A. RELATION OF HEREDITY AND ONTOGENY.

In studying heredity our attention must often be focused on the ontogenesis of the different characters, and we are sometimes inclined to regard the adult character as the product of the course of ontogenesis. But this is a superficial way of looking at things; the determiners of all characters are in the germ-plasm and together they direct the development of one part after another in orderly succession; a modernized form of the pre-formation doctrine seems logically necessary.

What do we know of the processes that take place in bringing the fertilized egg, freighted with its specific heredity, to its destination—the adult form? Modern embryological and cytological studies give us an insight into many of them. First of all, the egg has a certain organization that foreshadows something of its fate. Then cell-divisions begin, at first synchronous, but later becoming accelerated here and retarded there. Eventually (especially among animals) these cells become arranged into a membrane whose unequal growth in limited areas produces foldings. The folding of membranes, their stretching, local thickenings, or thinnings are largely the result of local inhibitions of water. Sometimes movements of individual cells occur out of the membranes into and through cavities or solid yolk-masses, and by the aggregation of such cells massive organs are sometimes formed. Local absorption of tissues already established may be effected in later life by such migratory cells. Membranes once established may form pockets or linear folds, as in gastrulation and gland formation; they may become perforated; two membranes may fuse along areas or lines and a perforation may even occur at the region of fusion. Linear strands or tubules may grow out, making connections, as nerves do, with distant organs; tubes may unite to form a network, or split lengthwise. Finally, membranes and masses undergo vacuolization, or masses may split apart or fuse together. Thus in the ontogeny that is proceeding under the control of heredity all is motion and change.

What are the factors that control all these movements—for these are the true factors of heredity? We do not know much about them, but we know some things. We know that cell-divisions occur at particular times and places under the influence of preceding division planes; but their normal occurrence may be interfered with by an abnormal chemical condition of the environment. We have reason for concluding that each developmental process is a "response"—a reaction of the living, streaming protoplasm to changing environment. The nature of the response to any stimulus probably depends on the chemical constitution of the protoplasm—and this is hereditary. In an important sense heredity is the control of ontogeny.

The specific characteristics are mostly those that appear late in ontogeny. The integumentary folds over the nasal bones of the chick appear on or about the tenth day. At that time it can be ascertained whether the comb is median, or multiple, or Y-shaped, or cup-shaped, or consists of 2 papillÆ. In the case of the single-comb the fold is linear and single; in the case of the pea-comb, linear and triple; in the case of the rose-comb, quintuple or irregularly wrinkled over the whole area; in the case of the Polish-comb, there is a pair of "pocket folds." In the single-combed fowl the single linear fold grows quickly to a great height and very thin, while in the pea-comb, with its additional pair of wrinkles, the median element is not so high as in typical single-combed races; in the pea-comb there is an additional folding stimulus and a reduced growth stimulus. In the heterozygote both stimuli are weakened; the lateral folds are usually much reduced—"are hard to make out," as I stated in 1906 (p. 35); and the factor that determines the continued growth (elevation) of the fold is weakened, so that the pea-comb—although "abnormally high" (1906, p. 35, figs. 20 and 21)—is not nearly as high as the single-comb of the Minorca (1906, fig. 4).

Two results are evident: first, each character in the heterozygous condition is reduced, and, second, each is much more variable than in the homozygous condition. Why is the character reduced? If the reaction to continued growth of the fold is strong in one race and weak in the other, then in the heterozygote that reaction, whatever its nature, is reduced. Why is the reduction in the response so variable? There is a variation in the irritability or other growing factor of the embryonic material that is destined to form the fold. Even Minorcas vary in the growth of the comb, and so do the Dark Brahmas. Let G be a constant element of the growth factor of the Minorca's comb; then G+a or G-a will indicate its variants. Let g be the growth factor of the Brahma's comb, and g+a and g-a its variants. Then the hybrids of these two races may be of the following types: Gg, Gg+a, Gg-a, Gg+2a, Gg-2a. This gives 5 varying conditions instead of 3 and greater extremes of variation.

