirahilis and obtained pink hybrids. The suggested explanation of this result was that a single dose of the red determiner gives pink while a double dose gives red. When Correns inbred these pink hybrids, he obtained the result presented in fig. 8, that is, I red 12 pink:i white. The mechanism in this case is quite evident. With this diagram in mind we shall consider some of the experiments of Nilsson-Ehle (6, 7) at the Swedish Experiment Station. He crossed two strains of wheat with red and white kernels. The Fj individuals had light red kernels, which of course suggests a repetition of the situation shown by Mirahilis in the experiment of Cor- rens. The F2 generation, however, showed a very dif- ferent result. The reds and whites appeared in the 72 Inheritance of Quanlilative Characters 73 ratio of 15:1; but in addition to this, among the 15 reds there could be distinguished varying degrees of redness. Nilsson-Ehle suspected that the 15:1 meant a dihybrid ratio, 16 individuals being necessary to give the ratio; so he constructed the tentative scheme shown in fig. 10. This shows a regular dihybrid ratio, except that the two factors in^'olved are similar. Applying the single dose and double dose conception, as used in the case of Corren's pink Mirabilis, we reach the following con- clusions: only no. i has four doses and therefore it is deep red; nos. 2, 3, 5, and 9 have three doses and are somewhat lighter red; nos. 4, 6, 7, 10, 11, and 13 have two doses and are still lighter red; nos. 8, 12, 14, and 15 have one dose and are very light red; while no. 16 alone has no dose and is the only pure white. This accounts for the 15:1 ratio, and the different shades of red. This is of course quite in accord with the Mendelian method of inheritance, only two assumptions being necessary: (i) that dominance is absent, two doses having twice the effect of one; (2) that the independent similar factors are cumulative in their operation. This was Nilsson-Ehle's conception, and of course he tested it by further experimental work, the results consistently confirming his assumptions. Since it is important to fix this conception clearly in mind, another type of diagram may represent the facts even more clearly. The proportion of the individuals showing the various degrees of redness in the F2 is graphically recorded in fig. 11, each dot representing one dose of the factors in question. Continuing these investigations, Nilsson-Ehle next discovered a new strain of red-grained wheat, which. 74 Outline of Genetics when crossed with the pure white strain, yielded Fi hybrids of an intermediate degree of redness as before. Fig. io. — Diagram illustrating Nillson-Ehle's explanation of 15: 1 ratio obtained in F2 generation from cross between red-grained and white-grained wheat. The F2 generation, however, showed a different situation. Reds and whites were obtained in the proportion of 63 : 1 ; the 63 reds as before falhng naturally into differ- Inheritance of Qua ntita live Characters 75 ent groups on the basis of degree of redness. Applying the same conception as before, Nilsson-Ehle discov- ered that in this case he was dealing with a trihybrid situation. Without constructing the usual Mendelian checkerboard, which would have to be extensive enough Pure Red VVKite Grades of Pink Fig. II. — Another method of visualizing Nillson-Ehle's 15:1 ratio (see fig. 10). for 64 individuals, the situation as it appeared in the F2 generation may be represented by lig. 12. If this graph be surmounted by a cur\'e, we recognize the regular "probabilities curve," exactly the kind used by biome- tricians to represent fluctuating variations about a specific type. 76 Outline of Genetics #• While P"'^ ^^^ Intcrnicdiato Grades Fig. 12. — Diagram illustrating Nillson-Ehle's 63:1 ratio I nJierilance of Quanlilaliue Characters 77 This conception of cumulative factors, therefore, has far-reaching significance. For a long time, biologists have recognized individual quantitative variation within the species. Darwin depended upon it as the basis of his theory of evolution; in fact, ever since Darwin's Origin of species, individual variation has been funda- mental in our conceptions. To account for this univer- sally recognized phenomenon, Darwix proposed his transportation hypothesis and Weismann offered his germinal selection, both of which were unsatisfactory explanations. Aside from these two attempts to explain individual variation, no other comprehensive scheme had been presented. Biologists had simply recognized the fact of individual variation without any clear con- ception of the mechanism. The importance of this new theory, therefore, is obvious. It is an ingenious explanation of the inherit- ance of quantitative characters and of the existence of individual variations. Furthermore, the theory has not been developed through meditation, but has its basis in scientific experiments. It is imaginative to a certain extent, as is every other valuable theory, but unlike most such theories, it has a substantial foundation, namely, Mendel's law. The importance of the possible role of cumulative factors in explaining individual variation, which in turn may be the basis of a certain type of evolution, has been emphasized because its importance has perhaps not yet been sufliciently appreciated. It promises to be one of the most important theories of biology, but of course will bear further testing by in- vestigators. 78 Outline of Genetics The doctrine of cumulative factors was further devel- oped by Emerson and East (5) in their work with corn. They were able to explain some of the ratios obtained by assuming three or four separately inherited cumula- tive factors, just as Nilsson-Ehle had done. They obtained other ratios, however, which required more independent cumulative factors to explain. Some idea of the extent of these investigations may be gained by noting the list of plant characters whose inheritance they explained on the basis of cumulative factors: number of rows, length of ear, diameter of ear, weight of seed, breadth of seed, height of plant, number of stalks per plant, earliness of flowering. In all of these cases breeding gave the same characteristic results. A cross between extreme parents gave hybrid progeny inter- mediate as to the character in question; and in the F2 generation the two extremes reappeared, along with all gradations of intermediates. The relative frequencies of these classes always resembled the normal probabili- ties curve. Nilsson-Ehle had been able to put his F2 inter- mediates into rather definite classes, corresponding to the number of doses of the determiner each had received. Emerson and East, however, could not do this with such exactness. Their results showed all gradations, but they could not distinguish any definite groups; that is, gradation was continuous and complete. In other words, they could not tell with certainty from out- ward appearance just how many doses of cumulative factors an individual contained. Their results, there- fore, do not seem so clear and striking as those of Nils- son-Ehle, but they are by no means vague and uncer- Inheritance of Quantitative Characters 79 tain. For example, even if they could not say definitely that a certain individual had three doses, they could always say approximately how many doses it had; and the breeding results always confirmed the idea of a num- ber of cumulative factors at work. For example, a plant with three doses may vary with respect to the character in question. It may approach the condition of the plant with four doses or it may vary toward the two-dose condition. Such variation may be explained by outside influences. Any classification of the F2 indi- viduals on the basis of the number of doses is more or less obscured by the influence of outside factors which are uncontrollable, or at least uncontrollable as yet in breeding work. Emerson and East have visualized these outside factors and discussed them. In order to explain this discussion, however, we must recall a feature of genetics which has previously been mentioned. Plant varia- tions in the largest sense fall under two categories, those due to (i) differences in gametic constitution, and (2) those due to responses to environment. The first cate- gory is the basis of all iMendelian conceptions, while the second category includes such variations as are usually thought not to be inherited, being acquired characters. This category is now commonly called fluctuating varia- tions. An illustration will make these two categories clear. Assume that a plant with a determiner for tallness usu- ally becomes 6 feet, while one without this determiner becomes 3 feet. The 6-foot plant, however, grown in good soil becomes 6.5 feet, while in poor soil it is 5.5 feet. In inheritance, of course, the 6.5- and 5.5-foot plants 8o Outliiie of Genetics behave exactly alike; the same is true of 6-foot plants. It must be evident, therefore, that a classification of F2 individuals on the basis of the number of doses might well be slightly obscured. If outside influences were lacking, the F2 situation could be represented by fig. 13 ; Fig. 13. — Diagram illustrating distribution of phenotype classes in Yi population from cross involving cumulative factors. Practically same diagram as fig. 12, and interpreted in same way; short rectangle at left indicates that very few plants of population contain maximum number of doses; short rectangle at right indicates that very few plants contain minimum number of doses; plants with intermediate number of doses most numerous, as indicated by tall rectangle in middle (see also fig- 14). but when outside influences are active, it may be repre- sented by fig. 14. It will be seen from this last diagram that not all individuals belonging to a particular size class may have the same number of doses; that is, con- ditions surrounding the development of a smaller-dosed individual may be so much better than those for a larger- dosed individual that they may exchange size classes in the result. In this way, the results of germinal constitu- Inheritance of Quanlitalive Characters 8i tion may be somewhat obscured by the varying external conditions of growth. Another factor that may obscure these results is what is called physical correlation. For example, a corn plant of small size, but with the hereditary capacity for producing large ears, could not fully express this A A K K \ \ \ ^ K Fig. 14. — Diagram illustrating how environmental inlluences may obscure phenotype classes of F2. Overlapping of phenotype classes makes possible that two apparently identical plants might actually have a different number of doses; diagram also shows that while breeder could not recognize whether a plant had two or three doses, he could distinguish between plants of two and four doses, etc. Thus intelligent selection could be effective. capacity. It could not produce as large ears as if it had been a large-sized plant. Such are some of the conditions or factors that tend to obscure results in the F2 generation and give rise to ratios hard to interpret. The weaker the influence of these factors the more clearly do the phenomena of cumulative factors come out. The total result of this phase of the work of Emerson and East, in spite of obscuring conditions that have arisen, has been to 82 Outline of Genetics strengthen greatly the conception of cumulative factors. A summary of their conclusions is stated as follows: ^'When one is dealing with quantitative characters, that is, those produced by cumulative factors, he is confronted by exactly the same principles of Mendelian inheritance as have long been known to apply to quali- tative characters. With quantitative characters, how- ever, the problem is more complex, due chiefly to two things: (i) we are usually deahng with more factors, and factors cumulative in their operation; (2) the signifi- cance of the breeding results is usually somewhat obscured by the natural fluctuations due to response to uncontrollable factors in the environment." In connection with the cumulative factor concept, a modifying statement should be made as to the mecha- nism involved. Heretofore it has been assumed that we are dealing with numerous, separately inherited factors, absolutely identical in their nature, cumulative in their effect. No doubt one might regard with sus- picion such a seemingly artificial mechanism. Probably it would be easier to believe if it were modified in the following manner. Instead of assuming that the numer- ous factors are identical in function, we may assume that each of these factors has its own peculiar function, but that that function plays a part, directly or indirectly, in developing the quantitative character in question. For example, suppose height is the character. One of the factors determines the development of numerous nodes; another increases the amount of chloroj)hyll ; another determines the size and \'igor of the root system; another brings early germination and a long growing season. Such factors, although not identical, will be Inheritance of Quantitative Characters 83 cumulative in increasing the height of the plant. Of course a single dose of one type of factor may not bring the same increase in height as would a single dose of one of the other types, and therefore the mathematics of the situation will be slightly modified. The fundamental mathematical system, however, will remain the same, and we will have the satisfaction of dealing with a natural mechanism rather than an artificial one. A few of the applications of the cumulative factor hypothesis are worth considering. Assume that a practical breeder crosses two extreme parent types in the hope of obtaining a hybrid com- bining the desirable characters of the two parents. If the material is corn, he might use one parent with large grains but few in num- ber, while the other parent had many grains but small ones. Such quantitative characters as these would be determined by cumula- tive factors, and the hybrid would be intermediate with respect to both of these characters, that is, the grains would be of medium size and medium number. No matter how many crosses he made, he would always get this result, and not the desired combination of large grains and many of them. Suppose now that these intermediate hybrids are inbred in the hope of obtaining the desired combination among the individuals of the F2 generation. It will be realized that the chances of obtain- ing a plant combining the two extreme characters of large grains and numerous grains would depend upon the number of factors that enter into the make-up of these quantitative characters. Assume that there are five factor pairs in each case. The mathe- matics of the situation would show that in order to get the desired pure type from a cross between two parents, each having their desirable character determined by five pairs of cumulative factors, it would require 100 acres of corn to have an even chance of getting one such individual in the Fj generation. It is altogether unlikely that any farmer would use 100 acres and a corresponding amount of labor on such an extreme chance. Even an agricultural experi- ment station would not feel justified in conducting such an experi- ment. 84 Otitline of Genetics The question arises whether there is any way of avoiding this impossible situation. The escape is suggested by the fact that time can take the place of numbers. East has shown that by growing looo individuals in the F2 generation, 100 in the F3, and 50 in the F4, one stands as much chance of getting the desired combination as by growing 250,000 in the F2, provided an intelli- gent selection is made in each generation. In other words, one who understands the mechanism of the inheritance of quantita- tive characters will grow only 1000 individuals in the F2 genera- tion, and will select for seed only those individuals with the most favorable combination of factors. In this way, by intelligent selection, factors are 'Spiled up" in the right direction from year to year. In a few years the desired result will be reached without the necessity of ever growing a very large number of individuals. Such work is practicable at experiment stations, and it is the kind of work that a number of them have been doing. Even the farmer is able to accomplish this. Although his selection of individuals is not quite as intelligent as that of a scientific breeder, he is at least selecting in the right direction and making some advance. A little more time and a little more acreage would bring him very close to the desired goal. A further application of the cumulative factor hypothesis may be considered. The practice we have been discussing under the title of "inheritance of quantitative characters" seems to be little more than what has already been called artificial selection, which is the oldest of all methods of plant breeding. It is a method that was thought to be discredited entirely by the work of De Vries and JoHANNSEN when they discovered ''elementary species" or ''pure lines," and demonstrated that artificial selection could never result in any large or permanent improvement. In consequence of this, artificial selection, as the most important method of secur- ing desirable races, gave place to pedigree culture at a number of experiment stations. The older method was not entirely aban- doned, for it had its uses, but many regarded it as a medieval method of breeding. The artificial selection which we have been describing, however, is distinctly different from the method prac- ticed by the early breeders. In brief statement, the difference is as follows. Inheritance of Quantitative Characters 85 The selection proposed is preceded by an intelligent hybridiz- ing, and after that genotypes rather than phenotypes are se- lected; that is, the selection is made on the basis of germ plasm rather than body plasm. 'J'his would be a sufficient reason for the superiority of the new method of artiiicial selection as compared with the old. A little further analysis will make the difference clearer. In the old method of artificial selection, the breeder, in the first place, is dealing with such germinal variations as happen to appear in his crop; and, in the second place, he is dealing with those fluctuations which appear as responses to the environment. When he selects a large plant to use for seed, that plant may be large on account of its germinal constitution; but, on the other hand, it may be large because it is growing in a less crowded place or a place more heavily fertilized than the others. In that case, the large plant might not furnish good seed. The plant breeder of the old method undoubtedly made such unfortunate selections frequently; that is, he selected on the basis of external appearance, and external appearance is very often a poor index of hereditary capacity. Furthermore, he would not keep his lines pure, but would deal constantly with an unmanageable mixture of good and mediocre types. Intelligent selection is based on germinal con- stitution only^by keeping careful pedigree records a selection of genotypes is possible — and therefore its results are quicker and surer. It is really a pedigree culture rather than a mass culture method. Another phase of the subject should be considered. When a plant breeder is trying to improve his crops by selection for quan- titative characters, although he uses the old method of selection, he is likely to be making some gain, as the experience of hundreds of years has shown. The germinal constitution of his crop plants is masked by fluctuations, of course, but this mask is not complete. Most of the plants he selects are bound to possess high numbers of factors of the right kind, and he probably rejects most of the plants with few factors. In any event, he has generally succeeded in the long run in getting a somewhat improved race. A summarized statement of this situation may be helpful. Our recently developed knowledge of the inheritance of quanti- 86 Outline of Genetics tative characters seems to justify artificial selection, but it does not justify the old blind method of selection. It emphasizes the need of intelligent selection, and shows how such selection can be made. In order to do this, one must understand the mechanism of inheritance involved, and must make his selection on the basis of genotype rather than phenotype. All along the line, strains must be discarded which, though recommended by the phenotype of one of their ancestors, are to be condemned on the basis of their breeding capacity. Selection is always to be made on the basis of breeding capacity, that is, genotype. The situations just considered enable one to understand two phenomena which have been baffling scientific plant breeders for some years. The races of plants improved by artificial selection have usually reverted to type when selection ceases. This fact was recognized for a long time, but was first pointed out clearly by De Vries (3) . Since then we have always expected this result, that no improvement will maintain itself, but will run back unless the selection is continuous. When a practical breeder announces that he has developed by selection a new race which continues to breed true without further selection, we are inclined to disbelieve him, for we know that only elementary species breed true. We explain that the practical breeder bases his selection on fluctuations, and therefore his new race is bound to revert to type. It is obvious now that there is a flaw in this argument. The practical breeder may be basing his selection on fluctuations, but at the same time he may be piling up cumulative factors in the right direction. Thus he might eventually secure a race con- taining all the cumulative factors. Such a race would be a homo- zygote and could not help breeding true. Most of the claims of artificially improved races that breed true may be false, but it should be remembered that such a thing is possible, and may be "stumbled upon accidentally," even with unscientific breeding. There is another phenomenon which has been much discussed, and which can now be explained in the same way. This is the so-caUed "fixation of hybrids." For years breeders have made promiscuous crosses and then begun artificial selection with the F2 generation. Eventually they have secured a pure-breeding new type. It will be remembered that it was in this way that Inheritance of Quantitative Characters 87 East worked with the quantitative characters in corn, and the explanation is the same. In addition to the practical value of the concei)tion of cumu- lative factors, the theoretical value is worth considering, for it explains things that have been very vaguely understood. This conception suggests that the origin of species by natural selection in the way described by Darwin, a method which for some time has been thought impossible, may actually be possible within limits. Of course natural selection in a certain sense has always been accepted, almost as generally as the fact of evolution. The point in dispute is as follows. Darwin used as the basis of natural selection those small individual variations which we have come to call fluctuations, the same kind of variations the old plant breeder used in his artificial selection. Darwin claimed that such varia- tions could be piled up until the result would be a new species. It was in 1900 that De \'ries showed in convincing way that this kind of variation never resulted in a new species; at best it only developed a race which approached the boundary of the species and never crossed it. Moreover, such a race would revert to type rapidly as soon as some slight change in conditions set up a new standard for selection. This argument, confirmed by experi- ment, has been generally accepted. We now know that individual variations are not always mere fluctuations or responses, but may be due to varying doses of cumulative factors. A selection on this basis may very well result in a new race that breeds true; and a race that breeds true is De Vries' definition of a new species. To reestablish Darwin's ideas on the origin of species is certainly an important considera- tion. The situation illustrates how genetics and evolution are tied up together, so that neither one of them can be appreciated fully without some knowledge of the other. A few words may be said in reference to the reversion of an old race to its original specific type. De \'ries outlined the situation clearly, and his conclusions are generally accepted. It is doubtful, however, whether it has ever been understood, since no one has ever devised a reasonable mechanism for such a rever- sion. The conception of cumulativ'e factors supplies this mecha- nism. A new race, developed by natural or artificial selection 88 Outline of Genetics among individual differences, means the piling up of cumulative factors in a given direction. Stop the selection and the old plants with the small numbers of factors are allowed to survive, reproduce, cross with the others, and eventually bring back the species to the original average condition. One very seldom has any occasion to work out problems on cumulative factors, since here the phenotypes do not show up as clearly as they do in connection with the other factor types. Any such problems, however, could readily be solved by some such method as the following. Remember that we are dealing with a dominance absent situation; and represent the number of doses as exponents attached to the numbers which indicate the frequen- cies of the different classes. AaBhCcXAABbcc. The A set gives a ratio of i with two doses : i with one dose, and should be represented as i^: i'. The B set gives i with two doses: 2 with one dose: i with no dose, and should be represented as 12:2^:1°. (i^*: iOX(i^:2': 1°) equals 14:33 32:1^ The C set gives i with one dose:i with no doses, or 1^:1". (14:33:32: 1') x(i':i°) equals 15:44:63:42; I^ The final result is i with five doses: 4 with four: 6 with three: 4 with two:i with one. During the last decade, the mechanism of cumulative factors has been invoked to explain a great many of the phenomena of genetics. One noted instance of this will be worth considering, as it has a very important bearing upon one of the fundamental con- cepts in connection with the mechanism of inheritance. A few years ago geneticists might have been grouped into two schools: "mutationists," who beUeved in the introduction of new hereditary units by mutation alone, maintaining that the hereditary genes were invariable and could not be modified by selection; and ''selectionists," who beheved that the genes could be modified by selection. The most prominent figure among the selectionists was Castle, and the main experimental evidence upon which he based his view was as follows. Castle (i) isolated a race of rats which had a black and white coat pattern known as "hooded" (the black pigmented area hav- Inheritance of Quantitative Characters 89 ing the location and general shape of a hood). This hooded pat- tern bred approximately true and behaved as a simple IMendelian recessive in crosses with rats of the ''wild" type. These facts naturally led Castle to beUeve that hooded was a simple Men- delian unit character, represented in the germ plasm by a single gene. Castle then commenced selection. For twelve generations selections were made from this new race without a single outcross, that is, every generation was inbred (brother and sister matings) , thus insuring the constant purity of the stock. In one series selection was made for an increase in the extent of the pigmented areas; in another series selection was made for decrease in the extent of these areas. The result was that the areas in the one series steadily increased, while in the other they steadily decreased. Castle pointed out that: (i) with each selection the amount of regression ("running back") grew less; that is, the effects of selection became more permanent; in other words, in each suc- ceeding generation there was a decreasing tendency to revert to the original average type; (2) advance in the upper limit of varia- tion was attended by a like advance of the lower limit. The total range of variation, therefore, was not materially changed, but there was a progressive