Chapter XV: Part 15

The case would be the same if an island were suddenly peopled with the whole snail-fauna of a neighbouring continent, with which a land connexion had arisen. If the island had previously been free from snails, all the species of the mainland would be able to exist there in so far as they were able to find suitable conditions of life, but each species would speedily take complete possession of the area peculiarly suited to it, so that none of their fellow migrants would be impelled, or would even find it possible, to adapt themselves to new conditions and thus to become variable and split up into varieties. If Ireland were at present free from snails, and if a land connexion between it and England came about, then the snail-fauna of England would probably migrate quite unvaried to Ireland, and in point of fact the snail-fauna of the two islands, which were formerly connected, is almost the same. For the same reason the fauna of England, as far as terrestrial snails are concerned, is almost the same as that of Germany.

On the other hand, it may be almost regarded as a law that an individual migrant to virgin territory must become variable. This could not be better illustrated than by the geographical distribution of terrestrial snails, which emphasizes the fact that a striking wealth of endemic species is to be found on all oceanic islands. Moreover, the fact that the number of these endemic species is greater in proportion to the distance of the island from the continent, indicates that the variability sets in more intensively and lasts longer in proportion to the small number of species which become immigrants in the island, and in proportion to the number of unoccupied areas which are open to the descendants of the immigrant species. This is undoubtedly the reason why the Sandwich Islands do not possess _a single species_ which occurs elsewhere, and the segregation of the unknown ancestral form into many species and several (four) sub-genera is also to be interpreted in the same way. There was probably in this case only one immigrant species, which found a free field, and adapted itself in its descendants to all the conditions of snail-life which obtained there, and in doing so split up into numerous and somewhat markedly divergent forms. But the number of different forms is much greater than the number of distinctive habitats, as Gulick indicates and substantiates in detail, for similar areas, if they are relatively isolated from one another, are inhabited not by the same forms, but by different though nearly related varieties, and this depends on the fact that from the species which was in process of varying a different combination of variations would be sent out at different periods, and the temporary isolation would result in the evolution of special local varieties.

But I do not believe that this would continue for all time. I rather think that these--let us say--representative varieties would diminish in numbers in the course of a long period. For the isolation of single valley-slopes or of particular woods is not permanent, individuals are liable to be carried from one to another in the course of centuries as they were at the beginning of the colonization of the isolated woods; forests are cleared or displaced by geological changes, connexions are formed between places, which were formerly separated, and in the course of another geological period the number of representative varieties, and probably even of species, will have diminished considerably,--the former will have been fused together, the latter in part eliminated. Even now Gulick speaks regretfully of the decimation of rare local forms by their chief enemies, the mice.

But even if the number of endemic forms in insular regions diminishes from the time when they were first fully taken possession of, it nevertheless remains a very high one, for even now Madeira possesses 104 endemic terrestrial snails, the Philippines have more than the whole of India, and the Antilles as many as the whole American continent.

Many naturalists believe that each isolated variety must diverge further and further from its nearest relatives as time goes on. Although I entirely admit that this is possible, for I have endeavoured to show that variational tendencies which have once arisen in the germ-plasm go on in the same direction until they are brought to a full stop in some way or other, yet I cannot admit that this must always be so. The species which has been carried to a strange area need not always contain particular variational tendencies in its germ-plasm, and need not in every case be impelled to such variations by the influence of new conditions. We know species which have made their way into new regions, and, without varying at all, have held their own with, or even proved superior to, the species which were already settled there. Many cases of this kind are known, both among plants and animals; these have been brought by man, intentionally or by chance, from one continent to another, and have established themselves and spread over the new area. I need only recall the evening primrose (_Œnothera biennis_[28]), whose fatherland is Virginia, but whose beautiful big yellow blossoms now display themselves beside nearly every river in Germany, having migrated stream-upwards along the gravelly soil; or the troublesome weed (_Erigeron canadense_), which is now scarcely less common in our gardens than in those of Canada; or the sparrow (_Passer domesticus_), which was introduced into the United States to destroy the caterpillars, but which preferred instead to plunder the rich stores of corn, and in consequence of these favourable conditions increased to such an extent that it has now become a veritable pest, all imaginable means for its extirpation having been tried--as yet, however, with no great results.

[28] This was written before the appearance of the researches which De Vries has made on the variations of _Œnothera_ in Europe. Thus the illustration may not be quite apposite, for it seems to remain undetermined whether the 'mutations' which occur in Holland do not also occasionally appear in America. See end of lecture xxxiii.

In all these cases the migration is certainly of recent date, and it is quite possible that, when a longer time has elapsed, some variations will take place in the new home, but in any case these instances prove that an immigrant species can spread over its new area without immediately varying.

