There are several points of considerable theoretical importance connected with Panmixia, which were omitted from the text, in order to avoid distracting attention from the main issue which is there under consideration. These side issues may now be appropriately presented in the form in which they were published in Nature, March 13, 1890[140]. After stating, in almost the same words, what has already been said in Chapter X, this paper proceeds, with the exception of a few verbal alterations, as follows.

"There is, however, one respect in which Professor Weismann's statement of the principle of panmixia differs from that which was considered by Mr. Darwin; and it is this difference of statement—which amounts to an important difference of theory—that I now wish to discuss.

"The difference in question is, that while Professor Weismann believes the cessation of selection to be capable of inducing degeneration down to the almost complete disappearance of a rudimentary organ, I have argued that, unless assisted by some other principle, it can at most only reduce the degenerating organ to considerably above one-half its original size—or probably not through so much as one-quarter. The ground of this argument (which is given in detail in the Nature articles of 1873-1874) is, that panmixia depends for its action upon fortuitous variations round an ever-diminishing average—the average thus diminishing because it is no longer sustained by natural selection. But although no longer sustained by natural selection, it does continue to be sustained by heredity; and therefore, as long as the force of heredity persists unimpaired, fortuitous variations alone—or variation which is no longer controlled by natural selection—cannot reduce the dwindling organ to so much as one-half of its original size; indeed, as above foreshadowed, the balance between the positive force of heredity and the negative effects of promiscuous variability will most likely be arrived at above the middle line thus indicated. Only if for any reason the force of heredity begins to fail can the average round which the cessation of selection works become a progressively diminishing average. In other words, so long as the original force of heredity as regards the useless organ remains unimpaired, the mere withdrawal of selection cannot reduce the organ much below the level of efficiency above which it was previously maintained by the presence of selection. If we take this level to be 80 or 90 per cent. of the original size, cessation of selection will reduce the organ through the 10 or 20 per cent., and there leave it fluctuating about this average, unless for any reason the force of heredity begins to fail—in which case, of course, the average will progressively fall in proportion to the progressive weakening of this force.

"Now, according to my views, the force of heredity under such circumstances is always bound to fail, and this for two reasons. In the first place, it must usually happen that when an organ becomes useless, natural selection as regards that organ will not only cease, but become reversed. For the organ is now absorbing nutriment, causing weight, occupying space, and so on, uselessly. Hence, even if it be not also a source of actual danger, 'economy of growth' will determine a reversal of selection against an organ which is now not merely useless, but deleterious. And this degenerating influence of the reversal of selection will throughout be assisted by the cessation of selection, which will now be always acting round a continuously sinking average. Nevertheless, a point of balance will eventually be reached in this case, just as it was in the previous case where the cessation of selection was supposed to be working alone. For, where the reversal of selection has reduced the diminishing organ to so minute a size that its presence is no longer a source of detriment to the organism, the cessation of selection will carry the reduction a small degree further; and then the organ will remain as a 'rudiment.' And so it will remain permanently, unless there be some further reason why the still remaining force of heredity should be abolished. This further (or second) reason I found in the consideration that, however enduring we may suppose the force of heredity to be, we cannot suppose that it is actually everlasting; and, therefore, that we may reasonably attribute the eventual disappearance of rudimentary organs to the eventual failure of heredity itself. In support of this view there is the fact that rudimentary organs, although very persistent, are not everlasting. That they should be very persistent is what we should expect, if the hold which heredity has upon them is great in proportion to the time during which they were originally useful, and thus firmly stamped upon the organization by natural selection causing them to be strongly inherited in the first instance. For example, we might expect that it would be more difficult finally to eradicate the rudiment of a wing than the rudiment of a feather; and accordingly we find it a general rule that long-enduring rudiments are rudiments of organs distinctive of the higher taxonomic divisions—i.e. of organs which were longest in building up, and therefore longest sustained in a state of working efficiency.

"Thus, upon the whole, my view of the facts of degeneration remains the same as it was when first published in these columns seventeen years ago, and may be summarized as follows.

