Proceedings of the Washington Academy of Sciences
Vol. VIII, pp. 197-403. February 13, 1907.

ASPECTS OF KINETIC EVOLUTION
By O. F. COOK

6. THE CONSTITUTION OF SPECIES.

Astronomy is reckoned as queen among the sciences because it has demonstrated that definite and orderly relations exist [357] amidst the apparently hopeless disorder of the stars. The ancients, grouped the stars into constellations, but modern science shows us systems ruled by laws of mathematical precision.

Biology has remained longer in the constellation stage. Species are still discussed, even by evolutionists, as though they were mere chance aggregates of organisms, at once too familiar and too diverse to be formally defined.

It may well be that no coherent definition can be made for species as mere aggregations or constellations of organisms; the idea itself is vague and essentially unscientific. The primary error was that of treating the species as a morphological group, whereas the true evolutionary species is a physiological system. Like a stellar system, it may contain a large number of different individual members, and even different kinds of members. The unity of the species does not depend upon the organisms being all alike. It is necessary only that they remain within range of mutual influence through interbreeding, which is the biological analogue of gravitation.

A species, that is, a normal, natural, evolutionary species, is a large, coherent group of freely interbreeding organisms. But with species, as with stars, all systems are not alike. There are suns, satellites, planets, asteroids, nebulae, variable stars, doubles and comets, in vast diversity of sizes and combinations.

In biology, as in astronomy, the most familiar things have proved very deceptive. The sun, moon and stars appear alike to revolve around the earth, from east to west. It was at first an extremely heterodox idea that the earth revolves around the sun. Moreover, neither of the apparent motions gave any intimation of the third order of motion, that of the system as a whole. In a similar way we have taken it for granted that the evolution of species could be explained by the motions we have been able to detect among our domesticated plants and animals. We are now learning that these types of life are not reliable examples of evolutionary systems, that their motions are often retrograde or degenerative instead of progressive and constructive. Nor are abnormal evolutionary conditions entirely confined [358] to domesticated organisms. Among the millions of biological systems many have wandered from the path of progressive evolution and are on the way to extinction. As with the motions of the heavenly bodies, nature herself has deceived us, or rather she has given us new riddles to read.

The motion of species is not like that of the stars, in simple geometrical figures. The evolutionary progress of species is accomplished by the weaving of an intricate fabric of lines of descent through the free interbreeding of the component organisms. The simple, normal and typical constitution of a species may be thought of as a huge but simple network of uniform texture. All the organisms are diverse, but the diversity is merely individual and indiscriminate, so that the network has a uniform texture.

THE SPECIFIC CONSTITUTION OF LIVING MATTER.

Inorganic matter exists in a variety of conditions or physical states, gaseous, liquid, colloidal, crystalline, granular or amorphous. The properties of matter depend upon these conditions or states quite as much or more than upon the chemical composition or ultimate nature of the materials of which they are composed. There are laws of gases, liquids and crystals because the different substances behave very much alike in the same physical states. Indeed, the same physical states of different substances are generally very much more alike than the different physical states of the same substance.

In a similar manner the qualities of living matter are to be associated and described with reference to its various states or conditions. Chemically it is a mixture of water and of small quantities of numerous substances and compounds. Physically it is a jelly or colloid. Biologically it manifests such powers as growth, digestion, motion and reproduction. Morphologically it consists of cells or protoplasmic units with a more or less differentiated internal structure, and a power to combine or associate into organisms.

For evolutionary purposes the chemical, physical and organic points of view do not suffice. It is necessary to recognize that living matter shows still another unique property, another kind [359] of constitution, the specific. A species is quite as concrete a phenomenon as a crystal. Both are collections or aggregates of smaller units, and the units have in both cases definite and necessary relations to each other on which the existence and further development of the crystal or the species depend.

It is true that many valuable evolutionary data have been secured from captive or domesticated plants and animals, but the results of this whole class of experiments indicate very definitely that evolutionary phenomena under these conditions are degenerative and not constructive. We are driven back to study the constitution of species in nature, to gain a clear understanding of the organic conditions which make possible genuine developmental progress, a true organic evolution.

No theory or evolutionary interpretation can hope for permanence which leaves out of account this primary fact that organisms normally exist in large groups of freely interbreeding individuals, the groups commonly called species. Domesticated varieties of plants exist without interbreeding and a few species in nature are supposed to propagate only by vegetative methods, by parthenogenesis or by self-fertilization, but no genus, family or order appears ever to have developed without the association of the individual organisms into interbreeding groups or species. The only exceptions, if any, are among the bacteria and other extremely simple forms of life which have failed to develop either a specialized nuclear structure in the cells themselves or an ability to associate and differentiate to form compound cellular organisms.

The reigning popularity of laboratory methods of research may permit small welcome for the suggestion of a method of evolution which requires the extensive equipment of nature and can not be demonstrated in cages or gardens, except by negative results, like those already well known. This disappointment need not continue, however, any longer than may be necessary to perceive that while experiments with domesticated species lose in apparent general significance under the new interpretation, they gain greatly in definiteness. If they do not show us how the fabric of normal evolutionary descent is woven, they at least teach us how it may be unravelled. This knowledge is of [360] great value, not only to help breeders in the making of useful domestic types, but also to students of the general problem.

Domesticated plants and animals furnished the most effective arguments for the theory of organic evolution, for although the ancestral wild types of many cultural species are still unknown, and may have become extinct, there can be no doubt that thousands of their varieties have originated in domestication, and that similar varieties continue to arise under the eyes of the cultivator and breeder. Domesticated plants and animals have supplied, too, nearly all the materials for evolutionary experiments, and it is also with them that evolutionary theories must find, ultimately, their practical application.

A false or inadequate theory, though avowedly based on studies of domesticated species, may be quite as injurious to agricultural progress as another drawn from facts ascertained from useless wild species. Any idea worthy of general credence will bear the test of application to both classes of phenomena. A theory is merely a way of thinking about things, and is useful if it enables us to see, or even to suspect, causal connection between facts previously unassociated. One theory is better than another if it brings important facts into relation, and is considered established as a law or doctrine when it accomodates all the facts of the field it was designed to cover. The distinction frequently attempted between "theoretical" and "practical" investigations of evolution is quite fictitious, as in other fields of knowledge.

By a curious perversity of language the designation "pure science" is often applied to accumulations of knowledge not yet refined enough to be useful for practical purposes. The talk of discrepancies between theory and practice amounts to a kind of fiction, a euphemistic way of saying that an inadequate theory may not be wholly worthless as an indication of relations not yet adequately understood.

For establishing the general fact of variation and thus demonstrating the possibility of an evolutionary and continuous creation, the variations which have arisen under domestication afforded the most pertinent and convincing testimony. No biologist now doubts that evolution has taken place and still [361] continues, but there is, nevertheless, a very wide and very practical divergence of opinion regarding the nature and causes of the evolutionary process. In the study of this question it becomes important to realize that the evolutionary condition of cultural species differs from that of wild types because of the much greater degree of inbreeding to which the former are commonly subjected.

The constitution of species has a practical bearing upon agriculture, not because the domesticated plants and animals have not been studied from an evolutionary standpoint, but for the very opposite reason, that they have been considered too exclusively, so that the important differences existing between them and wild species have been overlooked. Ideas drawn from domesticated varieties have been projected into nature at large, and this made it only the more impossible to appreciate the fact that grave differences exist between wild and domesticated groups of organisms.

Evolutionary science has gained much from the study of domesticated plants and animals, and may gain still more in the future. The objection is only to the use of such studies and results as an exclusive basis of interpretation of the facts of nature. All that happens in domestication may also happen in nature, for domestication is, after all, only a department of nature. It does not follow, however, that nature is fully mirrored in domestication; the mirror is too small. It shows us only the conditions in which constructive evolution does not take place, even in nature.

The recognition of the fact that evolution is a phenomenon depending upon the specific constitution of living matter has been delayed, no doubt, by the difficulties which have been encountered in the field of taxonomy. In the recent decades naturalists have faltered in the task of nomenclature set by Linnaeus. To merely describe and give names to the millions of evolutionary unit groups of organisms which occupy the surface of our planet is a work much too vast for the present resources of science. The temptation of weariness has been to shorten it by passing over the apparently useless redundancy of slightly different groups, or by declaring that all is vanity of [362] merely abstract conception, that species do not exist, and can not be defined.¹

This method of reasoning was very popular in mediaeval times and was then, reduced to the neatly pious formula: "Credo quia absurdum."