Similarly, it must be admitted that species which have belonged to two continents ever since Tertiary times need not have diverged since that time, and we know, for instance, thirty-two species of nocturnal Lepidoptera which are common to North America and to Europe and yet exhibit no differences, while twenty-seven other nocturnal Lepidoptera are, according to Grote, represented in America by 'vicarious' species, that is, by species which have varied slightly in one or other of the two areas, perhaps in both.

To sum up: we must undoubtedly admit that isolation has a considerable influence in the evolution of species, though only in association with selection in its various grades and modes, especially germinal selection, natural selection, and sexual selection. We can say generally that each grade and mode of selection will more readily lead to the transformation if it be combined with isolation. Thus germinal selection may call forth slight divergences in colour and marking, which will be permanent if the individuals concerned are in an isolated region. In isolation these variations will increase undisturbed, and in some circumstances will be intensified by sexual selection, so that the male sex will vary alone in the first place, though the female may follow, so that ultimately the whole species will be transformed. Finally, the most marked effect of isolation is seen when individual members of a species are transferred to virgin territory which offers unoccupied areas, suitable not to one particular species alone, but to many nearly related species, so that the immigrant colony can adapt itself to all the different possibilities of life, and develop into a whole circle of species. But we saw that such an aftergrowth of new forms, whether varieties, species, or even genera, may far exceed the number of different kinds of localities, if there be relative isolation between the different groups of immigrants within the insular region, as happens in the case of slow-moving animals like the terrestrial snails, or of small singing-birds, to which each island of a little archipelago is a relatively isolated region (Galapagos).

We may thus fully recognize the importance of local isolation without regarding the absence of crossing with the members of the species in the original habitat as the sole cause of species-formation, without setting 'isolation' in the place of the processes of selection. These last, taken in the wide sense, always remain the indispensable basis of all transformations, but they certainly do not operate only in the form of personal selection, but, wherever indifferent characters are concerned, in that of germinal selection. Here, too, we see the possibility of reconciliation with those naturalists who regard transformations as primarily dependent upon internal forces of development. The fact is that _all variations depend upon internal causes_, and their course must be guided by forces which work in an orderly way. But the actual co-operation of all these forces and variations is not predetermined, but depends to a certain extent upon chance, for of the possible modes of evolution the one which gains the upper hand in the play of forces at the moment is alone followed, the better are everywhere preferred, from the most minute vital units of the germ-plasm, up to the struggle between individuals and between species.

LECTURE XXXIII

ORIGIN OF THE SPECIFIC TYPE

Transition species of Celebes snails, according to Sarasin--Possible variations in the shell due to nutrition--Natural selection plays a part--Germinal selection--Temporary transitions between species--The fresh-water snails of Steinheim--How do sharply-defined species arise?--Nägeli's Developmental Force--The species a complex of adaptations--Adaptive differences between species--Adaptive nature of specific characters--The case of Cetaceans--Of birds--Additional note: the observations and theories of De Vries.

Our study of the influence which geographical isolation may have in transforming old and giving rise to new forms of life has led us naturally to a much more important problem, that of the origin of species as more or less sharply defined groups of forms, and I wish to make the transition to this problem by discussing another case of species-splitting effected in association with, or, as is usually said, _through_ isolation. The naturalists Paul and Fritz Sarasin, well known through their excellent studies on many components of the tropical fauna, have published in their latest work interesting discoveries in regard to the terrestrial snails of Celebes. These observations show that on this island a great transformation of snails has taken place, even since the later Tertiary period. A large number of new species of snail have arisen on this island since that time, and this, as the authors show to be probable, in association with the receding of the sea, that is, with the elevation of the island further out of the water, and thus with the increase of its surface. The modern terrestrial snails show chains of forms connected in many ways so that a series of species is connected by transition forms, and therefore does not really consist of separate species at all, although the extremes would seem to be separate species if they were studied by themselves without taking the transition forms into account. The state of things is exactly as if a Tertiary snail had spread from any small area over the whole island, and had been transformed slowly and in a definite direction in accordance with its distance from its starting-point. It is thus that we must interpret this discovery; we have here, beside each other in space, and indeed often disposed along geographical lines, the individual stages of a phyletic process of transformation, which has reached different levels at different places. One of the longest of these chains of forms is that of _Nanina cincta_, which runs across the island from east to west, and, beginning with the smallest and most delicate forms, ascends through many intermediate stages to the giant form _N. limbifera_. Such chains of forms have been previously recognized; thus Kobelt described one in the case of the Sicilian land-snails of the genus _Iberus_, and other cases are recorded in literature, but in all instances they refer to areas which must be regarded as isolated for the snails, and which have been colonized from a single starting-point.