"The cessation of selection when working alone (as it probably does during the first centuries of its action upon structures or colours which do not entail any danger to, or perceptible drain upon, the nutritive resources of the organism) cannot cause degeneration below, probably, some 10 to 20 per cent. But if from the first the cessation of selection has been assisted by the reversal of selection (on account of the degenerating structure having originally been of a size sufficient to entail a perceptible drain on the nutritive resources of the organism, having now become a source of danger, and so forth), the two principles acting together will continue to reduce the ever-diminishing structure down to the point at which its presence is no longer a perceptible disadvantage to the species. When that point is reached, the reversal of selection will terminate, and the cessation of selection will not then be able of itself to reduce the organ through more than at most a very few further percentages of its original size. But, after this point has been reached, the now total absence of selection, either for or against the organ, will sooner or later entail this further and most important consequence, a failure of heredity as regards the organ. So long as the organ was of use, its efficiency was constantly maintained by the presence of selection—which is merely another way of saying that selection was constantly maintaining the force of heredity as regards that organ. But as soon as the organ ceased to be of use, selection ceased to maintain the force of heredity; and thus, sooner or later, that force began to waver or fade. Now it is this wavering or fading of the force of heredity, thus originally due to the cessation of selection, that in turn co-operates with the still continued cessation of selection in reducing the structure below the level where its reduction was left by the actual reversal of selection. So that from that level downwards the cessation of selection, and the consequent failing of heredity, act and react in their common work of causing obsolescence. In the case of newly added characters, the force of heredity will be less than in that of more anciently added characters; and thus we can understand the long endurance of 'vestiges' characteristic of the higher taxonomic divisions, as compared with those characteristic of the lower. But in all cases, if time enough be allowed under the cessation of selection, the force of heredity will eventually fall to zero, when the hitherto obsolescent structure will finally become obsolete. In cases of newly added and comparatively trivial characters, with regard to which reversal of selection is not likely to take place (e.g. slight differences of colour between allied species), cessation of selection is likely to be very soon assisted by a failure in the force of heredity; seeing that such newly added characters will not be so strongly inherited as are the more ancient characters distinctive of higher taxonomic groups.

"Let us now turn to Weismann's view of degeneration. First of all, he has omitted to perceive that 'panmixia' alone (if unassisted either by reversed selection or an inherent diminishing of the force of heredity) cannot reduce a functionless organ to the condition of a rudiment. Therefore he everywhere represents panmixia (or the mere cessation of selection) as of itself sufficient to cause degeneration, say from 100 to 5, instead of from 100 to 90 or 80, which, for the reasons above given, appeared (and still appears) to me about the most that this principle can accomplish, so long as the original force of heredity continues unimpaired. No doubt we have here what must be regarded as a mere oversight on the part of Professor Weismann; but the oversight is rendered remarkable by the fact that he does invoke the aid of reversed selection in order to explain the final disappearance of a rudiment. Yet it is self-evident that the reversal of selection must be much more active during the initial than during the final stages of degeneration, seeing that, ex hypothesi, the greater the degree of reduction which has been attained the less must be the detriment arising from any useless expenditure of nutrition, &c.

"And this leads me to a second oversight in Professor Weismann's statement, which is of more importance than the first. For the place at which he does invoke the assistance of reversed selection is exactly the place at which reversed selection must necessarily have ceased to act. This place, as already explained, is where an obsolescent organ has become rudimentary, or, as above supposed, reduced to 5 per cent. of its original size; and the reason why he invokes the aid of reversed selection at this place is in order to save his doctrine of 'the stability of germ-plasm.' That the force of heredity should finally become exhausted if no longer maintained by the presence of selection, is what Darwin's theory of perishable gemmules would lead us to expect, while such a fact would be fatal to Weismann's theory of an imperishable germ-plasm. Therefore he seeks to explain the eventual failure of heredity (which is certainly a fact) by supposing that after the point at which the cessation of selection alone can no longer act (and which his first oversight has placed some 80 per cent. too low), the reversal of selection will begin to act directly against the force of heredity as regards the diminishing organ, until such direct action of reversed selection will have removed the organ altogether. Or, in his own words, 'The complete disappearance of a rudimentary organ can only take place by the operation of natural selection; this principle will lead to its diminution, inasmuch as the disappearing structure takes the place and the nutriment of other useful and important organs.' That is to say, the rudimentary organ finally disappears, not because the force of heredity is finally exhausted, but because natural selection has begun to utilize this force against the continuance of the organ—always picking out those congenital variations of the organ which are of smallest size, and thus, by its now reversed action, reversing the force of heredity as regards the organ.

