Quick Links
-Search Website
-Have A Question?
-Wallace News
-About This Site

Misinformation Alert!
Wallace Bio & Accomplishments
Wallace Chronology
Frequently Asked Questions
Wallace Quotes
Wallace Archives
Miscellaneous Facts

Bibliography / Texts
Wallace Writings Bibliography
Texts of Wallace Writings
Texts of Wallace Interviews
Wallace Writings: Names Index
Wallace Writings: Subject Index
Writings on Wallace
Wallace Obituaries
Wallace's Most Cited Works

Taxonomic / Systematic Works
Wallace on Conservation
Smith on Wallace
Research Threads
Wallace Images
Just for Fun
Frequently Cited Colleagues
Wallace-Related Maps & Figures

Alfred Russel Wallace : Alfred Wallace : A. R. Wallace :
Russel Wallace : Alfred Russell Wallace (sic)

Letters to "Nature" Regarding George
Romanes's Theory of Physiological Selection
(1886-1890): S390, S395, S428, and S429

Editor Charles H. Smith's Note: The prominent biologist George Romanes is remembered for several things, including his theory of physiological selection, which Wallace objected to. In these four letters to Nature Wallace briefly outlines why. See also S389 and S527. Original paginations indicated within double brackets. To link directly to this page, connect with: http://people.wku.edu/charles.smith/wallace/S390AND.htm

Physiological Selection and the Origin of Species (S390: 1886)

    [[p. 467]] As Mr. Romanes has referred to my article in the current number of the Fortnightly Review, and stated that he is prepared to answer what he terms "the very obvious exceptions" which I have taken to his theory, I shall be glad to be allowed to state, very briefly, what those exceptions are, and to give an illustration of one of the more important of them.

    (1) Mr. Romanes makes a great deal of the alleged "inutility of specific characters," and founds upon it his extraordinary statement that, during his whole life, Darwin was mistaken in supposing his theory to be "a theory of the origin of species," and that all Darwinians who have believed it to be so have blindly fallen into the same error. I allege, on the contrary, that there is no proof worthy of the name that specific characters are frequently useless, and I adduce a considerable series of facts tending to prove their general utility.

    (2) In support of his view as to the swamping effects of intercrossing, Mr. Romanes objects to the assumption of Darwin, "that the same variation occurs simultaneously in a number of individuals," adding: "Of course, if this assumption were granted, there would be an end of the present difficulty"; and his whole argument on this branch of the question rests on the assumption being false. I adduce evidence--copious evidence--that the supposed assumption represents a fact, which is now one of the best-established facts of natural history.

    (3) Mr. Romanes states, as the special feature of his physiological varieties, that "they cannot escape the preserving agency of physiological selection." He gives no particle of proof of this, while I show that, on the contrary, it is hardly possible for them to survive to a second or third generation. It is on this point that I wish to give an illustration. Mr. Romanes speaks of his supposed variations as "showing some degree of sterility with the parent form," while continuing to be fertile "within the limits of the varietal form"; but I hold that any such variety (beyond single individuals) can hardly exist, while he has adduced no proof whatever of their existence. To show the improbability of their existence, let us suppose a definite case.

    In a given species there is born an individual, A, which is infertile with the bulk of the species, but fertile with some few individuals of the opposite sex, a, b, c. Let there be a second individual, E, born from other parents in another part of the area occupied by the species, and fertile only with e, f, g. Other individuals, K, P, R, &c., may have similar relations, each infertile with the bulk of the species, fertile only with a few individuals which may be termed their physiological complements. Now each of these, separately, is a physiological variety, but the whole set, A, E, K, P, R, do not form one, but five distinct varieties. To form one variety all of them must be fertile with the same identical set of individuals of the opposite sex, and this seems to me to be so highly improbable that it must not be assumed till rigidly proved. Yet there is not one passage in Mr. Romanes' paper to show that he recognised this difficulty; on the contrary, he always speaks as if any number of separate physiological variations within one species must necessarily form one variety. It will easily be seen that the chances against any single variety of this nature being preserved are overwhelmingly great. For, first, at least two of the complementary individuals must survive to the breeding-season, and the chances against this are measured by the fertility of the species. If it produces ten young each year, the chances are between nine and ten to one against any one of them surviving. The chances against the two complements surviving will be about ninety to one; and then there remains the chances against the two meeting at the breeding-season, for, by the assumption, there is nothing whatever to bring them together but chance, and this may be any number of thousands to one.