We have now to inquire whether and how we can explain the origin of these chains of forms. The cousins Sarasin tell us how they at first attempted to refer the differences between the individual links of such a chain to the diverse influence of the external conditions of life, but in vain; neither the height above sea-level nor the character of the soil was sufficient, and natural selection was no more so; 'for why should a high _Obba_-form twisted like a beehive be either better or worse equipped for the struggle for existence than a smaller and flatter one?' It is true that we do not understand why, but this does not seem to me any reason to doubt that natural selection should be regarded as one of the causes of the divergence of these species, for we could not answer the same question in regard to any of the other structural differences between two species of snail, for the simple reason that we have far too little knowledge of the biological value of the parts of a snail. Or could any one tell of what use it would be to a snail-species to have the horns slightly longer, the foot somewhat narrower, the radula beset with rather larger or more numerous teeth? We might indeed imagine many ways in which it might be of advantage, but we are not in a position to say definitely why, for instance, longer horns should be better for one species than for another, and yet we do not believe that the structure of snails is less well adapted to the life of each species than that of any other animals. The snail's structure is certainly built up of hundreds and thousands of adaptations, like that of every other animal species, but while in many others we can, at least in part, recognize the adaptations as such, we cannot do so at all in regard to the snail. Simroth has pointed out that the spiral asymmetrical shell bears a relation to the one-sided opening of the genital organs, but that only states the general reason for the coiling of the shell. In studying the differences in the shell one is apt to think of its external appearance alone, of the protection which it affords to the soft internal organs of the easily wounded animal; perhaps also of the distribution of weight, which must be different in a high tower-like structure and a low flat spiral; possibly, too, of the varied obstacles and resistance the snail has to encounter in creeping into clefts and holes, or among a tangle of plants, according to the form of its shell; but is it not also conceivable that the form of the shell has been determined by its contents? As Rudolph Leuckart taught, the snail may be regarded as composed of two parts, one of which is formed by the head and foot, the other by the so-called 'visceral sac': the former may be called the animal half, because it chiefly contains the dominant organs--the nerve-centres, almost the whole mass of muscle, and the sense-organs; the latter the vegetative half, since it contains the main mass of the nutritive and reproductive systems--the stomach and intestine, the large liver, the heart, the kidneys, the reproductive organs, and so on. The vegetative half of the animal is always concealed within the shell; would not therefore any great variation in the size of liver, stomach, intestine, and so on, bring with it a variation in the size and form of the shell, as well as in the expansion or contraction of its coils? And might not such variations become necessary because of some change in the food-supply? It is only a supposition, but it seems to me very probable that becoming accustomed to a new diet, less easily broken up and dissolved and of diminished nutritive value, would cause modification not only of the radula and jaw-plate, but also of the stomach and the liver, the intestine and the kidneys, whose activity is closely associated. The stomach must become more voluminous, the liver which yields the digestive fluid must become more massive, and so forth. I will not follow this hypothetical example further, for I merely wished to recall the fact that the snail shell, to the form of which no biological significance can be commonly attributed, is actually a sort of external cast of the visceral sac, and consequently dependent on the variations to which that is liable in accordance with the conditions of its life. To give precise proofs for such processes is certainly not yet possible, for we do not even know with certainty what the diet of the various species of snail is, much less the difference between the modes of nutrition in two varieties, or the nutritive value of the materials used, or the changes in secretion, absorption, assimilation, and excretion which must be brought about by these differences. But we can at least see that variations in nutrition must be enough in themselves to give rise to new adaptations in the size, constitution, and mutual adaptation of the internal vegetative organs, and we cannot overlook the possibility that the form and size of the vegetative half, and therefore the form and size of its secretion, the shell, may also be caused to vary[29]. The fact that we cannot recognize, for instance, the beehive shape of an _Obba_ as an adaptation, is thus no proof that it is not one. But let us assume for the moment that it is not, and that it cannot be referred to natural selection any more than the other variations in the Celebes chain of forms, and we may further admit that they cannot be referred to sexual selection, still less to some 'inherent principle of perfecting,' not only because there is no question of perfecting in the matter, but because such a mystical principle is outside of the scope of natural history and its principles of interpretation.

[29] That this suggestion was not unjustified is evident from a recent contribution by Simroth ('Ueber die Raublungenschnecken,' _Naturwissenschaftliche Wochenschrift_, December 8 and 15, 1901). In this paper the author, who is an expert as regards the biology of Gastropods, shows that a change of diet may evoke many kinds of changes in the structure of the food-canal, which may indirectly compel changes in the shell. Thus in a small indigenous snail, _Daudebardia_, the pharynx has grown so enormously in thickness and length in adaptation to the predatory mode of life, that the head and the anterior part of the body can no longer be retracted within the shelter of the shell. For this reason, and also because of the snail's habit of following earthworms into their burrows, the shell has been shunted far back and obliquely downwards. It has at the same time markedly changed in its shape, as may still be verified by comparing the form of the shell in the young stages with that of the adult.

But that transformations in a definite direction can and must arise from fresh disturbances in the equilibrium of the determinant system, that is from germinal selection, we have already shown.

