The first line of direct evidence in favour of organic evolution which I shall open is that which may be termed the argument from Classification.

It is a matter of observable fact that different forms of plants and animals present among themselves more or less pronounced resemblances. From the earliest times, therefore, it has been the aim of philosophical naturalists to classify plants and animals in accordance with these resemblances. Of course the earliest attempts at such classification were extremely crude. The oldest of these attempts with which we are acquainted—namely, that which is presented in the books of Genesis and Leviticus—arranges the whole vegetable kingdom in three simple divisions of Grass, Herbs, and Trees; while the animal kingdom is arranged with almost equal simplicity with reference, first to habitats in water, earth, or air, and next as to modes of progression. These, of course, were what may be termed common-sense classifications, having reference merely to external appearances and habits of life. But when Aristotle laboriously investigated the comparative anatomy of animals, he could not fail to perceive that their entire structures had to be taken into account in Order to classify them scientifically; and, also, that for this purpose the internal parts were of quite as much importance as the external. indeed, he perceived that they were of greatly more importance in this respect, inasmuch as they presented so many more points for comparison; and, in the result, he furnished an astonishingly comprehensive, as well as an astonishingly accurate classification of the larger groups of the animal kingdom. On the other hand, classification of the vegetable kingdom continued pretty much as it had been left by the book of genesis—all plants being divided into three groups, herbs, shrubs, and trees. Nor was this primitive state of matters improved upon till the sixteenth century, when gesner (1516-1565), and still more cÆsalpino (1519-1603), laid the foundations of systematic botany.

But the more that naturalists prosecuted their studies on the anatomy of plants and animals, the more enormously complex did they find the problem of classification become. Therefore they began by forming what are called artificial systems, in contradistinction to natural systems. An artificial system of classification is a system based on the more or less arbitrary selection of some one part, or set of parts; while a natural classification is one that is based upon a complete knowledge of all the structures of all the organisms which are classified.

Thus, the object of classification has been that of arranging organisms in accordance with their natural affinities, by comparing organism with organism, for the purpose of ascertaining which of the constituent organs are of the most invariable occurrence, and therefore of the most typical signification. A porpoise, for instance, has a large number of teeth, and in this feature resembles most fish, while it differs from all mammals. But it also gives suck to its young. Now, looking to these two features alone, should we say that a porpoise ought to be classed as a fish or as a mammal? Assuredly as a mammal; because the number of teeth is a very variable feature both in fish and mammals, whereas the giving of suck is an invariable feature among mammals, and occurs nowhere else in the animal kingdom. This, of course, is chosen as a very simple illustration. Were all cases as obvious, there would be but little distinction between natural and artificial systems of classification. But it is because the lines of natural affinity are, as it were, so interwoven throughout the organic world, and because there is, in consequence, so much difficulty in following them, that artificial systems have to be made in the first instance as feelers towards eventual discovery of the natural system. In other words, while forming their artificial systems of classification, it has always been the aim of naturalists—whether consciously or unconsciously—to admit as the bases of their systems those characters which, in the then state of their knowledge, seemed most calculated to play an important part in the eventual construction of the natural system. If we were dealing with the history of classification, it would here be interesting to note how the course of it has been marked by gradual change in the principles which naturalists adopted as guides to the selection of characters on which to found their attempts at a natural classification. Some of these changes, indeed, I shall have to mention later on; but at present what has to be specially noted is, that through all these changes of theory or principle, and through all the ever-advancing construction of their taxonomic science, naturalists themselves were unable to give any intelligible reason for the faith that was in them—or the faith that over and above the artificial classifications which were made for the mere purpose of cataloguing the living library of organic nature, there was deeply hidden in nature itself a truly natural classification, for the eventual discovery of which artificial systems might prove to be of more or less assistance.