    There are, no doubt, other possible cases in which the physiological variety might be continued, but, as I have shown in my [[p. 468]] paper, the chances against it are always very great. Here, then, are three objections to Mr. Romanes' theory which seem to me to be weighty and fundamental; yet he says, in effect, that he anticipated, and is prepared to answer, them. This, I must say, puzzles me; because in the whole of his lengthy paper, occupying seventy-five pages, I cannot find any adequate recognition of their existence, or any attempt whatever to answer them.

    My apology for writing this is that I am shortly leaving England, and wish the readers of Nature, who may not have seen the Fortnightly, to be aware of the character of the objections which Mr. Romanes declares that he anticipated, but apparently thought of too little importance to require any discussion in his paper.

*                *                *                *                *

Mr. Romanes on Physiological Selection (S395: 1887)

    [[p. 366]] I have just seen Mr. Romanes's article in the Nineteenth Century, and his letter specially replying to myself in your issue of January 13 (p. 247). I do not propose to continue the discussion, but ask leave to make a few observations on some features of his reply in both the article and the letter.

    On the question of the "inutility of specific characters," he appeals to authority against me, and especially to Darwin's very cautious remarks, which seem to me to support my view much more than they do those of Mr. Romanes; but in any case this is a matter in which I decline to accept authority as an infallible guide. The impossibility of proving a negative is proverbial, but my opponent declares that his negative--the uselessness of specific characters--wants no proving, but must be accepted till in every case the affirmative is proved. Here, again, is a canon of criticism the validity of which I wholly deny.

    As to the swamping effects of intercrossing, there is again an appeal to authority, and Mr. Romanes now explains away (in the Nineteenth Century) what he had said about "simultaneous variations," by asking me to show such variations as the occurrence of an incipient spur on a duck's foot or horn on the head of a racehorse, in the belief, apparently, that these are the class of characters which are distinctive of closely-allied species! Such a demand, seriously made, appears to me so preposterous as to render further discussion of the matter with such an adversary out of the question.

    The argument to show that the supposed physiological variations would be perpetuated, seems to me as weak and unsatisfactory as ever. The question is really not worth further discussion till the required variations are proved to exist in the requisite abundance and possessing the peculiar relations to each other and to the rest of the species which would alone give them any chance of survival.

    I now leave the question, as between myself and Mr. Romanes, to the consideration of those naturalists who may be able to bestow upon it the requisite time and attention.

*                *                *                *                *

Dr. Romanes on Physiological Selection (S428: 1890)

    [[p. 79]] In his two latest articles dealing with this subject, Dr. Romanes has made certain statements as to my position in regard to it which call for a brief notice on my part.

    In his original paper, and in the summary of it published in Nature, Dr. Romanes adduced variations in regard to fertility and sterility as the fundamental fact in physiological selection. A few quotations will show this. He says: "It becomes almost impossible to doubt that the primary specific distinction (meaning sterility) is, as a general rule, the primordial distinction" (Nature, vol. xxxiv. p. 339). Again, he enforces this as against Darwin's view that sterility was a consequence or concomitant of other differences, as follows: "My theory, on the other hand, inverts this order, and supposes the primary distinction to be likewise (in most cases) the primordial distinction" (l.c., p. 363). This is very clear, but to show that he limited the term "physiological selection" to the results supposed to arise from this phenomenon, we have his reply to Mr. Galton, who urged a fact also dwelt upon by Darwin--the psychological disinclination to mate between many varieties--as an important factor in the differentiation of species: "Now I have fully recognized this principle as one amongst several others which is accessory to, although independent of, physiological selection" (l.c., p. 407). A little further on he again states his fundamental fact thus: "If my theory is true, it must follow, as Mr. Galton says, that such unions would be more or less sterile, and, as this sterility is itself the only variation which my theory supposes to have arisen in the first instance, ex hypothesi we can have no means of observing whether or not the individuals which present this variation 'consort with outsiders,' or with those individuals which do not present it" (l.c., p. 407). As if to leave no possible doubt as to the special point of his new theory, he again enforces it in the following passage: "And forasmuch as the sexual separation arises only by way of a variation locally affecting the reproductive system, when the variation is first sexually separated, it will in all other respects resemble its parent stock, and so be able to compete with it on equal terms" (l.c., p. 408).