Even if the changes of form with which we are here dealing had really no biological importance, they might quite well have been brought about by germinal selection, and only one thing remains obscure, that is, why the different stages on the path of distribution of a species are at a different level of evolution and not all at the same level. Why have not all been transformed? Why have some colonies remained near the ancestral form, while others have varied only a little, and others again a great deal? This cannot be explained by any assumption of an internal power of development, and the explanation can only be found in germinal selection associated with isolation, since the internal processes in the germ-plasm can quite well run a different course in different colonies. Nevertheless I am inclined to infer from these differences in the individual colonies of these chains of forms, that natural selection in the accepted sense has also played a part in the evolution of these snail varieties.

Such series of forms are especially interesting, because they show us the process of species-formation in its different stages beside each other in space, and thus simultaneously. They represent, so to speak, a horizontal branch of the genealogical tree of the species, as the Sarasins well express it, that is, a series of species arising from each other, which do not break off, but are all capable of life at the same time, and so exist simultaneously on different areas; they are species adapted to different localities, not to different times. The same is true of the snails of other isolated regions, except that the chains of forms are usually not simple, but split up into several chains of forms arising from one ancestral form, and under certain circumstances each of these may break up into two or more diverging series. The great number of related species in Madeira, or in the Sandwich Islands, compels us to this assumption, although the branching of the genealogical trees can no longer be demonstrated with certainty.

This splitting up of forms into several series on a varied insular region shows us once more that it is germinal selection alone which forms the basis of all transformations, and that there is not, as earlier naturalists, especially the botanists Nägeli and Askenasy, maintained, any peculiar impelling Force of Development innate in organisms. If there were such a force, a species would be obliged to go on continuously in the same direction, exactly like the Sarasins' chains of forms, but no breaking of the species into one or many forms could occur. But this breaking up into series is easy to understand when we take germinal selection into consideration, for the germ-plasm contains many ids and determinants, and each of these can enter upon new variations, so that one colony can vary in this direction, another in that, and a great diversity of forms living in isolation must, or at least may be the result, as we see in the case of the Sandwich Islands.

Let us delay a moment over the Sarasins' case of the Celebes snails. We are dealing here with series of forms in regard to which the ordinary conception of species fails us, for they contain varieties whose extremes are as far apart as distinct species usually are, which are not, however, distinct, since they are connected with one another by one and often by several transition forms. Thus we can only break them up into two or more 'species' by an arbitrary division at one place or other. The phenomenon itself is not new to us; we have seen that even Lamarck and Treviranus made use of similar series of forms, connected by transition stages, in their attack upon the old theory of creation, and sought to prove by means of these that the idea of species is an artificial one, read into nature by man, and not innate in nature, and that the forms of life were only apparently fixed and sharply defined, being in reality in process of slow transformation. Such beautiful and convincing examples as we now possess were not available at that time, but it might even then be said that it was the easier to make a new species the fewer examples one had to deal with, and the more difficult the more numerous these became, because with the number of individuals, especially if they come from a wide area, the number and diversity of the divergences increases also, so that in many cases, as in that of the Celebes snails, it becomes impossible to draw a line between the different species.

There are, however, many animal and plant forms which do not show such marked divergences, but rather exhibit a great harmony of individuals even in detail, and the conception of species is more readily applied to these. It would certainly be foolish to give it up, since we should then lose all possibility of arriving at any sort of orientation among the enormous wealth of forms in nature. But at the same time we must not forget that these 'typical' species only appear so to our short-sighted vision--short-sighted as far as time is concerned--and that they are connected from long-past times with 'species' which lived at an earlier date, by just such transition stages as connect the Celebes species of to-day, which are all living at the same time. The world of life on the earth only presents at any given time a 'cross-section of its genealogical tree,' and according as its branches grow out vertically or horizontally we receive an impression of typical, sharply defined species or of circles or chains of forms. In the first case the evolution of new species was associated with the dying out of the horizontal branches, and the end-twigs of the branch stand beside each other now apparently isolated and sharply defined; in the other case only a portion of the ancestral species has been transmuted, and the other part continues to live alongside of the species derived from it, and perhaps repeats the process of giving off a varied race of descendants.

The last thirty years have yielded much palæontological evidence of the successive stages of species-transformation. In quietly deposited horizontal strata of the earth's crust, lying one above another, the whole phyletic history of a group of snail-species has repeatedly been found in historic order, the oldest in the deepest layer, the youngest in the uppermost, and the numerous and often very divergent 'species' of a particular deposit are connected by transition forms in the intermediate strata. From the point of view of time, therefore, these are not 'typical' species, but circles of forms in a state of variability.

The most beautiful of such cases are the _Planorbis_ species from the small lacustrine deposits of Steinheim in Swabia, the _Paludina_ strata of Slavonia, and various groups of Ammonites.

These cases have been described and discussed so often that I need only refer to their most essential features.