"Now the oversight here is in not perceiving that the smaller the disappearing structure becomes, the less hold must 'this principle' of reversed selection retain upon it. As above observed, during the earlier stages of reduction (or while co-operating with the cessation of selection) the reversal of selection will be at its maximum of efficiency; and, as the process of diminution continues, a point must eventually be reached at which the reversal of selection can no longer act. Take the original mass of a now obsolescent organ in relation to that of the entire organism of which it then formed a part to be represented by the ratio 1:100. For the sake of argument we may assume that the mass of the organism has throughout remained constant, and that by 'mass' in both cases is meant capacity for absorbing nutriment, causing weight, occupying space, and so forth. Now, we may further assume that when the mass of the organ stood to that of its organism in the ratio of 1:100, natural selection was strongly reversed with respect to the organ. But when this ratio fell to 1:1000, the activity of such reversal must have become enormously diminished, even if it still continued to exercise any influence at all. For we must remember, on the one hand, that the reversal of selection can only act as long as the presence of a diminishing organ continues to be so injurious that variations in its size are matters of life and death in the struggle for existence; and, on the other hand, that natural selection in the case of the diminishing organ does not have reference to the presence and the absence of the organ, but only to such variations in its mass as any given generation may supply. Now, the process of reduction does not end even at 1:1000. It goes on to 1:10,000, and eventually 1:8. Consequently, however great our faith in natural selection may be, a point must eventually come for all of us at which we can no longer believe that the reduction of an obsolescent organ is due to reversed selection. And I cannot doubt that if Professor Weismann had sufficiently considered the matter, he would not have committed himself to the statement that 'the complete disappearance of a rudimentary organ can only take place by the operation of natural selection.'

"According to my view, the complete disappearance of a rudimentary organ can only take place by the cessation of natural selection, which permits the eventual exhaustion of heredity, when heredity is thus simply left to itself. During all the earlier stages of reduction, the cessation of selection was assisted in its work by the reversal of selection; but when the rudiment became too small for such assistance any longer to be supplied, the rudiment persisted in that greatly reduced condition until the force of heredity with regard to it was eventually worn out. This appears to me, as it appeared in 1873, the only reasonable conclusion that can be drawn from the facts. And it is because this conclusion is fatal to Professor Weismann's doctrine of the permanent 'stability' of germ-plasm, while quite in accordance with all theories which belong to the family of pangenesis, that I deem the facts of degeneration of great importance as tests between these rival interpretations of the facts of heredity. It is on this account that I have occupied so much space with the foregoing discussion; and I shall be glad to ascertain whether any of the followers of Professor Weismann are able to controvert these views.

"George J. Romanes."

"P.S.—Since the above article was sent in, Professor Weismann has published in these columns (February 6) his reply to a criticism by Professor Vines (October 24, 1889). In this reply he appears to have considerably modified his views on the theory of degeneration; for while in his Essays he says (as in the passage above quoted) that 'the complete disappearance of a rudimentary organ can only take place by the operation of natural selection'—i.e. only by the reversal of selection,—in his reply to Professor Vines he says, 'I believe that I have proved that organs no longer in use become rudimentary, and must finally disappear, solely by 'panmixia'; not through the direct action of disuse, but because natural selection no longer sustains their standard structure'—i.e. solely by the cessation of selection. Obviously, there is here a flat contradiction. If Professor Weismann now believes that a rudimentary organ 'must finally disappear solely' through the withdrawal of selection, he has abandoned his previous belief that 'the complete disappearance of a rudimentary organ can only take place by the operation of selection.' And this change of belief on his part is a matter of the highest importance to his system of theories as a whole, since it betokens a surrender of his doctrine of the 'stability' of germ-plasm—or of the virtually everlasting persistence of the force of heredity, and the consequent necessity for a reversal of this force itself (by natural selection placing its premium on minus instead of on plus variations), in order that a rudimentary organ should finally disappear. In other words, it now seems he no longer believes that the force of heredity in one direction (that of sustaining a rudimentary organ) can only be abolished by the active influence of natural selection determining this force in the opposite direction (that of removing a rudimentary organ). It seems he now believes that the force of heredity, if merely left to itself by the withdrawal of natural selection altogether, will sooner or later become exhausted through the mere lapse of time. This, of course, is my own theory of the matter as originally published in these columns; but I do not see how it is to be reconciled with Professor Weismann's doctrine of so high a degree of stability on the part of germ-plasm, that we must look to the Protozoa and the Protophyta for the original source of congenital variations as now exhibited by the Metazoa and Metaphyta. Nevertheless, and so far as the philosophy of degeneration is concerned, I shall be very glad if (as it now appears) Professor Weismann's more recent contemplation has brought his principle of panmixia into exact coincidence with that of my cessation of selection."