change in the point about which the variation occurred. In other words, it was like the progressive shifting of the center of a circle; the diameter of the circle did not change but the position of the circle, determined of course by its center, was gradually changing. These were the two important facts which Castle brought out and they have been stated approx- imately in Castle's own words. Fig. 15 will help make the situation clear. The average amount of variation in any one generation of the pure stock (the di- ameter of the circle referred to) is indicated by ,^-£->. Of course, even "pure stock" varies somewhat, since no two individuals are exactly alike, biology recognizing what is called "individuality." The point is that the comparatively small variation in a pure stock is not due to germinal differences, but to responses called out by varying external conditions, such as nutrition, light, etc. These response variations, usually called tluctuations, vary with different individuals, but the hereditary capacity of all of them remains go Outline of Genetics the same. A selection on the basis of fluctuations within a pure line, therefore, should not result in any permanent improvement ; in fact, it has been demonstrated many times that no such improve- ment can be effected in this way. When selection is made, - Pure-Bred Generations / Selection Begins The Control Fig. 15. — Diagram illustrating Castle's selection experiment with hooded rats. however, among varying doses of cumulative factors, an entirely different situation is faced, for in such a case we are not dealing with a pure line. The significance, therefore, of Castle's results may be real- ized. He bred his original pure line for many generations and found that it varied only within very narrow limits; and these Inheritance of Qiianiilatlve Characters 91 slight variations he regarded as mere lluctuations. I-'urthermore, he found that the character of his pure Hne behaved in crossing as a simple unit character and that no complex factors were involved. With this evidence he should not have been able to effect any permanent changes by selection, but this is exactly what he did. Selecting in opposite directions, he developed two new strains, the boundaries of the new strains being distinct from one another and distinct from the boundaries of the original strain, that is, the non-selected type that he started with. Castle's next step was significant. He crossed each of his new strains with the same wild race, the result being that each of his new strains behaved as a simple and distinct recessive unit. The high pigmentation strain ''came out of the cross" with the characteristic high pigmentation; the low pigmentation strain came out with the characteristic low pigmentation. The conclusion from this series of experiments may be given in Castle's words, as follows: ''The conclusion seems to me unavoidable that in this case selection has modified steadily and permanently a character unmistakably behaving as a simple Men- delian unit." The importance of this conclusion is evident. Mendelism had been based upon the conception that unit char- acters could not be modified. Mendelians of the "mutationist" school had granted only two possible methods for the origin of new races: (i) by recombinations of existing characters through hybridizing; (2) by the sudden and complete dropping out of an existing unit or the equally sudden addition of a new unit, both of which possibilities might arise from mutation. No "mutation- ist" would grant, however, the possibility of modifying an existing unit character, the thing which Castle claimed to have done, basing his claim upon well-controlled experimental breeding. If Castle's contention were true, it would result in the fundamental modification of Mendel's law. The whole mechanism would have to be modified to take into account new fields of variation that had not been thought to exist. The statements of the "mutationists" in reference lo these experiments should be considered. They attempted to explain Castle's results through the cumulative factor mechanism. The claim was made that Castle had started with a character 92 Outline of Genetics that had fluctuated continually, never having been brought to as small a variability as have most other characters. The question was raised whether Castle's assumption that this variability was merely due to fluctuation was altogether justified. Might not the variability have been due to varying doses of cumulative factors? Suppose for the moment that this were the case; it would not be surprising that Castle could develop two diverse strains by selection, for selection would result in piling up the cumulative factors in one direction or the other. Castle's rejoinder was that if this were a cumulative factor situation, why had none of the extremes appeared in the non-selected stock, which instead bred approximately true ? The answer was made that the extremes did not appear in the pure bred stock merely because of the mathematical limitations. If one is deahng with six cumulative factors, and the so-called pure stock has an inter- mediate number of doses, there could not be much chance of getting out the extremes in a single generation. It would be necessary to secure over 4000 progeny to have an even chance of getting one such extreme; or about 50 progeny to get anything that would very noticeably approach the extreme. It would seem, therefore, that Castle's chances to determine this would be very small. Rats certainly do not produce 4000 progeny in a single gen- eration; in fact, they produce much less than 50; therefore Cas- tle's "pure stock " went on in the intermediate condition, and only by selection could he pile up the factors and reach either extreme. Thus far the explanation seemed satisfactory. Castle showed, however, that the coat pattern condition behaved in crosses as a simple Mendelian unit; that is, it did not split up into complex ratios, but came out as a recessive in a regular 3 : i ratio. This really involved no difliculty. Suppose Castle crosses one of his pure strain rats having the hooded character with another race that has some pattern character that conceals the hooded character. If this other character is a .simple Mendelian one, the result of the cross would be the ordinary monohybrid ratio; that is, in the F2 generation from such a cross the ratio of hooded to non-hooded (with the " hood-conceahng " character) would be 1:3, which, in fact, is exactly what Castle got. At the same time, the amount of pigmentation, determined by numerous cumulative Inheritance of Quantitative Characters 93 factors, might go on in the same intermediate condition, unaffected by the cross. The rehition of pattern to non-pattern is merely a simple monohybrid system temporarily superimposed upon the other more complex system without permanently affecting it, any more than any inhibitory factor permanently affects the factors it inhibits, or a dominant permanently affects a recessive. It was in this way that the mutationists attempted to explain away Castle's results. Castle did not at first admit the ade- quacy of this explanation, but continued to maintain that he had modified a unit character by selection, and some geneticists agreed with him. This question might be raised. Why cling so strongly to the cumulative factor hypothesis and force Castle's results into this conception ? Is there anything sacred about a unit character that it should not be modified just as complex chemical molecules may be modified in certain reactions'? Why not admit that Mendelian factors may be modified, and explain Castle's results in this way ? The reason is that when we begin to admit that unit characters and single MendeUan factors may be modified, the whole con- ception of inheritance becomes chaos. The great advantage of the factor hypothesis is that it furnishes the clearest method of describing breeding results. East (4) makes an eloquent state- ment on this point. *' Taking into consideration all the facts, no one can well deny that they are well described by terminology which requires hypo- thetical segregating units, as represented by the term 'factor.' What then is the object of having the units vary at will ? There is then no value to the unit, the unit itself being only an assumption. It is the expressed character that is seen to vary; and if one can describe these facts by the use of hypothetical units, theoretically fixed, but influenced by the environment and by other units, sim- plicity of description is gained. If, however, one creates a hypo- thetical unit by which to describe phenomena, and this unit varies, he really has no basis for description." The question was finally settled in a very neat way by some critical breeding experiments that Castle himself performed on these same hooded rats (2). The degree of pigmentation on the high pigmentation strain was designatetl as +3.73 (in terms of Q4 Outline of Genetics certain arbitrary units), while the low pigmentation strain was -2.63. Critical examination revealed the fact that the cross between the +3.73 strain and the wild race brought a slight reduc- tion in the amount of pigmentation as it appeared in the extracted hooded "recessives" in the F2. Repeated recrossing of these extracted individuals with the wild race finally resulted in extracted hooded rats of the grade +3-04. No further reduction was pos- sible in this way. These results could be explained by the follow- ing assumptions. The hooded pattern is modified in degree of pigmentation by a varying number of doses of cumulative factors (as the ''mutationists" had previously maintained). The wild race is characterized by having a certain number of doses of these cumulative factors. The repeated crossings and extractions mentioned above would eventually result in producing rats which had the hooded pattern plus that number of doses of cumulative factors which was characteristic of the germ plasm of the wild race. Since it was found that repeated crosses with the wild race could bring the degree of pigmentation down to +3-04 and no lower, it was felt that +3.04 was the degree of pigmentation which would be produced by that number of cumulative factors which was characteristic of the wild race. The critical test of these assumptions could be made through a similar manipulation of the low pigmentation strain. If the assumptions were correct, the low pigmentation strain should be raised finally to +3-04 by repeated crossing with the wild race. Castle performed this experiment and got exactly this result, one of the families from the low pigmentation strain (-2.63) being finally brought up to -I-3.05. These results naturally caused Castle to change his views on the matter, and served rather generally to establish the views of the ''mutationists." The situation depended for its interpre- tation upon the cumulative factor mechanism. (Here it was felt that the cumulative factors were not primarily responsible for the production of the character in question, but served merely to modify the degree in which it expressed itself. Other cases of the same sort have been encountered elsewhere, the mechanism at play being commonly referred to in the literature as "multiple modifying factors.") InJieritance of Quantitative Characters 95 LITERATURE CITED 1. Castle, W. E., The inconstancy of unit characters. Amcr. Nat. 46:352-362. 1912. 2. , Piebald rats and the theory of genes. Proc. Nat. Acad. Sci. 5 : 1 26-130. jig. i. 1919. 3. De Vries, H., Species and varieties, their origin by mutation. Chicago. 1905. 4. East, E. M., The Mendehan notation as descriptive of physio- logical facts. Amer. Nat. 46:633-655. 191 2. 5. Emerson, R. A., and East, E. M., The inheritance of quantita- tive characters in maize. Bull. Agric. Exper. Sta. Nebr. no. 2 pp. 120. figs. 21. 1913. 6. Nilsson-Ehle, H., Einige Ergebnisse von Kruzungen bei Hofer und Weizen. Bot. Notiser 1908:257-294. 7. , Kreuzungsuntersuchungcn an Hafer und Weizen. Lands. Univ. Arsskr. N.S. II. 5:1-122. 1909. CHAPTER VI LINKAGE The fundamental mechanism of inheritance which was proposed by Mendel, and which was later supported by cytological studies, has been confirmed time and again by breeding experiments. Its scope is consid- erably enlarged by the factor hypothesis, but its basic concepts are not altered. It will now be necessary to consider some well-established facts of inheritance which can be interpreted only by analyzing still further the hereditary mechanism. It has been assumed that the chromosomes are the bearers of the hereditary units or genes. (The term ''gene" is used where it is not intended to imply whether the hereditary unit acts as a factor or determiner.) This has been warranted by the fact that the distribution of the chromosomes in inheritance fits exactly into the Men- delian scheme. In the cases that have been considered in the last few chapters, the genes have always been located on separate chromosomes, with the result that they have been passed on in inheritance quite independ- ently of each other. The intensive study of inheritance that has been made during the last decade, however, has revealed cases where the total number of genes known for the organism exceeds the number of chromosome pairs. The obvious conclusion is that more than one gene may be carried on a single chromosome. If this is true, it should result in decided modifications of the breeding 96 Linkage 97 results. Conversely, the occurrence of a certain type of breeding result would serve as a clear indication that more than one gene may be carried on a single chromo- some, and thus as a further confirmation of the belief that the chromosomes are the bearers of hereditary characters. In 191 1, results of just this sort were obtained in corn by Emerson (2), who stated: ''This is an example of a feature w^hich is probably very widespread in the plant world, but of which at present we know little." Long before any further important work was done along this line among plants, however, Morgan (5) published the results of his very careful and intensive breeding experiments with the fruit fly. His ideas have had a profound influence upon subsequent work in genetics. He has given us a more accurate picture of the hereditary mechanism and one that fits the facts bettei than any previously proposed. In simplest terms the picture is this. Each chromosome is a rodlike structure, and numerous genes are arranged in a line along this rod. Thus Morgan further analyzes the germ plasm by accu- rately locating the genes. (He does not attempt any description in physico-chemical terms of the genes them- selves or of the exact relation they may hold to the chromosomes on which they are carried.) We cannot discuss here the many ways in which this fundamental conception has cast light upon work in genetics. Suffice it to say that it has resulted in a new ''school" of geneticists whose experiments have been more intensive, more exact, and in s(^mc ways more "fundamental" than those of any previous school. To date most of the linkage work has been done with the 98 Outline of Genetics fruit fly, but a great deal of information is now being accumulated on linkage in corn, and numerous scattered demonstrations of the phenomenon have been made in other organisms as well. Only a rather simple explana- tion will be attempted here, to bring out merely some of the fundamental principles of the phenomenon. When first considering Mendel's law, the state- ment was made that more than one determiner might be located on a given chromosome. As yet we have © © © © Tall Red Parent © © © © Dwarf White Parent 4 Possible Gametes .-. F, Shows 9:3:3.1 Ratio Fig. 1 6. — Diagram showing normal dihybrid behavior when no linkage is involved. considered no such case, but linkage involves exactly this situation. In connection with some of Mendel's original crosses fig. i6 will be recalled. In this case a double dominant mates with a double recessive, and the result is a dihybrid ratio in the F2 generation, following the production of four types of gametes by the Fi plants. Suppose, however, that the determiner T and the deter- miner R are carried on the same chromosome, the situa- tion would be as represented in fig. 17. Here the Fj individuals produce only two types of gametes, as in a monohybrid, so that the F2 presents what amounts to Linkage 99 a monohybrid ratio, the tall individuals always being fed flowered, and the dwarf individuals always white flowered. The results obviously arise from the fact that T and R are linked, being located on the same chro- mosome, as are also / and r. These linkage results, taken by themselves, might seem to recommend the following interpretation. It is really the chromosome itself that is the important and indivisible unit in inheritance, while the distinction "LINKAGE" Dwarf White Parent Only 2 Gametes Possible .'. F2 Shows 3 : i Ratio Fig. 17. — Diagram showing "dihybrid" behavior when genes are linked. Fi produces only two types of gametes. between the various genes on a single chromosome is purely arbitrary and unnecessary. In other words, why need we assume that T and R are distinct genes, when we would be equally justified in assuming that tallness and redness are merely two of the effects })roduced by the same chromosome ? This latter assumption may appear attractive, but it becomes clearly im])ossil)lc when some of the further breeding results arc considered. Following out the foregoing example (merely as an illustration), when it was discovered that tall individuals always had red flowers, this fact was explained as linkage. loo Outline of Genetics The inference was that in these same cultures there could never appear a tall plant with white flowers nor a dwarf plant with red flowers, for if there were linkage, and the chromosome were the indivisible unit in inheritance, it would be impossible for tallness and redness to become separated. As a matter of fact, it was soon recognized that these 'impossible" individuals did actually occur. Small numbers of tall whites and dwarf reds regularly appeared among the same cultures in which the linkage of tallness and redness had been demonstrated, and the work had been done under such conditions of control that there could have been no experimental error. This new fact demanded an explanation, for with such chromosomes as TR and tr it would be impossible to obtain a tall white individual so long as the individuality of the chromosome was maintained. When chromosomes were examined with the modern lenses they were found to show all kinds of tangled contortions during the reduction division, and accordingly the scheme shown in fig. 1 8 was devised. These five stages represent phases that an allelomorphic pair of chromosomes may go through during reduction division. This pair of chromo- somes, which would normally lie side by side (i), may at times come to lie across one another (2). In this position the middle regions of the chromosomes are in contact and are conceived of as fusing (3) . The spindle fibers from each pole then lay hold of this compound chromosome and the pull comes in the direction of the arrows shown in the figure. This results in the break indicated in (4). Finally, two new chromosomes separate from the old compound chromosome, as indicated in (5). Thus T becomes linked with r, and later, when a mating Linkage lOI occurs between two gametes, each of which contains such a chromosome, the result is a tall, white-llowered R T 2 3 t ) R R Fig. iS. — Illustrating how crossing over may occur I02 Outline of Genetics individual. In the same way and with equal likelihood, dwarf, red-flowered individuals may appear. This scheme serves to account for the occurrence of the ^'exceptional" individuals in linkage cultures. The whole phenomenon is known as crossing over. It has been practically impossible to provide a direct demon- stration that the chromosomes behave in exactly this manner during the reduction division, but there has accumulated an enormous mass of indirect evidence from the breeding results to support this view. Evidently the chromosome is not the indivisible unit in inheritance, but is divisible according to a rather regular scheme. Whole sections may be evenly exchanged between the members of an allelomorphic pair of chromosomes. Once the phenomenon of crossing over had been identified, investigation was undertaken to determine the regularity and frequency of the phenomenon. It was discovered that the amount of crossing over that took place between a given pair of genes had a constant value. For example, lo per cent of the crossing over could be depended on to occur between T and R in every experiment involving these two determiners. The exact cross-over value is of course computed from the breeding results obtained. In the present example, a cross-over value of lo per cent between T and R would work out as follows. In the reduction division (in the Fi hybrid which results from tall red X dwarf white), crossing over takes place in lo per cent of the cases, while crossing over fails and the original linkage relation- ships are maintained in 90 per cent of the cases. As a result, four types of gametes are produced in the follow- ing frequencies: 45 per cent TR, 45 per cent /r, 5 per Linkage 103 cent Tr; 5 per cent tR; or 9 TR, 9 Ir, i 7>, i IR. 'J1ie F^ population which results from the random matings among this assortment of gametes is represented in fig. 19, the phenotypic ratio being 281 tall red: 19 tall white: 19 R R 81 ^T) Cl R fT\ R rT\ R R R T R 81 vly ^f t R Ji^ 9 9 ^^T7^ R R R R II 12 R t R t t tut 13 14 16 • Fig. 19. — Showing F2 population produced by random mating of gametes of Fi in a case of linkage with 10 per cent crossing over. dwarf red: 81 dwarf w^hite. In such cases, of course, the original investigator has for his data only this final pheno- typic ratio, and from these data must compute the amount of crossing over that has taken place. In actual practice this computation would be simplified by the use of a formula. As a matter of fact, the necessity of I04 Outline of Genetics using any such formula can usually be avoided through the following expedient. Instead of inbreeding the Fi, it can be back crossed with the double recessive parent (dwarf white). This parent race is perfectly homozy- gous, so that it produces only the one type of gamete (in spite of crossing over, which must be taking place here also). The results of this back cross are represented in fig. 20. It is obvious that the phenotypic ratio obtained (9 tall red:i tall white: i dwarf red: 9 dwarf T R R Fig. 20 — Population resulting from mating of gametes of Fi (shown above) with gametes of recessive parent (only one type of gamete, shown at left) in a case of linkage with lo per cent crossing over. white) corresponds exactly to the ratio among the types of gametes produced by the Fj. In this way the cross- over value is quite apparent, and no computation neces- sary. Back crosses wdth the recessive parent will always provide results which are easier to interpret than are the F2 ratios. As investigations were made of additional pairs of linked genes, it was discovered that each such pair had a characteristic and rather constant cross-over value. For example, assuming that a third gene A is located on the same chromosome w^ith T and Ry by means indicated Linkage 105 above a test is made as to the cross-over \'alue between A and T. This value is discovered to be 20 per cent. With these two cross-over values determined, consider- able interest now becomes focused upon the relations of A and R. Suitable tests reveal the fact that the cross- over value between A and R is 30 per cent. When three such cross-over values as these are considered together, Morgan's theoretical scheme of the 'linear arrange- ment of genes" is immediately suggested, for these results can best be interpreted on the following basis. The genes are arranged in a line on the chromosome, and the frequency of crossing over between any two genes depends upon their distance apart. Using arbi- trary units to correspond with the actual percentage of crossing over, the three genes employed in the example may be arranged in the order ATR, with A and T 20 units apart, T and R 10 units apart, and A and R, there- fore, 30 units apart. (The chromosome map first de- vised, e.g., fig. 18, would then have to be modified some- what to fit these new facts.) It is in this manner that Morgan and his students have been able to construct rather startling chromo- some maps, indicating in a very exact way the relative position and spacing of scores of genes on a single chro- mosome. Striking confirmation of the scheme appears in the following fact. The fruit fly possesses four chro- mosome pairs, one of which is visibly much shorter than the others. Breeding results reveal that the many known genes are associated in just four ''linkage groups," one of the four groups containing a much smaller number of genes than the other three, and being distinctly ''shorter," as mapped from the cross-over values. In io6 Outline of Genetics truth, this scheme of Morgan's has such an enormous mass of data to support it that, for all practical purposes, it may be regarded as an established fact. Cross-over values in the fruit fly may run as high as 80 per cent, and in one of the related species much higher values have been reported (Lancefield 4). Such a high cross-over value seems rather surprising, for it represents a case where the cross- overs are much more frequent than the non-cross-overs. It should be reahzed that a cross-over value of 50 per cent, where cross-overs and non-cross-overs are equally frequent, would give exactly the same breeding results as if the two genes in question were located on different chromosome pairs. Similarly, cross- over values higher than 50 per cent would give the same breeding results as though the linkage were reversed; that is, if M and N were 80 units apart, the breeding results of a single experiment involving M and N only would seem to indicate that M was linked with n and m with N. It is evident, therefore, that these higher cross-over values are computed from a considerable set of ex- periments. Every newly discovered gene is carefully tested with at least two other genes whose position is already known, and thus the new gene is accurately placed on the chromosome. The cases that have just been discussed are known as single cross-overs; the two chromosomes of the pair come to lie across one another at a single point, and a single break with the subse- quent rearrangement is sufficient to account for the results. In view of the physical mechanism which seems to be responsible for these cross-overs, it is not surprising to find that there may sometimes occur double cross-overs. In these cases the two chromosomes come to lie across one another at two points, and a break takes place at each point, with the corresponding exchange of chromosome regions. This amounts to an even exchange of corresponding zones from the middles of the chromosomes, with the two end zones on each chromosome remaining as before. Inas- much as crossing over of any sort is detected only through its effect on the breeding results, double crossing over can be demon- strated only in experiments that involve observations on at least three genes that are rather widely separated on the same chromo- Linkage 107 some. As would be expected, double cross-overs are inucli less frequent than single cross-overs. There have also been reported very rare cases of triple cross-overs. The question arises whether the frequency of crossing over depends strictly and only upon the real distances between the genes on a chromosome. In the first place, it is theoretically possible that certain regions of a chromosome may, owing per- haps to physical peculiarities, lend themselves more readily to crossing over than do other regions of the same chromosome. (A suggestion to this effect appears in some of Muller's work, see chapter on ''Mutation. ") This possibility is, of necessity, ignored in the chromosome maps, which are constructed purely on the practical assumption that crossing over is proportional to distance. In the second place, it is recognized that certain special influences, such as temperature and age of the organism, may modify the normal frequency of crossing over. Dependable values on crossing over are to be obtained, therefore, only when such conditions are absolutely standardized (Sturtevant 6). Further, one of the outstanding and unexplained peculiarities of this phenomenon is that, in the fruit fly, crossing over takes place freely in the female, while none whatsoever takes place in any of the chromosome pairs of the male. It is interesting to note that in organisms where the female is heterozygous for sex — female homozygous for sex in the fruit fly — exactly the reverse is true, crossing over taking place in the male but not in the female (Tanaka 7). In corn, where both sexes are represented on the same individual, there is no such restriction, for crossing over takes place with approximately equal frequency in microsporogenesis as in megasporogenesis (Emerson and Hutchinson 3). Finally, it has been demon- strated that the presence of certain special genes will modify the frequency of crossing over, so that it has been possible to isolate races of the fruit fly in which an extremely high or an extremely low rate of crossing over takes place (Sturtevant 6, Detlef- SEN l). io8 Outline of Genetics LITERATURE CITED 1. Detlefsen, J. A., Is crossing over a function of distance? Proc. Nat. Acad. Sci. 6:663-670. 