PROCEEDINGS OF THE WASHINGTON ACADEMY OF SCIENCES
VOL. IX, PP. 189-240. JULY 31, 1907.

MENDELISM AND OTHER METHODS OF DESCENT.1
By O. F. COOK.

SYNOPSIS
Introduction 189
The nature of experiments in descent 191
Parent organisms produce parent gametes 192
Parent gametes produce conjugate organisms 192
Conjugate and exjugate reproduction 193
Conjugates produce perjugates 195
Summary of the five elements of descent 197
Mendelism as a method of descent 198
Primary methods of descent 198
Synopsis of methods of descent 199
Diagram of methods of descent 216
Position of Mendelism as a method of descent 218
Mendelian "principles" not demonstrated 218
Mendelian inheritance of mutations 220
Problems of expression and transmission 224
Relation of Mendelism to normal diversity of descent 230
Evolutionary implications of Mendelism 232
Conclusions 238

POSITION OF MENDELISM AS A METHOD OF DESCENT.

No claim of completeness is made for these condensed summaries and tables. They are introduced at this time only to show the position of Mendelism among other methods of descent. It has been customary to describe Mendelism as a form of heredity, but in reality it constitutes a rather wide departure from the primary concept of heredity, as developed in evolutionary literature. Heredity usually implies the production of like by like, which many have considered the normal process of descent.

Mendelism is one of the methods of descent in which unlike produce unlike. Mendelism has aroused special interest in the scientific world largely because it seemed to contradict the earlier inferences from the idea of heredity, by showing that contrasted differences are preserved, and not reduced to a uniform intermediate average. Instead of being a form of heredity, Mendelism is a specialization of heterism; it is one of the methods of increasing diversity of descent, which sustains the efficiency of the processes of sexual reproduction. The preservation of differences inside the species by means of sex-inheritance is one of the most familiar phenomena of descent, but the intimate resemblance between Mendelism and sex-inheritance has not been adequately appreciated.

MENDELIAN "PRINCIPLES" NOT DEMONSTRATED.

To explain these phenomena of polar inheritance the Mendelian doctrine of pure germ-cells was invented, but the facts [219] do not require so extensive an assumption to be made. It is not necessary to suppose that gamete-formation is the reverse of fertilization. The facts indicate only that the perjugate gametes are definitely polarized or predisposed to the expression of one or the other of the divergent parental characters. It may not be entirely inconceivable that pure germ-cells could be formed on the plan of the Mendelian theory if the incompatibility of a hybrid were great enough to prevent mitapsis, but germ-cells which remained pure in this manner would constitute a very abnormal and exceptional phenomenon, like parthenogenesis or apogamy, instead of being a general method of descent.

If the parental protoplasms separated, instead of completing the fusion of the chromatin, the parental characters should segregate in the original groupings and not be found in all varieties of combinations as happens whenever the parents differ by two or more Mendelian characters. The theory of incompatibility is also excluded by the fact that the varieties which afford the best examples of Mendelism are usually closely related members of the same species, and the crosses are entirely fertile.

The Mendelian "principle of gametic purity" has sometimes been described very appropriately as "the law of disjunction ", since it rests on a denial of the fact of mitapsis, and holds that no permanent fusion takes place between the gametes which are brought together in Mendelian experiments. The presence in a germ-cell of the element or unit representing one of the divergent characters is supposed to exclude the unit which represents the alternative character.

But even the phenomenon of dominance, the primary fact of Mendelism, shows that the representation of a character in a germ-cell and the bringing of it to expression in an organism are two entirely distinct processes. In building up the conjugate structure the two gametes have a wide range of methods of coöperation. In Mendelism the character of one gamete dominates completely, so that the tendency of the other gamete to the expression of a different characteristic is completely concealed and held in abeyance. And yet Mendelian experiments [220] prove conclusively that the recessive character may not be lost or diminished by this period of suppression, but may have equal representation in the expression-tendencies of the gametes of the next generation.

Single gametes, like conjugate gametes, may retain and transmit characteristics which they do not express. The reappearance of the blue plumage of the wild pigeons in Darwin's crosses of white varieties, the return of the red and yellow feathers of the wild jungle-fowl to Professor Davenport's black and white chickens, and the uniform production of black-eyed mice as a result of crossing two pink-eyed varieties by Darbishire, are conspicuous examples of this fact, that characters may be brought again to expression after passing in latent condition through many generations of germ-cells.

With this fact definitely established, that it is possible for germ-cells to transmit characters for many generations without bringing or tending to bring them into expression, the Mendelian "laws of disjunction," "purity of gametes," "alternative inheritance," or whatever they may be called, become at once superfluous and inadequate. They are inadequate because they do not explain how a character-unit could once get back into gametes after being left out. They are superfluous because the facts indicate only that gametes have a tendency or polarity of expression of particular characters, though they may at the same time transmit other characters which can regain the power of expression in subsequent generations.