Those who have not persevered beyond this stage of skepticism and satisfied themselves of the existence of species in nature, can have little use for an interpretation based on the recognition of species as definite entities, consisting not merely of aggregates of individual organisms, but also of fabrics of interwoven lines of descent.

The difficulty in defining species is the lack of clear perceptions, not only of the nature and constitution of species, but also of the fact that several diverse types of phenomena are being covered by the word. Under such circumstances a general definition of species, however framed, could afford only a fictitious unification of expression, the ideas and implications covered by the term remaining essentially diverse and often quite contradictory. This confusion affords, however, no justification of a failure to use the term in one or another of the explicit senses of which it is capable, nor of a refusal to define the usage of the term in any particular connection.

The difficulty of defining the term species has arisen mostly from the fact that the phenomenon is a physiological one, whereas the general supposition has been that it is morphological. The idea that species are "founded on identity of form and structure," as the dictionaries say, is still widely prevalent, and is one of the tenets of evolutionary belief upon which Professor De Vries especially insists.

The impracticability of a morphological definition of species arises from the fact that it is impossible to set definite limits to the extent of the variability or diversity which is to be permitted in the species. Identity of form and structure makes an excellent definition; the objection to it is that no such species seem to exist in nature, or as Professor De Vries says, "* * * purely uniform species seem to be relatively rare." ² In some groups [363] all the members of the species are closely similar, but in others they may be extremely unlike, as when the specializations of sex and polymorphism have been developed. There is no need, however, that we define species as a morphological term, since species are not caused nor constituted by the likeness or unlikeness of the component organisms. Indeed, it is unlikeness rather than likeness that conduces to the prosperity of the species.

The species in nature is constituted by the fact that the component individuals breed together. For evolutionary purposes a species is a group of interbreeding organisms; nothing more is required, nothing less will suffice. Species are units of organic evolution; organisms continue to exist and to make evolutionary progress only in large groups of freely interbreeding individuals. Groups of organisms which do not interbreed are no longer species; they no longer have the typical and essential evolutionary constitution of living matter.

Whether the individuals are alike or different does not in the least affect the specific unity of a group if the organisms are associated in nature on a basis of free interbreeding. If the groups have ceased to interbreed, whether by reason of geographical barriers, or of structural or instinctive incompatibility, they are no longer a unit of evolution, no matter how close the external similarity may appear.

Natural species are not the only groups of organisms to which the name is applied, but since all other so-called species are mere parts or fragments of natural species, a recognition of natural species must precede a true appreciation of the more or less artificial subdivisions of species.

These evolutionary facts are quite independent of the old taxonomic idea that the limits of species could be determined by ascertaining whether the animals or plants can interbreed. The evolutionary question is whether they do interbreed. Groups able to interbreed perfectly will still follow divergent courses of evolution, if kept apart. On the other hand, the failure of the extreme members of the same species to interbreed would not destroy the unity and coherence of the group.¹ [364]

The exclusion of the domesticated plants and animals from use as illustrations of the true methods of evolution may appear to withdraw the subject from the consideration of all who do not have intimate acquaintance with species in nature. There remains, however, an excellent and very familiar example of evolutionary conditions, that of man himself. The genus Homo has achieved in a relatively brief period a wide divergence from its simian relatives. This progress in development has been coincident with the achievement of a world-wide distribution and with free interbreeding throughout the area of distribution, except as hindered by geographical barriers. Moreover, a further close analogy is to be found in the development of the human individual personality by a complex network of contacts with other members of a social group. Without such social contacts the intellectual development was limited to automatic instincts; with socialization new lines of evolution became possible, just as conjugation opened the road to the development of compound organisms, and the further various stages of advance in prolonged conjugation made possible higher and higher types of cellular structures.

LONGITUDINAL AND TRANSVERSE SECTIONS OF SPECIES.

Longitudinal sections of species show differences along lines of descent. They include what are commonly called life-histories, based on studies of the progressive changes of form and of methods of existence by which individual organisms follow each other in lines of descent.

Transverse sections of species show differences and relations between lines of descent, that is, the internal bionomy of the species. The objects of study are not the methods of development or the physiology of individuals as such, but the nature and relations of the different kinds of individuals which exist in the species. The individuals of a species which are alive at any one time may be thought of as affording a cross-section or end view of the network of descent.

Some of the facts of the constitution of species can be understood best from longitudinal sections, some from cross-sections, and many can be best thought of by keeping both aspects of the network in mind. [365]

DIVERSITY IN LENGTHS OF CONJUGATE PERIODS.

The patterns of longitudinal sections of the networks of descent of different species are determined by the longevity of the individual organisms. In popular language it might be said that the generations of some species overlap while those of other species do not. Many species, both of animals and of plants, are strictly annual. All of the adults die in the fall, and the species exists in the winter only in the form of eggs, spores or seeds. These hatch or germinate in the spring and all the new individuals grow to a simultaneous sexual maturity, interbreed, reproduce and die. All the members of the species are in nearly the same condition at the same time and the figure of descent is simple and regular.

A few species, such as the bamboos among the plants, preserve this complete simultaneity, although living through a considerable series of years. Flowers and fruits may be produced only at rare intervals of two or three decades. All the plants of the species reproduce at the same time and then die. But in nearly all groups the lengthening of the life of the individual organism means the overlapping of the generations and the simultaneous existence of many different forms or stages of the species.

Such a statement is not adequate, however, for a scientific description of the complexities of overlapping descent; for the word generation has been used with a great diversity of meanings. In the lowest unicellular organisms each independent cell-individual is a generation. In the next stage, where the cells are joined into simple and relatively undifferentiated structures, the word generation may well denote the interval between two successive conjugations, or rather the structure which is built up between the ending of one conjugation and the ending of the next. But even this definition fails us as we go higher in the scale of existence and find plants and animals which build two or more organic structures between successive conjugations.

In some cases there is a succession of two kinds of cellular structures, one structure being built up before the formation of the sex-cells, before conjugation commences, and another structure after conjugation has commenced. The former is [366] built of simple nonconjugate cells, the latter of double or conjugate cells. The nonconjugate structure corresponds to the "generation" of the simpler types of organization. The conjugate structure is a new feature intercalated into the previous life-cycle, which it often completely overshadows. The conjugation period of many organisms, and especially of the highest groups, both of animals and of plants, is now very much longer than the part of their life history which corresponds to a whole generation in the lower groups. For tracing homologies between the higher and the lower groups it is still possible to talk of the period between conjugations as a generation, but most of the generation is now occupied by the conjugation period, the life-time of the double-celled phase of organization. This corresponds merely to the fertilized egg-cell or oospore of the lower algae which do not build up any structures of conjugate cells.

In other cases, which are properly to be called alternation of generations, the diversity of the two interconjugational forms has been brought about by vegetative propagation, which replaces or supplements the sexual reproduction of the species. Alternation of generations, that is, of two forms of organic individuals in the same species, may take place either in the conjugate or in the simple or nonconjugate period of the "generation." Thus in the mosses and liverworts vegetative propagation is frequent in the simple-celled phase, while in the ferns and flowering plants it appears in the conjugate period. Vegetative propagation is often described as a purely asexual process, but this is not true of the higher plants, since the conjugate phase is wholly a sexual phenomenon, a part of the sexual process of conjugation.

It may therefore be held that the term generation, as popularly used with reference to the higher plants and animals, does not correspond to what is meant by generations among the lower groups. The period of the life-history which constitutes a generation among the more primitive types of life is so brief as to remain practically unnoticed among the highest. Conversely, the conjugate period which is so short and unimportant as not to complicate the question of generations in the lower groups is [367] lengthened to cover nearly all the activities of the species in higher types of life.

Among the lower groups the overlapping of the generations appears to be a mere coincidence and serves no important evolutionary purpose, but among the higher types it is a condition of the utmost significance, since it has permitted the development of parental instincts and of the numberless devices and habits by which the eggs or seeds or the young individuals are protected and nourished through periods of helplessness. The lengthening of the embryonic and juvenile periods has been necessary to permit the development of large and highly specialized organisms. The overlapping of the generations is also a prerequisite for the development of social habits and instincts, and especially in the transmission of the postnatal inheritance on which the development of human culture and civilization depends. Civilization has been developed and has persisted only among those races in which the family unit of social organization was maintained, so that the children secured the advantage of long and intimate contact with their parents and were thus able to acquire, transmit and accumulate in the race the collective experience and progress of the component individuals and families. Thus the aborigines of tropical America who live mostly in separate and isolated families have built up numerous primitive civilizations, while the natives of tropical Africa who live only in villages have never developed civilizations. Indian children are the constant associates and helpers of their parents while the children of an African village are herded among themselves in little troops or squads like the street waifs of our slums. Even our highly developed systems of formal education have this serious defect and danger, that they tend to disconnect the generations, and to throw the young into premature and reactionary forms of social organization instead of permitting them to grow gradually into their normal places in the general fabric of the community.