LinnÆus, for example, expressly says—“You ask me for the characters of the natural orders; I confess that I cannot give them.” Yet he maintains that, although he cannot define the characters, he knows, by a sort of naturalist’s instinct, what in a general way will subsequently be found to be the organs of most importance in the eventual grouping of plants under a natural system. “I will not give my reasons for the distribution of the natural orders which I have published,” he said: “you, or some other person, after twenty or after fifty years, will discover them, and see that I was right.”

Thus we perceive that in forming their provisional or artificial classifications, naturalists have been guided by an instinctive belief in some general principle of natural affinity, the character of which they have not been able to define; and that the structures which they selected as the bases of their classifications when these were consciously artificial, were selected because it seemed that they were the structures most likely to prove of use in subsequent attempts at working out the natural system. This general principle of natural affinity, of which all naturalists have seen more or less well-marked evidence in organic nature, and after which they have all been feeling, has sometimes been regarded as natural, but more often as supernatural. Those who regarded it as supernatural took it to consist in a divine ideal of creation according to types, so that the structural affinities of organisms were to them expressions of an archetypal plan, which might be revealed in its entirety when all organisms on the face of the earth should have been examined. Those, on the other hand, who regarded the general principle of affinity as depending on some natural causes, for the most part concluded that these must have been utilitarian causes; or, in other words, that the fundamental affinities of structure must have depended upon fundamental requirements of function. According to this view, the natural classification would eventually be found to stand upon a basis of physiology. Therefore all the systems of classification up to the earlier part of the present century went upon the apparent axiom, that characters which are of most importance to the organisms presenting them must be characters most indicative of natural affinities. But the truth of the matter was eventually found to be otherwise. For it was eventually found that there is absolutely no correlation between these two things; that, therefore, it is a mere chance whether or not organs which are of importance to organisms are likewise of importance as guides to classification; and, in point of fact, that the general tendency in this matter is towards an inverse instead of a direct proportion. More often than not, the greater the value of a structure for the purpose of indicating natural affinities, the less is its value to the creatures presenting it.

Enough has now been said to show three things. First, that long before the theory of descent was entertained by naturalists, naturalists perceived the fact of natural affinities, and did their best to construct a natural system of classification for the purpose of expressing such affinities. Second, that naturalists had a kind of instinctive belief in some one principle running through the whole organic world, which thus served to bind together organisms in groups subordinate to groups—that is, into species, genera, orders, families, classes, sub-kingdoms, and kingdoms. Third, that they were not able to give any very intelligible reason for this faith that was in them; sometimes supposing the principle in question to be that of a supernatural plan of organization, sometimes regarding it as dependent on conditions of physiology, and sometimes not attempting to account for it at all.

Of course it is obvious that the theory of descent furnishes the explanation which is required. For it is now evident to evolutionists, that although these older naturalists did not know what they were doing when they were tracing these lines of natural affinity, and thus helping to construct a natural classification—I say it is now evident to evolutionists that these naturalists were simply tracing the lines of genetic relationship. The great principle pervading organic nature, which was seen so mysteriously to bind the whole creation together as in a nexus of organic affinity, is now easily understood as nothing more or less than the principle of Heredity. Let us, therefore, look a little more closely at the character of this network, in order to see how far it lends itself to this new interpretation.

The first thing that we have to observe about the nexus is, that it is a nexus—not a single line, or even a series of parallel lines. In other words, some time before the theory of descent was seriously entertained, naturalists for the most part had fully recognised that it was impossible to arrange either plants or animals, with respect to their mutual affinities, in a ladder-like series (as was supposed to be the type of classification by the earlier systematists), or even in map-like groups (as was supposed to be the type by LinnÆus). And similarly, also, with respect to grades of organization. In the case of the larger groups, indeed, it is usually possible to say that the members of this group as a whole are more highly organized than the members of that group as a whole; so that, for instance, we have no hesitation in regarding the Vertebrata as more highly organized than the Invertebrata, Birds than Reptiles, and so on. But when we proceed to smaller subdivisions, such as genera and species, it is usually impossible to say that the one type is more highly organized than another type. A horse, for instance, cannot be said to be more highly organized than a zebra or an ass; although the entire horse-genus is clearly a more highly organized type than any genus of animal which is not a mammal.