MENDELIAN INHERITANCE OF MUTATIONS.

Belief in the Mendelian assumptions of character-units and pure germ-cells as general laws or principles of descent has been greatly strengthened by contemporaneous activity in the investigation of the phenomena of mutative variation. Mutative variations commonly obey Mendel's "laws." It is not strange that those who have accepted mutations as genuine species should consider the Mendelian "principles of inheritance" as of universal application, and of fundamental significance for evolution.

That there are very direct relations between mutations and [221] Mendelism can not be denied. Indeed, the relations appear to be still more intimate than even the advocates of the two theories have supposed. Varieties which have arisen by mutation may not only behave in a Mendelian manner when crossed, but mutations may actually represent Mendelian phenomena, at the time of their first appearance.

The nature of the protoplasmic combinations and readjustments which take place during the process of chromatin fusion (mitapsis) is not yet known, but the indications favor the commonly accepted view, that new variations usually occur at mitapsis. As mutations are relatively infrequent phenomena, it would only be in the rarest cases that two newly mutated gametes would meet in conjugation. The general rule must be that new variant gametes conjugate with gametes of the unmutated parental type, and thus have from the first the reproductive status of Mendelian crosses.

There are two ways for such a new gamete variation to gain visible expression in a plant or animal. If the mutation be dominant, it can make itself manifest in the first generation. If it be abeyant, or recessive, expression can be secured only in later generations, when it becomes possible for two of the recessive gametes to be brought into conjugation.

Mutative variations which have arisen under conditions of restricted descent tend to disappear when they are brought into contact with the wild type or with a less inbred stock, but new mutations are often dominant and prepotent when bred within their own group of immediately related organisms. Among Professor Davenport's hybrids the frizzled plumage was dominant, while the silky plumage was recessive.

That mutations may be hybrids already, at the time of their first expression in the conjugate organism, would explain differences of behavior between different kinds of new variations. A dominant variation would not be represented in all the progeny of the second generation. Even though the new variation were a plant which could be fertilized by its own pollen, at least a quarter of the second generation would express the character of the recessive gamete grandparent.

Recessive variations, though they can not secure expression [222] in the first or conjugate generation, may be expected to be constant or "true to seed" when they have once gained expression. In the language of special students of Mendelism, such a mutation is already a "pure recessive." To secure equally uniform expression of the dominant character it may be necessary to extract the "pure dominants," that is, to select the progeny of parents which produce offspring of the dominant type only.

A dominant mutation will usually make its first appearance as a single individual. A new variation occurring at mitapsis would be represented by only one female gamete, or by four male gametes. A recessive variation may secure simultaneous expression in numerous individuals of the same generation, or may continue to gain apparently independent expression in successive generations. The method of narrow breeding or self-fertilization used by De Vries and other experimenters to insure the purity of their stocks, is well calculated to bring such abeyant or recessive variations into expression.

These considerations may have practical significance in experiments for inducing new mutative variations by hybridization. The first or conjugate generation does not represent the final result of the cross, but only the conjugate relations of the gametes. The variations may be expected to take place in the mitapsis which precedes the formation of the sex-cells for the next generation, but only the dominant variations can gain expression in the second or perjugate generation of the cross.

For a recessive variation to be expressed it would be necessary for two of the recessive gametes to unite in conjugation, and this could not occur in the second generation, in any sexually differentiated species. To insure the expression of the recessive variations which the cross may have induced, it would be necessary to self-fertilize the perjugate generation. The usual custom of breeders of plants propagated by buds or cuttings is to make their selections and close the hybridizing experiment with the second generation. The expression of the recessive variations is thus entirely prevented, for no third generation is produced. With seed-propagated species the diversity caused by the recurrence of reversions continue for numerous generations, until the desired characters have been "fixed" by a course of persistent selection. [223]

The degenerative nature of many valuable variations of domesticated plants and animals is well known, and also their generally recessive status. It might, therefore, be worth while to take into account in experiments with plants this means of bringing recessive characters into expression by the self-fertilization of the perjugate individuals of crosses from which desirable variations are desired.

1Cook, O. F., 1904. The Vegetative Vigor of Hybrids and Mutations. Proc. Biological Society, Washington, 17: 83

Numerous analogies between hybrids and mutative new variations have been described and discussed in other places, but the present indications of the possible identity of the phenomena had not been perceived.1 Mutations not only behave like hybrids, but they may actually be hybrids, the mutative variation having taken place in only one of the gamete partners of the conjugation which has given rise to the variant organism. The partial or complete sterility of many mutations may be quite the same phenomenon as the sterility of many hybrids, and may sometimes be due to the incompatibility of very diverse gametes.