The _Planorbis_ strata of Steinheim were first investigated, from the point of view of the theory of descent, by Hilgendorf (1866). He described nineteen different varieties, which, as they are all connected in chronological succession with each other, he grouped together under the name of _Planorbis multiformis_. These little freshwater snails are found in millions in the strata of the former lake-basin of Steinheim, and they are arranged in so orderly and regular a manner that two observers, working independently and at different times, succeeded in building up the genealogical tree in almost the same way. According to Alpheus Hyatt, the later investigator, all the forms are derived from one ancestral form, _Planorbis lævis_, from which four different series have descended, one of them splitting up again into three subordinate series.

All the individual members of these series are connected by intermediate forms in such a manner that a long period of constancy of forms seems to be succeeded by a shorter period of transformation, from which again a relatively constant form arises.

We see, therefore, that the idea of species is fully justified in a certain sense; we find indeed at certain times a breaking up of the fixed specific type, the species becomes variable, but soon the medley of forms clears up again and a new constant form arises--_a new species_, which remains the same for a long series of generations, until ultimately it too begins to waver, and is transformed once more. But if we were to place side by side the cross-sections of this genealogical tree at different levels, we should only see several well-defined species between which no intermediate forms could be recognized; these would only be found in the intermediate strata.

The problem we have now to discuss is, how it comes about that relatively sharply defined species exist which are connected with ancestral forms further back, but which form among themselves an exclusive, more or less homogeneous, host of individuals. How does it happen that we everywhere find a specific type, and not an endless number of individual forms connected with one another in all directions?

This would require no further explanation if a phyletic evolutionary force impelled the forms of life to vary in a definite manner, and thus to become transmuted into new forms in the course of generations. In that case the whole genealogical tree of the organisms on the earth must have been potentially contained in the lowest moneron, so that, given time and the most indispensable general conditions of existence, the living world just as we know it must have resulted. Nägeli was the first to express this view, and he followed it out consistently, not even hesitating to deny the existence of all processes of selection, and to represent the whole of evolution as a process conditioned by this phyletic force, which would have given rise to the world of organisms which has actually arisen, even if the conditions of life at the different periods of the earth's history had been other than they were. I have always combated this idea, without however overlooking that it is based upon facts which--at that time at any rate--gave it a certain justification. We cannot pass it by without giving some other interpretation of the facts. Following Nägeli, the botanist Askenasy championed this view of 'variation in a different direction,' which gives rise to new forms; and in more recent times Romanes, Henslow, and Eimer expressed similar views, and--although they did not actually dispute the existence of processes of selection--they attributed a much less important rôle to them, and referred the phyletic genealogical tree of organisms in the main to other and internal causes.

Like Nägeli himself, his followers have laid stress upon the fact that natural selection cannot be the cause of the evolution and succession of particular species, because the differences which separate species from species are not of an adaptive nature, and therefore cannot depend upon selection; but if the step from one species to the next succeeding one does not depend upon adaptation, then the greater steps to genera, families, and orders cannot be referred to it either, since these can only be thought of as depending upon a long-continued splitting up of species. Genera, families, and all higher groups must be recognized as conventional categories, not as real divisions existing in nature itself. Even Treviranus and Lamarck maintained that the differences between genera depended just as much upon our estimate, our intellectual convenience, as do the differences between species. All forms were originally connected, though they may not be so now, and if the species are really not distinguished by adaptive characters, then neither are any other grades of our classificatory system, neither order nor classes, since they all depend originally on the transmutation of species. It was therefore quite consistent of Nägeli to seek the mainspring of organic evolution, not in adaptation, but in an unknown evolutionary force. Thus he refused to recognize adaptation as a consequence of selection, but regarded it, as Lamarck had done, as the direct effect of external conditions, and as an entirely subordinate factor in the transmutation of forms.

Nägeli and his modern successors conceive of phyletic evolution as depending upon definitely directed variation, resulting from internal causes and occurring at definite times, which of necessity causes the existing form to be transformed into a new one. To them the species appears, so to speak, as a vital crystallization, or to use Herbert Spencer's phraseology, as an equilibrium of living matter, which becomes displaced from time to time, and passes over into a new state of equilibrium, being transmuted into a new species, something like the pictures in a kaleidoscope. The species is thus something conditioned from within, which must be as it is and could not be otherwise, just like a crystal which crystallizes in one particular system and not in another; it must be just thus or it could not be at all. From the point of view of this theory it would be easy to understand that the thousands or millions of individuals composing a species all agree in essentials--that a specific type exists.