1920. 2. Emerson, R. A., Genetic correlation and spurious allelomor- phism in maize. Connecticut Agric. Exp. Sta. Bull. 167:1-142. 1911. 3. Emerson, R. A., and Hutchinson, C. B., The relative fre- quency of crossing over in microspore and megaspore develop- ment in maize. Genetics 6:417-432. 1921. 4. Lancefield, D. E., Linkage relations of the sex-linked char- acters in Drosophila obscura. Genetics 7:335-384. 1922. 5. Morgan, T. H., Heredity and sex. New York. 1914. 6. Sturtevant, a. H., Contributions to the genetics of Droso- phila melanogaster. HI. Inherited linkage variations in the second chromosome. Carnegie Inst. Wash, Publ. 278 : 305-341 . 1919. 7. Tanaka, Yoshimaro, Genetic studies on the silkworm. Jour. Coll. Agric. Sapporo 7:129-255. ph. 4. 1916. CHAPTER VII MUTATION When De Vries ''discovered" the phenomenon of mutation in Oenothera Lamarckiana, he stated that mutations were quahtative, discontinuous, constant changes in the germ plasm (see p. 6). These three fundamental characteristics still hold true, but some of De Vries' other ideas have been considerably qualilied by later work. The critical analysis of the germ plasm that has been effected during the last decade has made it possible to describe mutation in a much more exact way than before, and to describe it in terms of the Mendelian mechanism. For convenience our discussion of this general subject will be put into the form of a classification. (The sequence followed in this classihcation is that of the increasing magnitude of the "area" of the germ plasm affected by the change.) I. gene changes I. Locus changes. — These are changes restricted to a single locus of one of the chromosomes, so that they involve only one gene, without affecting even its nearby neighbors. Usually they are effective on only one chro- mosome of a pair, without affecting the corresi)onding locus of its allelomorpliic mate. Consequently, the change first appears in the heterozygous condition. (Baur estimates that such changes originate in the heterozygous condition four hundred times as frecjuently 109 no Outline of Genetics as in the homozygous.) They are mostly "loss" mutations and recessive to the previous condition. Only a very few dominant or "gain" mutations have ever been reported. Baur (i), working with Antirrhinum, concludes that changes of this sort take place more frequently in the vegetative tissues than in connection with gametogenesis . The earlier work on the fruit fly indicated that the locus changes took place late in gametogenesis, since only one individual of the new type would appear in a pro- geny. Later investigation, however, has indicated that the change may take place at almost any point in onto- geny (Bridges 4, Muller 8). (There are also indica- tions that changes of this sort may take place in purely somatic tissue, although in such cases, of course, the modi- fication cannot be perpetuated. See also chapter viii on this matter.) Zeleny (12) states that there is no periodicity to these mutations, thus refuting one of the early ideas of De Vries. The same investigator demonstrates that reverse mutations are more frequent than original muta- tions. (This, however, is simply because they are in the reverse direction, and not because of their recent origin.) In the case of these reverse mutations, the changes are always full jumps back to the original starting-point, and never result in an intermediate con- dition; nor will the selection of extreme types at all modify the rate at which these reverse mutations occur. MuLLER and Altenburg (10), who have conducted a critical examination of the fruit fly for mutations occurring on the first and second chromosomes, state that the vast majority of locus changes have a lethal or semilethal effect when present in the homozygous Mutation III (recessive) condition. (It is obvious, therefore, that a critical search for these mutations must in\'olve a very special technique. These authors are in possession of such a technique through their intimate knowledge of the linkage groups on the chromosomes in question, and their ability to detect the absence of certain expected classes.) On one chromosome they uncovered the start- ling fact that 50 per cent of the mutations were located in a restricted region at one end of the chromosome, which amounted to about 2 per cent of its length as charted from cross-over values. (It is an open question whether this indicates a highly mutable region of the chromosome, or whether cross-over values are an inaccurate index of length.) The most promising phase of Muller's work arises from his critical study of the rate of mutation. Consider- ing the whole length of the first chromosome of the fruit fly, one mutation occurs in 106 gametes. For the second chromosome the corresponding value is one in 175 gametes. Zeleny states that locus changes occur as frequently in one sex as in the other. Having estab- lished these constants, Muller is now investigating the possibility of modifying the normal rate of mutation. Already he has been successful in depressing the rate one- half by means of low temperatures. Eventually such knowledge may be turned to some practical \'alue. Two further points should be mentioned about the locus changes. Variation of this type has been encoun- tered (or at least identified) much more frequently than have any of the other types of changes mentioned below. The term "mutation" is usually restricted by geneti- cists to apply to locus changes. 112 Outline of Genetics The presence and absence hypothesis has been very generally accepted because of its value in simpHfying our concepts and pro- viding the most convenient scheme of notation. At the same time it is recognized that this hypothesis may not strictly represent the truth in all cases. There are two outstanding types of evidence that can hardly be interpreted by means of the presence and absence concept. One of these will be discussed later in another connection (p. 151). The other is as follows. In some cases other possibihties may be realized in connection with a single locus than merely the presence or absence of a given gene. For example, at a given locus on one of the chromosomes in corn, a condition W may exist, which results in colorless pericarp. In other plants that same locus may bear the gene F, variegated pericarp, and in still others S, completely colored pericarp. S is dominant to V and V to W (see also p. 119). As a matter of fact, these three are simply representatives of a series of ten different conditions that may be present at a given locus. Such cases are spoken of as systems or series of multiple allelomorphs, and it would be difficult to harmonize them strictly with the presence and absence hypothesis. The relation of systems of multiple allelomorphs to mu- tation is significant. It would be possible to arrange the genes involved in any system of multiple allelomorphs in a series, placing at the top the one which was dominant to all the others, and at the bottom the one that was recessive to all the others. This series, however, in no way reflects the order in which such genes have originated by mutation. In the fruit fly, there is a famous series of multiple allelomorphs for eye color, ranging from white through progressive steps in intensity of coloration; but it is not true that white first mutated to a light shade, which later mutated to the next darker shade, and so on; nor is it true that this series of mutants came off in a regular sequence down the scale of color intensity. In short, within a series of multiple allelomorphs the mutants come off in discontinuous rather than continuous series. It follows that mutations cannot be "led along" in a given direction by means of selection. Further, although it is common that all the genes in a series of multiple allelomorphs affect the same general character, exceptions Mutation 113 to this have been noted. Muller (9) cites a case in which the different mutant genes at the same locus may cause either shorten- ing of wing, eruption on thorax, or a lethal effect. At the same time, although we are thus repeatedly encounter- ing evidence on the discontinuity of mutation, it is possible that there is an underlying continuity of a sort that we are not in a position to measure. A statement of JVIorgan's (6) bears on this point. "Evidence is fast accumulating that common genes probably undergo analogous mutation in related species, the direction being conditioned by the physico-chemical constitution of the gene and not by some hypothetical directive force." 2. CoMPLEXMUTATiONS. — It is perhaps surprising that, in spite of the numerous cases of locus changes that were being discovered, there were for a long time no clear cases of mutations involving simultaneous changes in several neighboring factors in one region of a chromosome. Nilsson-Ehle (ii) now claims to have such a case, and calls it ''complexmutation." Normal wheat mutates to bearded speltoid, involving a simultaneous change in two closely linked genes. Among the F2 progeny of normal A' mutant appear a few bearded normal type and beardless speltoid, but only a few, due to the very close linkage of the two mutated genes. In another case the same investigator claims that three linked factors have mutated simultaneously. 3. Deficiency. — A rare phenomenon has been described by Bridges (3), working on the fruit fly. "Deficiency" as he calls it, is something more extensive than a simple locus change (and probably more extensive than the complexmutations). It is a "regional mutation," involving an "inactivation" of a portion of a chromosome, so that the genes on that region of the chromosome are rendered ineffective (nor can crossing over take place in that region). ii. chromosome changes ('' chromosome aberrations") I. Duplication.— Bridges (3) describes another rare type of change in the germ plasm, to which he gives the name "duplica- tion." Judging from the very unusual breeding results obtained, some abnormality in connection with mitosis has resulted in the 114 Outline of Genetics appearance of an extra piece of chromosome which dupHcates in its gene content a known region of one of the normal chromosomes. An organism in this condition is really triploid with respect to a part of one of the chromosome sets. 2. Non-disjunction. — ^This phenomenon, made fa- mous through the classic work of Bridges on the sex chromosomes of the fruit fly (chap, xiii), may prove to be a fairly common occurrence. In an irregular reduc- tion division, one of the chromosomes fails to '' disjoin" properly from its mate. As a result, one or two gametes are formed with an extra chromosome, and others which lack this chromosome. The latter fail to function, but a mating of the former with a normal gamete would produce a zygote with an extra chromosome. Blakes- LEE, Belling, and Farnham (2) have discovered this phenomenon in the jimson weed. Datura. The normal diploid number of chromosomes in this form is twenty- four. Twelve different '^ mutants" have been discovered with twenty-five chromosomes each. This seems to indicate that each of the twelve chromosomes (haploid) has failed to disjoin at least once in history. These twelve new forms are abnormal in their vegetative fea- tures, notably low in fertility, and tend to revert to the normal diploid ancestor. 3. Tetraploidy. — A hurried or incomplete mitosis will sometimes result in the simultaneous duplication of all of the chromosomes. This phenomenon has been observed several times in culture, and there are indica- tions that it has taken place frequently in the past. A general survey of the chromosome counts in our exist- ing plants and animals emphasizes the fact that the haploid number is much more frequently an even number Mutation 115 than an odd one. This, together with the fact that there are several species groups in which the chromo- some count of some of the members is just twice that of the others, suggests that tetraploidy may ha\'e played a considerable role in evolution. Tetraploidy commonly, but not always, brings gigantism. Blakeslee now puts the finishing touches on this tetraploidy conception by more work on Datura. In addition to the abnormal forms with twenty-five chromo- somes, he has discovered one completely triploid (thirty- six chromosomes) and one tetraploid form (forty-eight chromosomes). These latter both seem to be in a "better-balanced" condition than the non-disjunctional (twenty-five chromosome) forms, since they are more ''normal" with respect to their vegetative features and fertility. . The beauty of the situation arises from the fact that the tetraploid type contains a previously known Mende- lian factor. In normal diploid forms a hybrid of the composition Aa will give a 3: i ratio of purple flowered and white flowered in the F2. The tetraploid hybrid AAaa gives gametes in the ratio 1 AA:^ AA; i aa. Chance matings among these results in an F^ of 35 purple: I white. The F3 and later generations behave according to expectations on this basis. As stated before, the term "mutation" is now commonly restricted to locus changes. The author has not discovered the conventional term to include all of the foregoing cases unless it be merely "germinal variations." The bearing of these phenomena upon evolution miglil be considered briefly. Until a few years ago the general belief on evolution included the following notions: inheritance of acquired characters has been exploded; Darwinian variations are rather ii6 Outline of Genetics dubious as a basis for explaining evolution; but mutation, with natural selection among the mutants, will doubtless account for most of the facts. Now, in view of the more accurate knowledge of the mutation phenomenon that has been developed in recent years, the adequacy of mutation in explaining evolution must be considered more critically. First of all, it is evident that "complexmutation," "defi- ciency," and "duplication" could have played no important part in evolution, merely on account of the extreme rarity of these phenomena if for no other reason. Locus changes are sufficiently common, but consider the quahty of the mutants which result! In practically all cases the change is a "loss" mutation, and surely evolution cannot be accounted for on such a basis! One might merely regard this as evidence of the "trial and error" method by which nature operates, only rarely making those "gains" which must serve as the basis of progressive evolution. A few "gain" mutations have been reported, but there is reason to suspect that even these may be merely "reverse" mutations, regaining that which had previously been lost. Furthermore, the locus changes that have been reported, be they losses or gains, have seemed consistently non-adaptive. In short, it is difficult to imagine how progressive evolution can be accounted for either through single locus changes or through the accumulation of numerous locus changes. One can readily admit that such changes may account for the multiphcation of varieties or even species "on the same level," but can hardly be convinced that "our larger phylogenetic edifices have been erected from such building blocks." It is quite likely, however, that our knowledge is still too limited to visualize the evolution of the ages in terms of what we have seen happening during a very few years. Non-disjunction is out of the question as a basis for evolution- ary progress. The resulting "unbalanced" forms are clearly abnormal, and it is very doubtful whether they could permanently perpetuate themselves under the most favorable conditions, much less survive under conditions of sharp competition and environ- mental stress. Tetraploidy might well account for a certain amount of evolu- tionary progress, and we have good evidence that it has actually Mutation 117 (lone so in the past. 'J'he tetraploid form has more doses of desir- able growth factors than had its diploid ancestor, and as a rule is visibly more vigorous in one respect or another. A theoretical limitation, however, applies here also. Tetrai)loidy involves merely a quantitative gain, a multiplication of genes already present. One feels that not much progressive evolution could take place without the appearance of qualitatively new genes and the production of distinctly new adaptive characters. It would be safe to state that the foregoing considerations have caused many biologists to feel less certain in explaining evolution than they were a few years ago. This loss of faith in mutation, taken together with recent discoveries on inheritance of acquired characters (see chap, ii), is causing many to seek an explanation of progressive evolution in Lamarckian terms. It is of some interest to note that the original ''classic" examples of mutation provided by De Vries are probably not gen- uine cases. It had long been suspected by some that Oenothera Lamarckiana was a hybrid and its ''mutants" merely extracted recessives, but it was difficult to account on this basis for the very small number of "mutants" that were thrown every generation. MuLLER (7) probably deserves the credit for solving this vexing problem. In the fruit fly he discovered an essentially true- breeding hybrid race and explained it by a system of balanced lethal factors. These factors assert their lethal effect only when they occur in the homozygous recessive condition. In this race of flies, two such factors are present in heterozygous condition on the same pair of chromosomes, the dominant members of the heterozygous sets being on the opposite chromosomes of the pair. Such a hybrid continues to breed true as such, since any attempt to segregate brings the homozygous recessive condition of one or the other lethal with resulting death to the progeny. The reces- sives of any heterozygous set of genes on this same chromo- some pair will remain concealed when this stock is allowed to inbreed. Occasional crossing over will cause the appearance of a few of these recessives (in predictable frequencies), like the "mu- tants" thrown by Oe. Lamarckiana. In fact, De Vries himself now subscribes (5) to an explanation fundamentally similar to the preceding. About one-half of the ii8 Outline of Genetics seeds of Oe. Lamarckiana are empty. De Vries explains this by saying that Lamarckiana produces two kinds of gametes, the typical or laeta and the veliUlna. Each gamete has a lethal factor closely linked with the character factor. Heterozygous combina- tions give good seeds, homozygous give sterile. If one of the two lethal factors becomes "vital," the Oe. laeta or Oe. velutina muta- tion appears. LITERATURE CITED 1. Baur, Erwin, Mutationen von Antirrhinum majus. Zeit. Induct. Abstamm. Vererb. 19:177-193.^^^. 10. 1918. 2. Blakeslee, a. F., Belling, John, and Farnham, M. E., Chromosomal duplication and Mendelian phenomena in Datura mMidMis. Science 52:388-390. 1920. 3. Bridges, Calvin B., \'ermilion-deficiency. Jour. Gen. Physiol. 1:645-656. 1919. 4. , The developmental stages at which mutations occur in the germ tract. Proc. Soc. Exp. Biol, and Med. 17:1-2. 1919. 5. De Vries, Hugo, Phylogenetische und gruppenweise Artbil- dung. Flora 11-12:208-226. 1918. 6. Morgan, T. H., Evolution by mutation. Sci. Monthly 5: 46-53. 1918. 7. MuLLER, H. J., Genetic variability, twin hybrids, and con- stant hybrids, in a case of balanced lethal factors. Genetics 3: 42 2-4gg. fig. I. 1918. 8. , Further changes in the white-eyed series of Drosophila and their bearing on the manner of occurrence of mutation. Jour. Exp. Zool. 31:443-473.^^. 3. 1920. 9. — -, Variations due to change in the individual gene. Amer. Nat. 56:32-50. 1922. 10. — , and Altenburg, E., A study of the character and mode of origin of eighteen mutations in the X chromosome of Drosophila. Anat. Rec. 20:213. 1921. 11. Nilsson-Ehle, H., Multiple allelomorphe und Komplexmu- tationen beim Weizen. Hereditas 1:227-311. 1920. 12. Zeleny, Charles, The direction and frequency of mutation in a series of multiple allelomorphs. Anat. Rec. 20:210-211. 1921. CHAPTER VIII BUD VARIATION The outstanding feature of bud variation is that we know very little about it. It is true that quite a number of cases of bud variation have been investigated, but it could hardly be said that altogether satisfactory explanations of the phenomena have as yet been provided. The relation between bud variation and the JVIendelian mechanism is difficult to visualize with much clearness, nor is it easy to interpret the various cases in terms of each other. Bud variation may be defined as variation originating in vege- tative tissue. Such variation might involve merely (i) 'fluctua- tion," a response to environmental stimulus, or it might involve (2) a change in the genetic constitution of the parts affected. Cases of type (i) need not concern us here, since such variations are not inherited. As for type (2), this should be subdivided into: (a) cases in which the variation involves both somatic and germinal tissue, and in which, therefore, the variation will be heritable through seed; and (b) cases in which the variation involves somatic tissue alone, the variation not being heritable through seed. With these distinctions in mind, we may consider a classification of the phenomena of bud variation, which is based primarily upon the ideas of Emerson (6). I. SOMATIC MUTATION OF GENES This may be illustrated by some of the findings of Emerson in corn. An illustration of variation involving both somatic and germinal tissue is provided by the behavior of pericarp color. 5 is a gene which results in self- (completely) colored grains, being dominant to V which produces variegated grains, and which in turn is dominant to W which produces colorless grains. (These are three members of a series of "multiple allelomorphs"; see p. 112.) Corn of the formula VW, and which should, therefore, have all the grains variegated, will at times have some grains that 119 I20 Outline of Genetics are self-colored. Further breeding reveals that somatic mutation has occurred in the tissues concerned with the formation of these grains, such that VW has become SW. The change has taken place not only in the somatic tissue of the grains in question, but also in the germinal tissue within the grains, and is inherited accordingly. It is concluded that the recessive variegation gene V has mutated to its dominant self-color allelomorph S. Without attempting a discussion of the breeding tests upon which EiviERSON based his conclusions, it will be worth while at least to mention some of the other peculiarities of this phenome- non. V mutates to 5" rather frequently, but IF never mutates to S. V in the heterozygous condition (VW) mutates to S five times as frequently as when it is in homozygous combination (VV). The mutation takes place late in ontogeny much more often than in early ontogeny. As a result there are many more cases where small patches of self-colored grains appear on variegated ears than where large patches appear. In corn of the VV formula, only one of the V genes ever mutates to 5 at a given time. Reverse mutations, S changing to F, have also been noted. This same material provides also an example of somatic muta- tion which involves the soma alone and not the germinal tissue. In situations essentially similar to those described above, there may appear on the variegated ears a few aberrant grains which are apparently self-colored only on the crown of the seed. This character has been designated as "dark-crown," and it is notable that it is never inherited. Microscopic examination of the dark- crown and of the fully self-colored seeds indicates that in the former the epidermis alone is colored, while in the latter the epidermis alone remains colorless. The conclusion seems warranted, there- fore, that the two types of variation are fundamentally the same, both being true gene mutations, and that the non-inheritance of the dark-crown type is due to the accident that it occurs in the epidermal tissue outside the germ tract. II. SOMATIC SEGREGATION It has been pointed out by several investigators that bud variations appear much more frequently in plants that are hetero- zygous for the genes concerned than in plants which are homozy- Bud Variation 121 gous. This may be true, but it would not be safe to conclude that there is any mechanism ever provided in somatic tissue which corresponds to the normal Mendelian mechanism for segregation during gametogenesis. If "somatic segregation" ever takes place, it is through the operation of some quite different mecha- nism, as indicated in the examples given below. 1. Chromosome elimination. — A theoretical illustration would be as follows. A plant heterozygous for linked genes, A-B a-b, has an irregular mitosis take place in some part of its somatic tissue. One of the daughter-nuclei of this mitosis fails to receive its full complement of chromosomes, the A-B chromosome hav^ing somehow been eliminated. This cell and its progeny, haploid now with respect to this chromosome pair, which is represented only by the a-b chromosome, will form tissue in which the recessive char- acters a and b will become manifest. This would be the principle underlying somatic segregation through chromosome elimination. As a matter of fact, there are really only two clearly demonstrated cases of this sort, and both of these are limited to rather special situations. In one of these cases, " gynandromorphism " in the fruit fly, a special chromo- some set is involved, the sex chromosomes. This will be taken up, therefore, in the chapter on "Sex determination." In the other case, endosperm "mosaics" in corn, a special triploid tissue, the endosperm, is involved. This case will be discussed in the chap- ter on "The endosperm in inheritance." 