Variations which come true to seed from the very first may still behave as recessives when crossed, since recessive variations do not gain expression except when the gametes are of the same kind. Other variations, though they may not come true, may be dominant, or even prepotent. Thus narrow-leaved mutations of cotton (okra-cottons) which do not come true from seed may show, nevertheless, a definite and uniform prepotency when crossed outside the parent type.

The dominant and recessive relations are not, as some have supposed, evidences of incompatibility or failure of the gametes to accomplish normal and complete conjugation. Indeed, the present perceptions enable a definite distinction to be drawn between terms which are not truly synonymous, though often so used. Dominant and its alternative, recessive, have reference to the relations of gametes as shown in the characters of conjugate organisms. The character which secures expression is called dominant, while that which remains abeyant or is not expressed is called recessive. [224]

Prepotency, on the other hand, has reference to the results of completed conjugation, as represented in gametes. Characters which secure a stronger or more general expression in the new gametes are called prepotent, those which tend to decline or to disappear may be called subpotent. Characters which have equal polarity, so that they each secure an unmodified representation in behalf of the gametes may be called equipotent. If these polarities show equal strength in the conjugate relation, so that half of the conjugate organisms also represent each of the divergent characters, we would have a complete case of equipolar inheritance.

Characters which are transmitted without expression-polarity, either in conjugates or in gametes, are called latent. An abeyant or recessive character fails of expression only when joined with a dominant gamete polarized toward the expression of another character, but a latent character does not come to expression even when two gametes are of the same kind, though it may still reappear many generations afterward as in atavism and reversion. Latency might be described as character-abeyance in gametes. As a character represented in one of a pair of conjugate gametes may fail to secure expression in the conjugate organism, so the power of bringing a certain character to expression may remain latent in a gamete, as long as the polarity of expression continues to be directed toward another character. Just as the transmission of a character through a conjugate organism has no necessary relation to its expression, so the transmission of a character by a gamete has no necessary connection with a polarity or adjustment toward the expression of the character.

PROBLEMS OF TRANSMISSION AND EXPRESSION.

A curiously vivid analogy between the expression of characters and the expression of ideas is worthy of note at this point. A suggestive expression often enforces the complete logical elaboration of an idea, even in the face of contradictory facts. It is commonly supposed that the phenomena of alternative expression of divergent parental characters remained entirely unconsidered in the interval between the publication of Mendel's [225] work in 1866 and their rediscovery by De Vries and Correns in 1900. But in 1889 Galton published his book on "Natural Inheritance" including a chapter on eye-color in the human species, in which the following statement occurs:

1Galton, Francis, 1889. Natural Inheritance, pp. 138-139.

"Stature and Eye-colour are not only different as qualities, but they are more contrasted in hereditary behaviour than perhaps any other common qualities. Parents of different Statures usually transmit a blended heritage to their children, but parents of different Eye-colours usually transmit an alternative heritage." 1

It is evident enough from the context that the author used these words merely to describe the facts which he had ascertained regarding differences of eye-color among members of the same families, and not with the intention of proposing a new theory of the mechanism of descent. Nevertheless, it launched an expression which could be converted readily, and even unconsciously, into a theory of alternative transmission of characters, and thus undoubtedly did much to lend popularity to the "rediscovery of the Mendelian principles." Galton appears to have had no intention of suggesting that the transmission of one character inhibits the transmission of another character. He predicated transmission only from the standpoint of the parents, and as shown in the characters of the immediate offspring, and fully understood that these offspring would be able to transmit characters which their own bodies did not express. But the Mendelian theories have discarded such reservations and distinctions, and have gone over bodily to the assumption of alternative inheritance in the most complete sense of the words. They predicate the existence of character-unit particles and of germ-cells which receive and transmit particles representing only one of a pair of contrasted parental characters. But no such method of inheritance has been proven, and no such assumptions are required, or even permitted, by the facts of Mendelism.

Inheritance has been used unconsciously in two different senses, to cover and confuse two distinct phenomena, the transmission of characters and the expression of characters. [226] Regarding the real natures or details of the processes of transmission and expression, we know as little as we did of inheritance, but it is a measure of progress to learn that there are two factors instead of one. Transmission inheritance is not reciprocal or alternative like expression inheritance.

The relations which may be supposed to determine the characters of conjugate organisms are obviously very different from those which determine the characters of the gametes. The characters of a conjugate and those of the gametes it produces may both be results of relations assumed by the same original parent gametes, but the relations in the two cases are entirely distinct, as the results themselves testify. The conjugate results are large, highly complex organisms, while the exjugate results are minute, simple gametes.