In view of these facts, therefore, the system of classification which was eventually arrived at before the days of Darwin, was the system which naturalists likened to a tree; and this is the system which all naturalists now agreed upon as the true one. According to this system, a short trunk may be taken to represent the lowest organisms which cannot properly be termed either plants or animals. This short trunk soon separates into two large trunks, one of which represents the vegetable and the other the animal kingdom. Each of these trunks then gives off large branches signifying classes, and these give off smaller, but more numerous branches, signifying families, which ramify again into orders, genera, and finally into the leaves, which may be taken to represent species. Now, in such a representative tree of life, the height of any branch from the ground may be taken to indicate the grade of organization which the leaves, or species, present; so that, if we picture to ourselves such a tree, we may understand that while there is a general advance of organization from below upwards, there are many deviations in this respect. Sometimes leaves growing on the same branch are growing at a different level—especially, of course, if the branch be a large one, corresponding to a class or sub-kingdom. And sometimes leaves growing on different branches are growing at the same level: that is to say, although they represent species belonging to widely divergent families, orders, or even classes, it cannot be said that the one species is more highly organized than the other.

Now, this tree-like arrangement of species in nature is an arrangement for which Darwin is not responsible. For, as we have seen, the detecting of it has been due to the progressive work of naturalists for centuries past; and even when it was detected, at about the commencement of the present century, naturalists were confessedly unable to explain the reason of it, or what was the underlying principle that they were engaged in tracing when they proceeded ever more and more accurately to define these ramifications of natural affinity. But now, as just remarked, we can clearly perceive that this underlying principle was none other than Heredity as expressed in family likeness,—likeness, therefore, growing progressively more unlike with remoteness of ancestral relationship. For thus only can we obtain any explanation of the sundry puzzles and apparent paradoxes, which a working out of their natural classifications revealed to botanists and zoologists during the first half of the present century. It will now be my endeavour to show how these puzzles and paradoxes are all explained by the theory that natural affinities are merely the expression of genetic affinities.

First of all, and from the most general point of view, it is obvious that the tree-like system of classification, which Darwin found already and empirically worked out by the labours of his predecessors, is as suggestive as anything could well be of the fact of genetic relationship. For this is the form that every tabulation of family pedigree must assume; and therefore the mere fact that a scientific tabulation of natural affinities was eventually found to take the form of a tree, is in itself highly suggestive of the inference that such a tabulation represents a family tree. If all species were separately created, there can be no assignable reason why the ideas of earlier naturalists touching the form which a natural classification would eventually assume should not have represented the truth—why, for example, it should not have assumed the form of a ladder (as was anticipated in the seventeenth century), or of a map (as was anticipated in the eighteenth), or, again, of a number of wholly unrelated lines, circles, &c. (as certain speculative writers of the present century have imagined). But, on the other hand, if all species were separately and independently created, it becomes virtually incredible that we should everywhere observe this progressive arborescence of characters common to larger groups into more and more numerous, and more and more delicate, ramifications of characters distinctive only of smaller and smaller groups. A man would be deemed insane if he were to attribute the origin of every branch and every twig of a real tree to a separate act of special creation; and although we have not been able to witness the growth of what we may term in a new sense the Tree of Life, the structural relations which are now apparent between its innumerable ramifications bear quite as strong a testimony to the fact of their having been due to an organic growth, as is the testimony furnished by the branches of an actual tree.