Before passing on it may here be noted that, to any one who believes in the inheritance of acquired characters, there is open yet another hypothetical cause of degeneration, and one to which the final disappearance of vestigial organs may be attributed. Roux has shown in his work on The Struggle for Existence between Parts of an Organism that the principle of selection must operate in every constituent tissue, and as between every constituent cell of which an organism is composed. Now, if an organ falls into disuse, its constituent cells become worsted in their struggles with other cells in the organism. Hence, degeneration of the disused organ may progressively increase, quite independently of any struggle for existence on the part of the organism as a whole. Consequently, degeneration may proceed without any reference to the principle of "economized nutrition"; and, if it does so, and if the effects of its doing so are transmitted from generation to generation, the disused organ will finally disappear by means of Roux's principle.

The long communication above quoted led to a still longer correspondence in the pages of Nature. For Professor Ray Lankester wrote[141] to impugn the doctrine of panmixia, or cessation of selection, in toto, arguing with much insistence that "cessation of selection must be supplemented by economy of growth in order to produce the results attributed to panmixia." In other words, he denied that panmixia alone can cause degeneration in any degree at all; at most, he said, it can be but "a condition," or "a state," which occurs when an organ or part ceases to be useful, and therefore falls under the degenerating influence of active causes, such as economy of nutrition. Or, in yet other words, he refused to recognize that any degenerative process can be due to natural selection as merely withdrawn: only when, besides being withdrawn, natural selection is reversed, did he regard a degenerative process as possible. As a result of the correspondence, however, he eventually[142] agreed that, if the "birth-mean" of an organ, in respect either of size or complexity of structure, be lower than the "selection-mean" while the organ is useful (a fact which he does not dispute); then, if the organ ceases to be useful, it will degenerate by the withdrawal of selection alone. Which, of course, is merely a re-statement of the doctrine of panmixia, or cessation of selection, in somewhat varied terminology—provided that the birth-mean be taken over a number of generations, or not only over a few following the selection-mean of the structure while still in its highest state of efficiency. For the sake of brevity I will hereafter speak of these "few following" generations by the term of "first generations."

It remains to consider the views of Professor Lloyd Morgan upon the subject. In my opinion he is the shrewdest, as well as the most logical critic that we have in the field of Darwinian speculation; therefore, if possible, I should like to arrive at a full agreement with him upon this matter. His latest utterance with regard to it is as follows:—

"To account for the diminution of organs or structures no longer of use, apart from any inherited effects of disuse, Mr. Romanes has invoked the Cessation of Selection; and Mr. Francis Galton has, in another connexion, summarized the effects of this cessation of selection in the convenient phrase 'Regression to Mediocrity.' This is the Panmixia of Professor Weismann and his followers; but the phrase regression to mediocrity through the cessation of selection appears to me preferable. It is clear that so long as any organ or structure is subject to natural selection through elimination, it is, if not actually undergoing improvement, kept at a high standard of efficiency through the elimination of all those individuals in which the organ in question falls below the required standard. But if, from change in the environment or any other cause, the character in question ceases to be subject to selection, elimination no longer takes place, and the high standard will no longer be maintained. There will be reversion to mediocrity. The probable amount of this reversion is at present a matter under discussion[143]."

So far, then, Professor Lloyd Morgan is in complete agreement with previous writers upon the subject. He does not doubt that the cessation of selection must always be a cause of degeneration: the only question is as to the potency of this cause, or the amount of degeneration which it is capable of effecting.