2. Chimaeras. — A chimaera is a plant in which some of the tissues have all of the characteristics of one variety or species, while the rest of the tissues on this same plant are characteristically those of a different variety or species. The most famous chimae- ras are the "graft-hybrids" of Solanum produced by W'inki.kk (lo). This investigator made grafts of two distinct species of this genus, the tomato and the nightshade. After the tissues of stock and scion had been given time to "weld" together, Wink- ler cut the stem in such a way that the exposed cross-section was made up partly of tissues of the stock and partly of scion tissues. From this cut surface, adventitious buds would arise, and at times these buds came at the exact point where stock and scion tissues were in contact. Such buds developed branches which 122 Outline of Genetics were sectoral chimaeras, the tissues of one side of the branch being those of the tomato, while the tissues of the other side of the branch were night-shade tissues. Such sectoral chimaeras would not infrequently later produce branches that were periclinal chimaeras, having the tissues of one species inclosed within an envelope of the other. That these were really periclinal chimaeras was established by chromosome counts (tomato and nightshade having different chromosome numbers), and by the fact that seedlings produced by them were always of the species of the subepidermal tissue from which the gametes arise. The periclinal chimaeras in turn were observed at times to produce branches wholly of one or the other of the parent-species, a performance which may well be regarded as a type of somatic segregation. Fundamentally, the same behavior has been observed in certain "natural" periclinal chimaeras (notably in types of Pelar- gonium, Baur 2), involving white (deficient in chlorophyll) and green tissues. The manner of origin of these natural chimaeras is unknown, but it is quite possible that they arose as somatic mutations. A very interesting case has been reported by Bateson (3) in Bouvardia, which presumably may be something of the same sort as the foregoing. Varieties of Bouvardia that are maintained true to type by propagations from stem cuttings produce plants with very different flower form, size, and color when propagated by root cuttings. Since in normally produced buds of the stem both the epidermis and the deeper lying tissues are maintained through direct cell lineage, while the roots produced by stem cuttings arise from the plerome and break through the periblem and dermatogen, forming these parts anew, sprouts that develop from the roots must have the genotype of the stele rather than that of the cortex or epidermis. It is clear that classes i and 2 represent distinct types of somatic segregation, the first arising as the result of irregular chro- mosome distribution and the second from a segregation on the part of tissues as a whole. Both might well be regarded as anom- alies, since they are to be explained by irregularities in the com- mon plant program. There remains to be considered one more type of somatic segregation, and here, although no such finely balanced Bud Variation 123 mcchaaism for segregation is involved as that of the reduction division, at least the segregation is effected with some regularity. 3. Cytoplasmic segrkgation.— Numerous cases of apparent segregation of cytoplasmic elements have been i)rovi(led in i)lants. All of them involve visible effects on chlorophyll and all show non- Mendelian inheritance. (Cytoplasmic segregation is the conclud- ing item in Emerson's classification of bud variation. In order better to bring out the relationship between cytoplasmic segrega- tion and certain other plant phenomena, this item will be taken up as a part of the following classification [from Winge 9] of cases of chlorophyll inheritance.) Chlorophyll Lvheritaxce I. IMendelian inheritance, the characters being "carried in" the nucleus. Quite an array of cases of chlorophyll deficiency have been found to be inherited according to the normal JMendelian mecha- nism. In this class have been noted albino, pale green, yellow, and variegated types which are (usually) inherited as simple Mendelian lecessives to the normal green condition. II. Non-Mendelian inheritance, the characters being carried in some extra-nuclear portion of the gametes (Emerson's cyto- plasmic segregation). I. Biparental inheritance, the male as well as the female par- ent contributing (presumably) both cytoplasm and plastids to the zygote. A. The chlorophyll character governed by the distribution of the plastids themselves. An example of this type of thing is provided by the work of Baur (2) on certain types of Pelargonium, where the following behavior has been noted. If a white-leaved plant (white-leaved branch, see below) and a normal green-leaved plant are crossed (either way), the resulting hybrid illustrates what has been called by some ''particulate inheritance"; that is. the hybrid is varie- gated, showing irregular patches of green and white. If one of these white patches completely includes a bud. there will probably be produced by that bud a completely white branch. The flowers of this branch, when self-fertilized, give rise through their seeds 124 Outline of Genetics to white individuals only, and would evidently continue to breed true to the white condition if white individuals could be matured. In like manner the variegated hybrid may give rise to a pure green branch, which would start a line of pure green individuals. A suggested explanation of this situation is that the white condition results from the occurrence of purely colorless plastids in the tissue, while the green condition has the normal green chloroplasts. A cross between the two types will introduce into the hybrid zygote a mixture of green and white plastids ("plastid primordia"); and the same result will be obtained whichever way the cross is made, since the male as well as the female parent con- tributes plastids to the zygote. During somatogenesis in the plant which develops from such a zygote, there will be an inevitable segregation of green and white plastids, since there is no mecha- nism provided for a perfectly even distribution to daughter-cells of those cell components which lie outside the nucleus. If the number of plastids per cell be not too large, sooner or later, through the operation of the laws of chance, cells will arise which contain plastids entirely of one sort or the other, and these will produce tissues which are pure green or pure white. B. The chlorophyll character governed by the distribution of other and finer cytoplasmic elements than the plastids them- selves. Again male and female parents both contribute the effective extra-nuclear elements to the hybrid zygote, and again a tendency tow^ard irregular segregation appears during somatogenesis of the resulting plant. In this case, however, the effective units are so small and numerous that a complete segregation of units of the two types is never achieved, but merely the production of rela- tively paler and relatively greener regions on the plant. The behavior of Ikeno's (7) albomaculaia type of Capsicum might be interpreted on this basis. Paler and greener patches occur on the albomaculata plants, and the average paleness of the whole individual may be greater or less. Since this "average paleness" of the parent is reflected in the nature of the offspring, from any sort of a cross, it is felt that such a parent produces gametes having a characteristic proportion of green and white elements or units, whatever these elements or units may be. Bud Varialio}i 125 2. Maternal inheritance, the male parent contributing only a nucleus to the make-up of the zygote. In cases of this sort the source of pollen makes no dilTerence so far as the chlorophyll character is concerned. Consequently, since this maternal inheritance cannot be accounted for by par- thenogenesis in the plants that were used, it has been concluded that the seat of the character in question is in something that the female parent regularly contributes and the male parent never contributes. This idea is supported by certain cytological evidence that suggests the fact that, in some plants at least, the male nucleus is '' stripped clean of its cytoplasm" at the time that it is discharged from the pollen tube into the embryo sac. A similar distinction to that made under i may also be applied here. A. The chlorophyll character governed by the distribution of the plastids themselves. It follows that pure green and white parts will at times segre- gate out in the variegated plants concerned. Correxs' (5) albomaculata type of Mirabilis is said to be an example of this sort of thing. B. The chlorophyll character governed by the distribution of other and finer cytoplasmic elements than the plastids themselves. It follows that absolutely pure green and pure white parts will never segregate out, but only relatively paler and relatively greener parts. WiNGE (9) cites some of his own work on a variegated type of Humulus as an example. In conclusion it should be said that opinion as to the seat of chlorophyll inheritance is by no means settled. It would doubtless be wise to regard the foregoing classification of W'in'ge's merely as a convenient form in which to arrange the available evidence; other investigators would certainl}' disagree with some of Win'ge's interpretations of the phenomena. As an example of a case which can liardly be forced into Winge's classification, An'dersox's (i) green and white variegated race of corn might be cited. Inheritance is strictly maternal, and pure green and white areas segregate out on tlie leaves. Presum- ably then this would fit into Win'ge's class II, 2. A. Hut a care- ful cytological investigation of this material by Randolph (8) has revealed that there can be no sharp segregation among green and 126 Outline of Genetics white plastids at play to account for the green and white areas; that it must be rather a matter of the ''general physiological condition " of the two types of tissue. It is interesting to note that in this case a type of somatic segregation occurs on the ear of the variegated plant, resulting in certain groups of seeds that will produce green seedlings, other groups of seeds that will produce white seedlings, and still other groups of seeds that will produce variegated seedlings. In short, the cases of chlorophyll inheritance on a non-Men- delian basis are still under considerable discussion; a perfectly clear interpretation of the phenomena is not as yet available. Of this much, however, we may be sure: there is such a thing as non- Mendelian inheritance, and it becomes manifest in connection with a type of character which, on other occasions, is inherited according to the normal Mendelian scheme. In any event, such cases should not be regarded as a violation of Mendel's law, but merely as something outside the scope of Mendel's law, since they are evidently transmitted by some extra-nuclear mechanism. In good part the known examples of non-Mendelian inherit- ance are limited to such cases of chlorophyll inheritance as have been cited above. There is another small group of cases, however, that must also be regarded as illustrating non-Mendelian inherit- ance, although in quite a different way. Bateson and his coworkers (4) have discovered certain instances (e.g., inheritance of doubleness in Matthiola) in which the male and female organs of the same plant differ in the factors they carry. A clear explana- tion of this phenomenon has not been provided, but whatever the explanation may turn out to be, it seems certain that it will provide an exception to the normal Mendelian mechanism. Such cases have led Bateson to suspect that plants, as genetic machines, differ fundamentally from animals, segregation being clearly connected with synapsis in animals but not always in plants. This difference in the machinery may be tied up with the fact that "in animals the rudiments of the gametes are often visibly sepa- rated at an early embryonic stage, whereas in the plant they are given off from persistent growing points." Bud Variation 127 LITERATURE CITED 1. Anderson, E. G., Maternal inheritance of chlorophyll in maize. Unpublished. 2. Baur, E., Das Wesen und die Erblichkeitsverhiiltnisse tier ''Varietates albomarginatae hort" von Pelargonium zonule. Zeit. Abst. und Vererb.-lehre 1:330-351. 1908. 3. Bateson, W., Root-cuttings, chimaeras and ''sports." Jour. Genetics 6: 75-80. 1916. 4. , The progress of Mendelism. Xature 104:214-216. 1919. 5. CoRRENS, C, \'crerbungsversuche mit blass (gelb) griincn und buntblatterigen Sippen bei Mirabilis Jala pa, Urtica pilulijcra und Lunaria annua. Zeit. Abst. und Vererb.-lehre 1:291- 299. 1908. 6. Emerson, R. A., The nature of bud variations as indicated by their mode of inheritance. Amer. Nat. 56:64-79. 1922. 7. Ikeno, S., Studies on the hybrids of Capsicum annuum. Part 11. On some variegated races. Jour. Genetics 6:201- 229. 1917. 8. Randolph, L. F., Cytology of chlorophyll types of maize. Bot. Gaz. 73:337-375. pis. 11-16. 1922. 9. WiNGE, O., On the non-Mendelian inheritance in variegated plants. Compt. Rend. Carlsberg 14:1-20. 1919. 10. Winkler, H., Die Chimarenforschung als IMethodc der experi- mentellen Biologie. Phys. IMed. Gesell. Wiirzburg. Jahrb. 1913, 1914. CHAPTER IX THE GAMETOPHYTE IN INHERITANCE Thus far the discussions have dealt with inheritance in sporo- phytes; in fact, genetics practically never considers ganietophytes, through which inheritance must pass from one sporophyte to the next. The reasons for this neglect are obvious. Practically all of our land vegetation is made up of sporophytes, and therefore practically all of our experimental material has been sporophytes. Furthermore, gametophytes are inconspicuous (out of sight in seed plants), hard to get at, hard to work with, and apparently of no economic importance. Besides, in animals, as is well known, the generation equivalent to the gametophyte of plants is repre- sented by only a few cell divisions in the maturation of gametes. In other words, the gametophyte has no significance as a discrete generation in the animal kingdom; and since inheritance in plants is of interest to the public chiefly because it throws some light upon inheritance in animals, there has been little demand for any knowl- edge of inheritance in gametophytes. It is not surprising, therefore, that very little study has been made of the gametophyte generation in inheritance. There are reasons for believing, however, that such a study might be very profitable. The gametophyte generation, with its haploid chromo- some number, would provide an interesting and critical test of the Mendehan mechanism of inheritance. Certain features of inherit- ance would be expected to differ radically from inheritance in sporophytes. A generalized example might be considered. Gametophyte Ab is characterized by exhibiting the A character but not the B. The reverse is true of gametophyte aB. A cross between the two would produce zygote AaBb, followed by a sporo- phyte of the same formula. The following gametophyte genera- tion would contain four types in equal numbers, AB, Ab, aB, and ab. This is the program that would have to be followed if the Mendelian mechanism were at play. One would expect, there- 128 The Gamelophyte in Inhentance 129 fore, that there would be the following outstanding character- istics of inheritance in ganietophytes. (i) There could be no possibility of dominance, since but one representative of an allclo- morphic pair could be present. Any discovery of blending inherit- ance in gametophytes (with respect to a monoh\'brid situation) would serve to cast doubt upon the Mendelian mechanism. (2) Clear segregation would appear in the first generation following the cross and the phenot^-pic ratios would be quite different from those encountered among sporophytes. 'I'he various factor t>pes might, of course, be expected to operate, but again the ratios obtained would be unusual. All of this, however, is little more than idle speculation, serving merely to point out discoveries that might be expected in the future. A study of inheritance in gametophytes might be profitable for other reasons also. Among those higher plants which have been the objects of genetical research, the sex act is a very well- insulated performance, occurring deep within the tissues. Fur- thermore, it is both prefaced and followed by quite a sequence of events, which we know must be taking place with considerable regularity but which we cannot control. Surely there would be much greater hope of any artificial manipulation of the sex act, making possible a more critical study of the germ plasm, in those organisms where the gametes themselves could be manipulated. It would seem that such things might be possible in those lower plants where the gametophyte is the dominating generation, although the technique necessary for such experiments would doubtless be difficult to develop. The actual work that has been done on inheritance in game- tophytes is practically nil. Pure line studies have been made in a number of thallophytes, selection has been attempted, and some mutations have been found, but none of the experiments has revealed anything critical on the matter of segregation of characters following a sex act. Transeau (2) has made some observations on ll\c green alga Spirogyra, which, while they did not involve any experimental work, were nevertheless quite suggestive. This author was famil- iar with several species of Spirogyra in their natural habitats, and noted several natural populations which were clearly mixtures of 130 Outline of Genetics two or more different species. Under such circumstances fila- ments of one species were observed in the act of crossing with filaments of another species. Furthermore, in such mixed popu- lations there were discovered some filaments which were evidently the products of previous crosses, for they clearly showed new com- binations of the characters of two species. It is to be hoped that exact experimental work will prove feasible with this genus. One instance of the clean segregation of characters in the gametophyte generation is to be found among angiosperms, Selling's semi-sterility among beans. This wdll be described later in the discussion of the general subject of sterility. For the most part the gametophyte of angiosperms seems to be merely an expressionless intermediate stage between succeeding sporo- phyte generations. As East (i) puts it: ''Modern discoveries tend more and more to show that the sole function of the gameto- phyte of the angiosperms is to produce sporophytes. The char- acters which they carry appear to be wholly sporophytic, the factors which they carry functioning only after fertilization." LITERATURE CITED 1. East, E. M., and Park, J. B., Studies on self-sterility. I. The behavior of self-sterile plants. Genetics 2:505-609. 191 7. 2. Transeau, E. N., Hybrids among species of Spirogyra. Amer. Nat. 53:109-119.^^5. 7. 1919. CHAPTER X STERILITY The subject of sterility in plants is only in part a matter of genetics. Many of the problems involved can be solved only by the physiologist, ecologist, or cytologist. Some phases of the subject, however, have been rather successfully interpreted in terms of genetics. Tentative outlines of the general subject will be presented, merely to show what parts of the problem are being attacked by the geneticist. First of all, sterility might be classified in terms of the effects produced : A. Sterility. Complete failure of the sex act. B. Semi-sterility. Failure of part of the pollen, or part of the ovules, or part of both. C. Self-sterility. Pollen and ovules functional in cross- fertilization but not in self-fertilization. A more comprehensive classification might then be arranged on the basis of cause, although such a classification, in our present state of knowledge, must be rather vague and uncertain. I. Environmental causes. (Merely a few examples will be indicated. This part of the subject properly belongs to the ecologist and physiologist.) The examples of environmental causes given below commonly result in complete sterility A, although under special circumstances situations corresponding to B or C might be set up. I. Conditions too moist. When species that have become adapted to relatively dry conditions are subjected, at the lime pollen is mature and shedding, to unusually moist conditions, the pollen grains may absorb enough moisture to swell up and burst prematurely, thus losing their usefulness. (The sex act might also be circumvented by hard rain coming immediately after pollen distribution, which 131 132 Outline of Genetics would wash the pollen grains off the stigmas, and thus result in a certain amount of sterility.) 2. Conditions too dry. Unusual drought at the period that the stigmas are receptive may dry the stigmatic surfaces to such a degree that pollen grains will not adhere and germinate. Conditions of this sort at times limit the setting of seed in such plants as corn. I and 2 provide an illustration of a principle which is frequently encountered in biology^ — -opposite extremes of conditions bringing a similar end result (although the intervening sequences of causes and effects will of course differ). Other illustrations of this same thing appear later in this classification. 3. Poor ''nutrition." This may so limit the development of plants that they fail to mature up to the point of effecting a sex act. 4. Good ''nutrition." A principle familiar to botanists is that the optimum growth conditions frequently maintain the plant in the purely vegetative phase, so that reproductive parts are not developed. (The plant physiologist describes this in terms of a carbohydrate : nitrogen ratio.) 5. Season too short. Plants adapted to a long growing season are unable to com- plete their normal life-cycle up to the point of successful reproduc- tion when grown in regions which have a short season. 6. Unusual light conditions. Flowering and fruiting of many kinds of plants is induced by exposure to specifically favorable length of day which varies widely with the species (see Allard and Garner i). Radical departures from the customary seasonal program in this matter may serve to inhibit flowering and fruiting. (It may be that the underlying causes involved here are similar to those of 3 and 4.) Other examples of environmental causes for sterility could doubtless be provided by the plant ecologist. II. Large evolutionary tendencies. A. Sterility. Within certain groups of plants, what is apparently the natural phylogenetic sequence of genera and species indicates Sterility i t,2> evolutioiuiry progress in the dirccLion of cslablisliiiig partheno- genesis. (The most notable example is i)rovi(le(l by the C'om- positae, although it would certainly be unsafe to conclude therefrom that the ultimate goal of plant evolution was parthenogenesis. Quite on the contrary, one feels that the Compositae have seriously handicapped themselves, so far as future possibilities are concerned, by a ''freakish'' evolutionary maneuver.) C. Self-sterility. In effect evolution among angiosperms has achieved self- sterility through the perfection of devices favoring cross- pollination. Here should be cited: floral adjustments insuring cross-pollination by insects; protandry and protogeny; dioecism. (B. Semi-sterility. This class has little more than a theoretical existence here, although doubtless some forms might be found exhibiting the ''incompletion" of some of the above-mentioned evolutionary tendencies, and thereby exhibiting what is in effect semi-sterility.) III. Phenomena of genetics. Here are included cases where the mechanism underlying the phenomenon is affected by breeding operations. A. Sterility. I. Wide crosses. Crosses between distantly related parents may be effective in producing first generation hybrids which may be notably vigorous individuals (see chapter on ''Hybrid vigor"), but quite sterile. A notable example of this is the cabbage-radish hybrid, which achieves astounding proportions, but is completely sterile (Gravatt 8). Here the loss in efficiency in the reproductive system is dis- tinctly not accompanied by loss in efiiciency in vegetative develop- ment. This peculiarity is clarified by the following idea. Wide crosses involve the fusion of relatively ''inharmonious" gametes, w^hich might be expected to produce disturbances in the ontogeny of the resulting individual. The grosser mechanism which regu- lates vegetative development can evidently weather such disturb- ances, while the more finely balanced mechanism of gamete formation is upset. 134 Outline of Genetics 2. Inbreeding. (This again might be used as an example of opposite extremes of causes producing the same end result, for inbreeding can surely be regarded as the antithesis of wide crosses.) Inbreeding commonly results in "loss of vigor," through the production of weakling and degenerate plants of various types (see chapter on "Hybrid vigor"). Frequently these degenerate types exhibit faulty and ineffective reproductive parts; Types of this sort have frequently been obtained through inbreeding corn. 3. Definite hereditary factors. A good example of this appears in the case of "tunicate" or "podded" corn. Plants homozygous for the tunicate factor are sterile, while the heterozygotes are partially sterile (Eyster 7). B. Semi-sterility. 1. Wide crosses. The hybrids produced by wide crosses are not in all cases completely sterile (see III, A, i), but merely show an abortion of part of the gametes, notably part of the pollen. This phenomenon is, in fact, of such general occurrence that the existence of a certain amount of defective pollen is frequently used as a criterion of hybrid origin. ]VIany plants in nature have been found to show this characteristic; and such plants have been called "crypt- hybrids," the impHcation being that they are hybrids that have resulted from natural crossing. 2. Inbreeding. Some of the degenerate plants that commonly appear as a result of inbreeding (see III, A, 2) are not completely sterile, but merely unsuccessful in setting more than a few seeds. This may be due to a failure of part of the pollen or part of the ovules or part of both. 3. Definite hereditary factors. Corn which is heterozygous for the tunicate factor (see III, A, 3) is partly sterile. Here also comes a very unique case, which will be described in some detail, since it not only provides an ideal example of sys- tematic semi-sterility through the operation of definite hereditary factors, but at the same time it provides an example of inheritance in the gametophyte generation. Sterility 135 Bej.ling (2) made a cross between two races of beans, bolli of which were completely fertile. 'I'he resulting hybrids were semi- sterile. Uniformly just one-half of the pollen grains ai)fK'ared empty and collapsed, while one-half of the ovules had no embryo sacs. The sterile pollen and ovules appeared in random distribu- tion with the fertile. Inbreeding the semi-sterile hybrids, Belling obtained an F, generation which showed the following features: one half of the plants had perfect pollen; the other half had a mixture of equal numbers of good and bad pollen grains in all their flowers. The plants which had perfect pollen also had perfect ovules, while the plants with 50 per cent sterile pollen also had 50 per cent sterile ovules. In the F3 generation all the descendants from the fertile plants had perfectly good pollen and ovules; but the progeny of the semi-sterile plants again split up into the two classes, fertile and semi-sterile, as before. Belling states his general conclusion as follows: "The e.x- planation of the random abortion of one-half of the pollen and one-half of the embryo sacs must apparently be by the segre- gation of Mendelian factors among pollen grains and embr>'o sacs individually, and not by the action of these factors on the zygotes." To make this situation clear a diagram (fig. 21) may be con- sidered. It enlarges a little upon Belling's original ideas as he stated them, and emphasizes the sporophyte-gametophyte rela- tionship. Pollen grains and embryo sacs are gametophytes in the sense that they include the male and female gametophytes, so that W'hen the diagram shows sterile gametophytes it is the same as saying that both pollen grains and embr}'o sacs are sterile. This, of course, is just what Belling found; whenever one-half the pollen grains in random distribution were sterile one-half the embryo sacs in random distribution were also sterile. It should be remembered that Belling started with two com- pletely fertile races. Suppose that the parent race . 1 had a factor A' whose absence brought sterility in the ganietophytes (pollen grains or embryo sacs). Race B had a different factor 1', with similar effect, but inheritetl independently. When Helling crossed these races, all of the resulting F, hybrids were semi- 136 Outline of Genetics PARENTi RACE ; SPOROPHYTE: A ® ( GAMET0PHYTES:1 (FERTILE) (FERTILE) PARENT) RACE f SPOROPHYTE: B GAMETOPHYTES:/ (FERTILE) (FERTILE) F, 1 SPOROPHYTE; GAMETOPHHES: (FERTILE) (FERTILE) (STERILE) (STERILE) Fig. 21. — Diagram illustrating Belling's explanation of semi- sterility. Stcrilltv 137 sterile. In other words, in every 1', plant one-hiilf the game- tophytes were sterile. It is easy to see why XO and OY arc fertile, also why 00 is sterile (lacking both factors); but why should XY be sterile when it has both factors ? Belling explains it by saying that ganietophytes are unHkc sporophytes in that they normally have single factors instead of double factors. The germinal capacity of a gametophyte is just one-half that of a sporophyte. It is as if a gametophyte were "supersaturated" by a double factor. Such a situation is abnormal for a gametophyte and brings ab- normal results. Therefore the gametophyte having the abnormal double dose {XY) is just as sterile as the gametophyte with no dose {00).