In the foregoing case I have assumed that the positive character is that of increased growth in the Minorca; but the positive character may be an inhibition to indefinite growth of the pea-comb. Heredity may be conceived of as exerting at all points a control on developmental processes—sometimes initiating and continuing this; but often, on the other hand, slowing down or wholly inhibiting that. The inhibition of a process is quite as positive a function of heredity as its initiation. The hair of a young rabbit grows until it attains a certain length and then the growth ceases. The growing character is a youthful, embryonic one; the new character is the stoppage of growth. Similarly the young feathers of birds grow continuously until something intervenes that stops the growth and dries up the sheath. Now, in Angora rabbits and long-tailed fowl the epidermal organ continues its embryonic growth indefinitely; the something that intervenes to stop growth is absent. There is no reason for regarding the long hair or long feather as a positive condition and short hair or feather as due to its absence.

Again, Mediterranean fowl have non-feathered shanks; but in Asiatics the feet are feathered like the rest of the body (except the soles and face). It has been assumed that boot is an additional character and should be dominant over absence of boot. But, on the other hand, we may well think of the capacity of producing feathers as general to the skin. From this point of view the real question is, what prevents feather production on the eyelids, comb, wattles, and shank? It seems equally probable that there is an inhibitor of feather-growth for these few areas as that every conceivable area of the body has its special stimulus factor for feather development; or even as that there is such a factor to each separate feather-tract. In the Minorca, then, the inhibitor of boot is present; in the Silkie a weak heterozygous inhibition appears; but in the Dark Brahma there is no inhibitor and feathers extend down from the heel over the whole of front and sides of the foot and even on the upper surface of the toes—just as they do over the anterior appendages.

The case of the rumpless fowl is important in relation to the hypothesis of inhibitors. Either tail-production depends on a special factor TT, which is diluted, as Tt, in the heterozygote; or else there is a tail inhibitor, II, which is diluted, as Ii, in the heterozygote. In F₂ we expect, on the one hypothesis, 25 per cent tt, giving no tail, and 25 per cent TT, giving tail; on the other hypothesis 25 per cent ii, giving tail, and 25 per cent II, giving no tail. Actually we get all tailed in some cases; in others 25 per cent with no tail. Which hypothesis best fits the facts? Which is the more probable—that the 25 per cent recessive no-tail should produce a tail (as it were, out of nothing) or that the 25 per cent dominant tail inhibitor should be ineffective, permitting the development of a tail? It is clear that the ontogenetic failure of an inhibitor is easier to understand than the development of a character that is not represented at all in the germ-plasm. This matter is treated in another connection in the next section. But the present point is that it is equally in accord with the facts to regard heredity as initiating and inhibiting processes. If, indeed, processes were not regularly inhibited, they must, when once started, go on indefinitely, as do the hairs of Angora goats and wonder-horses.

As we have seen, ontogeny is not completed at hatching or birth. Many characters are at that time undeveloped. Hence, not infrequently the recessive condition is at first seen and is only later replaced by the dominant condition. The reverse sequence will rarely be followed, because development rarely, except in cases of degeneration, moves backward. One of the familiar cases of this sort is human hair-color. In youth this is frequently flaxen, later it becomes light brown, and eventually it may become dark brown. Darwin gives a number of examples in his Chapter XII of Animals and Plants under Domestication. To these I may add some from my own experience. The hybrids between white and gray Java sparrows are at first light and later become of a slaty gray like the dark parent. Many black fowl gain white feathers as they grow older, and every fancier knows that birds with complex white-and-black patterns can usually be "exhibited" only once, on account of loss of "standard" coloration late in life. In these cases the advanced condition in the series of melanic colors appears only late in ontogeny.[13] Similarly Lang (1908, p. 54) finds that in snail hybrids often the young shells have the recessive yellow color, only later in life showing the dominant red color. This is, of course, no reversal of dominance in ontogeny, but mere ontogenesis of pigmentation. So in general, since the recessive condition is absence of the character or its low stage of development and the dominant condition is presence of the full character, the individual in ontogenesis may exhibit in succession the recessive and then the dominant character, but not in the reverse order.

