It has already been shown, while setting forth in the preceding chapter the antecedent standing of the issue, that in this respect the prima facie presumption is wholly on the side of the transmission, in greater degree or less, of acquired characters. Even Weismann allows that all "appearances" point in this direction, while there is no inductive evidence of the action of natural selection in any one case, either as regards germs or somas, and therefore, a fortiori, of the "all-sufficiency" of this cause[35]. It is true that in some of his earlier essays he has argued that there is no small weight of prima facie evidence in favour of his own views as to the non-inheritance of acquired characters. This, however, will have to be considered in its proper place further on. Meanwhile I shall say merely in general terms that it arises almost entirely from a confusion of the doctrine of Continuity as absolute with that of Continuity as partial, and therefore, as admitting of degrees in different cases—which, as already explained, are doctrines wide as the poles asunder. But, leaving aside for the present such prima facie evidence as Weismann has adduced on his side of the issue, I may quote him as a hostile witness to the weight of this kind of evidence per contra, in so far as it has already been presented in the foregoing chapter. Indeed, Weismann is much too logical a thinker not to perceive the cogency of the "appearances" which lie against his view of Continuity as absolute—although he has not been sufficiently careful in distinguishing between such Continuity and that which admits of degrees.
We may take it, then, as agreed on all hands that whatever weight merely prima facie evidence may in this matter be entitled to, is on the side of what I have termed moderated Lamarckianism: first sight "appearances" are against the Neo-Darwinian doctrine of the absolute non-inheritance of acquired characters.
Let us now turn to another and much more important line of indirect evidence in favour of moderated Lamarckianism.
The difficulty of excluding the possibility of natural selection having been at work in the case of wild plants and animals has already been noticed. Therefore we may now appreciate the importance of all facts or arguments which attenuate the probability of natural selection having been at work. This may be done by searching for cases in nature where a congenital structure, although unquestionably adaptive, nevertheless presents so small an amount of adaptation, that we can scarcely suppose it to have been arrived at by natural selection in the struggle for existence, as distinguished from the inheritance of functionally-produced modifications. For if functionally-produced modifications are ever transmitted at all, there is no limit to the minuteness of adaptive values which may thus become congenital; whereas, in order that any adaptive structure or instinct should be seized upon and accumulated by natural selection, it must from the very first have had an adaptive value sufficiently great to have constituted its presence a matter of life and death in the struggle for existence. Such structures or instincts must not only have always presented some measure of adaptive value, but this must always have been sufficiently great to reach what I have elsewhere called a selection-value. Hence, if we meet with cases in nature where adaptive structures or instincts present so low a degree of adaptive value that it is difficult to conceive how they could ever have exercised any appreciable influence in the battle for life, such cases may fairly be adduced in favour of the Lamarckian theory. For example, the Neo-Lamarckian school of the United States is chiefly composed of palaeontologists; and the reason of this seems to be that the study of fossil forms—or of species in process of formation—reveals so many instances of adaptations which in their nascent condition present such exceedingly minute degrees of adaptive value, that it seems unreasonable to attribute their development to a survival of the fittest in the complex struggle for existence. But as this argument is in my opinion of greatest force when it is applied to certain facts of physiology with which I am about to deal, I will not occupy space by considering any of the numberless cases to which the Neo-Lamarckians apply it within the region of palaeontology[36].
Turning then to inherited actions, it is here that we might antecedently expect to find our best evidence of the Lamarckian principles, if these principles have really had any share in the process of adaptive evolution. For we know that in the life-time of individuals it is action, and the cessation of action, which produce nearly all the phenomena of acquired adaptation—use and disuse in animals being merely other names for action and the cessation of action. Again, we know that it is where neuro-muscular machinery is concerned that we meet with the most conclusive evidence of the remarkable extent to which action is capable of co-ordinating structures for the ready performance of particular functions; so that even during the years of childhood "practice makes perfect" to the extent of organizing neuro-muscular adjustments, so elaborate and complete as to be indistinguishable from those which in natural species we recognized as reflex actions on the one hand, and instinctive actions on the other. Hence, if there be any such thing as "use-inheritance" at all, it is in the domain of reflex actions and instinctive actions that we may expect to find our best evidence of the fact. Therefore I will restrict the present line of evidence—(A)—to these two classes of phenomena, as together yielding the best evidence obtainable within this line of argument.
The evidence in favour of the Lamarckian factors which may be derived from the phenomena of reflex action has never, I believe, been pointed out before; but it appears to me of a more cogent nature than perhaps any other. In order to do it justice, I will begin by re-stating an argument in favour of these factors which has already been adduced by previous writers, and discussed by myself in published correspondence with several leaders of the ultra-Darwinian school.
Long ago Professor Broca and Mr. Herbert Spencer pointed to the facts of co-adaptation, or co-ordination within the limits of the same organism, as presenting good evidence of Lamarckian principles, working in association with natural selection. Thus, taking one of Lamarck's own illustrations, Mr. Spencer argued that there must be numberless changes—extending to all the organs, and even to all the tissues, of the animal—which in the course of many generations have conspired to convert an antelope into a giraffe. Now the point is, that throughout the entire history of these changes their utility must always have been dependent on their association. It would be useless that an incipient giraffe should present the peculiar form of the hind-quarters which we now perceive, unless at the same time it presented the correspondingly peculiar form of the fore-quarters; and as each of these great modifications entails innumerable subordinate modifications throughout both halves of the creature concerned, the chances must have been infinitely great against the required association of so many changes happening to have arisen congenitally in the same individuals by way of merely fortuitous variation. Yet, if we exclude the Lamarckian interpretation, which gives an intelligible cause of co-ordination, we are required to suppose that such a happy concurrence of innumerable independent variations must have occurred by mere accident—and this on innumerable different occasions in the bodies of as many successive ancestors of the existing species. For at each successive stage of the improvement natural selection (if working alone) must have needed all, or at any rate most, of the co-ordinated parts to occur in the same individual organisms[37].
In alluding to what I have already published upon the difficulty which thus appears to be presented to his theory, Weismann says, "At no distant time I hope to be able to consider this objection, and to show that the apparent support given to the old idea [i. e. of the transmission of functionally-produced modifications] is really insecure, and breaks down as soon as it is critically examined[38]."