DIFFERENT TYPES OF CELLULAR ORGANIZATION.

The complexity of the constitution of species can not be fully appreciated unless it be kept in mind that each individual of all [368] the higher types of life is itself a compact system or colony of cellular organisms, and that these compound units are not only different as to the aggregate cell-individuals, but there are different kinds of cellular organizations. Not only does endless diversity exist among the unicellular or single-celled types of life; there are also different manners and degrees of cell-association to make up the multicellular types. If the cells of the colony-individuals are alike, the organism is called isocytic, if unlike heterocytic.

If the cells which associate have no separating cell-walls the organism may be described as plasmodial, as in the Myxomycetes and in such algae as Caulerpa and Acetabularia. If the cells have the form of long slender filaments the organism is described as hyphal, as in the fungi; if built of definite cell blocks it is called cellular, in the strict sense. The fourth or highest type, found in the animals, combines the other three. Some cells remain quite free and unattached, like the red and white blood corpuscles; some tissues are still plasmodial, others hyphal, while still others, and these in the majority, have definite cellular structure.

Finally, the colony-individuals differ in being built of cells which are not conjugating (agamic cell-structures) or of those which are in conjugation (conjugate cell-structures). Of the latter there are two types, the first is that shown by the higher fungi which build colony-individuals of binucleate cells, formed before the nuclei have fused in conjugation (apaulogamic cellstructures). The second type of conjugate structure is that of the higher plants and animals whose bodies are built up of cells with the nuclei fused, but with a double number of chromosomes (paragamic cell-structures).

These facts are capable of a very definite graphic representation in our ideal longitudinal sections of specific networks of descent. Double-celled structures are the conjugate product of two lines of descent and their existence is to be shown in our diagram by double, closely parallel lines. The network which represents the method of descent of intermediate groups, such as the archegoniate plants (liverworts, mosses and ferns), may show single and double lines in almost equal proportions. Primitive [369] groups may show only single lines, higher groups only double lines, except at the actual points of junction where conjugation takes place.¹

In alternation of generation and metamorphosis the organism changes its external form without altering the figure of descent. Alternation of generations, like the differentiation of separate sexes, exists in simple-celled as well as in double-celled organisms. The phenomena are of an entirely different and minor order of significance compared with the diversities of the different types of cellular structure. Wonderful as the changes are, they are still of a merely morphological and adaptive character and do not indicate new evolutionary departures of the scope of the double-celled structures.

SPECIFIC CONSTITUTIONS MODIFIED BY SPECIALIZED HETERISM.

There are two principal groups or kinds of specific constitutions which can be studied or thought of as cross-sections of the networks of descent. These two series of special types of species arise through two forms of specialization of methods of descent. Instead of remaining uniform or homogeneous throughout, the network of descent becomes variously subdivided or separated into subspecific strands.

The first form of subspecific differentiation consists in specializations of heterism, that is, the establishment within the species of definite forms of diversity of descent, so that individuals are not merely different individually, but fall into two or more groups regularly distinguishable by definite characters. These groups are not formed by isolation, and their existence does not interfere with interbreeding, but usually has the contrary effect of encouraging or compelling interbreeding, since the members of the same group may be unable to interbreed with each other, but are specially adapted for interbreeding with the members of the other group or groups of which the species is composed.

SPECIES WITHOUT SPECIALIZATION OF HETERISM (ARROPIC).

The diversity of normal symbasic descent remains miscellaneous and unspecialized. The individuals may be more or less [370] obviously different, but the differences are fluctuating or completely intergraded, so that no definite alternatives of descent appear, and no distinct subspecific groups are indicated.

Individuals are all similar, equivalent and bisexual or hermaphrodite. None of the vertebrate or arthropod animals show this condition, but it appears to be very common among the lower animals and among plants. Species in which there are no specializations of heterism, no differentiated paths of alternative descent, may be called arropic species.

The arropic condition is not merely synonymous with hermaphroditism, through all arropic species are bisexual. The hermaphroditism of the lower groups of animals and of plants is a normal condition incidental to their more primitive organization. Among the higher groups which have attained sexual differentiation hermaphroditism has reference more definitely to abnormal cases of bisexuality. The arropic condition is also more definite and restricted than bisexuality, since organisms may be bisexual and still manifest some of the following forms of alternative heterism.

SPECIES WITH SPECIALIZATIONS OF HETERISM (ROPIC).

Specializations of heterism exist, and definitely alternative routes of descent are followed by different individuals. The individual members of species fall into distinct groups, but not as the result of segregation or of differences of environmental conditions. The group differences are usually such as to facilitate or to compel interbreeding between the groups.

The attainment of the ropic condition marks an important stage in the evolution of a species, very favorable, apparently, to its further development and to the greater and greater extension of the heteric specializations. The distinction is entirely concrete and practical, but there seems to be no suitable and convenient English word by which to designate it. The expressions alternation and alternative have been used too widely already, and would increase the confusion now existing as the result of identifying alternation of generations with phenomena of entirely distinct nature, such as the different kinds of cellular structures.[371]

Subsexual Species.— A species consisting of bisexual organisms divided into subsexes, that is, into groups differing in one or more characters, but not showing special adaptations to secure cross-fertilization.

The first stage of specialized heterism is represented by species which include two or more types or forms, merely for the sake of the diversity, as it were, and with no sexual diversification, that is, no adaptations, for securing cross-fertilization between the two forms. The differences appear to be of the same nature and to have the same symbasic utility as secondary sexual characters, but the utilization of them is still left to chance. Examples of subsexes are probably to be found in such species as Verbascum blattaria, Viola bicolor, and others in which plants of different castes live together indiscriminately. Antidromous or right-and-left-handed plants like cotton and Castilla, might also be recognized as affording instances of subsexual differentiation.

It often happens in zoology that the sexes of the same animals are at first described and named as two distinct species, but after their true relations have been ascertained one of the supposed species is, of course, rejected, no matter how diverse the sexes may be. Similarly, these subsexual forms need to be taken into account by the taxonomist. The criteria commonly applied to determine specific distinctness are not adequate, since it is possible for constant differences unconnected with sexual diversity, to exist inside the same species without in any way justifying the taxonomic subdivision of the group on the usual basis. There is, however, no reason why any established type of diversity like these subsexes should not be named and described separately, just as the sexes are treated separately when their characters are different.

Botanists are acquainted with numerous instances of diversity among the members of species which may prove to be subsexes; though it is also possible that the differences may belong to species which closer study may distinguish. Thus there are species of Actaea which have the berries either waxy white or crimson, and in about equal quantities. Numerous species of Delphinium have the flowers either pink or blue. In species [372] of Aconitum purple and creamy or greenish white flowers are described. Pink flowers also appear occasionally as definite variants of white-flowered species of Achillaea.

Semisexual Species.— A species consisting of bisexual organisms divided into semisexes, that is, into groups differing in characters which conduce to interbreeding between the groups.

This is the condition reached by many species in which the individuals are all bisexual, but differ among themselves in characters which insure, or at least facilitate, cross-fertilization. In the well known instance of Lythrum there are three castes of plants with short, medium, and long styles and filaments, and three different kinds of pollen grains and stigmatic papillae. A long-styled plant produces only short and medium stamens, and must be fertilized by pollen from long stamens, to be found only on other plants. The semisexes of the primrose were described by Darwin. Similar conditions are known in Oxalis, Houstonia, and many other genera.

Among plants, at least, it might appear that semisexual conditions are more advantageous than the next stage of completely differentiated sexes. Cross-fertilization is secured, but at the same time all individuals may produce seed, and not merely half of them. That complete sexual differentiation has been attained notwithstanding, and in so many different groups, affords an intimation of the importance of symbasic heterism in the structural economy of organisms. The fact loses none of its significance if we reflect that the complete separation of the sexes in plants reduces by half the facilities of the species for producing seeds. All individuals being stationary, the males can contribute to the welfare of species by none of the accessory habits which have been so richly developed among the animals. Indeed, it is by no means unlikely that the tendency of selective influence on many plants has been to keep them in the semisexual condition, sexually differentiated only far enough to secure cross-fertilization, but not far enough to preclude the production of seeds by all individuals.