B. DOMINANCE AND RECESSIVENESS.

If segregation is the cornerstone of modern studies in heredity, dominance forms an important part, at least, of the foundation. In any case, a critical examination of dominance is now required; the more so since its significance and value have often been doubted.

First, how is a dominant character to be defined? It has been defined both on the basis of visible results in mating and on the basis of its essential nature. On the basis of visible results in hybridizing dominant characters may be defined as Mendel (1866, p. 11) defined them: "jene Merkmale, welche ganz oder fast unverÄndert in die Hybride-Verbindung Übergehen." Bateson's translation (1902, p. 49) renders this passage: "those characters which are transmitted entire, or almost unchanged in the hybridization."

On the basis of the essential nature of the dominant character there has obtained a great diversity of definitions. Thus de Vries (1900, p. 85) suggested that the "systematically higher" character is the dominating one, and, again (1902, pp. 33, 145), that the dominant character is the phylo-genetically older one. Many have suggested that it is the positive or present character that dominates over the negative, latent or absent. This last idea has become the prevailing one and its history is worth summarizing.

As early as 1902, Correns used as Mendelian pairs, presence of coloring material and absence; also modification into yellow and no modification. In 1905, he extended somewhat this use of present and absent characters, k (keine) preceding the symbol of a character as a negative. Still he did not pretend to generalize the relation of dominance and recessiveness to be that of presence and absence. In 1903 (p. 146) de Vries stated that in very many cases Mendel's law held when one quality is active and the other latent, and that the active quality is dominant. His illustrations show that by activity he meant essentially presence, by latency absence from the visible soma. Bateson's third report (1906) applies presence and absence to several additional cases, and, at the International Genetics Conference of that year, Hurst developed the presence-and-absence hypothesis, favoring the view that the factor for absence is nothing at all, but finding that certain cases, such as Angora coat, offer a difficulty. At the same meeting I suggested that "a variation * * * that is due to abbreviation of the ontogenetic process, which depends on something having dropped out, will be recessive," a progressive variation dominant; and in 1908 I expressed the conclusion that "dominance in heredity appears when a stronger determiner meets a weaker determiner in the germ. The extreme case is that in which a strong determiner meets a determiner so weak as to be practically absent, as when a red flower is crossed with white." I suggested that in some cases of recessiveness of an apparent advanced condition, like Angora hair, the dominant factor is an inhibitor. In the last year or two the presence-and-absence theory has gained wide acceptance, but I still think the cases where there is dominance of the advanced condition over the less advanced—of the quantitatively well-developed over the quantitatively less well-developed—have not been sufficiently considered. In human hair-color any other hypothesis demands that there are many units in the higher grades of pigmentation and fewer in the lower grades and that the presence of the surplus factor in any other higher grade dominates over its absence in the next lower grade; but there is no evidence in human hair-color of distinct, discontinuous units in the common yellow-brown series. And, in ontogeny, the different grades of color form a continuous series whose development proceeds throughout early life and may even be stimulated to an advanced stage of darkening by disease. The cessation of color development may take place at any point, and this seems incompatible with the theory of unit-characters for the different grades of human hair-color. In the present paper, on the other hand, the characters dealt with are mostly unit-characters and their quantitative variations mostly heterozygotic. Even the case of the Silkie boot (table 31, C) referred to in an earlier paper[14] as illustrating recessiveness of the less advanced condition proves, on further analysis, to be a case of heterozygotism. It seems highly probable that the future will show that many more advanced or progressive conditions are really due to one or more unit-characters not present in the less advanced condition. In that case it will appear that there is perfect accord in the two statements that the progressive condition and the "present" factor are dominant.