Or, to take another illustration. Classification of organic forms, as Darwin, Lyell, and HÄckel have pointed out, strongly resembles the classification of languages. In the case of languages, as in the case of species, we have genetic affinities strongly marked; so that it is possible to some extent to construct a Language-tree, the branches of which shall indicate, in a diagrammatic form, the progressive divergence of a large group of languages from a common stock. For instance, Latin may be regarded as a fossil language, which has given rise to a group of living languages—Italian, Spanish, French, and, to a large extent, English. Now what would be thought of a philologist who should maintain that English, French, Spanish, and Italian were all specially created languages—or languages separately constructed by the Deity, and by as many separate acts of inspiration communicated to the nations which now speak them—and that their resemblance to the fossil form, Latin, must be attributed to special design? Yet the evidence of the natural transmutation of species is in one respect much stronger than that of the natural transmutation of languages—in respect, namely, of there being a vastly greater number of cases all bearing testimony to the fact of genetic relationship.

But, quitting now this most general point of view—or the suggestive fact that what we have before us is a tree—let us next approach this tree for the purpose of examining its structure more in detail. When we do this, the fact of next greatest generality which we find is as follows.

In cases where a very old form of life has continued to exist unmodified, so that by investigation of its anatomy we are brought back to a more primitive type of structure than that of the newer forms growing higher up upon the same branch, two things are observable. In the first place, the old form is less differentiated than the newer ones; and, in the next place, it is seen much more closely to resemble types of structure belonging to some of the other and larger branches of the tree. The organization of the older form is not only simpler; but it is, as naturalists say, more generalized. It comprises within itself characters belonging to its own branch, and also characters belonging to neighbouring branches, or to the trunk from which allied branches spring. Hence it becomes a general rule of classification, that it is by the lowest, or by the oldest, forms of any two natural groups that the affinities between the two groups admit of being best detected. And it is obvious that this is just what ought to be the case on the theory of descent with divergent modification; while, upon the alternative theory of special creation, no reason can be assigned why the lowest or the oldest types should thus combine the characters which afterwards become severally distinctive of higher or newer types.

Again, I have already alluded to the remarkable fact that there is no correlation between the value of structures to the organisms which present them, and their value to the naturalist for the purpose of tracing natural affinity; and I have remarked that up to the close of the last century it was regarded as an axiom of taxonomic science, that structures which are of most importance to the animals or plants possessing them must likewise prove of most importance in any natural system of classification. On this account, all attempts to discover the natural classification went upon the supposition that such a direct proportion must obtain—with the result that organs of most physiological importance were chosen as the bases of systematic work. And when, in the earlier part of the present century, De Candolle found that instead of a direct there was usually an inverse proportion between the functional and the taxonomic value of a structure, he was unable to suggest any reason for this apparently paradoxical fact. For, upon the theory of special creation, no reason can be assigned why organs of least importance to organisms should prove of most importance as marks of natural affinity. But on the theory of descent with progressive modification the apparent paradox is at once explained. For it is evident that organs of functional importance are, other things equal, the organs which are most likely to undergo different modifications in different lines of family descent, and therefore in time to have their genetic relationships in these different lines obscured. On the other hand, organs or structures which are of no functional importance are never called upon to change in response to any change of habit, or to any change in the conditions of life. They may, therefore, continue to be inherited through many different lines of family descent, and thus afford evidence of genetic relationship where such evidence fails to be given by any of the structures of vital importance, which in the course of many generations have been required to change in many ways according to the varied experiences of different branches of the same family. Here, then, we have an empirically discovered rule in the science of classification, the raison d’Être of which we are at once able to appreciate upon the theory of evolution, whereas no possible explanation of why it should ever have become a rule could be furnished upon the theory of special creation.

Here, again, is another empirically determined rule. The larger the number, as distinguished from the importance, of structures which are found common to different groups, the greater becomes their value as guides to the determination of natural affinity. Or, as Darwin puts it, “the value of an aggregate of characters, even when none are important, alone explains the aphorism enunciated by LinnÆus, namely, that the characters do not give the genus, but the genus gives the characters; for this seems founded on the appreciation of many trifling points of resemblance, too slight to be defined[1].”