But this conception can hardly be entirely correct, although there is some truth at its foundation, namely, that germinal variations which arise independently are the basal roots of all transmutation. But the species is not simply the result of these internal processes, it is not even mainly so; it is not the result of an internal, definitely directed developmental force, even if we attempt to think out such a force in a purely scientific or mechanical, instead of a mystical, sense. It seems clear to me that the species is not a life-crystal in the sense that it must, like a rock-crystal, take form in a particular way and in no other for purely internal reasons and by virtue of its physical constitution; the species is essentially a complex of adaptations, of modern adaptations which have been recently acquired, and of inherited adaptations handed down from long ago--a complex which might quite well have been other than it is, and indeed must have been different if it had originated under the influence of other conditions of life.

But of course species are not exclusively complicated systems of adaptations, for they are at the same time 'variation-complexes,' the individual components of which are not all adaptive, since they do not all reach the limits of the useful or the injurious. All transformations arise from a basis of spontaneous chance variations, just as all forest plants grow from the soil of the forest, but do not all grow into trees, the adaptive forms which determine the essential character of the forest; for many species remain small and low, like the mosses, grasses, and herbs; and these too have a share, though a subordinate one, in determining the character of the forest, which depends definitely, though only partially, on the loftier growths.

According to my view all adaptation depends on an alteration in the equilibrium of the determinant system, such as must arise from intra-germinal or even general fluctuations in the nutritive supply, affecting larger or smaller groups of determinants and causing variation in them to a greater or less degree. And these variations may be in a quite definite direction, persisted in for internal reasons, as we have already seen in the section dealing with germinal selection. These variations are the building-stones out of which, under the guidance of personal selection, a new specific type, that is, a new complex of adaptations, can be established. In this type many indifferent characters are involved, which are just as constant characters of the species as the adaptations.

The opponents of the selection theory have often urged against it this constancy of indifferent characters, but as soon as we cease to restrict the principle of selection to 'persons,' and extend it also to the lower categories of vital units, the occurrence of indifferent characters is easily understood. To illustrate characters of this kind, Henslow has recently called attention to the species of gentian, whose flowers have a corona split into five tips in some species and into six or seven in others, and we cannot possibly ascribe any biological significance to these specific characters. It is quite possible that they possess none; but did not even Darwin express his belief that many peculiarities of form 'are to be attributed to the laws of growth, and to the mutual influence of parts,' forces which he rightly refrained from including under 'natural selection' in his sense of the word, but which we now regard as an expression of intra-selection or of histonal selection? It is this, in our opinion, which brings about the co-adaptation of the parts to form a harmonious whole, which admits of the primary adaptations to the conditions of life being followed or accompanied by correlative secondary variations, and which plays an important part in directing the course of every individual development, and is therefore uninterruptedly active within the organism. We cannot analyse the factors precisely enough to be able to demonstrate in an individual case why the corona should be divided into four in one species of gentian and into five in another, but we can understand in principle that all adaptations of a species which are not primary are determined by the compelling influence of intra-selection. And we need not now rest content even with that, for we know that this intra-selection--as we have already seen--is active within the germ-plasm, and it is only a logical consequence of the principle of germinal selection to suppose that variations of definite determinants due to personal selection may in the germ-plasm itself give rise to correlative variations in determinants next to them or related to them in any way, and that these may possess the same stability as the primary variation. This seems to me a sufficient reason why biologically unimportant characters may become constant characters of the species. Correlation is not effected only in the perfect organism; it exists at every period of its life, from the germ till death, and what it brings about is quite as inevitable as what is evoked through adaptation by means of personal selection.

We can thus also understand that indifferent characters may be contained not only in individual ids of the germ-plasm, but also coincidentally in a great majority of them, as soon as we think of them as dependent upon the characters established through personal selection, for these must be contained in a majority of the ids.

But there is still another reason why indifferent characters should become stable, and that is the effect of general variational influences on all the individuals of a species, as, for instance, in many climatic varieties, and probably also in many cultivated varieties.

But even when we have fully recognized that, from the arcana of the germ-plasm, new minimal variations are continually cropping up, which are biologically indifferent, and nevertheless become variational tendencies, and may increase even to the extent of causing _visible_ differences, and that therefore varieties of snails or of butterflies, or of any animal or plant whatever, may originate through germinal selection alone, it cannot for a moment be supposed that the transmutation of species depends upon this process exclusively or even preponderantly. This was Nägeli's mistake, and that of his followers as well, that he ascribed to his 'principle of perfecting' the essential rôle in directing the whole movement of evolution, while the general structure of all species shows us that they are, so to speak, built up of adaptations. But adaptations could not be--or could only be fortuitously and exceptionally--the _direct_ result of an internal power of development, since the very essence of adaptational changes is that they are variations which bring the organism into harmony with the conditions of its life. We are therefore forced either to underestimate greatly the part played by adaptation in every organism--and that is what Nägeli did--or to leave the standpoint of natural science altogether, and assume a transcendental force which varied and adapted the species of organisms _pari passu_ with the changes in the conditions of life during the geological evolution of our earth. This would be a sort of pre-established harmony, through which the two clocks of evolution--that of the earth and that of organisms--kept exact time, although they had quite different and independent works!