^ In developing the F2 ratios of course only the fer- tile gametophytes function. XY and 00 are eliminated, so far as posterity is con- cerned, so that we have to deal only with the chance matings among the fertile gametes {XO and OY). According to the laws of chance there are four possible matings between these gametes (fig. 22). Out of the four resulting F2 sporophytes two would evidently pro- duce only fertile gametophytes and would remain fertile as long ^ Belling's rather awkward assumption to the efTect that the gametes with the double dose {XY) are non-functional on account of being "supersaturated" might be improved upon by the following. Assume race A has complementary factors A' and O for fertility, while race B has a similar set of complementary factors, () and Y, located on the corresponding chromosomes. The scheme then works out as before, gametes of the BO and AT' formulas both being non-functional for the same reason that a necessary pair of comiilemcntary factors is not present. ® Fig. 22. — Diagram showing how the F2 would be produced according to Belling's idea of semi-sterility. 138 Outline of Genetics as they were inbred. The other two are exactly Hke the original Fi hybrid and therefore semi-sterile, having one-half sterile gameto- phytes. The whole dynasty may be represented as follows: p F. Fertile X Fertile Semi -sterile F. 50% Fertile Fertile Fertile 50% Semi-sterile F3 50% Fertile ^^ Fertile 50% Semi-sterile 1 1 F, 50% Fertile 50% Semi-sterile This is a very ingenious scheme, and, like most others, should be tested by further experiments. To a certain extent it has already met this test, for Belling (3) has subsequently reported a few more generations in which the breeding results were entirely con- sistent with those of the earlier generations. Also he has dis- covered two new races of beans which give similar results. C. Self-sterihty. This is a phenomenon which has loomed up as a great practical problem during the last few years. Where pollen and ovules are entirely healthy, and functional in out crosses, but quite ineffective in bringing about self-fertilization, the condition of self-sterility is said to exist. This frequently appears in certain of our impor- tant horticultural plants, such as apples and plums, so that it has often been found necessary to include a mixture of several varieties within the orchard in order to insure effective pollination. IVIany investigators have been attacking this problem with various tech- niques, the most critical work from the point of view of geneticists being that of East on Nicotiana. By crossing self-sterile with self-fertile species, East (6) has demonstrated that self-sterility behaves in inheritance like a simple IVIendelian recessive to self-fertility; a single gene evidently determines the difference between the two conditions. East (4) has further made an attempt to analyze the relations of self- Slcrilily 139 sterile plants inler se. In his earlier cultures he had found that the self-sterile plants were consistently cross-fertile; that is, there might ])e quite a group of individuals each one of which would set seed when pollinated from any of the other individuals, but would not set seed when self-pollinated. East concluded that, when one is deahng with self-sterile plants (those lacking the gene for self-fertility), pollen is effective only when it comes from a source that has a somewhat different germinal constitution from that of the ovules. If, therefore, a group of self-sterile plants is consistently cross-fertile, it is to be concluded that every individual of this group differs in some degree from every other individual of the group with respect to a certain set of factors that is effective in this connection. If this assumption is correct, it should be possible in the later generations to obtain groups of individuals all of the same genotype with respect to the effective factors. The individuals of any such group should then be cross-sterile with reference to each other. East actually obtained such groups among the later generations, thus supporting his assumptions on the relations of self-sterile plants inter se. An exact factorial analysis has not been possible as yet, but it is plainly a matter of Mendelian inheritance, and the general mechanism is rather clearly indicated. Much work remains to be done on the physiology of self- sterility, although a few interesting findings have already been made on that matter. It has been discovered (at least for a great many cases of self-sterility) that the problem is tied up with the growth of the pollen tube. Own pollen, quite healthy and func- tionable on foreign stigmas, will also germinate and start pollen tubes on own stigmas. Such tubes, however, are for some reason not successful in reaching the ovules. Assumptions were made, by various authors, that own stigmas poison own pollen tuljes or furnish them with inadequate nutrition. One author (Moore 9) has assumed that own stigmas provide own pollen tubes with too good nutrition, so that the tubes fatten but do not elongate (just as the hypha of a fungus will elongate more on a poor nutri- tive medium than on a good one). East (5) himself has done some critical work, however, that indicates the inaccuracy of all of the foregoing assumptions, and reveals an interesting phenomenon I40 Outline of Genetics that actually accounts for the results in the matter. Through careful sections of stigmas made at intervals, East discovered the following fact. Own pollen germinates on own stigmas just as readily as does foreign pollen, and the first increment of growth of the two types of tubes takes place at the same rate. After that own pollen tubes continue to grow steadily and "normally" at the same rate at which they started, but the rate of growth of foreign pollen tubes is continuously accelerated, as though they were receiving some stimulus which was ineffective on own pollen tubes. The result is that own pollen tubes fail to reach the ovary before the stigma and style have decayed, while foreign pollen tubes, with their accelerated growth, ''reach the goal before the road has become blocked." LITERATURE CITED 1. Allard, H. a., and Garner, W. W., Flowering and fruiting of plants as controlled by length of day. U.S. Dept. Agric. Yearbook, 1920. pp. 377-400. 2. Belling, John, K study of semi-sterility. Jour. Heredity 5:65-75. 1914. 3. , A hypothesis of semi-sterility confirmed. Jour. Heredity 7:552. 1916. 4. East, E. M., and Park, J. B., Studies on self-sterility. I. The behavior of self-sterile plants. Genetics 2:405-609. 1917. 5. , Studies on self -sterility. II. Pollen tube growth. Genetics 3:353-366.^^5. 3. 1918. 6. East, E. ISI., Studies on self -sterility. III. The relation between self-fertile and self-sterile plants. Genetics 4:341- 345- 1919- 7. Eyster, W. H., The linkage relations between the factors for tunicate ear and sugary endosperm in maize. Genetics 6: 209-240. 192 1. 8. Gravatt, F., a radish-cabbage hybrid. Jour. Heredity 5: 269-272. 1914. 9. Moore, C. W., Self-sterility. Jour. Heredity 8:203-207. 1917. CHAPTER XI THE ENDOSPER]\I IN INHERITANXE We have dealt chiefly with inheritance in the sporo- phyte, in connection with which most of the work in plant genetics has been done. Brief mention has been made of inheritance in the gametophyte, on which there has been very little work. It is appropriate now to con- sider inheritance in the endosperm. This classihcation raises the question as to the nature of the endosperm. It was at one time generally regarded as belonging to the gametophyte generation, but since the discovery of "double fertilization" in 1898 many have claimed that it belongs to the sporophyte generation. On the basis of chromosome numbers, it is neither, so that there is also the claim that endosperm is neither sporophyte nor gametophyte; at least we are justified in considering inheritance in endosperm as a separate topic. As might be inferred, endosperm shows some features character- istic of a gametophyte, others characteristic of a sporo- phyte, and still others peculiar to itself. Judgment as to its nature, therefore, will depend on which of these fea- tures is emphasized. It is generally believed that angios})erms have been derived from gymnosperms, and it is natural therefore to explain angiosperm structures by the corresj)onding structures of gymnos])erms. 11ie gyninosj)erni and angio- sperm ovules are contrasted in fig. 2^, whicli will assist in the following discussion. In gymnosperms the situa- 141 142 Outline of Genetics tion is clear. After the germination of the megaspore, everything within the old megaspore wall is gametophyte Archegonium Egg Old Megaspore \ { ^^ \__ \- Fusion Nucleus Wall, Inclosing Female Gametophyte _ Old Megaspore Wall, Inclosing Female Gametophyte Gymnosperm Angiosperm OVULE Gymnosperm Embrya Megaspore Wall Endosperm SEED Angiosperm Embryo Megaspore Wall Endosperm Fig. 23. — Diagram contrasting young ovules and mature seeds of gymnosperms and angiosperms. tissue; fertilization affects the egg only, resulting in a sporophyte embryo. In the seed, therefore, the embryo is imbedded in nutritive tissue which is evidently the The Endosperm in InJicritance 143 vegetative body of the female gametophyte, and this tis- sue is called the endosperm. In angiospcrms, however, a new situation introduces doubt. It can be said as before that after the germination of the megaspore everything within the megaspore wall is female gametophyte tissue, but it cannot be said that fertilization affects the egg only, for one of the sperms fuses as regularly with the fusion nu- cleus as does the other sperm with the egg. It will be re- membered that the fusion nucleus is formed by two nuclei, which have migrated from each end of the sac, so that when the sperm enters into the fusion there is a triple fusion. After fertilization the fertilized egg, of course, forms the ombryo sporophyte, but usually every nucleus of the old gametophyte disappears except the fertilized fusion nucleus, which then forms the endosperm in which the young sporophyte is imbedded. For this reason the fertilized fusion nucleus is usually called the endosperm nucleus. A comparison of the angiosperm and gymnosperm seeds reveals the following contrast (fig, 23). In the appearance of their essential structures, they are exactly alike, and on that basis some might claim that the endo- sperm of angiosperms is the same as that of gymnosperms, that is, gametophyte tissue. The opposing claim is that, although the gymnosperm endosperm is gametophyte tissue, the situation in angiosperms is essentially dif- ferent. In angiosperms, the endosperm docs not arise from morphologically unmodified gameto})hyte tissue, as in gymnosperms, but entirely from the en(losj)erm nucleus, and this nucleus is clearly the product of fusion of male and female nuclei. With such an origin, the endosperm nucleus is comparable with the zygote, and 144 Outline of Genetics the endosperm tissue is sister to the embryo sporophyte. In other words, in angiosperms the endosperm and embryo are twins. This means that the endosperm of angiosperms belongs to the sporophyte generation, although of course it is a distinct individual which pro- duces no progeny. The embryo sporophyte is a parasite upon its twin and devours it. It will be recognized that there is some reason for both of these claims. Is there any way of testing the claims, that is, of distinguishing between sporophyte and gametophyte tissue ? The cytological distinction, based on chromosome count, is that the sporophyte is 2X tissue and the gametophyte is x tissue. Applying this test, it is found that endosperm tissue is neither x nor 2X, but 30;, as might be expected from the triple fusion. The conclusion involves several possibilities, as follows: 30: is evidently nearer 2X than x, and therefore endosperm tissue is more like sporophyte than gametophyte tissue; but on the other hand two of the x's have come from the female gametophyte, and therefore two-thirds of the endosperm is female gametophyte. On the basis of predominance, therefore, endosperm tissue is more like female gametophyte tissue than anything else. Finally, there is a third alternative, and that is that the 30; con- dition deserves to be set apart in a category by itself, which would mean that endosperm is neither gameto- phyte nor sporophyte. These are the claims and the evidence as to the angiosperm endosperm. Opinion is not settled, but the facts are clear. This prepares for a consideration of the bearing of this situation upon genetics. The geneti- cist is not much concerned about the exact morphologi- The Endosperm in InJieritance 145 cal or physiological nature of endosperm, but he is much concerned about its behavior in inheritance. J*erhaps the phenomena of endosperm inheritance may help to decide whether endosperm is gametophyte or sporophy te or neither. Certain races of corn have yellow endosperm, while in other races it is white (colorless). If a cross is made with pollen from the yellow endosperm race on the silks of the white endosperm race, what results would be expected ? We could assume that yellow is dominant over white, since yellow is probably due to the presence of a factor which is absent in white. In making such a cross, therefore, we should expect a hybrid embryo to be formed which would show the dominant character of yellow endosperm when this embryo becomes a plant bearing ears the next season. On the contrary, we tind the dominant yellow character appears the same year that the cross is made. The cross, of course, puts the yellow endosperm factor in the young hybrid embryo, but we cannot imagine that this embryo passed the character out into the endosperm that surrounds it. The real mechanism is as follows. Some time after this phenomenon was discovered in 1872, it was named xenia (in 1881), the definition of the term being the direct effect of foreign pollen upon the endosperm. At the time of its discovery the mechanism involved in xenia was not understcK>d. Later, double fertilization was discovered, and this furnished the neces- sary mechanism. A pollen grain from the yellow endo- sperm race contains two male gametes, and each gamete contains the factor for yellow endosjHTin. One of the gametes fertilizes the egg and produces a hybrid embryo, 146 Outlme of Genetics which, in the next generation, behaves as a heterozygote for yellow endosperm. The other male gamete fertilizes the fusion nucleus and produces the endosperm nucleus, which therefore contains the factor for yellow, the result being that the endosperm is yellow, although the ovule belongs to the white race. Xenia means, therefore, that the endosperm is a hybrid as well as the embryo, and the ''triple fusion" involves the transmission of hereditary characters. Fertilization of the fusion nucleus is just as essential as fertilization of the Qgg, and so far as inherit- ance is concerned the endosperm and embryo are sister- sporophytes. The exact function of double fertilization is not clearly understood. Nemec (7) has sought to account for endosperm hybridization as an adaptation which results in a better adjustment of the composition of the reserve food supply to the needs of a hybrid embryo. Xenia throws considerable light on the nature of endosperm. Because of its behavior in inheritance, geneticists would naturally regard the endosperm as a sporophyte, an abnormal sister to the embryo. The phenomenon of xenia is not limited to the case of yellow endosperm, but appears in connection with quite a number of endosperm characters. The red- grain and purple-grain characters in corn, which were employed to illustrate types of factor interaction, are also governed by this mechanism. In these cases, how- ever, an additional detail appears. A section of a grain of corn appears in hg. 24. There is first the pericarp or "seed coat," which is the ovary wall, belonging to the old sporophyte, and therefore does not concern us. Within this is a thin alcurone layer, which is the outer TJie Endosperm in Tuheritance 147 layer of endosperm, while the bulk of the seed consists of the starchy endosperm. Since aleurone is endosj)erm, colors peculiar to it would show xenia in inheritance. This was the case in East's red and purple corn, the colors being located in the aleurone layer. There is another phase of the situation to which attention should be called. By pollinating the silks of a white-grained individual with pollen from a red- grained individual, xenia is secured, the resulting grains erm mbrvo Cross-Section of Com-Seed Fig. 24. — Diagram of corn seed being red like those of the pollen parent. In the reciprocal cross, however, that is, pollinating silks of a red-grained individual with pollen from a white-grained individual, a different result is obtained. The resulting grains are not white like those of the pollen parent, but red like those of the ovule parent. There is no xenia, therefore, for the pollen has no immediate effect upon the develop- ing endosperm. This seeming difliculty, however, is easily explained. When the pollen parent is white and the ovule parent is red, the endosperm gets its characters from both parents, and since red is dominant ()\'er white the resulting endosperm will be red because the female nuclei that entered into the triple fusion carried the factor for red endosperm; and therefore the pollen from the white ])arent seemed to ha\c no cIUh t. 'i1u' mecha- nism works in all cases, but, owing to dominance, xem'a appears only in certain cases. 'Jliere is no need to dis- 148 Outline of Genetics cuss all of the Mendelian situations in which xenia may occur. An understanding of the underlying mechanism should enable us to analyze such cases and reach a con- clusion as to the expected results. A law which East (i) has formulated in reference to xenia is pertinent: ''When two races differ in a single visible endosperm character, in which dominance is com- plete, xenia occurs only when the dominant parent is male (pollen parent). When the two races differ in a single endosperm character, in which dominance is incom- plete, or when they differ in two characters (factors), both of which are necessary for the development of the visible difference, in both of these cases xenia occurs when either parent is male." This may be called the law of ''normal" xenia. What may be called "abnor- mal" xenia should now be considered. In connection with some of his work on sweet and starchy corn, East (i) was able to distinguish two distinct races of starchy corn. In one race the starch occurred in a loose powdery or floury condition, while in the other race it was compacted into a hard, flinty, or so-called corneous condition. The two races, therefore, may be spoken of as floury and corneous races of starchy corn. East made various crosses between these two races to discover the method of inheritance of the two endo- sperm characters. Naturally such characters would be expected to show xenia. In the following description, therefore, when the Fi generation is referred to, both the hybrid embryo and the hybrid endosperm surrounding it will be included. When East used the floury race as the pollen parent and the corneous race as the ovule parent, the Fi genera- Tlie Kudos perm iji TfiJfcritance 149 tion was all corneous. When he made the reciprocal cross (corneous pollen and lloury ovule), the Fj genera- tion was all floury. This result certainly suggests maternal inheritance, for in both cases it is the character of the ovule parent that is transmitted. If it is assumed that this is a case of maternal inheritance, two problems are encountered: (i) to prove that this behavior is not due merely to parthenogenesis; (2) to discover the mechanism to explain maternal inheritance in this case. In the first place, East established the fact that there was no possibility of parthenogenesis. Continuing his investigation, he inbred the Fi generation in each case and examined the F2 progeny. If he were dealing with a case of maternal inheritance, what should the F2 genera- tion show? It should be exactly the same as the Fi generation, for in true maternal inheritance a race will go on forever breeding true to the maternal character, whether it is self -pollinated or cross-pollinated. If this had been a case of true maternal inheritance, East should have obtained the following results: Floury X Corneous Corneous X Floury (ovule parent) n|/ (ovule parent) i' Fi Floury Corneous Fa Floury Corneous etc. etc. Actually, however, he obtained the following results : Floury X Corneous Corneous X Floury (ovule parent) ^ (ovule parent) \1^ Fi Floury Corneous ^ Floury : I Corneous h Corneous : ^ Floury 150 Outline of Genetics The conclusion is that this is not a true case of mater- nal inheritance. East offers a very reasonable explana- tion of these results, based upon the peculiarities of double fertilization. These characters appear super- ficially to be maternal for the following reasons. The endosperm nuclei are 30:, 2x from the female and x from the male. In the characters under discussion, the presence of two factors always dominates the presence of one factor; thus corneous female crossed with floury male produces progeny that are all phenotypically corne- ous, while floury female crossed with corneous male for the same reason produces progeny which are all pheno- typically floury. The mother always determines the character of the hybrid endosperm because there are always two female nuclei to predominate over the single male nucleus. In the embryo, however, this predomi- nance does not occur, for there only a single female nucleus has fused with the single male nucleus. When this hybrid embryo matures, therefore, it is evident that it will produce gametes of two sorts, 50 per cent corneous and 50 per cent floury. Since the female is really the only decisive factor, so far as endosperm is concerned, the ratios appearing among the female gametes in the Fj generation will be the ratios that will appear also in the F2 endosperms. In other words, 50 per cent of the F, endosperms will be corneous and 50 per cent floury, no matter what may be the source of the pollen. It is obvious that the explanation of this peculiar form of apparently maternal inheritance depends entirely upon a clear conception of the phenomenon of triple fusion. Conversely, this type of inheritance indicates that the triple fusion, instead of being merely a meaningless TJie Eiidosper))! in TjiJicrUance 151 cytological peculiarity, is really significant in inherit- ance. The foregoing case, which is the only endosperm character that is inherited in this peculiar way, provides a conspicuous criticism of the presence and absence hypothesis (see also pp. 51, 112). Whichever of the two characters be assumed to be due to the presence of a gene, a demonstration is provided to the effect that two absences will dominate one presence. In spite of this outstanding exception, the terminology of the presence and absence hypothesis is retained by practi- cally all geneticists, and is, in fact, employed even in connection wath the case of corneous and floury endo- sperm. Webber (8), in 1900, experimenting on xenia in corn, uncov- ered some interesting anomalies. Pollen from a red-grained race, applied to silks of a white-grained race, should result in solid red grains if xenia is normal. Of course Webber actually obtained this result in the vast majority of cases, but occasionally there appeared two other types of grains : (a) white grains covered with numerous, irregular patches of red, commonly called ''mottled"; {b) grains of which a large and uninterrupted area of the aleurone was pure white, while the remaining area of the aleurone was just as pure red, commonly called "mosaic." For these cases he con- structed an ingenious explanation. Normally, the second male nucleus fuses with the fusion nucleus, and the result is a solid red grain. In some cases, how- ever, the second male nucleus (i) does not join with the other two; while in still other cases the second male nucleus (2) fuses with but one of the polar nuclei, leaving the other polar nucleus to act independently. Either of these irregularities, Webber felt, would serve to account for the anomalous grains, for in either case female nuclei would be left to act independently in the formation of part of the endosperm. That part of the endosperm would neces- sarily show colorless aleurone, since the female parent to the cross 152 Outline of Genetics could contribute no color. The remainder of the endosperm, produced by the male nucleus (either (i) independently or (2) in conjunction with one of thepolars), would, of course, show the red aleurone characteristic of the male parent. The arrangement of the red and white areas, sometimes accord- ing to the mottling plan and sometimes according to the mosaic, might be explained by the usual method of endosperm formation. Endosperm formation begins with free nuclear division, the result- ing nuclei being free in the cytoplasm of the embryo sac. The cell walls are not formed for some time; sometimes not until nuclear division is completed. Before a large number of free nuclei have appeared they move from the central region of the sac and usually become placed near the wall, where free nuclear divi- sion continues. When walls begin to appear, separating the nuclei, wall formation begins at the periphery of the sac and extends toward the center, in what is called centripetal growth. This program, which is common in seed plants and is known to occur in wheat, is doubtless the program in corn. If, then, the second male nucleus fails to unite with the fusion nucleus and each divides separately, when their progeny nuclei move out to the periphery of the sac the nuclei of male and female origin may well become more or less mixed. In their further division, there w^ould be groups of cells of male origin interspersed among groups of female origin. The result would be red and white areas on the mature grain, intermingled as irregular blotches, giving the mottled effect (a). On the other hand, if the daughter-nuclei of the male and female components migrated en bloc to the wall of the embryo sac, and no mixing occurred between nuclei of the tw^o types, the result would be the production of anomalous grains of the mosaic type {b). These ingenious proposals of Webber's helped to focus the attention of other investigators upon the problem imposed by the occurrence of anomalous grains of these two types. Webber's conception of the mottled grains (a) was shown to be fallacious by the experiments of Kempton (6) and of Emer- son (4). It was found that mottled grains, instead of being anomalies as Webber had believed, would appear in considerable numbers and with dependable regularity under the proper con- The Endosperm in Inheritance 153 ditions. The required conditions were peculiar and interesting. If the R factor enters the cross with the male parent only, a mottled aleurone results; if the R factor enters with the female parent only or with both parents, solid red is the result. Thus there is a pheno- typic distinction between grains having but one dose of the R factor (male parent only) and those having two (female parent only) or three doses (both parents). Furthermore, this phenome- non will not appear in all races of corn, but only in those which contain a dominant factor for mottling {S, for "spotted" aleurone, as Kempton calls it). Webber's explanations, (i) and (2), however, might still apply to the anomalous grains of the mosaic type (6). Proposi- tion (i), that the second male nucleus fails to fuse, and acts inde- pendently in endosperm formation, was proved to be impossible by some of the experiments of East (2). Factors R and C must be present simultaneously for the production of red aleurone. A cross between two colorless types, CCrr and ccRR, therefore, should produce only red grains. Even here, however, aberrant grains sometimes appear, part of the grain being red and the rest colorless. Failure of the second male nucleus to fuse with the female polar nucleus in such a case would result in a grain which was entirely colorless, a thing which never occurred. It is only by fusion of male and female nuclei that any part of the aleurone can be red. The experiments on this point were suf- ficiently extensive to demonstrate that the second male nucleus never fails to affect a fusion with at least one of the female nuclei. There yet remained, however, Webber's possibility (2), fusion of the second male nucleus with only one of the female polars, the other female polar acting independently. This last possibility was disproved by Emerson' (3) in the following inter- esting manner. A colorless, sugar>' type, CCrrsusu, was used as female parent in a cross with a colorless, starchy type, ccRR^uSu. The resulting grains were red, starchy, save for a few aberrant grains which were red in part and colorless in part, but starchy throughout. Webber's proposition (2) fails here, since fusion of the second male nucleus with only one of the polars would pro- duce grains which were red, starchy in part (from male nucleus 154 Outline of Genetics fused with one polar) and colorless, sugary in part (from independ- ent polar). These critical experiments served to disprove Webber's prop- ositions and proved that the normal program of double fertiliza- tion is invariable in corn. The occurrence of the occasional anomalous mosaic grains, however, remained to be explained. "Somatic mutation" was invoked by some as an explanation, but proved unsatisfactory for a number of reasons. Emerson (5) has finally obtained critical evidence which indicates a very satisfactory explanation of the phenomenon. The factor wx for waxy endosperm {Wx, corneous endosperm) is known to be carried on the same chromosome with the C factor. A cross was made between a colorless, waxy female parent, c-wx c-wx, and a red corneous male parent, C-Wx C-Wx (the R factor being present in both parents). The resulting triploid endosperm was of the formula c-wx c-wx C-Wx. If non-disjunction (passing of both halves of a divided chromosome to one pole) occurred in connection with the third of these chromsomes, one of the result- ing nuclei would be diploid for this chromosome set, c-wx c-wx, and the other tetraploid, c-wx c-wx C-Wx C-Wx. Endosperm produced by the former should be colorless, waxy; endosperm produced by the latter should be red, corneous. Emerson obtained aberrant grains which were of exactly this constitution, the colorless areas being at the same time waxy and the red areas corneous. This experiment, considered together with the pre- vious ones, indicates that occasional non-disjunction is the expla- nation of these aberrant grains. (The frequency of these particular aberrant grains is one in 423, and one may expect non-disjunction to take place in connec- tion with some one chromosome in the corn endosperm in about one of every fourteen grains. Direct cytological demonstration is to be hoped for. Non-disjunction is known to occur at times elsewhere in the plant and animal kingdoms. Possibly the trip- loid nature of endosperm furnishes an especially favorable condi- tion for its occurrence.) This fascinating series of experiments shows how features of the morphological and cytological program in a plant may be demonstrated in a very convincing way through the indirect evi- TJic Endosperm in TnJicrilance 155 dcncc provided by c.