The definition of dominance on the ground of results meets at the outset with a difficulty the germ of which is observable in Mendel's cautious statement "ganz oder fast unverÄndert." Even Mendel observed that the hybrids between white-flowered and purple-red flowered peas have flowers less intensely colored than the darker parent. The experiments of the last seven years have shown that the "dominant" character is often very greatly changed—indeed, in extreme cases a blending of characters may occur—in the first generation. Correns (1900 b, p. 110) very early stated that in a certain set of crosses between good species the hybrids showed the character of both parents, only reduced, but in varying degrees. Bateson and Saunders (1902, p. 23) found in crossing two forms of Datura that—

Although the offspring resulting from a cross between any two of the forms employed are usually indistinguishable from the type which is dominant as regards the particular character crossed, yet in other cases the intensity of a dominant character may be more or less diminished either in particular individuals or in particular parts of one individual. In Tatula-Stramonium cross-breds the corolla is often paler in color than that of the dominant parent (as has already been noticed by Naudin), but even in the palest specimens the deep blue color of the unopened anthers leaves no doubt as to the presence of the dominant color element. * * * The occurrence of intermediate forms was also occasionally noticeable in the fruits. Among the large number of capsules examined, there were some of the mosaic type, in which part of the capsule was prickly and the remainder smooth, while others, suggesting a blend, were more or less prickly all over, but the prickles were much reduced in size, and often formed mere tubercles.

Bateson and Saunders further showed (1902, p. 123) that in the case of comb and extra-toe in poultry "the cross-bred may show some blending and * * * the intensity of the dominant character is often considerably reduced."

Correns (1905, p. 9) pointed out that there was known, even at that time, a complete series of cases at one extreme of which one determiner completely hindered the appearance of the other, while at the opposite end of the series the hybrid showed an intermediate condition, both determiners appearing with equal strength.

The following year, in my first report on Inheritance in Poultry, I laid great stress on the imperfection of dominance, and this phenomenon has become more striking and clear in the subsequent years, until in the present paper it is recognized as the key to the explanation of many apparently anomalous types of heredity.

The first case in the present work in which imperfection of dominance is considered is that of the hybrids between I and oo comb. Here median comb is mated with no-median. Each somatic cell of the hybrid—at least in the comb region—has only half the full determiner for median comb. The determiner is weakened, and so the median comb is imperfectly developed, namely, at the anterior end of its proper territory. The weakening varies much in degree in the heterozygote. The median comb may be reduced to 70 per cent of its normal length or it may not develop at all.

The second case of imperfection of dominance is that of polydactylism. Extra-toe mated to normal gives extra-toe in 73 per cent only of the offspring in the case of the Houdans. Any trace of 6 toes (on one or both feet) is found in only 12 per cent of the hybrid offspring from a 6-toed Silkie parent. Certainly dominance here is very like blending.

The third case of imperfection of dominance is that of syndactylism. No syndactyls were noticed in F₁. My first conclusion was that syndactylism is recessive; but later studies have shown that it is dominant and that all matings of two syndactyl parents yield about 56 per cent syndactyl offspring.

Rumplessness gives an illustration of how dominance may be so weak as to be absent altogether; so that from F₁ alone the erroneous conclusion is drawn that it is recessive; indeed, in one strain, only faint traces of the character made their appearance in successive generations.

Finally, winglessness is a character which appears not to be inherited at all. Nevertheless our experience with rumplessness leads us to suspect that winglessness also is an impotently dominant character.

Looking at the matter frankly and without prejudice, the question must be answered: Has not the whole hypothesis of dominance become reductio ad absurdum? What visible criterion of dominance remains, where dominance fails completely? All the usual statistical landmarks of proportional appearance in successive generations being lost, can one properly speak of dominance and recessiveness at all?

Amid the general ruin of criteria, however, one means of detecting dominance remains. That extracted character which in F₂ or subsequent generations shows in homologous[15] matings in some families a wide range of variability is dominant, while that extracted character which constantly, in all homologous matings, shows no or very little variation is recessive.