But that the determining significance of adaptations in organic forms is underestimated even now is evidenced by the continually repeated statement that species differ, not in their adaptive characters, but in purely morphological characters, whereas it is obvious that we are far from being able to estimate the functions of a part with sufficient precision to be able to say definitely whether the differences between two nearly allied species are or are not adaptations to different conditions. The same is true with regard to the other side of the problem--the conditions of life. These are often to all appearances identical in two allied species, but even where they are visibly different it is often difficult to assert that the differences between the two species can be interpreted with certainty as adaptations to the specific conditions of life. At an earlier stage we discussed the protective coloration of butterflies, and we saw that the forest butterflies of the Tropics frequently mimicked a dry leaf on their under surfaces. In the various regions of the extensive forest districts of the Orinoco and the Amazon in South America there are fifty species of the genus _Anæa_ alone, and in the resting pose all these bear a most deceptive resemblance to a leaf, yet each of them differs from the rest in the mingling of its colours, its brilliance, and usually in markings when these are present. If we wished to be able to decide whether these specific differences were of an adaptive nature or not, we should first of all require to know in what kind of forest two neighbouring species lived, and in what places, among what sort of leaves, they were in the habit of settling. Even then we should at best only know whether the species _A_ was better protected, as far as our own eyes were concerned, among the leaves of the forest _A´_ than the species _B_, and conversely; but we could not tell whether they _required_ this protection, or whether the species _A_, if transferred to the forest _B´_, would be more frequently discovered and destroyed by its enemies than in its own forest-home, and that alone could prove the difference to be biologically important, that is, to have selection-value. The difficulty, indeed the impossibility, of arriving at such decisions can perhaps be better illustrated by an example from our indigenous fauna. No one doubts that the upper surface of the anterior wing in the so-called banner-moth (_Catocala_) possesses a very effective protective colouring; by day the moths rest with wings spread out flat upon tree trunks, wooden fences, walls, &c., and they are so excellently suited to their environment that they are usually overlooked both by man and animals. But each of the twelve German species of _Catocala_ has a special protective colouring; in _Catocala fraxini_ it is a light grey, in _Catocala nupta_, a dark ash-grey, in _Catocala elocata_ rather a yellowish-brown grey, in _Catocala sponsa_ an olive brown, in _Catocala promissa_ a mingling of whitish-grey and olive brown, and so on. All these colourings are protective; but could any even of our most experienced and sharp-sighted entomologists prove that each of these different shades of colour depends upon adaptation to the usual resting-place of the particular species to which it belongs? And yet it is on _a priori_ grounds highly probable that this is the case. But even this would by no means dispose of the whole problem, for each of these protective colour schemes is composed of several, often many, tints; it must be so if they are to fulfil their end at all, for a uniformly coloured wing would contrast with the bark of every tree and with every wooden fence. The wing-surface must therefore bear on a lighter background a number of lines and streaks varying from brown to black, and usually running zigzag across the wing; beside these are spots of lighter colour, which complete the deceptive picture. This 'marking' of the wing is similar in all twelve species, and yet in each it is different in detail. It is constant in each, and thus is a specific character. But who would venture to undertake the task of proving that each of these streaks, spots, zigzag lines, &c., is or is not adaptive--that the details are necessary adaptations to the resting-place which had become habitual to the species, or, on the other hand, simply expressions of the variational tendencies of the elements of marking, depending upon germinal selection? This would be an impossible task, and yet we are here dealing with a character which, _as a whole_, is undoubtedly adaptive; in many of the differences between other species even that is not certain.

It seems to me, therefore, hardly reasonable to talk of the 'insufficiency of natural selection' because we are not able to demonstrate that the minutiæ of specific characters are adaptational results. Personal selection intervenes whenever the variations produced by germinal selection attain to selection-value; and whether we can determine the exact point at which this takes place in individual cases is, as I have said before, theoretically quite indifferent.

Moreover, there are cases in which we _can_ prove that specific differences are of an adaptive nature. When, of two nearly related species of frog, the spermatozoon of one possesses a thick head and that of the other a thin head, and when at the same time the micropyle through which alone the spermatozoon can make its way into the ovum is wide in the first species and narrow in the second, we have before us a specific character which is obviously adaptive.