-ircful breeding; cxi)criincnts. where it would be rather hopeless to effect any sueh coin^ineing demonstration through direct morphological or i\-(ologi(al examination. LITERATURE CITED 1. East, E. M., and Hayhs, H. K. Inheritance in maize. Conn. Agric. Exper. Sta. Bull. no. 167. pp. 142. pis. 25. 191 1. 2. East, E. M., Xenia and the endosperm of angiospcrms. Bot. Gaz. 56:217-224. 1913. 3. Emerson, R. A., Anomalous endosperm development and the phenomenon of bud sports. Zeit. Induk. Abstamm. \'ererb. 14:241-259. 1915. 4. — , A fifth pair of factors, Aa, for aleurone color in maize, and its relation to the Cc and Rr pairs. Cornell Univ. Agric. Exp. Sta. Mem. 16. pp. 231-289. 1918. 5. , Genetic evidence of aberrant chromosome behavior in maize endosperm. Amer. Jour. Bot. 8:411-424. fig. i. 192 1, 6. Kempton, J. H., Inheritance of spotted aleurone color in hybrids of Chinese maize. Genetics 4:261-274. Jigs. 3. 1919. 7. Nemec, B., Das Problem der Befruchtungsvorgange. Berlin. 1910. 8. Webber, H. J., Xenia, or the immediate effect of pollen in maize. U.S. Dept. Agric. Bull. no. 22. pp. 44. pis. 4. i9CX3. CHAPTER XII HYBRID VIGOR The phenomenon of hybrid vigor has already been referred to. It is a matter so intimately related to genetics, particularly plant genetics, both on the theo- retical side and in connection with practical breeding, that it will be worth while to consider it in some detail here. The first record of observations on hybrid vigor is that of KoLREUTER in 1776, who states that crossing results in an increase of general vegetative luxuriance and in an increase in the facility of vegetative propagation and viability. Later Gartner discussed the same phenome- non but gave no important new ideas. Finally, hybrid vigor attracted the attention of Darwin (4), who states that crossing hastens the time of flowering and maturing and increases the size of the individual. He adds the very important fact that it is not mere crossing that gives the stimulus, but crossing forms that differ in the constitution of their sex elements ; in other words, cross- ing between different flowers on the same plant gives no advantage, nor does crossing two individuals which are gemiinally identical. He assumed (incorrectly, see p. 161) that any effective germinal dift'erence was to be accounted for by the fact that the parents had been growing under different environmental conditions.^ ^ It is probably Darwin who is responsible for bringing hybrid vigor to the attention of botanists, although the modern popular impres- sion might be that Burbank deserves the credit because of his experience in producing some remarkably fast-growing, large, and vigorous hybrids. 156 Hybrid Vigor i^y Even ]\Iendel's classic pea hybrids supplic-d furllier instances of increase in size resulting from crossing. ''Stems of I foot and 6 feet in length yielded without exception hybrids which varied in length between 6 feet and 7I feet" (see East and Jones 6). Among the modern investigators of hybrid vigor. Shull, East, and Jones have contributed much toward an explanation of the phenomenon. Skull's (12) conclusions up to the year igio may be summarized as follows. His work was entirely with corn, and the conclusions contained some very significant points. 1. "The progeny of every self-fertilized corn plant is of inferior size, vigor, and productiveness, as com- pared with the progeny of a normally cross-bred plant derived from the same source." In general this con- clusion would be admitted by everyone, but it raised one question. It was known that when two races have been inbred for many generations they frequently "run out,'' gradually losing their vigor. In such a case a cross between the two races tends to restore the original vigor. The remaining question, however, is whether the same result may be effected by a cross between two inbred races which have not run out, but remain in normal \igor. Shull answers that hvbrid vigor is exhibited when both parents are above the average condition as well as when they are below it. 2. Another question which naturally ari.^es is as follows. When these crosses are made it is of course the Fx generation that shows the hybrid vigor. If the Fi generation is inbred, what is the status of the Fj and later generations with reference to vigor? SiiULL 158 Outline of Genetics answers this question in the following general way. ''The decrease in size and vigor which accompanies self- fertilization is greatest in the first generation and becomes less and less in each succeeding generation, until a con- dition is reached in which there is (presumably) no more loss of vigor." The facts involved in this statement may be represented in fig. 25. In this figure, it can be seen clearly that the great loss of vigor comes immediately after self-fertilization again begins. After that, self- fertilization brings additional loss in vigor, but this loss Parent Races F, F, F, Fg etc. Fig. 25. — Illustrating status of hybrid vigor in Fi and later genera- tions. Vigor represented by height of rectangles. is less with each succeeding generation. It is as though a very definite limit were being approached and each generation goes down one-half of the remaining distance toward that Hmit. Just why and in what way this limit is approached will be considered later in connection with the work of East and Jones. 3. ''A cross between sibs (sister and brother) within a self-fertilized family shows little or no improvement over self-fertilization in the same family." This, it will be noticed, is simply carrying a little further the point that Darwin originally discovered. We realize Hybrid Vi^or 159 that an inbred race should be homozyjjjous; therefore all the indi\'i(luals involved would ha\e the same f^erminal constitution. A cross between any two such individuals would really not be effective in producing a hybrid, so that it would not be surprising that such a cross fails to bring hybrid vigor. 4. ''A cross between plants belonging to two self- fertilized families results in a progeny of as great vigor, size, and productiveness as are possessed by families that have never been self-fertilized." The conclusion from this is that inbreeding results in no permanent loss of vigor. A race may "run out'' if inbred continu- ously, but when crossed with another race it immediately seems to regain all the original vigor. It is as though all germ plasm contains the potentiality of developing vig- orous individuals. This potentiality, however, cannot express itself until the proper combination of conditions arises, and this proper combination seems to be connected in some way with hybridizing. 5. ''Reciprocal crosses between two distinct self- fertilized families are equal" in producing hybrid vigor. When reciprocal crosses are equal it suggests a Mendclian phenomenon. Is it possible that hybrid vigor may be explained in terms of Mendelism ? These are five "laws" of hybrid vigor presented by Shull, in 1910. It should be noted that they are not hypotheses but observed facts. The hypotheses were developed later when more of the facts were in. A practical suggestion made by Siiui.i. in connection with hybrid vigor is of interest. Ciranted that hybrid vigor is an estab- Ushcd fact, the question of its practical use in connection with crop plants should be taken into account. If a farmer after years i6o Outline of Genetics of work has finally developed a desirable new strain of corn by selection, he is not likely to favor hybridizing with some other strain in any wholesale way. He must preserve his pure strain at all costs. Shull has suggested the following solution of this practical problem, as indicated in fig. 26. Two desirable strains {A and B) are developed. One small plot (I) is planted entirely with A, and at some distance another small plot (II) is planted with A and B in alternating rows. Plot I is used only to perpetu- ate A in pure condition. In plot II all the .1 plants are detasseled. The silks of these .1 plants, therefore, are pollinated by B pollen only, and the resulting grains in the .1 rows are all bound to be hybrids. Using these grains as seed for the crop, hybrid vigor A B A B A B PLOT I PLOT 11 Fig. 26. — Shull's scheme of planting for making practical use of hybrid vigor in corn. will be obtained. At the same time both .4 and B are perpetuated in the pure condition, since the B rows in plot II are always self- pollinated, as there is no other pollen in that neighborhood. This is a very simple solution of the problem, without necessitating laborious hand pollination. The investigations and conclusions of East (5) may next be considered. Shull did his work entirely with corn, but East investigated the problem in a more whole- sale way. After assembling an extensive collection of data, he made the summarizing statement that 59 out of 85 angiosperm crosses showed a noticeable increase in Hybrid Vii^or i6i vigor. East of course did not conliniR' to investigate all of these 85 types, but concentrated ui)()n two repre- sentatives. Corn was selected as represent in<( s[)ecies normally cross-fertilized in nature, while tobacco was used to represent those species generally self-fertih'zed in nature. East's results with corn need not be discussed in detail, for they confirmed Shull's results in every point. It was found that crosses between plants of aj)pi"()xi- mately the same genotype resulted in little or no hybrid vigor, even in cases where the two parents to the cross had been grown under different environmental conditions (thus correcting Darwin's misconception, see p. 156). It was also observed that some crosses resulted in rela- tively less hybrid vigor than others. From such results East developed a very significant and useful Mendelian interpretation of hybrid vigor. His proposition is that hybrid vigor is proportional to the number of factors in which the parents differ. This situation may be \'isu- alized from the following diagram. Parents !• i AABBCCDDXAABBCCdd = AABBCCDd = \iii\c hybrid vigor AABBCCDDX AABBccdd = AABBCcDd = more hybrid vigor AABBCCDDX AAbbccdd = .1.1 BhCcDd = still more hybr id vigor AABBCCDDX aabhccdd = AaBbCcDd = mo<.i hybrid vigor It is the Fi of course that shows the vigor, but what index can be obtained from the germinal formula of the Fi generation as to the amount of hybrid \ igi)r that it will show? It is evident that this index lies in the fact that hybrid vigor is proportional to the innnber of factors 1 62 Outline of Genetics that are in the heterozygous condition in the Fi genera- tion. Thus in the first case shown in the diagram there is only a single heterozygous set {Dd), and the result is little hybrid vigor. Following down the diagram it will be noted that 2,3, and 4 of these heterozygous sets bring an increasing amount of hybrid vigor. These are the facts that lie at the basis of East's theory which he calls heterozygosis. This term should not be confused with heterosis, which is commonly used to express merely the fact of hybrid vigor. We shall now consider how this conception of hetero- zygosis serves to account for the phenomena that Shull had previously discovered in connection with hybrid vigor. 1. The fact of hybrid vigor. — Heterozygosis suggests that hybrids are vigorous on account of the heterozygous sets of factors that they contain. 2. The decrease in vigor after self-fertilization begins again.^ — The greatest loss in vigor comes between the Fj and F2 generations. Thereafter the loss becomes gradu- ally less each generation, approaching a definite limit beyond which no further loss in vigor occurs. Heterozy- gosis explains this as follows: AABBCCDDXaabhccdd = AaBhCcDd. In this case the Fi generation is 100 per cent heterozy- gous, all four factor sets being heterozygous, and there- fore it is very vigorous. In later generations, as is well known, more or less homozygous sets will be split off. Introducing homozygous sets into some individuals will reduce the aggregate heterozygous condition of the whole Hybrid Vigor 163 population to something less than too per cent; there will therefore be a corresponding loss in \ig(jr. If the genotype of the F^ population be considered (a simpler example, AABBXaabb, will suffice), some very clear conclusions may be drawn. The F2 population is heterogeneous with respect to hybrid vigor, in sharp contrast with the F,, where all the individ- uals showed the same amount of hybrid vigor. In the F^ there will be one genotype which is heterozygous with respect to all of the factor pairs involved (as was the F,), and which, therefore, shows the maximum amount of vigor. There will be other geno- types which are homozygous with respect to all the factor pairs, and show no vigor. And there will be still other genotypes which are parti}'' heterozygous and partly homozygous, and show an intermediate amount of vigor. This heterogeneity of the F, gen- eration with respect to amount of hybrid vigor is in agreement with the actual experimental results. If the average vigor of the whole F^ population be computed, in terms of relative numbers of factor sets in the heterozygous condition, this will be found to have a value of 50 per cent, in con- trast with the o per cent of the original grandparental generation and the 100 per cent of the Fj. On the same basis the F3 will be found to have 25 per cent, the F4 12.5 per cent, and so on, exactly one-half of the vigor being lost with each succeeding generation of inbreeding. This serves to account for Shull's observation that the greatest loss in vigor is between the Fi and F, generations. Thereafter the loss gradually approaches the limit when the per- fectly homozygous condition is reached for the whole population, and then there can be no more loss in vigor. 3. A cross between sister and brother elTects nothing. — This is evident, for it introduces no heterozygosity. 4. "A cross between ])lants belonging to two self- fertilized families results in a i)ri)geny of as great \igor, size, and productiveness as are possessed by families that have never been self-fertilized." Heterozygosis accounts for this by showing that a cross between two 164 Outline of Genetics pure lines may bring into the hybrid a maximum number of heterozygous sets, quite as many as are present in cross-fertilized families. 5. Reciprocal crosses are equivalent.^ — This would obviously follow from any Mendelian hypothesis such as heterozygosis. East next studied tobacco as representing those species which are generally self-fertilized in nature. It is a common impression that tobacco is a striking exception in the matter of hybrid vigor. In tobacco crosses the hybrid progeny, instead of being more vigorous, are fre- quently less vigorous than either parent. East admits that there are certain cases of this kind, but points out a number of other cases which are quite " normal" in show- ing hybrid vigor. In any event, the tobacco situation strongly suggests the idea that hybrid vigor appears less prominently in species that are generally self-fertilized in nature than in species normally cross-fertilized. It may be that the "subnormal" tobacco hybrids are products of such wide crosses that hybrid vigor can no longer operate (see p. 169). The phenomenon of hybrid vigor appears also in a great many other plants. It has of course been noted most frequently in cultivated forms, but there is also some evidence as to its occurrence among wild plants. Not only has it been observed among many angiosperms, woody as well as herbaceous, but also among gymno- sperms and pteridophytes; and there is even some slight evidence that hybrid vigor occurs in the sporophyte of the bryophytes (see Britton i). As for the exact nature of the phenomenon, quite a number of features arc involved. Primarily, hybrid Hybrid Vliior 165 \iuj()r amounts to an increase in the si/r of tells, as well as multiplication in the number of cells; in oilier words, an increase in the ]X)wer of assimilation. \'ial)ilit\' of seeds is increased, and the more rapid growth and earlier maturity of the seedlings is quite noticeable. 'I'ime of flowering and maturing is hastened, although in many cases increased longevity has been brought about. One sees a distinct increase in the size of the roots. In the stem there is no increase in the number of nodes, but the internodal development is striking. (The 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.) Usually the stem growth is greater than the leaf growth, but the increase of the latter can be definitely traced. The size of the flower is usually not affected, nor is there any change in the size of small fruits, such as tobacco. In fleshy fruits, however, such as tomato and egg plant, there is a marked increase. On the individual plant there are distinctly more flowers and fruits, and in some cases separate inflorescences are longer, as in the ears of corn. (Total yield in corn has, in some crosses, been increased 100 per cent or more.) Endurance against unfavorable environmental factors and resistance to disease have also been frequently noted as properties of hybrids. Facility of vegetative pro])agation is increased. (Moreover, there is no evidence to i)ro\e that plants lose any of their hybrid \ igor in long continued vegetative multiplication through innumerable generations.) In general, there is similarity between hybrid \ igor and the effect of a good environment. Those characters which arc the quickest to be mcKliiied by external factors 1 66 Outline of Genetics also show the greatest change on crossing. There is at least one difference between the two, however; in time of maturity, environment and hybrid vigor have some- what opposite effects. Generally speaking, favorable growing conditions tend to delay flowering and maturing, while conditions w^hich tend to stunt the plants tend, like hybrid vigor, to hasten them (East and Jones 6). There seems little doubt that hybrid vigor is also manifested in the animal kingdom. One might reasonably expect this from the fact that the principles of inheritance are fundamentally the same in plant and animal kingdoms, and hybrid vigor is a matter of inheritance. As a matter of fact, there are many cases among the records of professional animal breeders which might be cited as evidence of hybrid vigor. It seems equally evident, however, that this is not so general a phenomenon among animals as among plants; and it should be noted that many zoologists refuse to recognize in hybrid vigor anything like a general law, pointing out cases among animals in which hybridizing apparently results in loss of vigor. It is rather to be expected that such a general phenomenon as hybrid vigor must have played a part in the evolution of the plant kingdom. A few suggestions follow (from East and Jones 6) . 1. Fixation of characters favoring cross-fertilization. — "Vari- ations must have appeared that favored cross-fertilization. Those plants producing a cross-fertilized progeny would have had more vigor than their self-fertilized relatives. The crossing mechanism could then have become homozygous and fixed, while the advantage due to cross-fertilization continued." 2. Fixation of sex act itself. — "Some means of favoring union of dissimilar spores occurred as a chance variation. Through the combination of somewhat different qualities this new dual product, the zygote, was better enabled to develop and reproduce. Its survival coefficient was high. The tendency for union of spores persisted and became characteristic of the species." 3. Preservation of undesirable characters in cross-fertilized species. — "In self-fertilized species, new characters that weakened Hybrid Vit^or 167 the individual woidd have been immediately eliminated. Oidy strains that stood by themselves, that survived on their own merits, would have been retained. On the other hand. weak, j^enotyfxjs in cross-fertilized species were retained through the vigor that they exhibited when crossed with other genotypes. The result is, therefore, that self-fertilized strains that have survived competi- tion are inherently stronger than cross-fertilized strains. On this account weak genotypes may often be isolated from a cross- fertilized species that as a whole is strong and hardy." 4. Rise of the sporophyte generation. — The commonly ac- cepted interpretation of hybrid vigor is based upon a Mendelian mechanism that would be effective only in the diploid generation. In the evolution of the plant kingdom, the haploid gametophyte generation has been superseded in dominance by the diploid sporo- phyte generation. Hybrid vigor may help to account for this. Some recent investigations have extended the scope of hybrid vigor in an interesting and significant way. The work was done originally by Collins and Kemptox (3), and later confirmed and extended by Jones (8). In brief, it is as follows. If corn sporophytes exhibit hybrid vigor, will the endosperm also show the same phenomenon ? Endo- sperms, as has been stated, are genetically equivalent to sporophytes in several ways. If crossing increases vigor and size of sporophytes, therefore, it might be expected to increase the size of the endosperms also. Furthermore, the endosperms have considerable advantage over sporophytes as material for such investigation. We say that hybrid sporophytes are more vigorous than pure bred sporophytes, but just how much more vigorous cannot be stated with exactness. In order to demonstrate this clearly, it would be necessary to have the hybrid and the pure bred stock growing side by side in e.xactly the same conditions, but the conditions cannot be controlled with exactness. The environmental factors affecting the size and vigor of a corn plant are numerous, complex, and to a large extent 1 68 Outline of Genetics uncontrollable. Thus two different plants, growing side by side, might be in a distinctly different environment without the fact being recognized. It cannot, therefore, be said with much cer- tainty just how much hybrid vigor a given plant shows when there are so many unknown factors that might affect size and vigor. On the other hand, if it is claimed that the endosperm of one grain shows a given amount of hybrid vigor as compared with the grain that grows next to it upon the same ear, the statement would be more exact, for the two endosperms have developed under con- ditions which are unquestionably much more constant than the conditions surrounding the different sporophytes in a corn field. Jones selected a plant with white endosperm and pollinated it with a mixture of its own pollen and pollen from a yellow endosperm race. In the resulting ear, therefore, he had a mixture of yellow endosperm grains and white endosperm grains. The former grains of course wxre hybrid, since the yellow factor was introduced by the foreign pollen, while the white endosperm grains must have resulted from own pollen and were homozygous. In this way, Jones obtained side by side in the same ear endosperms obviously hybrid and endosperms obviously homozygous. When he weighed these two types he found that the hybrids exceeded the homozygotes in weight by from 5 to 35 per cent. He made the reciprocal cross, using the same mixture of yellow^ and white pollen on silks of the yellow race. Of course all the resulting endosperms were yellow, but the hybrids, which had the yellow factor only from the female side, were distinctly lighter yellow than the homozygotes, which had the yellow factor from both male and female sides. Weighing these two types, Jones obtained the same results as before, the hybrids exceeding the others in weight by an average of 20 per Hybrid Vigor 169 cent. This is really the clearest demonslralion of iiybrid vigor that has ever been pro\'ide(l, for the conditions of the experiment were ideally constant. It is interesting to note in this connection that there is no selective action favoring foreign pollen when these pollen mix- tures are applied. In fact, the results indicate that own pollen is successful in bringing about fertilization in a slightly greater number of cases than is foreign pollen. It has been stated that the amount of hybrid vigor varies directly with the width of the cross. Of course this statement applies only within certain limits. The situation is somewhat clarified by considering the following series of cases which is arranged with respect to width of cross. 1 . Parents so diverse that cross cannot be made. 2. Cross possible but seed obtained fails to germinate. E.\- ample, certain Nicotiana crosses. 