Sexual Species.— A species consisting of unisexual organisms, or divided into two sexes, male and female, so that interbreeding between the sexes is necessary to reproduction. [373]

The complete separation of species into two sexes is the condition obtaining in all the higher animals, both vertebrates and arthropods, as well as in many of the lower animals, and in numerous plants. It has been found recently that even among the moulds and other lower fungi the plant body, or mycelium, is of two kinds, and that spores are produced only when these are brought together.

Secondary sexual characters are of two kinds, or may be so considered: (1) Those which are accessory to reproductive processes, or assist in caring for the seeds, eggs, or young, such as the mammae of the higher animals; (2) those which are merely the result of accumulation of differences which add to the heterism or internal diversity of the species, such as the manes, beards, tail-feathers or sexual differences of color or form which are of no use in reproduction or in the environmental relations of the species.

The environmental uselessness of many sexual differences is an obvious and well known fact. Not only do the two sexes generally occupy exactly the same environment with equal success, but the presence or absence of many sexual characteristics may have no practical significance for the individual. Some varieties of mankind are beardless; some have beards only late in life, and some have beards in early manhood, but cut them off without appreciable detriment. The uselessness of such characters is shown even more strikingly in certain species of beetles. Some of the males are scarcely distinguishable externally from the females, while others have the head or thorax fantastically modified by the growth of long, heavy, antler-like processes. It is easy to understand that for all the males to be thus encumbered might be a serious handicap to the species.

It may be that selection will help to explain why such features commonly pertain to the male sex. Great diversity among the females would interfere with recognition by males unless their instincts were modified in a corresponding manner. Moreover, variation is the more practicable in the male sex because the extent of the coordination necessary among the bodily organs is not so great. Variation, which in the females might [374] have occasioned serious functional derangements or might have too greatly increased the difficulties of existence, can be tolerated by the males without injury to the species.

That secondary sexual characters are often so completely without function, in the ordinary sense of the word, does not mean that they are of no value to the organism. With reference to the environment they are often worse than useless, but in the physiology of descent they may have an important function. The existence of two sexes doubles, as it were, the symbasic effect of cross-fertilization, by permitting the accumulation of two sets of variations, a second reason for the more rapid progress made by sexually diversified organisms.

What has been called organic evolution has been thought of too exclusively from the environmental side. Evolution has an internal as well as an external function; it has a bearing upon the quality of organisms, as well as upon quantity. Species are advantaged not only by characters which give them a wide range and permit the propagation of large numbers, but it is of equal importance that the vitality of the species be maintained through the provision of adequate diversity of descent, as assured by sexual specialization and by the access of new variations.

The doctrine of sexual selection was invented by Darwin to explain the so-called secondary characters, differences admittedly useless from the environmental standpoint, the two sexes of a species being subject, generally, to identical external conditions. And yet there is everywhere manifest a tendency to the further accentuation of sexual diversities, which are by no means confined to man, or to the higher animals in which esthetic instincts have been attained.

Viewed as specializations of heterism, secondary sexual characters have an obvious and general utility, though of an internal nature. A species with two separated sexes is the stronger because it can accumulate two lines of variations. Symbasic interbreeding becomes, as it were, doubly effective, and the stimulus of diversity can be utilized for a much longer period than if the character were to spread to all the members of the species.

If the present interpretation of the facts be correct, we have [375] in the familiar phenomenon of sex an example of a fundamental evolutionary principle which has thus far escaped formal recognition. Heterism is a concrete property or requirement for constructive evolution, though left quite out of account in theories which have thought to explain organic development by external influences of environment, or by internal "mechanisms of heredity."

Sex specialization in species corresponds to paragamy in cells; the sustained diversity of the associated sexes is curiously analogous to the prolonged separation of the parental chromosomes. Sexuality supplements paragamy, and both serve the same purpose of increasing the vitality of the individual organisms and the coherence of the specific networks of descent.

Supersexual Species.— A species consisting of organisms of two sexes, but with one or both sexes again subdivided into two or more kinds of individuals.

That the uses of the diversities of the sexes are not limited merely to the reproductive functions, is well shown by the fact that specializations of heterism are sometimes carried beyond the stage of definite sexuality. Thus there are, among the sexually differentiated higher animals and birds, numerous instances of the existence of two color-forms, indifferently intermingled, but not intergraded. It has been found, for example, that there are in eastern North America two kinds of screech-owls, red and gray, which are not separated geographically or in breeding.

The following reference to the occurrence of leopards of two colors in the Malay region may serve as a sample of many similar observations among the mammals.

"Many of the hunters I have met, and some of the authors I have read, appear to consider the black leopard a distinct species, but it is simply a freak of the ordinary spotted leopard, just as the silver and the black fox are freaks from the common red. In a litter from a red vixen I have seen a silver among red pups; and I met a man in the jungle where lower Siam meets the Malay Peninsula who had found a black among the spotted leopard's cubs, upon which, however, the spots, of course, are not very clearly defined until they become older."

. . . "I noticed after I got its pelt off, that in the sun it had [376] a kind of watered silk appearance, as a result of the deeper black of the spots, which, though invisible, were really there just the same."¹

In a similar case of supersexual dichromatism in a chrysomelid beetle experiments showed that the two color-forms could be separated and established as uniform varieties by selective breeding.² The mating of black individuals produced only black offspring in the first generation, while matings of spotted individuals continued to give a proportion of black offspring until the third generation.

SPECIFIC CONSTITUTIONS MODIFIED BY RESTRICTED DESCENT.

This is the second form of diversity of constitutions revealed by cross-sections of networks of descent. Unlike the specializations of heterism, the members of groups formed by restricted descent do not, of course, breed together, for it is in this that the restriction of descent consists. The specializations of heterism are in accord with the evolutionary advancement of the species, but the groups formed by restricted descent are removed from the conditions of free interbreeding and of normal evolutionary progress. They represent, instead, the different stages of a process of deterioration.

Symbasic Species.— Species with descent unrestricted, consisting of large numbers of diverse individuals freely interbreeding in a broad, continuous and regular network of descent.

A species is not merely an aggregation of organisms, whether alike or different; the organisms are connected by a completely interwoven fabric of lines of descent. Such plants as Portulaca oleracea, Poa pratensis and Ceratodon purpureus, may serve as examples of very widely distributed symbasic species.

Porric Species.— Species made up of partially segregated subspecies. The cross-section of the network of descent, instead of showing a rounded or regular form, is irregular, or partially subdivided into arms or branches.

Widely distributed species, but locally diversified, like the [377] European Helix hortensis, afford the best examples of this type of intraspecific diversity. The quail, or Virginia partridge, a nonmigratory bird widely distributed through eastern North America from New England to Guatamala, shows many local subspecies connected by series of imperceptible gradations. The sugar maple of eastern North America has several geographical subspecies.

Stenic Species.— Species consisting of stens, that is, of narrowly segregated subspecies, domesticated varieties, or breeds, propagated by sexual reproduction.

As a result of propagation by narrow breeding, the individual members of a sten are much more nearly uniform than those of normal symbasic species, or even than those of geographical subspecies. As purely stenic species may be mentioned those which do not exist any longer in the wild state, but are made up of many local domesticated varieties. The domesticated animals fall here, except as they may represent hybrids of different wild species. Of domesticated plants the Indian corn or maize is the best example, since it has retained a complete system of cross-fertilization, which many domesticated plants have lost.

Very small, closely localized natural species, like the remarkable Hawaiian land-snails upon which Gulick has based his theory of evolution by isolation, represent essentially the same condition of restricted descent as domesticated stenic varieties.

Linic Species.— Species composed of separate, parallel or slightly diverging lines of descent, propagated by autogamy or parthenogenesis, and not united into a network.

Wheat and barley are perhaps the most conspicuous examples of linic species among domesticated plants, though many other species are autogamous, with more or less consistency. Strict line breeding is not possible, of course, among the sexually differentiated higher animals, but is sometimes approached by what is called in-and-in breeding of closely related individuals.

Line-bred organisms are extremely uniform, even more so than stens. Self-fertilization involves only the combination of gametes of the same origin and probably of very nearly identical nuclear configuration; at least there is even less variation. Linic species occur in nature as in the well-known instances [378] of Hieracium upon which Nägeli based his theory of evolution in a definite direction. The persistence by parthenogenesis of the individual differences of transplanted specimens was accepted as proving that variation held to definite directions.