Now it is evident, without comment, of how much value aggregates of characters ought to be in classification, if the ultimate meaning of classification be that of tracing lines of pedigree; whereas, if this ultimate meaning were that of tracing divine ideals manifested in special creation, we can see no reason why single characters are not such sure tokens of a natural arrangement as are aggregates of characters, even though the latter be in every other respect unimportant. For, on the special creation theory, we cannot explain why an assemblage, say of four or five trifling characters, should have been chosen to mark some unity of plan, rather than some one character of functional importance, which would have served at least equally well any such hypothetical purpose. On the other hand, as Darwin remarks, “we care not how trifling a character may be—let it be the mere inflection of the angle of the jaw, the manner in which an insect’s wing is folded, whether the skin be covered with hair or feathers—if it prevail throughout many and different species, especially those having very different habits of life, it assumes high value; for we can account for its presence in so many forms, with such different habits, only by inheritance from a common parent. We may err in this respect in regard to single points of structure, but when several characters, let them be ever so trifling, concur throughout a large group of beings having different habits, we may feel almost sure, on the theory of descent, that these characters have been inherited from a common ancestor; and we know that such aggregated characters have especial value in classification[2].”

It is true that even a single character, if found common to a large number of forms, while uniformly absent from others, is also regarded by naturalists as of importance for purposes of classification, although they recognise it as of a value subordinate to that of aggregates of characters. But this also is what we should expect on the theory of descent. If even any one structure be found to run through a number of animals presenting different habits of life, the readiest explanation of the fact is to be found in the theory of descent; but this does not hinder that if several such characters always occur together, the inference of genetic relationship is correspondingly confirmed. And the fact that before this inference was ever drawn, naturalists recognised the value of single characters in proportion to their constancy, and the yet higher value of aggregates of characters in proportion to their number—this fact shows that in their work of classification naturalists empirically observed the effects of a cause which we have now discovered, to wit, hereditary transmission of characters through ever-widening groups of changing species.

There is another argument which appears to tell strongly in favour of the theory of descent. We have just seen that non-adaptive structures, not being required to change in response to change of habits or conditions of life, are allowed to persist unchanged through many generations, and thus furnish exceptionally good guides in the science of classification—or, according to our theory, in the work of tracing lines of pedigree. But now, the converse of this statement holds equally true. For it often happens that adaptive structures are required to change in different lines of descent in analogous ways, in order to meet analogous needs; and, when such is the case, the structures concerned have to assume more or less close resemblances to one another, even though they have severally descended from quite different ancestors. The paddles of a whale, for instance, most strikingly resemble the fins of a fish as to their outward form and movements; yet, on the theory of descent, they must be held to have had a widely different parentage. Now, in all such cases where there is thus what is called an analogous (or adaptive) resemblance, as distinguished from what is called an homologous (or anatomical) resemblance—in all such cases it is observable that the similarities do not extend further into the structure of the parts than it is necessary that they should extend, in order that the structures should both perform the same functions. The whole anatomy of the paddles of a whale is quite unlike that of the fins of a fish—being, in fact, that of the fore-limb of a mammal. The change, therefore, which the fore-limb has here undergone to suit it to the aquatic habits of this mammal, is no greater than was required for that purpose: the change has not extended to any one feature of anatomical significance. This, of course, is what we should expect on the theory of descent with modification of ancestral characters; but on the theory of special creation it is not intelligible why there should always be so marked a distinction between resemblances as analogical or adaptive, and resemblances as homological or of meaning in reference to a natural classification. To take another and more detailed instance, the Tasmanian wolf is an animal separated from true wolves in a natural system of classification. Yet its jaws and teeth bear a strong general resemblance to those of all the dog tribe, although there are differences of anatomical detail. In particular, while the dogs all have on each side of the upper jaw four pre-molars and two molars, the Tasmanian wolf has three pre-molars and four molars. Now there is no reason, so far as their common function of dealing with flesh is concerned, why the teeth of the Tasmanian wolf should not have resembled homologically as well as analogically the teeth of a true wolf; and therefore we cannot assign any intelligible reason why, if all the species of the dog genus were separately created with one pattern of teeth, the unallied Tasmanian wolf should have been furnished with what is practically the same pattern from a functional point of view, while differing from a structural point of view. But, of course, on the theory of descent with modification, we can well understand why similarities of habit should have led to similarities of structural appearance of an adaptive kind in different lines of descent, without there being any trace of such real or anatomical similarities as could possibly point to genetic relationship.