In order to gain clearness as to the significance of natural selection in the restricted sense, that is of personal selection, it seems to me much more important to study the different groups of animals and plants with special reference to what they undoubtedly exhibit in the way of adaptation. For that reason I discussed different groups of adaptations in detail in some of the preceding lectures, although, or rather because, they all teach us that every part of every species, whether animal or plant, even every secretion, and indeed every habit, every inherited instinct, is subject to adaptation to the conditions of life. It seems difficult to refuse to admit that this is the natural impression which this study conveys, and it is strengthened as our knowledge increases; _that every essential part of a species is not merely regulated by natural selection, but is originally produced by it_, if not in the species under consideration at the time, then in some ancestral species; and, further, that every part can adjust itself in a high degree to the need for adaptation. It was not without a purpose that I discussed the phenomenon of mimicry so fully, for it, above all others, teaches us how great a power of adaptation the organism possesses, and what insignificant and small parts may be transformed, in a remarkable degree, in accordance with some actual need. We saw that a butterfly might assume a colouring which diverged entirely from that of its nearest relatives, but which caused it to resemble an immune species of a different family, and thereby protected it more effectively from persecution. Such a case can no more be due to a dominating phyletic force than to a chance and sudden displacement of the state of equilibrium of the determinant system; it can depend only on natural selection, that is, on a sifting out of the diverse variations offered by germinal selection, and the unhampered expression and augmentation of those favoured.

But it is not only these minute variations, insignificant in relation to the whole structure of the animal, which can be determined by natural selection. The same applies to the phyletic evolution as a whole; even that is not directed by the assumed internal principle of development.

Adaptations, from their very nature, can only depend upon selection, and not upon an internal principle of evolution, since that could take no account whatever of external circumstances, but would cause variations in the organism altogether independently of these. Thus, in considering the origin of any of the larger groups of animals, we may exclude a phyletic power as the guide of its evolution as soon as we can prove that all its essential structural relations, as far as they diverge from those of nearly related groups, are adaptations. We may not be able to do this for nearly all of the animal groups, and it will hardly be possible in regard to a single group of plants, because our insight into the _biological_ significance of characters, which means more than the functional significance of the individual parts, and their correlation as parts of a whole, is seldom sufficiently intimate or thorough. But among animals we can do this in regard to some groups; one of these is the order of whales or Cetaceans.

Cetaceans, as is well known, belong to the Mammalia, that is to say, to a class whose structure was built for life on the land. The ancestors of Cetaceans were similar to the other mammals, and possessed a coat of hair and four legs, and a body the mass of which was so distributed that it could be borne by those four legs. But all the modern Cetaceans live in the sea, and they have therefore entirely changed their bodily form; they have become spindle-shaped like fishes, well adapted for cleaving the water, but incapable of moving upon land. At the same time, their hind-legs have completely disappeared, and can now be demonstrated only as rudiments within the mass of muscle (Fig. 130, _Br_, _Tr_, _Fr_), while the fore-legs have been transformed into flippers, in which, however, the whole inherited, but greatly shortened, skeleton of the mammalian arm is concealed (_OA_, _UA_, _H_). The skin has lost its covering of hair so completely that in some cases no traces of it are demonstrable except in the embryo. All these changes are adaptations to an aquatic life, and could not have been produced independently of the influence of external conditions. But there is much more than this. A thick layer of blubber under the skin gives this warm-blooded animal an effective protection against being cooled down by the surrounding water, and at the same time gives it the appropriate specific gravity for life in the sea; an enormous tail-fin similar to that of fishes, but placed horizontally, forms the chief organ of locomotion, and for this reason the hind-legs became superfluous and degenerated. Similarly, the muscles of the ear have also disappeared, for the hearing organ of this aquatic type is no longer suited for receiving the sound-waves through an air-containing trumpet, but receives them by a shorter route from the surrounding water, directly through the bones of the skull. Remarkable changes in the respiratory and circulatory organs make prolonged submersion possible, and the displacement of the external nares from the snout to the forehead enables the animal to draw breath when it comes up from the depths to a frequently stormy surface. It would take a long time to enumerate all that can be recognized as adaptive in these remarkable aquatic mammals to a life in what to their ancestors would have been a strange and hostile element. Let us study particularly the case of the whalebone whales, for instance the Greenland whale, and we are at once struck by the enormous size of the head, which makes up about a third of the whole body (Fig. 130). Can this, which has such an important effect in determining the whole type of animal, be an outcome of some internal power of development? By no means! It is rather an adaptation to the mode of nutrition peculiar to this swimming mammal, for it does not, like dolphins and toothed whales, feed on large fishes and Cephalopods, but on minute delicate molluscs--Pteropods and pelagic Gastropods, on Salpæ, and the like, which often cover the surface of the Arctic Ocean in endless shoals, sometimes extending for many miles. To enable the whale to sustain life on such minute morsels it was necessary that it should be able to swallow enormous quantities; teeth were therefore useless, and they have become rudimentary, and can only be demonstrated in the embryo as rudiments (dental germs) in the jaw; but in place of these there hang from the roof of the mouth-cavity great plates of 'whalebone,' a quite peculiar product of the mucous membrane of the mouth, the ends of which are frayed into fibres, and form a sieve-net for catching the little animals which are engulfed with the sea-water. The mouth-cavity itself has become enormous, so that great quantities of water at a time can be strained through the net of whalebone-plates.