3. Hybrid seed germinates, but resulting hybrid plants are so weak that they fail to reach maturity. E.xample, other Nicotiana crosses. 4. Hybrid plants mature and are extremely vigorous, but are sterile except possibly in back crosses. Example, cabbage-radish hybrid, an enormous but completely sterile plant. Example from animal kingdom, the mule. (On this matter see also chapter on "Sterility.") 5. Hybrid plants more vigorous than parents, and completely fertile. Example, corn crosses and many others. 6. Parents to cross so closely related that no protluction of hybrid vigor is noticeable. (An interesting phenomenon appears in certain wheat crosses, where it is found that the Fi endosperms are well developed in the fertile crosses, but shriveled in those crosses which are to produce sterile or partially sterile F, plants. Even in these latter cases, however, hybrid vigor appears in the vegetative parts of the F, plants. Sax ii.) Obviously, it is only within the limits of classes 4 and 5 that it can be said that hybrid vigor varies directly with the width of the cross. It is impossible to say where the species boundary tits 170 Outline of Genetics into the foregoing scheme, since species boundaries are more often matters of personal opinion than indices of crossabihty. The theory of heterozygosis claims that hybrid vigor appears in proportion to the number of factors in which the parents of the cross differ. This claim should be considered briefly. Is heterozygosis really an explana- tion of the phenomenon of hybrid vigor? It seems obvious that it is not. It was known that hybrids were vigorous because they were hybrids. Heterozygosis states that hybrids are vigorous to the degree that their parents differed in hereditary factors; in other words, this is merely a statement that hybrids are vigorous because they are hybrids, with the addition that the more hybrid a hybrid is the more vigorous it is. It follows, therefore, that heterozygosis is not an explana- tion of hybrid vigor, but merely a restatement of the phenomenon in Mendehan terms, with the additional idea that there may be various degrees of hybrid vigor. It is not the intention to discredit heterozygosis as a valuable conception, but to point out that it is not a real explanation, merely a more intelligent statement of facts. Furthermore, heterozygosis is rather unsatisfactory in another way. It locks the door on any hope of origi- nating pure strains having as much vigor as first genera- tion hybrids. For these reasons it would seem desirable to seek an explanation of hybrid vigor along other lines. Such an explanation may be developed from the following con- siderations. In nature a "struggle for existence" occurs among species and individuals. There must occur also a struggle for existence among unit characters. If a unit char- Ilyhy'id l^ii^or 171 acter is umlcsinil^lc il is cliniiiiatt'd, Un- tin- iiiniinanl characlers, for undesirable recessives might well sur\ i\ l' hv escapinj^ natural selection while in heterozygous combination with their dominant allelomorphs. It follows that the domi- nant unit characters that have survived and a])pear in the plants of today are for the most part desirable ones. The question may be raised as to what constitutes a ''desirable'' character. It may be any one of a number of things, but is there not some feature which is common to all desirable characters ? The common feature of all desirable characters would seem to arise from their rela- tion to the vigor of the organism. Each desirable char- acter must add somewhat to the vigor of the plant that contains it, and associated with vigor are such things as size and productiveness. Is it not reasonable that those plants will be most vigorous which have in combination the greatest number of desirable characters ? 1'he plants which have the greatest combination of such characters are the hybrids. A diagram similar to that which was used to explain heterozygosis may be considered: Parents Fi A A BBCCDD X A A BBCCdd = A A BBCCDd = 1 it Uc hybric 1 vigor AABBCCDDXAAbhccdd = A A BbCcDd = si'iW more hybrid vigor In that explanation it was stated that the lirst case showed little hybrid vigor because it had only one hetero- zygous set {Dd), while the other case showed more hybrid vigor because it had three such heterozygous sets. Hybrid vigor, therefore, appeared in proportion to the 172 Outline of Genetics number of heterozygous sets in the hybrid. This dia- gram served the purpose of explaining heterozygosis, but it win now be discarded because it does not represent the most important result when two races are crossed. The important result is represented in the following diagram : Parents F« mbbuDDEEFF ] AaBbCcDdEeFf =...or^ hybrid vigor tifcCDDei } ^^imoDdeefj = less hybrid vigor The thought is that in each of these two cases the hybrid is more vigorous than either parent, not because it con- tains more heterozygous sets, but because it contains more dominant factors, which means more ''desirable" characters. For example, in the first case each parent contains three factors, the small letters representing merely the absence of factors. The Fi generation, there- fore, contains six factors, and for this reason is more vigorous than either parent. It is stated in the diagram that in the first case there is ''more hybrid vigor" and in the second case "less hybrid vigor," simply because hybrid vigor is a relative term. It represents merely how much more vigorous the hybrid is than either parent. In the first case the parents have three factors and the hybrid six, the increase being three, which measures the amount of hybrid vigor. In the second case each parent has two and the hybrid four; the increase, therefore, is only two, and for this reason there is less hybrid vigor in the second case than in the first. Assuming that the majority of dominant factors are desirable, and that desirable factors make for general Hybrid Vigor 173 vigor, it would follow tlial the most vigorous plant will be the one containing the greatest number of dominant factors. It has been shown that the j)lants contain- ing the greatest number of dominant factors are the hybrids; it is for this reason that hybrids are relati\ely vigorous. The following question may be raised. If it is granted that most desirable factors tend somewhat to increase the general vigor, do they all do this to the same degree ? The natural answer is in the negative, but this has no bearing upon the validity of the explanation. On the other hand, if heterozygosis be accepted for an explanation the question presents a dilhculty. Heterozygosis would suggest that Aa induces vigor, not because of any particular factor that it represents, but because it is a heterozygous set. It seems more reasonable and natural to suppose that certain factors induce more vigor than others. It is evident that the suggestion made above is that of a real explanation of hybrid vigor and not merely a restatement. Keeble and Pellew (id) suggested it in 1 9 10, and since that time it has been somewhat discussed in the literature, being referred to as ''the hypothesis of dominance (accounting for hybrid vigor)." At lirst statement the theory seems sound, but actually it does not fit the facts. The two outstanding objections to this theory of dominance are brought out in the })ublica- tions of Shull, Emerson, and East. I. If hybrid vigor were due to dominance, it would be possible in generations subsequent to the F, lo recom- bine in one race all of the dominant factors in the homo- zygous condition. Thus there could be isolated a race that was ''100 ])er cent vigorous," and since il would be homozygous, its vigor would not be lost by inbreeding. Actually, though, no one has (as yet) been able to "lL\" 174 Outline of Genetics hybrid vigor in this way; ''all maize varieties lose vigor when inbred." 2. Experience assures us that the distribution of individuals in the F2 generation with reference to hybrid vigor is represented graphically by a symmetrical curve, similar to the normal probabilities curve; the class con- taining the greatest number of individuals is that which shows the medium amount of hybrid vigor, while on either side of this class the fall in the curve is regular, reaching its lowest points in the two small extreme classes which show respectively greatest hybrid vigor and least hybrid vigor. According to the dominance hypothesis, the largest class of the F, individuals is that showing greatest hybrid vigor (if only a few effective factors are assumed, as was the case in the work of Keeble and Pellew), while the smallest class would be that showing least hybrid vigor. The curve representing such a situa- tion would be asymmetrical and strikingly different from that which actually occurs. For these two reasons the dominance hypothesis, as proposed by Keeble and Pellew, has been discarded. Although it is theoretically attractive, its failure to satisfy these two important details of the hybrid vigor situation has condemned it. Recently Jones (7) has ingeniously modified the dominance hypothesis so as to avoid these difficulties. The argument is essentially the same, with one very significant modification. Jones visualizes the situation as represented in fig. 27. In this case it is a question of linkage of dominants and recessives. The vigor of one parent is due to the two dominant factors A and D, while that of the other parent is due to the two dominant Hybrid Vigor 175 factors C and B. The hybrid is more \i^'()r{)us than either because it combines all four dominant factors. The attractiveness of this scheme Hes in the fact that it escapes the objections that were made to the older domi- nance hypothesis. Fig. 27. — Diagram to aid in visualising Jones's explanation of hybrid vigor by dominance of linked factors. I. The fact that 100 per cent hybrid vigor cannot be fixed is quite in accordance with Jones's scheme, for it is obviously impossible to isolate a race homozygous for all four factors, A, B,C, and D. As a matter of fact, it would be theoretically possible under this scheme to isolate just such a homozygous race. If crossing over took place during gamete formation by the F,. .1 and C might come to lie on the same chromosome. When a gamete containing such a chromosome mated with another gamete of the same sort, a race would thereby be established which was homozygous with respect to -1 and C. If a similar performance took place (cither simultaneously or in some subsequent generation) in the other chromosome pair, the race would also achieve homozygosity with respect to B and D, and would thereafter breed true for all four factors. It must be evident, however, that there would be but a remote chance of realizing this theoretical possibility. in;ismuch as there actually must be many more than two chromosome pairs involved, 176 Outline of Genetics and more than two effective factors on each chromosome. At the same time it is rather encouraging to believe that such a possibility exists, so that eventually we may be able to obtain a race that is homozygous for all or practically all of the desirable factors. 2. A simple mathematical demonstration will show that the distribution of F, individuals is quite what it should be, represented by a symmetrical curve similar to the curve of probabilities. It can be demonstrated rather rapidly that Jones's scheme will satisfy the requirements on distribution of F2 individuals with respect to hybrid vigor, and on the progressive loss of vigor in the later inbred generations. It was seen that the heterozy- gosis theory could account for these facts by the use of the simple example, AAbbXaaBB. As a matter of fact, fig. 27 becomes the exact mathematical equivalent of this example if we substitute the ^c chromosome of the dominance scheme for A of the heterozy- gosis scheme; the bD chromosome for b; the aC chromosome for a; and the Bd chromosome for b. In this way it will be discovered that the two schemes run exactly parallel in accounting for the facts in every generation. In fact, this new theory, ''the dominance of linked factors," seems altogether sound and natural. We should rather expect that each chromosome would bear several dominant factors conducive to vigor and several recessives as well. Recently Collins (2) has presented some interesting consid- erations bearing on this scheme of Jones's. Collins maintains that in explaining hybrid vigor one should place the emphasis on the suppression of deleterious recessive characters rather than on the accumulation of dominant growth factors. This is merely a change in the point of view. Collins further maintains, however, that the dominance scheme can really account for the facts with- out the assumption of linkage, provided a sufiicient number of effective factors be assumed. There is some rather good evidence to support these contentions of Collins. For the present, how- Hybrid Vigor 177 ever, Jones's explanation seems distinctly more serviceable than any other that has been offered. Skull's scheme to take practical advantage of hybrid vigor in such a plant as corn has one distinct drawback. As was brought out in chapter ii, the size of a corn plant is limited by the size of the seed which produces it, and this, in turn, is limited by the size of the mother-plant upon which the seed developed. Since Shull's races A and B were both inbred races, they must have been rather small. Consequently, whichever wa>' the cross was made, the seeds containing the Fi embr>'os would be limited in size by the small size of the female parent, so that the P', plants would get a poor start and would never be able to attain the size that would have been possible had they come from large seeds. Jones (9) suggests a way of overcoming this dilTiculty. Starting with four pure races, cross .1 with B and C with I). This will result in two vigorous Fi types, either one capable of develop- ing large seeds. Cross these two, and u double hybrid results which combines, in good part, the advantages of all four of the original races, and is not limited in size by starting from a small seed. Of course there is a certain drawback here also, since the generation used for the crop is an F2 with respect to the two original crosses, and has somewhat less than the maximum vigoi on that account. The relative advantages and disadvantages of such breeding plans can be evaluated only by experiment. Junes claims to have gotten better results from his double cross method than could be obtained from Shull's plan. From the discussion that has been preseiUeci in lliis chapter, one may safely conclude that the ])henomena which arise in connection with inhreedin*^ and outbreed- ing can be explained satisfactorily in terms of the Men- dehan mechanism of inheritance. It should l)e recog- nized that inbreeding is not injurious per se (^throuL^di mere fact of consangtiinity), but because it serves to isohite undesirable recessive tyi)es from a liybrid nu'xture. The proof on this point is that inl)rec'diiig in homozy- gous stock results in no deterioration, joxios (9) has 178 Outline of Genetics carried on through twelve generations the inbred corn cultures that were started by East. In the course of this experiment, a great number of undesirable recessive types have been thrown off. On the other hand, certain of the lines that have been isolated by this inbreeding are quite normal and healthy, though small in stature and yield. A point of homozygosity has been reached where further inbreeding brings no further loss in vigor. On the other hand, hybrid \'igor does not arise from the act of crossing per se, but merely through a combina- tion in the hybrid of the maximum number of desirable factors. On this point the proof lies in the fact that crossing brings hybrid vigor only when the parents to the cross differ in their germinal constitution. There is plenty of evidence on this point. In Jones's inbreeding experi- ments, a point of homozygosity has been reached where crosses between different individuals of the same line brings absolutely no hybrid vigor. In conclusion, attention should be called to the danger of confusing phenomena of hybrid vigor with those of cumulative factors. Both mechanisms may operate on some generalized quantitative character such as size, but the hereditary behavior is distinctly different. Cumula- tive factors bring an Fi which is no more variable than either parent-type and intermediate in size, and later generations which are highly variable. The average size of the whole population, however, is the same for every generation, including the parental and the Fi generations. The hybrid vigor mechanism also brings an Fi no more variable than either parent-type (as would any Mendelian mechanism for that matter), and later generations which Hybrid Vigor 179 are widely variable. In this case, however, the average size of the whole population is distinctly different in the different generations. In such a matter as size the hybrid vigor manifestations would he superimposed upon the cumulative factor manifestations. It might appear unsatisfactory and arbitrary to assume domi- nance of factors as essential to explaining hybrid vigor, and lack of dominance in the case of cumulative factors. It is quite likely, however, that the fundamental and ''natural" distinction between the two mechanisms lies in this very point. Where a number of factors interact in affecting some quantitative character and those factors show lack of dominance, a cumulative factor mechanism is thereby set up. A similar interaction where the factors are dominant brings into play the hybrid vigor mechanism. There is a difference between the two mechanisms simply because some factors show dominance and others do not. This idea may be reinforced by the following theoretical sug- gestion. Where the environment (using the term in its widest sense) imposes no limitation upon the degree to which a character may be expressed, it follows that two doses of a factor must have twice the effect of one; dominance is lacking. Where the environ- ment limits the expression of a character, and one dose of a factor results in a development of the character to this limit, two doses can affect nothing more; dominance is present. Furthermore, these environmental limitations may shift as the environment changes. Such an environmental shift could affect in no way the degree of development of those characters in connection with which there is no dominance, but would be expected to affect the degree of development of those characters where dominance occurred. According to this idea, we should expect non-domi- nance or cumulative factor characters to be of such a sort that the environment never affects the degree of their development; while dominance or hybrid vigor characters would be those which envi- ronmental changes could also modify. lH>r the most part this actually agrees with the facts (see p. 165). Further investigation will doubtless provide a more definite answer on this matter. i8o Outline of Genetics LITERATURE CITED 1. Brixton, E. G., A hybrid moss. Plant World 1:138. 1898. 2. Collins, G. N., Dominance and the vigor of first generation hybrids. Amer. Nat. 55:116-133. jig. i. 1921. 3. Collins, G. N., and Kempton, J. H., Effects of cross-pollina- tion on the size of seed in maize. U.S. Dept. Agric. Circular 124. 1913. 4. Darwin, C, The effects of cross- and self-fertilization in the vegetable kingdom. London, 1876. 5. East, E. M., and Hayes, H. K., Heterozygosis in evolution and in plant breeding. U.S. Dept. Agric, Bur. PI. Ind. Bull. 243. pp. 68. pis. 8. 1912. 6. East, E. M., and Jones, D. F., Inbreeding and outbreeding. Philadelphia. 1919. 7. Jones, D. F., Dominance of linked factors as a means of accounting for heterosis. Genetics 2:466-479. 191 7. 8. , Bearing of heterosis upon double fertilization. Bot. Gaz. 65:324-333. figs. 3. 1918. 9. ■ — , The effects of inbreeding and cross-breeding upon development. Conn. Agric. Exp. Sta. Bull, 207. 1918. 10. Keeble, F., and Pellew, C, The mode of inheritance of stature and of time of flowering in peas {Pisum sativum). Jour. Genetics i : 47-56. 19 10. 11. Sax, Karl, Sterility in wheat hybrids. I. Sterility relation- ships and endosperm development. Genetics 6:399-416. 1921. 12. Shull, G. H., Hybridization methods in corn breeding. Amer. Breeders Mag. 1:98-107. 1910. CHAPTER XIII SEX DETERMINATION Sex determination is very properly a part of the sub- ject of genetics. It should be realized, however, that a vast amount of investigation has been carried on in this field, and it will be possible here to take up only a limited number of representative cases. Since this subject has been investigated a great deal more thoroughly and for a great many more years in animals than in plants, it will be appropriate first to consider some of the findings of the zoologists. Until very recently, at least, there appeared two outstanding and seemingly quite contradictory views as to the basis of sex determination. 1. Some believed that sex is predetermined by the chromosome equipment that enters into the zygote. 2. Others believed that sex may be determined othcr- w^ise than by the chromosomes, the decisive factors being certain physiological conditions surrounding the unfer- tilized egg or the developing embryo. These two general views will ])e referred to as the chro- mosome theories and the physiological theories. Repre- sentative examples of each will be considered brietly. Chromosome theories. — A classic example of the simplest kind is to be found in the nematode worms. Fig. 28 will indicate how sex is determined in this case. Both male and female have ten chromosomes (com- monly called aulosomcs) to determine most of their 181 l82 Outline of Genetics somatic characters; but in addition there are extra chromosomes that determine sex, known as sex chromo- somes, or heterochromosomes . In this case the male con- tains only one sex chromosome, while the female con- tains two. At the reduction division, when the gametes MALE OOOO o • o oooo MATURE INDIVIDUALS i FEMALE DOOO ©•• O OOOO , — - — ■ > OOOO o • o oo oo oo oo o» ^o oooo Zygotes (This type of zygote (This type of zygote produces a male) produces a female) Fig. 28. — Illustrating behavior of sex chromosomes are formed, the chromosome equipment is reduced one- half. It is obvious that in the female each egg receives one sex chromosome, and therefore all eggs are alike in this feature. In the male, however, with one sex chro- mosome, at the reduction division the solitary sex chro- mosome goes to one pole, leaving the other pole without Sex Dclcrui'nuUiou 183 such a chromosome. As a result lluTc arc two kinds of sperms, one haU' conlaininj^^ a sex chromosome, the other half containing none. At fertilization, if an egg mates with a sperm having a sex chromosome the zygote con- tains two, and this will develop into a female, for females are characterized by two sex chromosomes. With a sperm of the other type, the zygote receives only one sex chromosome and must produce a male individual. As a result, males and females are produced in equal numbers, sex being determined by the type of sperm that enters into the sex fusion. Certain conclusions may be drawn from this mecha- nism of sex determination, which will serve to provide a sharp contrast with the corresponding conclusions that may be drawn from the physiological theories. a) The sex ratio will regularly be 50 per cent males: 50 per cent females. It would be rather hopeless to modify this ratio by artificial means. h) Sex is a qualitative matter, only two conditions being possible, strictly male and strictly female. Numerous instances of the sex chromosome mechanism have been discovered in the animal kingdom. Details ditler in the different cases, but the essential mechanism remains the same. In addition to the type of case described above, where the male- has only one member of the sex chromosome pair, there are in general three other possibilities. The male may have one largo chromosome (similar to the pair in the female) paired with a small one; the male may have two sex chromosomes of ai^proximately the same size but different in shape; or the male may have two sex chromosomes which are morphologically itientical, but physi- ologically different in their intluence on sex. In all of these cases the fundamental mechanism remains the same, the male being heterozygous for sex, so that two types of six^rms are protluced in equal numbers, and the sex of the olYspring depends upon which 184 Outline of Genetics type of sperm has effected fertilization. In the cases where the male has an unequal pair of sex chromosomes, that member of the pair which is similar to the equal pair of the female is known as the X chromosome, while the other chromosome of the male is the Y chromosome {XX is female, XY is male). Furthermore, although the male is usually the heterozygote for sex, there are some cases in which the female is the heterozy- gote. In such cases the sperms are all alike; two types of eggs are produced in equal numbers, and the sex of the offspring depends upon which type of egg has effected fertilization. This is the situation in the birds and in the Lepidoptera group of insects, while in practically all of the other known cases it is the male that is the heterozygote for sex. When genes are located on the X chromosome their method of inheritance is characteristic, being known as sex-linked inheritance. (This term should not be confused with sex-limited inheritance, which applies to cases where the genes are carried on the auto- somes in the usual manner, but can express themselves only in one sex, e.g., heavy beardedness in man.) Numerous examples of sex-linked inheritance are to be found in the fruit fly. The female fruit fly has a pair of X chromosomes, while the male has an X mated with a F, the two being distinguishable by their shape. Numerous genes are located on the X chromosome, but none have been identified on the F, which seems quite negli- gible in inheritance.^ "Vermilion" {v), a recessive to normal red eye (F), is a gene of this sort. When a vermilion female {iX) {vX) is mated with a red male {VX) F, all of the female progeny are bound to be red {vX) {VX), w^hile all of the male progeny will be vermilion {vX) F, as is explained by fig. 29. This has often been called "zig-zag" inheritance, since the character of the mother passes to the sons, while the character of the father passes to the daughters. All of the other possible matings work out just as one would expect from the mechanism that is involved. » Some very recent papers (Castle lo) have suggested that genes may actually be located on the F chromosomes in some animals. A peculiar "one-sided" type of inheritance results, since it is possible for such genes to be present only in the male. The author is not as yet certain that these findings will be "accepted" by other geneticists. Sex Dclcrminaliou i8 = In this connection it will be worth while to consider brielly some of the work of Bridges (5), which provides the "final" demonstration that the chromosomes are the bearers of hereditary characters, and that the sex chromosomes are really the effective Fig. 29. — Shaded X chromosomes carries gene (V) for red eye; unshaded X chromosome lacks this gene (i.e., condition v)] V chromo- some of male distinguished by shape. Individuals carrying l' are red eyed; those lacking it are vermilion eyed. This diagram shows how vermilion female (upper left) mated with red male (upper right) results in red daughters (lower left) and vermilion sons (lower right). units in determining sex. Occasional non-disjunction of the A' chromosome during gametogcnesis in a vermilion female results in the production of two abnormal types of eggs. (t'A') (rA') and O (fig. 30). The matings of these two abnormal types of eggs with 1 86 Outline of Genetics the two normal types of sperms from a red male will result in four possible types of zygotes: A {vX) {vX) egg with a {VX) sperm gives a {vX) iyX) {VX) zygote, which might be expected to produce a red female, but actually dies (according to Bridges' earlier statements). A {vX) {vX) egg with a Y sperm gives a {vX) {vX) Y zygote, which produces a vermilion female. 