Lastly, to adduce the only remaining argument from classification which I regard as of any considerable weight, naturalists have found it necessary, while constructing their natural classifications, to set great store on what Mr. Darwin calls “chains of affinities.” Thus, for instance, “nothing can be easier than to define a number of characters common to all birds; but with crustaceans any such definition has hitherto been found impossible. There are crustaceans at the opposite ends of the series, which have hardly a character in common; yet the species at both ends, from being plainly allied to others, and these to others, and so onwards, can be recognised as unequivocally belonging to this, and to no other class of the articulata[3].” Now it is evident that this progressive modification of specific types—where it cannot be said that the continuity of resemblance is anywhere broken, and yet terminates in modification so great that but for the connecting links no one could divine a natural relationship between the extreme members of the series,—it is evident that such chains of affinity speak most strongly in favour of a transmutation of the species concerned, while it is impossible to suggest any explanation of the fact in terms of the rival theory. For if all the links of such a chain were separately forged by as many acts of special creation, we can see no reason why B should resemble A, C resemble B, and so on, but with ever slight though accumulating differences, until there is no resemblance at all between A and Z.

I hope enough has now been said to show that all the general principles and particular facts appertaining to the natural classification of plants and animals, are precisely what they ought to be according to the theory of genetic descent; while no one of them is such as might be—and, indeed, used to be—expected upon the theory of special creation. Therefore, the only possible way in which all this uniform body of direct evidence can be met by a supporter of the latter theory, is by falling back upon the argument from ignorance. We do not know, it may be said, what hidden reasons there may have been for following all these general principles in the separate creation of specific types. Now, it is evident that this is a form of argument which admits of being brought against all the actual—and even all the possible—lines of evidence in favour of evolution. Therefore I deem it desirable thus early in our proceedings to place this argument from ignorance on its proper logical footing.

If there were any independent evidence in favour of special creation as a fact, then indeed the argument from ignorance might be fairly used against any sceptical cavils regarding the method. In this way, for example, Bishop Butler made a legitimate use of the argument from ignorance when he urged that it is no reasonable objection against a revelation, otherwise accredited, to show that it has been rendered in a form, or after a method, which we should not have antecedently expected. But he could not have legitimately employed this argument, except on the supposition that he had some independent evidence in favour of the revelation; for, in the absence of any such independent evidence, appeal to the argument from ignorance would have become a mere begging of the question, by simply assuming that a revelation had been made. And thus it is in the present case. A man, of course, may quite legitimately say, Assuming that the theory of special creation is true, it is not for us to anticipate the form or method of the process. But where the question is as to whether or not the theory is true, it becomes a mere begging of this question to take refuge in the argument from ignorance, or to represent in effect that there is no question to be discussed. And if, when the form or method is investigated, it be found everywhere charged with evidence in favour of the theory of descent, the case becomes the same as that of a supposed revelation, which has been discredited by finding that all available evidence points to a natural growth. In short, the argument from ignorance is in any case available only as a negative foil against destructive criticism: in no case has it any positive value, or value of a constructive kind. Therefore, if a theory on any subject is destitute of positive evidence, while some alternative theory is in possession of such evidence, the argument from ignorance can be of no logical use to the former, even though it maybe of such use to the latter. For it is only the possession of positive evidence which can furnish a logical justification of the argument from ignorance: in the absence of such evidence, even the negative value of the argument disappears, and it then implies nothing more than the gratuitous assumption of a theory.