The older interpretation of the facts of descent led us to reckon the gametes as products of the bodies of the plants or animals, minute propagating particles qualified to faithfully reproduce the same kind of a body, unless this were prevented by external circumstances. The present interpretation recognizes that new gametes are not so much the offspring of the conjugate body as of the gametes from which the body was formed. The gamete-producing cells are often related to the cells of the body of the " parent organism" only as they represent common offspring of the original gamete-pair, and even then there is the very serious difference that the body-cells, being devoted to vegetative purposes only, do not complete conjugation at all. The gamete cells do not correspond directly to any of the body cells because they have passed through mitapsis and belong already to a new generation, a new section of the network of descent.

Conjugate organisms are built up at the knots or junctions of the networks of descent. In the lower groups exjugate cells build up organisms on the threads of the network, between the knots. Conjugate organisms can be thought of as thickening the knots of the network, exjugate organism as thickening the threads. But in both cases vegetative specialization may bring a loss of reproductive powers and exclude the specialized cells from any direct representation in the subsequent network of [227] descent of the species. In neither series was there any primary or essential difference between the cells as body-cells or as germ-cells, but vegetative specialization has repeatedly, and in many independent groups, rendered a part of the cells sterile neuters whose function it is to nourish and protect the reproductive cells and the helpless offspring, instead of leaving any direct progeny of their own. Organisms are, after all, merely social organizations of cells, and there are different kinds of cell-organizations, just as there are different kinds of organizations among men and other social animals.

Generally speaking, the cells which compose the bodies of the higher organisms do not leave any descendants to perpetuate their characters. Except for the fact that the body-cells and the sex-cells are both descendants of the original gametes, there is no relation between them, just as the highly specialized castes of ants and termites are sterile, and are propagated anew in each generation from normal sexual insects, and not from any ancestors which possessed these specialized caste characters. Thus if the character-unit theory were to be logically elaborated the germ-cells would need to be thought of as containing, not merely units to represent the formal characters in which species and varieties differ, but units governing the characters and adjustments of every separate cell of the vegetative body. All the different kinds of cells which compose the body of a higher plant or animal represent as many cell castes, and there would need to be character-unit determinants to govern the formation and arrangement of these, quite as much as for the different kinds of organisms which the species may contain.

The germ-cells continue to transmit characters which can be brought into expression only through the medium of vegetative cells which do not take any part in the reproductive process, any more than soldiers, workers and nurses of a colony of ants or termites. The complexities of this problem of transmission transcend the powers of the human imagination, but there do appear to be any corresponding limitations in practice, for even characters which are not expressed may still be carried along, generation after generation. The greater the number of details the more effective and reliable, is this hereditary memory, if [228] only it be allowed a reasonable diversity of alternatives of expression. But if too closely restricted to the monotonous repetition of the same story in each generation it wanders from the hereditary pathway, though it may be recalled to the path by reminiscent contact with another wanderer from the same group.

It is very natural to infer that the characters of conjugate organisms are contained in the germ-cells, since the organisms are known to develop from germ-cells. But as yet we do not know how the characters are contained in the germ-cells. We have no concrete knowledge whatever of the nature of characters during the interval of transmission or pre-expressional existence. But we do know that the conditions of characterexpression are essentially different from those of charactertransmission. Dominance and potency are phenomena of expression, and afford no hint or basis of inference regarding the nature of the mechanism of transmission.

There is no longer any reason to believe in any general or fundamental difference between somatic and germinal tissues as alleged by Weismann. The bodies of the higher plants and animals simply represent different degrees of vegetative specialization of tissues once wholly reproductive. A distinction is to be drawn, however, between the transmission of characters by simple or exjugate cells and the expression of characters by conjugate or double-celled structures. Mendelism and other forms of polar descent give us no new knowledge regarding the structure of gametes. They concern only the expression relations which these gametes assume in the conjugate condition.

1Cook, O. F. Transmission Inheritance distinct from Expression Inheritance, Science, N. S., 25: 911. 1907.

Conjugate organisms are not merely resultants of the characters which the gametes were adjusted to bring into expression; they may show a reciprocal expression of the gamete polarities, or may have characters which neither of the gametes would have expressed if it had a partner of its own kind. If two magnets or magnetic compasses are brought together their directions will change, and if one overpowers the other both may take the same direction. Transmission is represented by the dial of the compass, expression by the needle. One organism [229] can express only one set of characters, though it may transmit to posterity an unlimited number of varying equipments.1

Conjugate organisms may give no hint of the presence of a character which the gametes are able, nevertheless, to transmit unimpaired through long series of generations. The behavior of two gametes in the conjugate phase is no criterion of the results of the conjugation, as represented by the new gametes of the next generation. The inheritance of expression-polarities is entirely distinct from transmission-inheritance. Expression-polarities may be mutually exclusive, so that polarity of one character implies the abeyance or non-polarity of expression of alternative characters, but there are no indications of such exclusiveness in the transmission of characters.