Likewise De Vries has made use of linic autogamous species of Draba to illustrate his conception of elementary species. The uniformity and stability of the line-bred plants has been taken to represent the normal condition of species, and the inference has been made that the species recognized in nature by taxonomists are generally composed of similar independent units, the effect of the method of propagation, to resolve the species into separate lines of descent, being left out of consideration.

Clonic Species.— Species consisting of separate lines of descent continued by vegetative propagation alone.

Clones, like lines, are propagated from single individuals, but by vegetative processes only, so that variation is almost completely avoided. Nevertheless, even vegetatively propagated plants are not completely uniform. Clonic groups of the same origin often show fine gradations of diversity, and occasional mutative variations are known.

Clones do not exist, of course, among the higher animals, but they are exceedingly numerous among plants. Several domesticated species now exist, as far as known, only in this form. The horse-raddish, sweet-potato, banana, arracacha, yautia and taro may be mentioned as seedless plants, but large numbers of others are nearly seedless or have varieties which are seedless.

THEORIES OF EVOLUTION BY RESTRICTED DESCENT.

It is a noteworthy fact that the earlier theories of evolution, including those of Darwin, Nägeli, Gulick and De Vries, have been based upon one or another condition of restricted descent. The kinetic theory is the only suggestion of a method of evolution applicable to conditions of unrestricted descent. The predisposition to see in restricted descent ideal conditions of evolution has been strengthened, if it has not been wholly supported, by the fact that it is only in restricted descent that the traditional ideal of heredity can be applied. Only narrow-bred organisms afford even an approximate identity of form and structure. [379]

De Vries, Gulick and Nägeli have given their chief attention to extreme forms of restriction, like those of Draba, Achatinella and Hieracium. Darwin kept much nearer to the consideration of natural conditions, though his doctrine of selection implies that evolutionary progress depends entirely upon the plan of causing species to change by restricting the descent of the component individuals. In the kinetic theory, it need scarcely be repeated, the result of selective restriction is not evolution, but specialization. The evolutionary motion would still take place if the selective restrictions of descent were not imposed.

COMBINED FORMS OF SUBSPECIFIC DIVERSITY.

Modifications of the constitution of species by specializations of heterism do not interfere with the attainment of the other form of diversity by restricted descent. Thus a sexual species may be partially segregated into geographical subspecies or may be narrowed still further into the stenic condition of domesticated varieties and breeds. Linic and clonic subdivisions of sexually differentiated species do not occur, of course, among the higher animals, being limited to the lower groups and to plants which have the power of sexual propagation or of parthenogenetic development. But even among the cultivated plants it does not appear that any sexually differentiated species has been resolved completely into the clonic condition. There are large numbers of clonic female varieties of figs and date-palms, but the male trees are usually recruited from chance seedlings, so that the network of descent is not entirely destroyed. The female half of the species is represented by vegetatively propagated clones, but on the male side miscellaneous individual diversity remains.

The existence of restricted subspecific groups may not interfere in the least with the maintenance of a normal specific network of descent. A widely distributed symbasic species may have a few porric subspecies as a result of the partial isolation of particular localities. Special conditions, such as an alpine climate, might restrict a part of a species to linic or clonic propagation while the remainder retained fully symbasic conditions of descent. Through the fabric of broadly diversified descent there may run narrowly compact strands composed of [380] linic or clonic individuals, which no longer share the symbasi.c interbreeding of the group and afford no true criterion of the conditions under which evolution goes forward. Just as most planets are attended by satellites, so species are sometimes found to be supplemented by small subspeciflc adjuncts, little species-like groups of organisms which some have taken for new or incipient species, but which stand in a permanently subordinate or retrograde relation to the evolutionary part of the species.

LIMITATIONS OF CLONIC PROPAGATION.

Vegetative propagation, whether in nature or in domestication, appears to conduce always to seedlessness. Some have thought to explain this fact by reference to the superiority of the asexual over the sexual propagation. This reasoning is scarcely adequate, in view of the fact that much larger numbers of species have retained their capacity of producing seeds, though regularly supplementing the sexual by the vegetative propagation. The greater probability is that the decline of sexual fertility in vegetatively propagated types is a symptom of deterioration, just as sterility is a frequent characteristic of abnormal variations or of hybrids.

The formation of the sex-cells, as we now know, is a highly specialized and complicated process, and it is easy to understand why it should be the first of the physiological functions to become deranged and inefficient. It is known also, from the behavior of hybrids and mutations, that vegetative vigor has no direct relation or apparent connection with reproductive vigor. Indeed, sterile hybrids and mutations often show great and notably superior strength and longevity, due, we may suppose, to the stimulation which attends new variations. This consideration may also explain why clonic and linic species usually appear to consist of definite groups of closely similar individuals. These groups may have originated by individual mutative variations of notable vegetative vigor, which have on this account survived or crowded out the weakening survivors of the original symbasic species or other variations less recent or less vigorous.

The disastrous effects of inbreeding among the higher animals have been known for centuries, and are taken into account [381] by all breeders. That the same principles apply to plants, has remained in doubt for two reasons: (1) The much less complex organization and less specialized tissues of plants render many of them less acutely dependent upon cross-fertilization. (2) The plants which have been longest under cultivation are not grown for their seeds and are propagated asexually, so that their decline in reproductive fertility has not diminished their economic value. No plant valued for its seeds has been propagated other than from seeds for any considerable period.¹ Numerous tropical root-crops and fruits, such as the sweet-potato, yam, agave, sugar-cane, banana, pine-apple, and breadfruit have been grown for thousands of years from cuttings, probably without the interposition of a single seedling generation. In a sexually propagated species inbreeding would have led long since to extinction, but these clonic varieties are still extremely vigorous. Nevertheless, such species do not form a real exception to the rule of deterioration under inbreeding, since a very large proportion of them, belonging to many and very diverse families, have shown this tendency towards seedlessness.

The reduction or elimination of the reproductive parts has been ascribed by some to selection, and by others to a supposed biological law of paucity which causes useless parts to disappear. No basis of fact has been shown, however, for either of these explanations; unassisted nature supplies us with instances like Sphagnum and Lunularia to which neither would logically apply, but which would be well accommodated in the view that continued asexual propagation, like other forms of isolation, weakens the reproductive powers. This law would also explain why the absence of sexual reproduction appears only as the character of aberrant species or genera, and has not been able to persist for a period long enough to permit the differentiation of organic groups of higher systematic rank. Botanists seem not to have ascertained the existence of any wild phanerogamous plant which is always and everywhere seedless. [382]

The opinion has long existed among horticulturists that varieties of fruit trees tend to deteriorate, but a biological explanation has been lacking thus far. The most prominent horticultural writer to defend such a view is Burbidge, who holds that budding and grafting are artificial and unnatural processes, for which propagation by rooted cuttings should be substituted. The analogy of the seedless tropical root-crops indicates that the use of cuttings would afford no protection against the gradual reduction of fertility, though the suppression of seeds in fruit trees may not be an undesirable symptom, except when it is accompanied by a deterioration in quality. Only a few horticultural varieties have been propagated as clones for more than a century, but the advance of sterility has already become appreciable to nurserymen, who are careful to plant seeds from seedling trees, in the belief that these germinate better and produce more vigorous stocks than the fruit of grafted clonic varieties.

That superior varieties are commonly deficient in vigor is thus explainable without reference to any special perversity of nature; such varieties may owe their reproductive debility to the fact that they have been more carefully and persistently propagated without crossing. Some varieties of peaches, for example, yield a very small percentage of viable seed. In France many attempts to secure seedlings of the "Alexander" have failed. This variety and the very similar "Amsden" appeared about the same time and are supposed to be seedlings of "Hale's Early," a variety also notably deficient in reproductive fertility, since only about ten per cent. of the seeds germinate. The seedlings of "Hale's Early" are also, as a general rule, very diverse, without close resemblance to the parent or to each other. The variety called "Hill's Chili" affords an instructive contrast, in that practically all the seeds germinate and about ninety per cent. of the seedlings come true to the parental type, leaving about ten per cent. of variations.¹

Obviously, the evolutionary status of these two varieties is very different; one is entering upon the stage of mutative aberration, [383] while the other is approaching that of complete sterility. Horticulturists have not uncommonly believed that the longer the succession of "grafted generations" of tree fruits the greater the likelihood of deviations from the type of the original seedling, but this idea seems not to have received scientific consideration or support, perhaps because it appeared to contradict the opinion of Darwin¹ and many other evolutionary writers who have held that characters can be permanently "fixed" by inbreeding, or close selective segregation, of which propagation by cuttings may be taken to be the extreme form. The kinetic theory of evolution permits us to understand, however, that the "fixity" to be secured either by inbreeding or by asexual propagation is only relative, and that its result in both cases is to predispose the organism to abrupt variations and reproductive debility.