I will now sum up the various considerations which have occupied us during the present chapter.

First of all we must take note that the classification of plants and animals in groups subordinate to groups is not merely arbitrary, or undertaken only for a matter of convenience and nomenclature—such, for instance, as the classification of stars in constellations. On the contrary, the classification of a naturalist differs from that of an astronomer, in that the objects which he has to classify present structural resemblances and structural differences in numberless degrees; and it is the object of his classification to present a tabular statement of these facts. Now, long before the theory of evolution was entertained, naturalists became fully aware that these facts of structural resemblances running through groups subordinate to groups were really facts of nature, and not merely poetic imaginations of the mind. No one could dissect a number of fishes without perceiving that they were all constructed on one anatomical pattern, which differed considerably from the equally uniform pattern on which all mammals were constructed, even although some mammals bore an extraordinary resemblance to fish in external form and habits of life. And similarly with all the smaller divisions of the animal and vegetable kingdoms. Everywhere investigation revealed the bonds of close structural resemblances between species of the same genus, resemblance less close between genera of the same family, resemblance still less close between families of the same order, resemblance yet more remote between orders of the same class, and resemblance only in fundamental features between classes of the same sub-kingdom, beyond which limit all anatomical resemblance was found to disappear—the different sub-kingdoms being formed on wholly different patterns. Furthermore, in tracing all these grades of structural relationship, naturalists were slowly led to recognise that the form which a natural classification must eventually assume would be that of a tree, wherein the constituent branches would display a progressive advance of organization from below upwards.

Now we have seen that although this tree-like arrangement of natural groups was as suggestive as anything could well be of all the forms o£ life being bound together by the ties of genetic relationship, such was not the inference which was drawn from it. Dominated by the theory of special creation, naturalists either regarded the resemblance of type subordinate to type as expressive of divine ideals manifested in such creation, or else contented themselves with investigating the facts without venturing to speculate upon their philosophical import. But even those naturalists who abstained from committing themselves to any theory of archetypal plans, did not doubt that facts so innumerable and so universal must have been due to some one co-ordinating principle—that, even though they were not able to suggest what it was, there must have been some hidden bond of connexion running through the whole of organic nature. Now, as we have seen, it is manifest to evolutionists that this hidden bond can be nothing else than heredity; and, therefore, that these earlier naturalists, although they did not know what they were doing, were really tracing the lines of genetic descent as revealed by degrees of structural resemblance,—that the arborescent grouping of organic forms which their labours led them to begin, and in large measure to execute, was in fact a family tree of life.

Here, then, is the substance of the argument from classification. The mere fact that all organic nature thus incontestably lends itself to a natural arrangement of group subordinate to group, when due regard is paid to degrees of anatomical resemblance—this mere fact of itself tells so weightily in favour of descent with progressive modification in different lines, that even if it stood alone it would be entitled to rank as one of our strongest pieces of evidence. But, as we have seen, it does not stand alone. When we look beyond this large and general fact of all the innumerable forms of life being thus united in a tree-like system by an unquestionable relationship of some kind, to those smaller details in the science of classification which have been found most useful as guides for this kind of research, then we find that all these details, or empirically discovered rules, are exactly what we should have expected them to be, supposing the real meaning of classification to have been that of tracing lines of pedigree.

In particular, we have seen that the most archaic types are both simpler in their organization and more generalized in their characters than are the more recent types—a fact of which no explanation can be given on the theory of special creation. But, upon the theory of natural evolution, we can without difficulty understand why the earlier forms should have been the simpler forms, and also why they should have been the most generalized. For it is out of the older forms that the newer must have grown; and, as they multiplied, they must have become more and more differentiated.