The reason for this difference in the inheritableness of the two conditions is easy to understand on the principles enumerated in the last section. A positive character has a real ontogeny. But, as we have seen, the development of any character may be interrupted at any stage. Most aberrations among organisms are due to a retardation or failure of normal development. In human affairs we recognize this tendency in the terms "degenerates" and "defectives" (constituting from 2 to 4 per cent of the population). Indeed, there are few persons who are not defective in some physical or psychical character. In cases where the commonest form of abnormality is due to a development in excess it seems probable that a normal restraining or inhibiting factor is defective or absent. On page 88 I tried to show how common in ontogeny such restraining and inhibiting factors are. Since ontogenetic processes are so often cut short by external conditions, we can understand the variability in the degree of development of positive characters.

On the other hand, whenever the fundamental hereditary stimulus or the material for a character is absent from the germ-plasm of both parents, then it can appear in none of the offspring; they will be practically invariable in respect to this condition. Only the ontogenetic fluctuations of other real characters may influence the defect. Consequently the absent state reproduces itself, the "recessive breeds true."

The considerations here presented bear upon the hypothesis of change of dominance. Bateson and Punnett (1905, p. 114) say of poultry: "The normal foot, though commonly recessive, may sometimes dominate the extra-toe character." This idea of occasional change in dominance has been expressed more than once in the literature. I think the phrase an unfortunate one. In my earlier report[16] I urged that a characteristic that is anywhere dominant is so without regard to race or species involved. If this is so it is clearly improbable that it should vary from individual to individual, or in the same individual at different times. Rather in view of the imperfection of dominance we should say that a dominant character sometimes fails to develop, in which case it is absent from the progeny; that is all. It is particularly apt to fail of development when dilute—heterozygous.

C. POTENCY.

Perhaps an apology is needed for introducing the much-abused word "potency"; but there is hardly another that can be so readily adapted to the precise definition I desire to give to it. The potency of a character may be defined as the capacity of its germinal determiner to complete its entire ontogeny. If we think of every character as being represented in the germ by a determiner, then we must recognize the fact that this determiner may sometimes develop fully, sometimes imperfectly, and sometimes not at all. When such a failure occurs in a normal strain a sport results.

Potency is variable. Even in a pure strain a determiner does not always develop fully, and this is an important cause of individual variability. But in a heterozygote potency is usually more or less reduced. When the reduction is slight dominance is nearly complete; but when the reduction is great dominance is more or less incomplete and, in the extreme case, may be absent altogether. The series of cases of varying perfection of dominance described in this work illustrate at the same time varying potency. The extreme case is that of the rumpless fowl. The character in this case is an inhibitor of tail development. This character has arisen among vertebrates repeatedly and has become perpetuated in some amphibia and primates, including man. In the case of our cock No. 117, the action of the inhibitor is very weak, so that in the heterozygote the development of the tail is not interfered with at all and even in extracted dominants it interferes little with tail development, so that it makes itself felt only in reduced size of the uropygium and in bent or shortened back. But in No. 116 the inhibiting determiner is strong. It develops fully in about 47 per cent of the heterozygotes and 2 extracted dominants may produce a family in all of which the tail's development is inhibited. In the case of the rumpless condition that arose apparently de novo in my yards, the new inhibitor showed an intermediate potency completely stopping the tail development in 1 out of 25 heterozygotes. These three cases afford a striking illustration of a variation in the potency of the same inhibiting character in different strains.

Not only is potency variable, but its variations seem, in some cases, to be inheritable. This we have seen to be the case with the Y-comb (p. 15); with the extra-toed condition of Houdans (p. 23); and with rumplessness (cf. offspring of No. 117 as compared with No. 116, p. 40). On the other hand, the extra-toed condition of Silkies, the grade of clean shank, and the degree of closure of nostril seem not to be inherited.