ORIGIN OF LINIC AND CLONIC CONDITIONS.

The occurrence of self-fertilization, parthenogenesis, and vegetative propagation in nature has undoubtedly caused many writers to suppose that these methods of descent represent truly normal evolutionary conditions. Indeed, no abnormality need be charged in the many cases where the species maintains at the same time the normal network of descent by sexual reproduction with free interbreeding. The abnormal condition supervenes when the species loses its network of symbasic descent and is resolved into disconnected lines. Such a condition may result whenever the normally sexual and symbasic reproduction becomes less effective than autogamous or purely vegetative methods of propagation. Thus, in such little plants as Draba and Viola, which have to avoid the competition of larger neighbors by blossoming early in the spring, the non-symbasic methods of propagation take on great importance, for insects are scarce and the weather often so inclement as to completely prevent the transfer of pollen.

Similarly, in alpine and arctic conditions, vegetative propagation is much safer, and usually much more successful than sexual reproduction. The short and treacherous seasons often prevent the ripening of seed. The formation of apogamic buiblets [384] instead of flowers is frequent among the saxifrages and other Arctic plants, though many similar instances are known in natives of temperate and tropical regions.

Wheat and barley, and to a less degree several other domesticated plants, have been unconsciously selected towards autogamy in a similar manner, by being cultivated far to the north of their original habitats. In unfavorable seasons only the autogamously fertilized seeds would ripen. The wild relatives of all these plants, so far as known, have facilities for cross-fertilization.

That autogamy and other forms of restricted descent conduce to the breaking up of species into small subspecific groups, is well shown among the cereals. The rye plant has retained and even accentuated its provisions for cross-fertilization, and has kept its position as a relatively normal coherent species, instead of falling apart into distinct varieties. Cross-fertilization has also been fully maintained in the corn plant, but here the large size of the seeds and their compact grouping on the ears greatly facilitate selection, and have favored the establishment of many local varieties.

RELATION OF LINIC TO CLONIC PROPAGATION.

The fact that reproductive fertility deteriorates more rapidly than vegetative vigor, when organisms are placed under conditions of restricted descent, is to be correlated with another phenomenon, discovered by Darwin, that autogamous fertilization is sometimes superior to more miscellaneous methods of narrow inbreeding. This fact has been generally accepted to mean that autogamy and heterogamy are both normal evolutionary conditions. In the kinetic interpretation it does not appear that autogamy is a truly normal and progressive state. The superiority of strict autogamy over more miscellaneous inbreeding appears explainable by analogy with parthenogenesis and vegetative propagation. All three processes can be viewed as methods of postponing deterioration from restricted descent, by omitting the nuclear readjustments which are required in normal sexual reproduction. When diversity of descent is no longer sufficient for normal readjustments, degeneration begins, in the form of mutative variations. These usually fall below the [385] parental standards, or at least diverge from them so seriously as to injure the commercial value of the crop, as strikingly shown in the tobacco varieties studied by Mr. A. D. Shamel.¹

Seed produced by autogamous fertilization yields plants of very much greater uniformity, and it is in this fact that their superiority lies. The plants were not better, as individuals, than some of those produced by the more miscellaneous breeding, but the tendency to degenerate variation had been avoided, or at least postponed.

Such facts do not appear to warrant any general contrast between cross-fertilization and self-fertilization, but only between narrow breeding and line breeding, and of these the line breeding appears to be superior because it constitutes an approximation to vegetative propagation and avoids the need of nuclear readjustments with inadequate diversity of descent. The union of two nuclei which are the autogamous progeny of the same individual organism, can hardly require any new adjustments to be made. The formalities of sexual reproduction are observed, but diversity of descent, which gives physiological value and evolutionary significance to the process, has been eliminated. Self-fertility and parthenogenesis, like vegetative propagation, have value only as means of avoiding, for a time, the normal results of restriction of descent, not because they represent normal evolutionary methods of organic succession.

DIVERSITY REACTIONS IN RESTRICTED DESCENT.

Efforts toward the selective improvement of domesticated plants and animals have been accompanied everywhere by the narrowing of the lines of descent, and often by close inbreeding. How far this abnormal condition is responsible for the results of experiments with domesticated species, and how far these results are of general evolutionary significance, remains to be considered. Most of our important food-plants were domesticated long before the period covered by human history or tradition, so that the general claim of selective improvement through thousands of years could not be denied, and has [386] continued to be accepted as a sufficient cause of the extensive modifications which have taken place.

The question has been debated at length on theoretical grounds, but without decisive results, since it appeared to lie outside the range of experimental determination, owing to the vast periods of time which have figured in the calculation. Fortunately, all plant cultures are not the same in method or in history, and the so-called Arabian coffee furnishes an instructive contrast with other domesticated species. Coffee has probably not been in cultivation much more than a thousand years, and has existed but a few centuries, or often only a few decades, in its present centers of production. It is not an annual, but a shrub, or small tree, the selective improvement of which would require more years than planters generally expect to give to the business. Plantations are generally large, and experiments with individual trees are difficult and time-consuming, so that it is only within recent years that the securing of improved varieties of coffee has received serious attention. The evolutionary factors of selection and of long periods of local influences of soils and climates are thus alike absent, and yet there is no lack of coffee varieties with abundant diversity in form, habit and color. Their general similarity consists only in being inferior in fertility to the parent type.

So much has been written upon the improvement of plants by domestication and selection that this inferiority of coffee varieties may seem exceptional, but the apparent anomaly disappears if we reflect that fruit trees and other horticultural plants supposed to have been greatly improved in domestication are not grown for the seeds, and hence complete fertility in the sexually reproductive sense has been a minor consideration or even a positive disadvantage; indeed, with many plants it has been one of the direct objects of selection to reduce the number of seeds or to eliminate them completely. More or less seedless abnormalities are valuable, for example, among the grapes, plums, and oranges. If coffee were cultivated as an edible fruit the new sorts would be of use, since thicker pulp and smaller seeds are frequent characteristics of the berries; indeed, a coffee which did not produce any normally developed seeds [387] was found in 1903 in Costa Rica. As ornamentals, some variations offer new colors and greater abundance of flowers, and the foliage and habit of the trees sometimes deviate strikingly from the normal or parent form. Unfortunately, the planters would find an advantage only in the direction of increasing the number, size, or weight of the seeds themselves, and they accordingly pronounce the new varieties worthless.

Similar abrupt variations of many cultivated plants and animals were studied and described by Darwin as "sports," but it was also known to him that such variations are relatively infertile and do not persist in the presence of the normal or less closely inbred types, so that it has remained for Professor De Vries to base upon such variations a general theory of evolution. The variations, or sports, chiefly studied by Professor De Vries are those of an evening primrose native in North America and escaped from cultivation in Holland, and thus accidentally segregated from the wild stock of its species. It belongs, like the coffee, to a family in which there are specialized provisions to assist cross-fertilization, so that the early manifestation of the effects of inbreeding might be expected.

The variations of OEnothera described by Professor De Vries seem to be closely parallel to those of coffee; most of them are conspicuously deficient in reproductive fertility, and some are quite sterile. This relative or complete sterility of sports, or variations secured by inbreeding, warns us that evolutionary inferences founded on this class of facts must be carefully revised, since it is obvious that organisms notably deficient in the power of reproduction can not be expected to have played a large role in the process of organic evolution. Nature, like the coffee-planters, requires seeds; reproductive efficiency is the first requisite of survival.

A general evolutionary significance of the phenomena of mutations becomes apparent when the facts are interpreted from the standpoint of normal heterism, that is, as reactions from the abnormal uniformity which is the first result of restricted descent. The diversity of mutations is greater than the diversity of normal heterism, but this is in entire accord with what we know of other physiological reactions of organisms. Mutations [388] are at once degenerative and reconstructive, just as the high temperature which attends many diseases of the human organism is at once an evidence of illness and an indication of constructive systemic reaction. Indiscriminate crossing of mutative varieties tends to restore the wild type of the species. Mongrel dogs are wolfish; mongrel pigeons, even of white ancestry, are blue; mongrel roosters become red in approximating the primitive game breeds; mongrel flowers are single and small.