Again, we have seen that there is no correlation between the importance of any structure from a classificatory point of view, and the importance of that structure to the organism which presents it. On the contrary, it is a general rule that “the less any part of the organization is concerned with special habits, the more important it becomes for classification.” Now, from the point of view of special creation it is unintelligible why unity of ideal should be most manifested by least important structures, whereas from the point of view of evolution it is to be expected that these life-serving structures should have been most liable to divergent modification in divergent lines of descent, or in adaptation to different conditions of life, while the trivial or less important characters should have been allowed to remain unmodified. Thus we can now understand why all primitive classifications were wrong in principle when they went upon the assumption that divine ideals were best exhibited by resemblances between life-serving (and therefore adaptive) structures, with the result that whales were classed with fishes, birds with bats, and so on. Nevertheless, these primitive naturalists were quite logical; for, from the premises furnished by the theory of special creation, it is much more reasonable to expect that unity of ideal should be shown in plainly adaptive characters than in trivial and more or less hidden anatomical characters. Moreover, long after biological science had ceased consciously to follow any theological theory, the apparent axiom continued to be entertained, that structures of most importance to organisms must also be structures of most importance to systematists. And when at last, in the present century, this was found not to be the case, no reason could be suggested why it was not the case. But now we are able fully to explain this apparent anomaly.

Once more, we have seen that aggregates of characters presenting resemblances to one another have always been found to be of special importance as guides to classification. This, of course, is what we should have expected, if the real meaning of classification be that of tracing lines of pedigree; but on the theory of special creation no reason can be assigned why single characters are not such sure tokens of a natural arrangement as are aggregates of characters, however trivial the latter may be. For it is obvious that unity of ideal might have been even better displayed by everywhere maintaining the pattern of some one important structure, than by doing so in the case of several unimportant structures. Take an analogous instance from human contrivances. Unity of ideal in the case of gun-making would be shown by the same principles of mechanism running through all the different sizes and shapes of gun-locks, rather than by the ornamental patterns engraved upon the outside. Yet it must be supposed that in the mechanisms assumed to have been constructed by special creation, it was the trivial details rather than the fundamental principles of these mechanisms which were chosen by the Divinity to display his ideals.

And this leads us to the next consideration—namely, that when in two different lines of descent animals happen to adopt similar habits of life, the modifications which they undergo in order to fit them for these habits often induces striking resemblances of structure between the two animals, as in the case of whales and fish. But in all such instances it is invariably found that the resemblance is only superficial and apparent: not anatomical or real. In other words, the resemblance does not extend further than it is necessary that it should, if both sets of organs are to be adapted to perform the same functions. Now this, again, is just what one would expect to find as the universal rule on the theory of descent, with modification of ancestral characters. But, on the opposite theory of special creation, I know not how it is to be explained that among so many instances of close superficial resemblance between creatures belonging to different branches of the tree of life, there are no instances of any real or anatomical resemblance. So far as their structures are adapted to perform a common function, there is in all such cases what may be termed a deceptive appearance of some unity of ideal; but, when carefully examined, it is always found that two apparently identical structures occurring on different branches of the classificatory tree are in fact fundamentally different in respect of their structural plan.

Lastly, we have seen that one of the guiding principles of classification has been empirically found to consist in setting a high value on “chains of affinities.” That is to say, naturalists not unfrequently meet with a long series of progressive modifications of type, which, although it cannot be said that the continuity is anywhere broken, at last leads to so much divergence of character that, but for the intermediate links, the members at each end of the chain could not be suspected of being in any way related. Well, such cases of chains of affinity obviously tell most strongly in favour of descent with continuous modification; while it is impossible to suggest why, if all the links were separately forged by as many acts of special creation, there should have been this gradual transmutation of characters carried to the point where the original creative ideal has been so completely transformed that, but for the accident of the chain being still complete, no one of nature’s interpreters could possibly have discovered the connexion. For, as we have seen, this is not a case in which any appeal can be logically made to the argument from ignorance of divine method, unless some independent evidence could be adduced in favour of special creation. And that no such independent evidence exists, it will be the object of future chapters to show.