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.