D. REVERSION AND THE FACTOR HYPOTHESIS.

The brilliant development of the factor hypothesis, only dimly fore-shadowed by Mendel[17] (1866, p. 38), clearly expressed by Correns (1892), applied to animals by CuÉnot, and further elaborated by Bateson and Castle and their pupils, has quite changed the methods of work in heredity. More forcibly than ever is it brought home to us that the constitution of the germ-plasm—not merely the somatic character—is the object of our investigation. With this principle fully grasped the existence of cryptomeres and the resolution of characters have become clearer. But the most striking result accomplished has been that of clearing up the whole range of phenomena formerly placed in the category of "reversion." No idea without a semblance of inductive explanation has been more generally accepted in the Darwinian sense both by professed biologists and practical breeders than this. Not only was the fact of recurrence of ancestral types in domesticated organisms accepted, but the idea that, in some way, hybridization per se destroyed the results of breeding under domestication was maintained.[18] Now we know that, under domestication, many races have been preserved that are characterized by a deficiency of a character or by a new, additional one, and that hybridization, by bringing together again those characters that are found in the ancestral species, may bring about again individuals of the ancestral type. There is nothing more mysterious about reversion, from the modern standpoint, than about forming a word from the proper combination of letters.

E. THE LIMITS OF SELECTION.

In the last few decades the view has been widespread that characters can be built up from perhaps nothing at all by selecting in each generation the merely quantitative variation that goes farthest in the desired direction. I have made two tests of this view, using the plumage color of poultry.

(1) Increasing the red in the Dark Brahma × Minorca cross.—The Dark Brahma[19] belongs to the group of poultry that contains a majority of characters derived from the Aseel type. Nevertheless, its plumage is closely related to that of the Jungle-fowl, from which it may be derived on the assumption that the red part of the pattern has become, for the most part, white. However, a little red remains on the middle of the upper feathers of the wing-bar. I crossed such a bird with a Black Minorca, and, as reported in my earlier work,[20] the offspring were all black, except that the males showed some red on the wing-bar. The amount of red varied in the different males, and I decided to test the possibility of much increasing the amount of the red by selection in successive generations. So I chose the reddest cock to head the pen. In this pen (No. 632) 222 chicks were produced and grew to a stage in which their adult color could be determined. Of these 222 chicks, 160, or 72 per cent, were black, without red; 24, or 10.8 per cent, were black with some red; 38, or 11.7 per cent, were typical Dark Brahmas, and 9 others, or 4.5 per cent, were modified Dark Brahmas.

The following year (pen 732) I bred a cock derived from the last year's pen, a bird that resembled much the male Dark Brahma (except that it was somewhat darker), to sundry hens, hybrids between the Dark Brahma and Minorca—some of the first and some of a later hybrid generation, but all black except that some of the 1906 birds had a little buff on the breast and the primaries. The F₁ (black) × F₂ (Dark Brahma) gave 51 per cent black offspring, 27 per cent with a black-and-red Game pattern, and 22 per cent with the Dark Brahma pattern devoid of red. Thus the third generation suddenly gave me a red-and-black Game-colored bird (plate 12)!

My interpretation of the foregoing results is as follows: The Dark Brahma gametic formula proves to be CIrnwx, whereas the Black Minorca is C(IR)Nwx, where (IR) is equivalent to, and merely a further analysis of, the J of the formula of the Minorca as given in earlier sections. The I stands for the Jungle pattern without red and R is the red element in that pattern. Obviously N and R are the differential factors, 4 kinds of gametes occur in F₁, and in every 16 offspring these factors are combined in the following proportions: 9 NR, 3 Nr, 3 nR, 1 nr (compare the distribution of color types in the 222 offspring of pen 632). The F₂ male selected as father of the next generation (in pen 732) was an extracted Dark Brahma in coloration and probably formed only 1 kind of gamete, nr; but the hens were heterozygous in respect to N and R. Consequently 4 kinds of zygotes are to be expected in F₃; and expectation was realized as indicated in table 68.

Table 68.