If transmission is to be thought of as accomplished by character-unit particles the indications are that the tendency is to accumulate as great a diversity of particles as possible, instead of excluding any because of incompatibility. Transmission is inclusive, not exclusive. The continued transmission of a character can not be ascertained, of course, unless it be brought into expression, but the possibilities of transmission without expression are now amply demonstrated. To assume that a character has failed of transmission because it remains unexpressed, is entirely gratuitous.

It remains true, of course, that the characters of the gametes can only become known through their expression in conjugate organisms, but this difficulty of interpretation does not absolve us from the recognition of two distinct factors or problems of inheritance. The problem of potency relates to the polarity or adjustment of unconjugated gametes toward or away from the expression of certain characters. The problem of dominance has reference to the adjustment of gametes in the conjugate relation. We have yet to learn why the polarity of one gamete overpowers the other and completely controls the development of the conjugate body, and how the abeyant polarity of the other gamete is able to reassert itself at the period of mitapsis and secure equal representation in the new generation of gametes. [230]

As bearing upon the solution of these problems of descent, the facts of Mendelism take on very great interest and importance. As yet they have served only to obscure the issues because they have been interpreted as phenomena of incompatibility or "disjunction." The facts of Mendelism take on far greater significance as soon as we cease to compare them to the abnormal results of incompatible hybrids, and perceive the far closer analogy with the alternative or polarized expression of sexual differences.

RELATION OF MENDELISM TO NORMAL DIVERSITY.

Only the gratuitous assumption of ideal uniformity stands in the way of a recognition of the fact that the normal tendency of evolution is not toward "identity of form and structure" inside specific lines, but toward the development of diversity among the members of species. Diversity between the members of the same species is often accentuated by differences of environment, but there are many differences, both individual and sexual which are independent of the environment, and which continue to appear, without regard to uniformity or diversity of environment. Mendelism, like sexuality, is a means of preserving a desirable diversity of descent inside the species. It is not a phenomenon of incompatibility, but rather of avidity, tending to strengthen conjugation by accentuating the unlikeness of the gametes.

Sexuality and Mendelism are both to be reckoned as phenomena of normal diversity of descent, or heterism. Sexuality is the more specialized form of heterism, because only sexually diverse gametes are united in conjugation, whereas like gametes can still conjugate in Mendelian descent, and thus miss the effect of the desirable diversity which in sex-inheritance is fully preserved. But sexuality and Mendelism are very closely alike in methods of inheritance. Each organism and each gamete represents the expression of one of a pair of contrasted characters, but not both. There is no tendency to combine the characters into an intermediate average. The series or scale of possible variations is represented by the two extremities, while the intermediate points come only rarely and abnormally into expression. [231]

As Mendelism may be considered a somewhat weaker form of sex-inheritance, so there are weak forms of Mendelism which serve to connect the definitely polarized forms of alternative descent with the other still less specialized methods of inheritance. Instead of the typical dominance of one parental character in the conjugate phase, there may be an intermediate or compromise, but this need not interfere with alternative descent in the gametes of the next generation. Or it may happen that the conjugate organism expresses a character shown by neither of the gametes when mated with their own kind. Definite alternative descent in the gamete progeny also grades off into intermediate expressions of the two characters, or into the expression of a whole series of intergraded intermediates, or some ancestral characteristic may reappear, though long dormant in both the parent varieties.

The extent to which Mendelism and other less accentuated forms of polarized inheritance exist in natural species is a subject which has received very little investigation thus far. The reasonable expectation would be that species will be found to differ very greatly in these respects, just as they show very great diversity of degrees of accentuation of alternative sexual characters. Narrow breeding, which tends to the accentuation of variations, probably tends also to the accentuation of polarity of expression. This agrees with the fact that Mendelism is so frequent a phenomenon in crosses of narrow-bred varieties, but so rare in crosses of undomesticated species or subspecies. A course of narrow breeding, such as experimenters in Mendelism prescribe as a means of ascertaining the " purity" of the stock by securing uniformity might well bring the plants or animals into condition for an exhibition of strict Mendelism, instead of showing somewhat less specialized forms of polar inheritance. Mutative variations which represent reactions from abnormal uniformity may be ready from the first to make the most of the new diversity by means of polar inheritance. Nor is it without significance for general evolutionary purposes that the natural selections of a dominant variation would not tend toward uniformity, but preserve and increase the heterism of the species. It is only by highly conscious selection of the "pure dominants" [232] that we can avoid the natural result of perpetual hybrid diversity.