6 Fig. 30. — Showing the four types of zygotes which result from mat- ings between non-disjunctional eggs of vermilion female and normal sperms of red male. Note particularly the second and third z\-gotes, which produce the "exceptional" individuals, vermilion female and red male. An O egg with a (VX) sperm gives a (VX) zygote, which produces a red male. An O egg with a Y sperm gives a Y zygote, which dies. Inasmuch as vermilion female crossed with red male normally gives only red daughters and only vermilion sons, the vermilion daughters and red sons which result from the non-disjunctional eggs appear as startling exceptions to the normal rule. The Sex Dclcrmi)ialion 187 occurrence of occasional individuals of these exceptional tyix's led Bridges to suspect that non-tlisjunction had taken j)lace. Later he effected a striking confirmation by cytological demonstration that the exceptional vermilion females possessed a Y chromosome, and that the exceptional red males had no Y chromosome. This work provides the linal convincing demonstration that the chromo- somes are the bearers of hereditary characters, since abnormalities in the distribution of a certain chromosome set arc accomjjanied by corresponding abnormalities in the distribution of those genes which were assumed to be located on that chromosome set. In the chapter on bud variation the phenomenon of ''chromo- some elimination" was discussed (p. 121). Morgan* and Bridges (17) have discovered this sort of thing in connection with the sex chromosomes in the fruit fly. An individual which starts its development as a normal female, A'A', has one of the A' chromo- somes eliminated from one of the daughter-cells at an early embryonic division. Tissues arising from this daughter-cell have only one A" chromosome and show the characteristics of the male sex, while the rest of the tissues are female. Individuals of this part female — part male type are known as gynamiromorphs. Physiological theories.— In 1906 Hertweg (14) performed some sex determination experiments with frogs. The eggs are laid free in the water before fer- tilization, so that they furnish unusually good material for such experiments. Normally the eggs are fertilized very soon after they are laid, with the result that the progeny consists of api^roximately 50 j)er cent males and 50 per cent females. Hertweg took some of these eggs and allowed them to overripen before fertilization took place, that is, he put aside some eggs as soon as they were laid and allowed them to remain unfertilized iov an unusually long period. While these eggs were standing in the water he found that they absorbed an unusual amount of water, and the obvious conclusion was that overripe eggs show high water content. He then allowed 1 88 Outline of Genetics these overripe eggs to be fertilized, and the resulting progeny were loo per cent males. His conclusion was that sex was not determined by the chromosome equip- ment, but by the physiological conditions of the egg, high water content resulting in males. This theory was confirmed in a striking way in 191 2 by Miss King (15), who performed the converse of HERT^VEG's experiments, using toads' eggs. Taking some newly laid eggs, she withdrew^ water from them; then allowed them to be fertilized, and the resulting progeny were 90 per cent females. The obvious conclu- sion is that eggs with low water content produce females. Finally, there is the remarkable work of Riddle (18) with pigeons. Hertweg and Miss King had found that sex is determined by the physiological factor of water content. Riddle has investigated the matter a little more fully, and from his analysis of the physiological conditions of male and female he gives the following contrasts : Male Female High percentage of water Low percentage of water Low percentage of fat High percentage of fat Low percentage of phosphorus High percentage of phosphorus High rate of metabohsm Low rate of metaboUsm It appears from this that high water, low fat, and low phosphorus are male attributes or conditions, while the female attributes are the reverse. The main feature of difference, however, to which the other contrasting con- ditions are subordinate, is that the male shows high metabolism and the female low metabolism. The idea is that any physiological conditions that affect water content, fat content, or phosphorus content, or through Sex Dctcrniiiititio)!- 189 these (or otherwise) the metaboh'c rate in the egg, will affect the sex of the resulting ])rogeny. Following these ideas, Riddle was able to control the sex ratio by various means. Furthermore, he makes the somewhat startling statement that sex is a quanti- tative phenomenon; that is, the difference between male and female is a difference in degree only. A diagram (fig. 31) will illustrate the situation. It rej)re.sents a graduated scale based on the physiological condition of High H,0 U. H,0 Males with Females witK Low Fat Very Male Some Female Some Male Veiy Female Males Characteristics Oiaractemtics Females H.(h F„ LowP / \/ \ H>(hP 'v ^'^ " High miI5 '^"'ALES L'ing the cultural conditions. He voices the belief of many other botanists when he draws the following conclusions: ''Sexuality is a state or condition not Mendelian in nature, but related to functional activity of the plant and profoundly influenced by cn\'ironment. Malcness and femaleness in hemp are probably controlled by the metabolic level of the cells, and sex reversal takes place when the metabolic level is decidedly changed or dis- turbed. Any tissue in its growth may be in a neutral state of varying degrees of intensity, and during its continued growth can pass from one state to the other without any reference to chromosome segregation or combination which are the ordinary causes of Mendel ian phenomena." The situation might be clarified somewhat b\ the following generaUzation. Not only are there relatively fewer plants than animals in the unisexual condition, but even in those plants that arc unisexual, this condition is not so completely ''established" as in animals. 'J'iie sex chromosome mechanism seems to operate only in organisms where the purely unisexual condition prevails and has prevailed for some time back in their i)hylo- genetic history. Many of the dioecious angiosj)enns, however, seem rather recently to have been deri\-e(l from ancestors which have the two se.xes represented in the same flower (or at least on the same plant). In these the dioecious condition seems not to have been firmlv estab- 196 Outline of Genetics lished ; a regular sex chromosome mechanism has not as yet been perfected. It is small wonder that bewildering sex conditions appear in these "imperfectly dioecious" angiosperms. Schaffner and others have pointed out how extrinsic factors may operate to determine sex in such forms. It is evident, though, that intrinsic hereditary factors may also play their part in such cases. This may be illustrated by some of the experiments of Strasburger (22) on Mercurialis (later confirmed by Yampolsky 23). Stras- burger had the idea that the pollen mother-cell develops pollen grains with stronger and weaker male tendencies, while the mega- spore mother-cell develops eggs with stronger and weaker female tendencies. It is therefore the algebraic sum of the two as they meet in fertilization that determines the sex of the progeny. If a pollen grain with strong male tendencies mates with an egg with weak female tendencies the resulting individual will be male, and simi- larly for the other combination. These assumptions are supported by the behavior of Mercurialis. This form has for the most part pure male and pure female individuals, but at times it throws intersexes of various grades. Certain plants are prevailingly female, but bear a few "weak" male flowers. In a plant of this sort, it would seem that the female tendencies are stronger than the male. When such a plant is inbred, using pollen from the weak male flowers on the stigmas of the strong female flowers, the resulting progeny is all female, which is in accordance with Stras- burger's theory. Other plants are prevailingly male, but bear a few weak female flowers, and inbreeding these results in all males. Finally, there are some plants which are evenly monoe- cious, half their flowers being strong males and the other half strong females. Inbreeding such plants yields a progeny which is 50 per cent male and 50 per cent female. It is obvious that from such results Strasburger would be convinced of his theory of male and female tendencies. Of considerable interest to botanists is some work that has been done on the sexual condition of Mucor and related genera of fungi. Blakeslee (2, 3) found three different sexual types of mycelia, two of which he called "plus" and "minus" strains. Sex Dclermi nation igy Although they looked alike in every particular, he concluded that they were sexually different for the following reason. Neither strain by itself is capable of producing zygotes, but when phis and minus strains are brought together sexual branches from the one meet sexual branches from the other and procs can mingle freely in the mycelium. Proceeding further with this material. Bi.akeslee (4) isolated numerous plus and minus strains, and found that they ditTereti in their sexual intensity, as computed in terms of the number of zygotes formed under standard conditions. Evidently some 198 Outline of Genetics strains are more strongly plus (female, as was later determined by indirect means) and others less strongly plus, and the same was true of the minus (male) strains. This strongly suggests a quanti- tative interpretation of sex. Recently Burgeff (9) has discovered some startling facts in connection with sex in some of the other genera of Mucorineae. Absidia shows the customary plus and minus strains, as does also Parasitella. This latter genus is a parasite upon other genera of the same family, and in connection with this parasitic habit there appears a remarkable situation. The plus strain of Parasitella will parasitize the minus strain of Absidia but not the plus strain of Absidia; while the minus strain of Parasitella will parasitize the plus strain of Absidia but not the minus strain of Absidia. The author concludes that the hypothetical sexual substance which distinguishes the plus and minus mycelia of Absidia is identical with the substance that induces parasitism, and that the parasitic relationship here has arisen as the result of an unsuccessful attempt at hybridization between the two genera. The discussion to date leaves the interpretation of sex determination in a distinctly unsettled condition. We find that in a great many animals and a very few plants a very definite sex chromosome mechanism oper- ates to determine sex; and that sex is a qualitative proposition, only the two conditions of strictly male and strictly female being possible. On the other hand, it is suggested by the sexual behavior of some animals and quite a number of plants that the general physiological condition is important in determining sex; and that sex is a quantitative matter, intersexes or sex intergrades being possibilities that are frequently realized. There are three possible conclusions with reference to these contradictory theories: (i) an acceptance of one and rejection of the other; (2) the claim that both amount to the same thing, that they express the same Sex Dctcrwiuation iqq fundamental facts in dilYcrcnt tcnns or by the use of different indices; (3) the chiim that both are true hut cover different territories, that one of tJiem cxphiins certain types of cases and the other explains other t>'])cs of cases. Until very recently the tJiird alternative seemed the most acceptable, inasmuch as the two t\7)es of sex deter- mining mechanism had never been clearly identiticd in the same organism. The recent work of Hkiugk.s (6, 7), however, sways opinion to the second of the foregoing alternatives, for it harmonizes the two contradictory views on sex determination to a degree that would hardly have seemed possible. An unexpected distribution in inheritance of known factors, which are located on the second and third chro- mosomes of the fruit fly, was explainable on the assump- tion that the female parent was triploid with respect to these chromosomes. Cytological examination proved that this was actually the case. The same grouj) of flies also exhibited some remarkable irregularities in their sex condition. A considerable group of intersexes occurred, as evidenced by the secondary sex characters and the condition of the gonads as well. (This was apparently a bimodal group , some of the intersexes being of a more '^female" t>pe and others of a more "male" t}T3e.) Cytological examination of these indi\iduals revealed that the second and third chromosomes were regularly present in a triploid condition, that the fourth chromosome was either diploid or tri])loitI, and that two X chromosomes were regularly present (witJi or without a Y chromosome). The situation is interpreted as follows : 200 Outline of Genetics ''It is not the simple possession of two X chromosomes that makes a female, or of one that makes a male. The preponderance of genes that are in the autosomes tends toward the production of male characters; and the net effect of genes in the X is a tendency to the production of female characters. The ratio of 2X:2 sets autosomes produces a female, while iX : 2 sets autosomes produces a male. An intermediate ration, 2X13 sets autosomes, produces an intermediate condition, the intersex." ''The fourth chromosome seems to have a disproportionately large share of the total male-producing genes; for there are indica- tions that the triplo-fourth intersexes are preponderantly of the 'male' type, while the diplo-fourth intersexes are mainly 'female' type." According to this conception, 3 A': 2 sets autosomes should be " superfemales " and iX:3 sets autosomes should be "supermales." Bridges has actually identified such types, both being sterile. It is certain that this conception will exert a far-reaching influ- ence upon the existing ideas of sex detemination. In the first place, it gives a somewhat more exact idea as to the elements effective in determining sex. Hitherto it had been thought, rather vaguely, that the X chromosome determines sex either per se or by virtue of some special factor which it contains. It is interesting now to realize that a number of factors may be influ- encing sex in one direction or the other, and perhaps that these are identical with factors which have previously been known as playing another role. A different rate of metabolism has commonly been associated with the two sexes; a study of the influence of specific factors on metabolic rate now becomes significant in this connection. In the second place, it furnishes an exact interpretation of intersexes on a chromosome basis. Hitherto intersexes have usually been interpreted in rather vague physiological terms, and have been used as an argument against the sex chromosome theory (or have been harmonized with the sex chromosome theory only by the assumption of some additional extra-chromosomal influ- ence — GoLDSCHMiDT 13). Bridges' Conception now paints a Sex Dclcrniiniition 201 quantitative picture of sex without calling upon any other cffcclive elements than the "orthodox" factors of inheritance that arc located on the chromosomes. Intersexes arc therefore accounted for by the same general mechanism as normally produces only pure males and females in the fruit fly. In the third place, the theoretical possibility of artificially controlling sex is illuminated. Such control should l>e possible to the degree that the ordinary heritable characters can successfully be duplicated artificially. Hridc;i:s acknowledges that thr en- vironment may affect sex within certain limits. .Mlhough sex is fundamentally a quantitative proposition. \ X : 2 sets autosomes provides such a considerable preponderance of male-inducing factors, and 2X12 sets autosomes provides such a preix)nderance of female-inducing factors, that only these two distinct qualitative conditions are visualized under ordinary circumstances. Hoth of the foregoing conditions are far from the point of ecjuilibrium between the opposite types of sex intluences. Under such cir- cumstances the minor influences of single factors in one direction or the other produce no appreciable elTect. .\s a matter of fact, a factor mutation in the germ plasm or an unusual combination of extrinsic physiological conditions might intervene to influence a male individual toward femaleness (or vice versa), but the individual is so preponderantly male that the etTects of these minor influences are not noticeable. On the other hand, in those individuals (the interse.xes) where the male-inducing and female-inducing factors are near the jxiint of equilibrium, the minor influences of single factors in one direc- tion or the other become noticeable. In such an intlividual an unusual combination of extrinsic physiological conditions may swing the individual more toward maleness or more toward female- ness, and these deviations will be observeil. This idea is borne out by the actual facts, since the inlluence of environmental conditions upon the grade of sex in Bridgks' intersexes is notice- able, but the same conditions do not produce noticeable clTecls upon the normal males and females. The intersexes, representing a condition near an equilibrium between opj)osite factor influ- ences, are more "responsive" to environmental diflerenci's. more "fluctuating" than are the normal males and females. 202 Outline of Genetics One might then assume that in organisms where the unisexual condition has existed for some time back in phylogeny, a definite sex chromosome mechanism has been established. This mecha- nism insures (normally) the production of two types of individuals in equal numbers, those which are preponderantly males and those which are preponderantly females. The grade of sex does not appear to fluctuate in response to varying environmental influences, since these influences are relatively insignificant in such cases. In other organisms, however, which have more recently been evolved from bisexual ancestors, a regular sex chromosome mecha- nism has not yet been perfected. The appropriate machinery is not yet at work to produce individuals which are preponderantly male and individuals which are preponderantly female in equal numbers. Instead, sex is being influenced by numerous factors which are distributed sporadically rather than in organized groups as in the fruit fly. The net effect of these factor influences is commonly near to the point of equilibrium, so that the organism is more responsive to environmental influences on sex grade. Under such conditions the sex grade and the sex ratio may be susceptible to a certain amount of artificial control through manip- ulation of the effective environmental influences. Finally, this work of Bridges' casts a new light upon the whole subject of unit characters. Careful investigation of flies which are triploid with respect to one or more chromosome sets, leads Bridges to draw the same general conclusions with regard to other so-called unit characters that he drew with regard to the character of sex. Many characters have their degree of develop- ment influenced, not merely by the presence or absence of certain single genes, but by the net effect of the influences of numerous genes. It is true that there is commonly one gene that exerts a greater influence on the character in question than do any other genes, and it is quite common that all the other genes may be constant in their presence or absence, so that only the effects of the one gene are noticeable, and we identify it as " the determiner" of the character in question. In such cases the equiUbrium of •opposing influences is normaUy being affected to a perceptible •degree only by the presence or absence of a single gene. Abnormal situations, however, may arise, as the result of non-disjunction of Sex Dclermi)hitiou 203 certain chromosome sets. This. I)y inlro'onia, sex determination in, 193 Bud variations, 119 Burbank, L., 150 BurgelT. II., 197, 198 Cabbage-radish hybrid, hybrid vigor in, 169; sterilily in. 133 Cannabis, sex determination in, 195 Capsella, alpine adaptation in, 28 Capsicum, chlorophyll inluTilancc in, 124 Castle, William i:., 2^, 88, 184 Chimaeras, 1 2 1 Chloroph\ll inheritance. 1 2^ Chromogen, 59, 64, 67 Chromosomes, 41 Chromosome aberrations, 113 Chromosome changes, 113 Chromosome elimination. 121, 187 Chromosome theories of sex deter- mination, iSi Chrxsomdid beetles, dilTerentia- tion of germ plasm and body plasm in, 16 Claussen, R. E., 12 • Collins, G. X., 1O7. 176 Complementary factors, 57, 67, 69, 70, 137 Complexnuitation. U3. no Compositae. parthenogenesis in Conscious effort, in animal evolu- tion, 2 Constant variations, o Continuity of the germ plasm, 15 Continuous variations. 4 Corn: aleurone layer in. 147; l)ud variation in. up; chlorophyll inheritance in, 125; crossing over in, 107; factor interactions 207 208 Outline of Genetics in, 57, 6i, 64, 68, 69, 70, 78; false inheritance of acquired characters in, 31, 177; hybrid vigor in, 157, 161; inheritance of acquired characters in, 18; inheritance of endosperm char- acters in, 145, 146, 148, 151; inheritance of quantitative e char- acters in, 78; linkage in, 97, 154; multiple allelomorphs in, 112, 119; non-disjunction in, 154; sterility in, 132, 134 Correlation, physical, 81 Correns, C, 49, 53, 72, 125, 193 Crossing over, 102, 175 Crypthybrids, 134 Cumulative factors, 67, 72, 177 Cytoplasmic segregation, 123 Czapek, P., and M. E., 59 Darwin, Charles, 3, 11, 77, 87, 156 Danvin, Erasmus, i Datura: non-disjunction in, 114; tetraploidy in, 115; triploidy in, 115 Deficiency, 113, 116 Degeneration through disuse, 3 Determinate variations, 9 Determiner, 42, 56 Detlefsen, J. A,, 107 DeVries, Hugo, 6, 11, 84, 86, 87, 109, no, 117 Dictyota, inheritance of acquired characters in, 33 Dihybrid ratio, 46 Diploid, 42 Discontinuous variations, 6 Diseases, inheritance of, 20 Dominance, 39; accounting for hybrid vigor, 1 73 ; failure of, 53 Doncaster, L., 14, 15 Dose, double and single, 44 Double crossing over, 106 Duplex, 50 Duplication, 113, 116 East, E. M., 57, 61, 64, 78, 84, 93, 130, 138, 148, 153, 157, 160, 166 Elimination factor, 22 Elodea, sex determination in, 192 Emerson, E, R., 61, 69, 78, 97, 107, 119, 152, 153, 154 Endosperm, inheritance in, 141; hybrid vigor in, 167, 169 Engler, Arnold, 32 Envdronment, inheritance of effects of, 23; role in evolution, 2, 3 Eyster, W. H., 134 Factor hypothesis, 56, 93 Farnham, ]\I. E., 114 Fern prothallia, sex in, 194 Fittest, survival of, 4 Fixation of hybrids, 86 Fluctuating variations, 4, 79, 85, 89, 119, 201 Four-o'clock, blending inheritance in, 53, 72; chlorophyll inherit- ance in, 125 Frog, sex determination in, 187 Fruit fly: crossing over in, 107; deficiency in, 113; duplication in, 113; gynandromorphs in, 187; _ linkage in, 97, 105; multiple allelomorphs in, 112; mutations in, no; non- disjunction in, 185; sex-lined inheritance in, 184 Funaria, sex determination in, 191 Gartner, C. F., 156 Gal ton, Francis, 12 Gametophyte, inheritance in, 128, 134; sex determination in, 190, 196 Garner, W. W., 132 Gene, 56, 96, 97 Gene changes, 109, 119 Genotype, 45 Geothe, J., i Goldschmidt, R., 200 Graft-hybrids, 121 Index 2oq Gravatt, F., 133 Guyer, M. F., 10, 25, 31 Gynandromorphs, 187 Hansen, E. C, 32 Haploid, 42 Hayes, H. K., 57, 61, 64 Hegner, R. W., 16 Herrick, C. J., 30 Hertweg, R., 187 Heterochromosomes, 182 Heterosis, 162 Heterozygosis, 162, 170 Heterozygote, 44 Homoz}-gote, 44 Hooded rats, Castle's selection experiments with, 88 Humulus, chlorophyll inheritance in, 125 Hutchinson, C. B., 107 Hybridization, evolution through, 9, II Hybrid Wgor, 156 Ikeno, S., 124 Immunity to disease, inheritance of, 20 Inbreeding, and hybrid vigor, 177; and sterility, 134 Independent unit characters, 38 Indeterminate variations, 9 Inhibitory factors, 64, 67, 69 Intersexes, 190, 194, 196, 199 Isolation, its role in evolution, 7, 1 1 Jennings, H. S., 24 Jimson weed: non-disjunction in, 114; tetraploidy in, 115; trip- loidy in, 115 Johannsen, W., 44, 84 Jones, D. F., 157, 166, 107, i74, 177 Jordan, David Starr, 7 Kceble, F., 173 Kempton, J. II . 152, 167 King, 11. I)., iss Kohircuter, J. G., 156 Lamarck. J., 2, 10, 11, 12. 10. 28 LancefKld, 1). E., 106 Lethal factors, 68, 69, no, 117 Light seed, effect of, 13, 14 Linear arrangement of genes, 07, 105 Linkage, q(\ 113, 174, 184 Locus changes. loq, ii^v •" Long, W. H., i^^ Lotsy, J. P., 9, 61 Lychnis, sex determination in. iq3 Malthus, T. R. s Marchal, I'^l.. and Km., igi Maternal inheritance, 125. 14Q Matthiola, factor interactions in. 70; non-Mendelian inheritance in, 126 Mayr, H., 31 ]Melandrium, sex determination in, 193 Mendel, Gregor, 37, 56, 157 Mercurialis, sex determination in, 196 Metabolic rate, in sex determina- tion, 188, 195, 200 ^Microorganisms, inheritance in, 24 Mirabilis, blending inheritance in, 53, 72; chlorophyll inheritance in, 125 Moditicalion of unit characters, S8 McMlifying factors, Oi, 04 Monohybrid ratio, 44 Moore, C. W., 131) Morgan. T. II., 97, nj, 187 Mucor, sex in, io<> Mule, as an example of hybrid vigor, lOo MulUr. 11. J , 110, 113, 117 2IO Outline of Genetics Multiple allelomorphs, 112, 119 Multiple modifying factors, 94 Mumis, E. M., 32 Mutation, 6, 109, 119 Mutationists, 88 Mutilations, inheritance of, 17 Nemec, B., 146 Neo-Darwinians, 13 Neo-Lamarckians, 13 Nicotiana: self-sterility in, 138; hybrid \^gor in, 164; wide crosses in, 169 Nilsson-Ehle, H., 72, 78, 113 Non-disjunction, 114, 116, 185, 199 Non-Mendelian inheritance, 1 23 Nulliplex, 50 Oenothera, mutation in, 6, 117 Origin of species, Darwin's theory, 3 Orthogenesis, 8, 11 Oxidase, 59, 64, 67 Parasitella, sex in, 198 Parasitism and sex, 198 Park, J. B., 138 Parthenogenesis, 133 Particulate inheritance, 123 Pea, jNIendel's experiments with, 37, 157 Pelargonium, chimaeras in, 122; chlorophyll inheritance in, 123 Pellew, C. W., 173 Phenotype, 45 Phycomyces, sex in, 197 Physical basis of heredity, 41 Physical correlation, 81 Physiological theories of sex de- termination, 181, 187 Pigeons, sex determination in, 188 Presence and absence hypothesis, 49, 112 Protoplasmic connections, 17 Puccinia, inheritance of acquired characters in, t,2) Purity of gametes, 40 Qualitative variations, 6 Quantitative characters, inherit- ance of, 72 Quantitative variations, 4, 72 Rabbits, inheritance of acquired eye defects in, 25, 31 Randolph, L. F., 125 Reciprocal crosses, 38 Reduction division, 42 Regression, 87, 89 Reversion, 59, 87, 89 Riddle, O., 188 St. Hilaire, G., i Santos, J. K., 192 Saunders, E. R., 70 Sax, K., 169 Schaffner, J. H., 195 Segregation, 40; cytoplasmic, 123; somatic, 120 Selection, 84, 88, 112 Selectionists, 88 Selective fertilization, 169, 193 Sex chromosomes, 182, 191, 192 Sex determination, 181 Sex-limited inheritance, 184 Sex-linked inheritance, 184 Shull, G. H., 157, 177, 193 Simplex, 50 Smith, E. A., 10 Solan um, graft-hybrids in, 121 Somatic mutation of genes, 119 Somatic segregation, 1 20 Sphaerocarpus, sex determination in, 190 Spirog>'ra, natural crosses in, 129 Sterility, 131 Index 211 Stocks, factor interactions in, 70; non-Mendelian inheritance in, 126 Strasburger, E., 190, 196 Struggle for existence, its r6le in evolution, 4 Sturtevant, A. H., 107 Superfemales, 200 Supermales, 200 Supplementary' factors, 61, 67, 68 Survival of the fittest, 4 Susceptibility to disease, inherit- ance of, 20 Sweet peas, reversion in, 59 Tanaka, Y., 107 Tetraploidy, 114, n6 Toads, sex determination in, 188 Transeau, E. N., 129 Trihybrid ratio, 48 Triple crossing over, 107 Triploidy, 113, 115, 199 Unit characters, 38, 89 Use and disuse, inheritance of effects of, 2, 19 Walter. H. K., on inheritance of ac(juired characters, 14, 17 Weismann, August, 9, 13, 15, 17, i'"^, 19, -^5. 77 W lieat: cumulative factors in, 72; complexmutation in, 113; hy- brid vigor in, 169 Wide crosses, 169 Wiedershcim, W., i8 Williams, J. Lloyd, 7,1^ Winge, ()., 123 Winkler H . i:'! A' chromosome, 184, 191, 192, 200 Xenia, 145 1' chromosome, 184, 191, 192 Yampolsky, C, 196 Yeast, inheritance of acquired characters in, 1,2 Zedebaur, E., 28 Zeleny, C, no, 1 1 1 Zig-zag inheritance, 184 ntOPERTY UBUIT N. C. State CoIIeflC PRINTED IN THK t.S.A.