Stronger evidence could scarcely be demanded for proving that the interbreeding of the members of a species is a measure of organic stability, not a stationary or uniform stability, but a stability of coherent symbasic motion.

EXAGGERATED HETERISM OF CLONIC HYBRIDS.

Further evidence that mutations are reactions from abnormally restricted descent may be drawn from the results of sexual reproduction among clonic varieties. The sexual offspring of plants which have been subjected to considerable periods of vegetative propagation always show a very large amount of individual diversity. This has caused them to be reckoned as hybrids, though in reality they represent a very distinct type of evolutionary phenomena. Each clonic variety is, after all, only an individual member of its species, and as such varieties have not been selected or bred to uniformity, in the sense of coming true to seed, they and their offspring might be expected to retain the original amount of heterism or normal individual diversity of the wild type of the species. As a matter of fact the sexual offspring of clones have an individual diversity of the order of mutations. The only difference appears to be that all the individuals may be mutants, instead of the relatively small percentages usually appearing in species which have been subjected to courses of selective inbreeding for the elimination of heterism.

DIVERSITY RELATIONS BETWEEN SUBSPECIES WITH RESTRICTED DESCENT.

As long as the diversity of the members of species appears either as the merely accidental or arbitrary result of environmental [389] influence or of mechanisms of heredity, both the theory and practice of evolution remain mysterious and contradictory. It is only after the physiological value of diversity in the constitution of species has been recognized that we begin to gain a definite appreciation of the practical bearings of evolutionary facts. With nature wrongly interpreted, the results of domestication and breeding were likewise obscured and distorted. As long as our reckoning was based on the false ideal of uniformity and stability of species, it was not possible to gain an orderly concept of even the simplest of evolutionary relations, or to escape from the confusion and contraditions which have left even the most concrete investigators in hopeless disaccord.

Among breeders of plants there exists the greatest possible diversity of opinion regarding the value of hybridizing as a means of securing new organic forms of superior agricultural utility. Some breeders have secured very valuable hybrids, while others have found hybrids of no use at all as a means of increasing the desirable characters of the species which they were seeking to ameliorate. To explain and reconcile this apparent contradiction is not only a matter of scientific interest in its bearing upon the general subject of evolution; it is also of much practical importance to be able to distinguish between the different kinds and combinations of subspecific groups and to avoid a waste of efforts upon methods and materials which do not promise useful results.

The time has not yet come for the establishment of absolute standards and criteria, if indeed such a time is ever to come. There are unforseen accidents, not only in the best regulated families, but in nature as well. It is the rarely unusual circumstance, the exception to all known rules, which may have great interest and potential value. The sterility of mules is one of the most invariable of the phenomena of hybridization, and yet fertile mules are not altogether unknown, nor is it certain that such an animal might not be a means of securing new and desirable variations of our equine stocks. Hybrids between the different species of bovine animals are generally fertile and readily made, but the establishment of a breed combining the blood of the buffalo and the domestic cow has proved difficult. [390]

For the practical breeder, as for the scientific investigator, nothing should be taken for granted until verified by actual experiment, but it is, nevertheless, useful to have, if possible, a system of interpretation by which results once attained can be understood, and proper discrimination made between the relative prospects of alternative fields of investigation. Selections, mutations, crosses and hybrids, have entirely different importance in different groups, depending upon the nature of the characters which it is desired to secure, and upon the adaptability of the species to different methods of propagation. In the amelioration of coffee, for example, mutations promise little because of their smaller production of seeds, but if the flowers or pulp of the berries were the valuable part, mutations would be as valuable as among other horticultural species.

Selection and hybridization have been thought of as two alternative methods by which evolution might be brought about, and the debate has continued as to which is the better. The question could never be answered in this form, for the assumption on which it is asked is a false one. The normal species, the unit of evolution, is neither stationary nor uniform. It not only makes a slow and gradual advance,, as a whole, but it manifests all the time a vast diversity among the different individuals. Some of this diversity is induced by the environment, but much of it is quite spontaneous and continues to appear even in a uniform environment.

The value of selection does not lie in any power to cause these inherent differences; it can only preserve them and prevent, as it were, the swinging back of the pendulum of normal diversity. The alert breeder seeks to catch it at its highest and to hold it steadily there. It cannot be held forever, as is now generally recognized. Sooner or later the selected type deteriorates, and shows itself inferior to some more recent selection which has lost less of the normal vigor of the species.

To hybridize selected varieties may serve merely to release the pendulum and allow it to swing back along the curve of normal diversity. The vast majority of the progeny are likely to be inferior to the parents in the special qualities which have made them valuable. Some of them may approach the standard, [391] but they seldom or never surpass it. The breeder concludes that hybridizing is a mistake and finds that much more can be accomplished by selection. This conclusion is quite correct if he is dealing only with long-domesticated strains of plants and animals, and if he wishes to obtain from them the greater accentuation of some character already specialized by selection. If the varieties are not too unlike, or too long selected, the result of crossing will be to restore the more normal but less desirable diversity. If the varieties crossed are somewhat more remote, the diversities may balance each other into a somewhat uniform intermediate average. Still longer selection may establish the specialized characters as definitely alternative, in the Mendelian sense, so that they do not combine again into a single hereditary pattern, but separate regularly into the two original components, as in the pea hybrids studied by Mendel, and the many other instances discovered by more recent investigators.

In none of these three cases or types of hybrids is there any reason to expect an increase of characters beyond the range of accentuation to be reached by selection; they all involve, instead, a lessening of the amplitude of diversity obtainable through selection. If the selective specialization of characters of a variety were a true step in the evolution of the species, these kinds of hybrids could be called reversions or retrogressions, since they appear to go backward and undo the results of selection. To call them reversions is very misleading, however, from the evolutionary standpoint, since the closely selected type, however useful, represents only a temporary and abnormal phenomenon, a holding of the pendulum of variation to one side, instead of permitting it to describe its normal vibrations, or to change its general position and point of support.

The simple analogy of the pendulum proves entirely inadequate as a means of illustrating the normal conditions and requirements of true evolutionary advances of specific groups, for we are not dealing then with vibrations of single characters, but with a complicated network, a veritable fabric of descent and of character-combinations. The pendulum analogy is appropriate only for the single lines or narrow strands of descent which selection separates from the web of the species, and [392] holds for a time at a point of high expression a character which averages much lower in the species at large.

MUTATIVE VARIATION OF SELECTED VARIETIES.

The only way in which the accentuation of such a narrowly selected character can be still further increased, beyond the range of normal variation of the species, is by abnormal variation; that is, by mutation. The narrow selection may be said to induce the mutations because it weakens and unbalances the hereditary tendencies of the variety, but the mutations are by no means limited to the character or quality for which the variety has been selected; they are likely to take any or all directions. Some of them are generally found to carry the breeder along the lines he desires to follow.

Are hybrids between selected varieties of the same plant or animal of no practical breeding utility? Yes, if it is desired to preserve or strengthen the vitality of the organism or to secure intermediate characters, or new combinations of characters already existing.

The general answer must be negative, if the purpose is to obtain new characters, or higher degrees for accentuation of characters already specialized by selection. Instead of securing a larger range of diversity, the contrary results are much more likely to be reached. It may even happen, if the varieties have been subjected to narrow selection, that the hybrid offspring, instead of being more variable than their parents, will actually be more uniform, the hybridization bringing them back, as it were, to the hereditary road from which they were beginning to wander towards mutative degeneration.

The mutations are as abnormal, of course, in the strictly evolutionary sense, as the narrow descent which induces them, but for agricultural purposes they may be very valuable, and often the abrupt change of form seems to lend them a remarkable vegetative vigor which greatly increases their productive capacity. This is notably the case among plants, and especially among those cultivated for their vegetative parts instead of for their seeds.¹ [393]

The facility with which many plants can be propagated from cuttings or by grafting often permits sterile mutations and crosses to be preserved and utilized for long periods of time. Among animals, on the other hand, mutations are of relatively small value. The higher organization of animals renders them liable to earlier and more serious deterioration from inbreeding, though there is great difference in the susceptibility of different kinds of animals.

BEHAVIOR OF DISCRIMINATE MUTATIONS.