EVOLUTIONARY IMPLICATIONS OF MENDELISM.

The general evolutionary significance of Mendelism as a method of descent does not lie in the assumptions of characterunits and pure germ-cells, but in the facts of polarized inheritance, which show how new variations may be preserved without isolation and become the means of evolutionary progress of species. There are, as we have seen, all degrees of accentuation of the phenomena of dominance and polarity. The narrow-bred varieties and mutations show them in the most extreme form, but sex-inheritance is very widespread, and the other less accentuated forms of specialized diversity will probably be found to be quite as generally distributed.

1"Such facts as I have cited above [melanic sports, etc.] . . . seem to lead me to the conclusion that some new characters may arise in nature suddenly . . . and persist as specific characteristics. The Mutation Theory in Animal Evolution, Science, N. S., 24: 558.

Hitherto we have been expecting species to be uniform, and have considered the presence of definite diversities in the same species as something exceptional and peculiar. The possibility that new variations may be preserved without isolation and thus contribute to the evolutionary progress of species, was recognized some months ago by Professor Davenport,1 but was then associated with the mutation theory of De Vries, with which it has no true relation. Professor De Vries is particularly explicit in denying any such process of gradual transformation, his theory being that the fact of mutation by itself constitutes the evolution of a new species, which remains unchanged until another period of mutation ensues.

"There is neither a gradual modification nor a common change of all the individuals. On the contrary, the main group remains wholly unaffected by the production of new species. After eighteen years it is absolutely the same as at the beginning, and even the same as is found elsewhere in localities where no mutability has been observed. It neither disappears nor dies out, nor is it ever diminished or changed in the slightest degree. . . [233]

1De Vries, H., 1905. The Evidence of Evolution. Smithsonian Report for 1904, p. 396.
     An even more detailed and explicit denial of any gradual transformation of species is contained in a paper by Professor De Vries in the Monist for January, 1907.
     "The conception of mutations agrees with the old view of the constancy of species. This theory assumes that each species has its birth, its life-time and its death even as an individual, and that throughout its life it remains one and the same. Thus it is only natural that wild species are almost always observed to be constant, since by a mutation they do not change themselves but simply produce a new type. This is allied to its ancestor as a branch is to a tree, the stem continuing its own growth, no matter how many branches it produces. Just so a species may produce quite a number of new forms without being changed itself in the least thereby. Among palaeontologists Scott has given forth this same view. According to his conception, species are derived from one another by small shocks. Each shock caused the old limits to be transgressed; but, after it, the new species remained unchanged until, perhaps after centuries, a new shock made it transgress its new limits. Each single type (be it species, subspecies or variety) is thus wholly constant from its first appearance and until the time it disappears, either after, or without, the production of daughter-species."

"Thence we must conclude that new species are produced sideways by other forms, and that this change only affects the product, and not the producer." 1

But if a new variation can arise suddenly and persist without isolation as a character of its species, the producer is changed as well as the product. Indeed, the new variation does not gain or maintain its existence as a separate product, but merely increases the diversity of the parent group.

Spontaneous transformation of species by the prepotency of new variations, under natural conditions of wide distribution and free interbreeding, is the method of developmental progress outlined in what has been called the kinetic theory of evolution. Previous theories had not recognized the possibilities of direct and spontaneous accumulation of new variations. Usually they had sought the agencies of evolution in the isolation of variations by selection or in other forms of environmental interference with an ideal condition of uniformity and stability. The mutation theory of De Vries reasserts this uniformity and stability as the normal condition of species, and seeks to explain evolutionary progress by sudden leaps from one species into another, and not by any gradual change, through prepotency [234] or otherwise. The degenerate nature of mutative variations of narrow-bred groups and the fact that they usually fade out entirely when brought into contact with the unrestricted descent of the wild type, are further practical reasons for the complete divergence between the mutation theory and that of kinetic prepotency. But even mutations are usually prepotent among their own equally narrow-bred relatives, and thus afford at least an illustration of the power of genetic variations to maintain themselves and gradually transform broad-bred species in nature.

Instead of compelling us to abandon the conception of evolution as a process of gradual change in the characters of a species, mutations and Mendelism in reality strengthen and reinforce this older idea of continuity. They show that progress does not depend upon environmental causes, either direct or indirect. The studies of mutations have made it certain that diverse variations occur, even under the same condition, and Mendelism demonstrates that such variations can also be preserved without isolation. Both Mendelism and mutation, in their typical forms, represent abnormal conditions of restricted descent, but they may still have great evolutionary interest if we accept them as evidence, or even as suggestions, that less accentuated forms of the same phenomena of inheritance are concealed among the more nearly indiscriminate diversities shown by the members of wild species in nature.