Taking, first, the case of bulk or size of an organ, as distinguished from its organization or complexity, we have seen that Weismann represents the cessation of selection—even if working quite alone, or without any assistance from the reversal of selection—to be capable of reducing a fully developed organ to the state of a rudiment, or even, if we take his most recent view, of abolishing the organ in toto.

Professor Lloyd Morgan, on the other hand, does not think that the cessation of selection alone can cause reduction further than the level of "mediocrity" in the first generations—or, which is much the same thing, further than the difference between the "birth-mean" and the "selection-mean" of the first generations. This amount of reduction he puts at 5 per cent., as "a very liberal estimate."

Here, then, we have three estimates of the amount of degeneration which can be produced by panmixia alone, where mere size or bulk of an organ is concerned—say, 3 to 5 per cent., 10 to 20 per cent., and 95 per cent. to 0. At first sight, these differences appear simply ludicrous; but on seeking for the reasons of them, we find that they are due to different views touching the manner in which panmixia operates. The oversights which have led to Weismann's extremely high estimate have already been stated. The reason of the difference between the extremely low estimate of Professor Lloyd Morgan, as compared with my own intermediate one, is, that he supposes the power of panmixia to become exhausted as soon as the level of mediocrity of the first generations has become the general level in succeeding generations. In my view, however, the level of mediocrity is itself a sinking level in successive generations, with the result that there is no reason why the reducing power of panmixia should ever become exhausted, save that the more reduction it effects the greater is the force of heredity which remains to be overcome, as previously explained. Thus the only question between Professor Lloyd Morgan and myself is—Does the level of mediocrity fall in successive generations under the cessation of selection, or does it remain permanently where it used to be under the presence of selection? Does the "birth-mean" remain constant throughout any number of generations, notwithstanding that the sustaining influence of selection has been withdrawn; or does it progressively sink as a consequence of such withdrawal?

In order to answer this question we had better begin by considering now the case of organization of structure, as distinguished from mere size of structure. Take any case where a complex organ—such as a compound eye—has been slowly elaborated by natural selection, and is it not self-evident that, when natural selection is withdrawn, the complex structure will deteriorate? In other words, the level of mediocrity, say in the hundred thousandth generation after the sustaining influence of natural selection has been withdrawn, will not be so high as it was in the first generations. For, by hypothesis, there is now no longer any elimination of unfavourable variations, which may therefore perpetuate themselves as regards any of the parts of this highly complex mechanism; so that it is only a matter of time when the mechanism must become disintegrated. I can scarcely suppose that any one who considers the subject will question this statement, and therefore I will not say anything that might be said in the way of substantiating it. But, if the statement be assented to, it follows that there is no need to look for any cause of deterioration, further than the withdrawal of selection—or cessation of the principle which (as we are supposing) had hitherto been the sole means of maintaining efficient harmony among all the independently variable parts of the highly complex structure.

Now, I hold that the same thing is true, though in a lesser degree, as regards degeneration of size. That there is no difference in kind between the two cases, Professor Lloyd Morgan implicitly allows; for what he says is—

"In any long-established character, such as wing-power in birds, brain-development, the eyes of crustacea, &c., no shortcomer in these respects would have been permitted by natural selection to transmit his shortcomings for hundreds of generations. All tendency to such shortcomings would, one would suppose, have been bred out of the race. If after this long process of selection there still remains a strong tendency to deterioration, this tendency demands an explanation[144]."

Here, then, deterioration as to size of structure (wings of birds), and deterioration as to complexity of structure (brain and eyes) are expressly put upon the same footing. Therefore, if in the latter case the "tendency to deterioration" does not "demand an explanation," beyond the fact that the hitherto maintaining influence has been withdrawn, neither is any such further explanation demanded in the former case. Which is exactly my own view of the matter. It is also Mr. Galton's view. For although, in the passage formerly quoted, Professor Lloyd Morgan appears to think that by the phrase "Regression to Mediocrity" Mr. Galton means to indicate that panmixia can cause degeneration only as far as the mediocrity level of the first generations, this, in point of fact, is not what Galton means, nor is it what he says. The phrase in question occurs "in another connexion," and, indeed, in a different publication. But where he expressly alludes to the cessation of selection, this is what he says. The italics are mine.