When mutations are crossed with other members of their own immediate group of related individuals they are generally prepotent. They do not tend to average away and disappear, but are repeated, or even accentuated, in a considerable proportion of each successive generation, and sometimes in all of them. Plant mutations which can be propagated by self-fertilization are often constant from the first, and have been thought by some to represent the formation of genuine new species.

When mutations are bred outside of their own group, and especially when they are crossed with the wild type of the species or with the variety which has not been long or closely selected, they are not prepotent, but recessive. The new mutative characters appear weaker than the others and may fade out and disappear entirely. The same result may be reached by indiscriminate interbreeding among the representative of two or more mutations or selective varieties. The ancestral characters of the wild type of the species reassert themselves, and may even reappear in crosses between varieties from which they have long been lost.

All these and other similar phenomena can be understood, or at least brought into rational relations, if we keep in mind the fact that crosses between the narrowly selected varieties or mutations of the same species tend to restore the original and normal conditions of free interbreeding. They tend, in other words, to repair and reconstruct the normal fabric of symbasic descent, and to reduce the strains and deteriorations caused by too close segregation, too little diversity, and too much inbreeding.

Instead of being monstrous or unnatural, these crosses are [394] more normal, more vigorous, and more fertile, than their parents. Why, then, are they called hybrids? Because we have been led astray by the theory of normally uniform and stationary species, in which it was made to appear that anything which interfered with identity of form and structure was essentially unnatural, like a cross between members of species which do not normally breed together, and which produce, when so bred, abnormal progeny. There are many groups in nature which are reckoned as species, but which are no farther apart than some of the varieties of cultivated plants, and which can breed together without difficulty or abnormality. For systematic purposes it is desirable to recognize each separate natural group of organisms as a species, and this can also be justified from evolutionary standpoints, for segregated groups are able to make evolutionary progress on distinct lines, and eventually to become different from other groups of common origin.

It often happens, however, that evolutionary progress is not consistent in the vegetative and reproductive parts of the organisms. Species which appear very distinct externally may, when brought together, breed freely and normally, while others whose bodily differences are difficult to detect may refuse to mingle or may produce only sterile or otherwise abnormal hybrids. While it is thus difficult or, it may be, impossible, to draw an absolute line of definition, or to restore the old distinction between hybrids and crosses, this does not justify us in ignoring the very wide and very practical differences between the extreme conditions of this series of phenomena.

ANALOGIES OF HYBRIDS AND MUTATIONS.

The phenomena which have the nearest and most genuine relations with hybrids are not crosses, but mutations. Hybrids and mutations can both be described in the same words, as aberrations from normal heredity. Both are due to the same cause, inadequate fertilization, which unbalances the organic equilibrium and gives rise to abrupt variation, usually in many directions at once. Mutations and hybrids show also a general deficiency of fertility. This is carried, very often, to the extreme of complete sterility, though there may be present at the [395] same time unusual vegetative vigor, analogous, in all probability, to the stimulation of energy of growth which appears in normal crosses and in prepotent new variations. Though no experiments are known to have been made with the idea of such a test directly in mind, the indications are that results of mutation and hybridization might prove in the same species almost identical, for many so-called false hybrids do not appear to be the results of a genuine and effective interbreeding, but seem rather to involve an approach to the phenomenon of artificial parthenogenesis, somewhat similar to the parthenogenetic development through chemical and mechanical stimuli, described by Loeb and others. The two nuclei of the supposed parents of the false hybrid do not appear to have united and combined the parental qualities, since the progeny shows no definite indication of the traits of one of the supposed parents, either in the first or in subsequent generations. The facts discovered by Guyer in sterile hybrid pigeons, that the parental chromatin elements remain separate and do not undergo a normal mitapsis, illustrates the possibility of false hybrids, especially in plants and in lower types of animals where parthenogenesis can take place. Such an abnormal and inadequate method of fertilization would explain extensive variations of the progeny, which well deserve to be called false hybrids. Nor is it unlikely that the same explanation may be found to apply to variable hybrids, even when they share the characters of the parents. The indications are that in different cases there are all possible gradations in the extent and efficiency of the combination of the parental elements, from that which affords mere stimulation to that which gives a fully intermediate result.

It does not follow, however, that the combination is normal or complete when the first generation is intermediate. The first generation may be intermediate under two nearly opposite conditions, as already noted. Crosses are intermediate when the parental elements are thoroughly congruous. Their combination merely restores a normal condition of symbasis, that is, provides a normal amount of diversity of descent. The first generation of hybrids is also intermediate when the parental elements are very diverse and antagonistic. Hybrids which [396] appear quite uniformly intermediate in the first generation may prove, nevertheless, to be completely sterile, as in the mule, whereas intermediate crosses between narrow varieties are always completely fertile, more so, it may be, than their more inbred parents. No distinction is to be drawn between crosses and hybrids which are uniformly intermediate and at the same time fertile, but there is a wide range of phenomena between an intermediate, fertile cross between narrow varieties and an intermediate sterile hybrid between diverse species. Next to the hybrids which are intermediate, but sterile, are those which are intermediate and fertile, but show diversity and partial sterility in the second generation, proving that the parental elements did not combine in a manner to afford a stable equilibrium of heredity. In another stage of hybridity, with less diversity of parents, the first generation is variable, which may be taken to mean that the parental elements are sufficiently similar to influence each other, instead of exerting a uniform degree of repulsion. Nevertheless, they do not combine readily, but form uncertain and extremely varied combinations.

The purpose of this enumeration is to show that with hybrids, as with crosses, there is a series of phenomena which can be described and interpreted in terms of diversity, using as a standard the normal diversity of the individuals of species in nature. In this way it is possible to avoid the ambiguities which have attended the use of the false and artificial standard of uniformity. From normal diversity there may be departures on either side, on the one to abnormal uniformity, on the other to abnormal diversity, and both of these can be reached, as we have seen, in several ways. Uniformity appears:

1. In closely selected varieties (stens).
2. In varieties or individuals propagated from cuttings or by other asexual methods (clones).
3. In the progeny of inbred saltatory variations (mutations).
4. In crosses between moderately inbred stenic varieties.
5. In first generation hybrids between species so remote as to combine with difficulty.

Likewise diversity greater than the normal may appear:

1. Among mutations from narrowly inbred varieties. [397]
2. Among crosses between individual clonic types, long subjected to vegetative propagation.
3. In a species or variety which has been placed in new and unwonted conditions (neotopic mutations).
4. Among crosses between narrowly inbred varieties (Mendelian hybrids).
5. Among hybrids between species not too remote to combine at all, but not sufficiently related to combine in a regular and uniform manner.

THE NATURE OF STERILE HYBRIDS.

A further distinction of fundamental significance remains to be added to the preceding, before the full range of the phenomena of interbreeding can be made apparent. The general impression has been that the development of a new individual represented the result of a combination of the two parental sexcells, but this is only partially true, especially among the higher plants and animals. The fusion of the parental sex-cells is carried through only two of the three stages of conjugation. Fertilization unites the outer, unspecialized protoplasms (plasmapsis) and also the nuclei (karyapsis), but the chromatin, the most highly specialized cell-substance, the citadel, as it were, of the life of the cells, remains distinct until after the new individual has developed, so that the body is not composed of simple, post-conjugational cells, but of double cells in a condition of prolonged conjugation.

The fusion of the chromatin granules, or ultimate sex-elements (mitapsis), may not take place until the new individual is mature and about to form new sex-cells of its own. The other cells of the body never reach mitapsis. The sterility of hybrids arises, it is now believed, from the inability of the sex-elements to pass this third and final stage of conjugation. It was always mysterious that hybrid combinations which could be made for one generation could not continue for a second or a third generation. This new appreciation of the nature of the process of conjugation makes it apparent, however, that hybrids are sterile because the parental elements do not make even one complete conjugation. There is thus a definite difference [398] between a sterile hybrid and a fertile combination, one which might have restricted the use of the term hybrid to the former. Sterile hybrids, like false hybrids, are scarcely to be reckoned as forms of conjugation. They are rather to be looked upon as more nearly allied to parthenogenesis, a development through stimulation merely, but without the possibility of forming new relations of heredity or of making new combinations of characters. Sometimes there is not even enough cooperation between the mismated partners of the cell-units to carry the organism through even the normal cycle of individual existence. Hybrids often refuse to grow up, or they may die suddenly and without apparent external cause.

The building up of each cellular organism involves a continued cooperation between the parental sex-elements, which may be thought of as persisting in all the cells of which the body is composed. Whenever this cooperation breaks down, or proves inadequate, the further development of the conjugate organism becomes impossible.