Evolution is a process of change in the characters of species, and not merely the origination of new species, either by mutative leaps or by Mendelian combinations of characters. Evolution is accomplished by the gradual, constructive synthesis of variations. Darwin may have been in error in supposing that evolution is actuated by selective agencies of the environment, or that the steps by which it is accomplished are infinitesimally gradual, but it has yet to be shown that he was in error regarding the general continuity of the motion as indicated by the continuity of the differences inside the species and also of the species themselves when viewed as members of genera, or of genera when grouped in families.

This fact of continuity convinced Darwin of evolution, and not natural selection and other theories by which he attempted [235] to explain evolution. It is not too much to say that all who have faced this question in the light of broad and thorough knowledge of species in nature, agree with Darwin regarding the generally continuous and the gradual character of evolution, no matter how much they may differ with him regarding methods and causes. It is at this point that taxonomic studies have a very practical bearing on evolution, and become an almost indispensable training and qualification for evolutionary judgment. The difference between mutations and natural species is not that they are less different than species, as Professor De Vries and his followers had feared, but that they are more different than species, and different in a different way. They lack the indiscriminate diversity, the protean flexibility of form which often renders it extremely difficult to find any completely diagnostic differences between species which are nevertheless truly distinct in nature. Mutations of the same species, on the contrary, commonly differ in numerous, definite and much more constant characteristics. The two phenomena, species and mutations, belong to essentially distinct categories, as unlike as crowds and drill-corps.

Mutation is not a period, but a condition. A narrow-bred group in the mutative condition all wears, as it were, the same uniform. A new mutation constitutes a definite change in this uniform. It would be very much easier to describe and classify the military uniforms of the various European nations than to find any equally diagnostic characters for the nations themselves. The nations are much more seriously and essentially different, of course, than their military uniforms, but the differences are of a different sort, and not to be stated in the same terms. Species have more essential evolutionary differences than mutations, though mutations are at the same time more definitely different. The phenomena are, in short, of different categories and incommensurable. Discussions can lead to no conclusions until these fundamental diversities of facts and concepts are appreciated.

Professor Davenport has assured us that "The question of the origin of species is that of the origin of specific characteristics or differential marks." But, unfortunately, there is no evolutionary [236] connection between the origination of species and the origination of marks or characters by which species may, or may not, be distinguished. Species may be quite distinct and may yet offer precious few diagnostic marks. On the other hand, one and the same species may manifest so much diversity of characters that numerous species could be distinguished were it not for the intergrading forms.

Though many have taken it for granted, nobody has shown that there is any relation between the origination of characters and the origination of species. Species do not have to be originated to originate characters, nor characters to originate species. Nor has it ever been shown that the origination of species has any causal relation to evolution. Species originate by isolation, that is, by the subdivision of older species. The groups attain differential characters after isolation, in the same way that characters would have continued to be attained if the original species had not been subdivided. The origination of the species does not cause the origination of characters, nor does the origination of characters cause the origination of species.

Nor is there any necessary relation between the discontinuity of characters which distinguish species and the discontinuity of character-diversity among the members of the same species. Species might differentiate in the most gradual manner, but having once differentiated their characters would be discontinuous, for this is the taxonomic criterion of specific differentiation. Nevertheless, discontinuous variations might spread gradually through a species and transform it. The discontinuity of the variations would not break the continuity of the evolution of the species. Discontinuous variations make the mutation theory abstractly conceivable, but they do not afford any evidence that it is the true method of evolution.

The origin of species is not evolution; evolution is the changing of characters in species. Not all changes of characters are evolution, but only changes of characters in species, or changes of the same nature as changes in species. Mutative changes of characters in narrow-bred domesticated varieties, and the strict Mendelian polarity of expression of such mutative differences, are not phenomena of evolution, but of degeneration. [237] They do not teach us the principles or methods of normal constructive evolution, except as we learn to interpret them conversely. The discontinuity of variation shown by the origination of these characters of degenerating domesticated varieties finds few complete parallels in nature. Moreover, these mutative characters, instead of being nicely preserved in nature, are wiped out by contact with the wild stock of the species, or even by free interbreeding among the domesticated varieties themselves. The prepotency of recent mutations corresponds to the prepotency of genetic variations in nature, but the evolutionary status of the mutations is very different from that of the genetic variations.

Discontinuous variations do not prove discontinuous evolution, but makes continuous evolution more practicable and probable. It is true that Darwin and many other believers in selection as the motive power of evolution have dwelt upon the efficacy of accumulating even infinitesimal differences, but there can be no objection to the accumulation of variations of more than infinitesimal size if the facts seem to justify such a view. The fundamental issue is between evolution, or gradual changes in species (whether by smaller or larger steps), and saltation, or sudden jumps from one species into another. De Vries has even been careful to specify that the jumps shall be sideways, on the ground that species once originated by mutations do not change or make further advances, all progress being confined to lateral displacements of mutative variation.

Mutations and Mendelism are very interesting phenomena, not only for breeders of domestic plants and animals, who have long known them under other names, but in adding, if somewhat indirectly, to the evidence that new variations are prepotent, and that evolution may be accomplished by the spontaneous accumulation of appreciable genetic variations, instead of depending upon the selective integration of infinitesimal effects of environmental influences. When this distinction between discontinuous variation and discontinuous evolution is once appreciated, it will become apparent that the mutation theory and the Mendelian "laws" which have been enlisted in its support are assumptions which the facts do not warrant. [238] But no cause is lost without contributing something to the progress of science.

The attempted resuscitation of the saltatory theory and the Mendelian "laws" has resulted in a new and more thorough survey of this part of the field of variation, the same ground which had been canvassed by Darwin's contemporaries and critics, Mivart and Galton. But the general implications of the facts do not appear to be very seriously altered. Darwin considered the possibilities of saltatory evolution, but dismissed the idea as of incidental significance only. Galton estimated the value of discontinuous variation much higher than Darwin, but was not misled into the hazardous conjecture that saltatory variations cause evolution by originating new species.

Darwin did not imagine that he had solved the problem of evolution in the sense of having definitely ascertained the nature of the motion or the causes of the variations by which it goes forward. He was solicitous rather to make the doctrine appear sufficiently reasonable to secure the serious attention of the scientific world. He did not have complete confidence in the foundations of his own colossal idea, and labored hard on natural selection and other unnecessary scaffolding to support it. Later criticism is removing these superfluous theoretical timbers, but the primary conception of evolution as a process of gradual change in the characters of species, not only remains intact, but stands forth with increased assurance of truth and permanence.

CONCLUSIONS.

A typical experiment in Mendelism, instead of involving two successive crosses or conjugations of gametes, includes only one such cross. The so-called first generation is built up by the vegetative subdivisions of the gamete parents, before conjugation is completed. The so-called second generation represents the first organisms produced after the completion of the conjugation of the gamete parents.

The difference of proportion between the two generations in the expression of divergent parental characters is to be explained by the peculiar methods of reproduction followed by the higher plants and animals, and by the fact of dominance or expression-polarity, [239] instead of by the Mendelian theory of alternative inheritance of character-unit particles.

Mutations do not differ from Mendelian hybrids in any essential respect, either at the time of their first appearance or in later combinations. The preservation of the new character by definitely reciprocal inheritance of expression-polarities is favored by the same conditions of restricted descent which induce the mutative variations.

The analogy of the Mendelian phenomena, applied to variations induced by crossing, shows that new characters which come to expression in the first or conjugate generation are not likely to be permanent. Dominant variations can gain expression in the second or perjugate generation, but recessive variations are not shown before the third generation, and may not be brought into expression until still later generations, unless the first perjugate generation is self-fertilized. As many variations of economic value behave as recessives, this fact is of practical significance in breeding experiments.

Two distinct phenomena have been confused in the Mendelian conception of inheritance, transmission and expression. The failure of a character to secure expression does not indicate that it has failed of transmission. Polarity, or reciprocal expression inheritance of divergent parental characters, explains the phenomena of Mendelism and related forms of descent without requiring the assumption of pure germ-cells or of character-unit particles.

There is no evidence that normal transmission-inheritance is a phenomenon involving the alternative admission or exclusion of character-units, or that characters are transmitted as particles or mechanisms. The process of transmission is independent and separate from the process of expression, which often yields polar or reciprocal results. This reciprocal polarity of expression-inheritance shows how new characters can be preserved and thus contribute to the normal diversity of a species or gradually transform it. Evolutionary advance can thus take place without selective or geographical isolation. The general evolutionary significance of Mendelism lies in its testimony to this fact, and not in the theories of inheritance by character-units and pure germ-cells. [240]

The accumulation of variations without isolation agrees with the other evidences of evolution by the continuity and gradual transformation of the characters of species. It shows that the phenomena of mutations and of Mendelism do not justify the theory of saltatory or discontinuous transformation by leaps from one species into another. The latter assumption requires the formation of a new species to preserve each new character, while the former permits simultaneous progress in the constructive combinations of many characters inside the network of descent of a single, undivided species.