    Now surely all this makes it absolutely clear that Dr. Romanes's theory of physiological selection, so far as it had any originality, was founded on the supposition of sterility-variation alone, arising in an otherwise undifferentiated species; and he claimed that such variations "cannot escape the preserving agency of physiological selection," and that "physiological selection must be quite as vigilant as natural selection, and it seizes upon the comparatively unuseful variation of sterility with even more certainty than natural selection can seize upon any useful variations" (l.c., p. 364).

    These last statements, by the truth of which alone the use of the term "selection" can be justified, I showed by two carefully considered cases to be absolutely unfounded, and the exact opposite of what must really occur (l.c., p. 467; and "Darwinism," p. 182). Having thus proved that "physiological selection," in the only form claimed by Dr. Romanes as original, does not exist, and that the only modes by which degrees of sterility between distinct species can arise are those discussed or suggested by Darwin himself, with the addition of the possible action of natural selection in increasing incipient sterility between slightly differentiated forms, will it be believed that I am accused of having appropriated the theory of physiological selection without acknowledgment! In the Nineteenth Century (May 1890, p. 831), Dr. Romanes says of me: "He presents an alternative theory to explain the same class of facts. Yet this theory is, purely and simply, without any modification whatsoever, a restatement of the first principles of physiological selection, as these were originally stated by myself." And now, in the October issue of an American magazine, The Monist, he has an article entitled "Mr. A. R. Wallace on Physiological Selection," in which the original main point, of sterility-variations alone leading to and constituting "physiological selection," is almost entirely ignored, and the various modes by which isolation is produced between incipient species or in which infertility arises in correlation with other divergent characters, are all claimed as forming part of the theory of physiological selection. He quotes from "Darwinism" my exposition of the effects of partial infertility arising between "two varieties in process of adaptation to somewhat different modes of life within the same area," to show "how unequivocal and complete is Mr. Wallace's adoption of our theory" (The Monist, No. I, p. II). "Our" refers to Mr. Gulick, who is taken into partnership by Dr. Romanes. And again he speaks of "the peculiar position to which he has eventually gravitated with reference to my views--professing hostility on the one hand, while reproducing them as original on the other" (l.c., p. 19).

    I have here confined myself to showing, by Dr. Romanes's own repeated and emphatic statements, what was the essential and original theory to which he gave the name of "physiological selection." The whole of this special doctrine I have argued against as unsound, because, on close examination, it proves to be quite inadequate to produce any such effects as are claimed for it. Whether I was right or wrong in doing so, I did, as a matter of fact, and do still, wholly reject this fundamental and essential part of the theory--the only part which had even a primâ facie claim to originality. I also totally reject the two subsidiary doctrines on which Dr. Romanes lays great stress as adjuncts of his theory--that of the inutility of a large proportion of specific characters, and that of the power of isolation alone "without the aid of natural selection" to produce new species; while, so far as I know, the only points in which I agree with him are those in which we both make use of Darwin's facts and adopt Darwin's explanation of them. Yet, notwithstanding this rejection of all that is special in his teachings, Dr. Romanes has the hardihood to assert that I claim them as my own; that I merely restate his theory "purely and simply, without any modification whatsoever"; and that my adoption of his theory "is unequivocal and complete."

    I leave it to others to characterize these extraordinary statements in the terms that fitly apply to them.

*                *                *                *                *

Dr. Romanes on Physiological Selection (S429: 1890)

    [[p. 150]] As Dr. Romanes now declares that the essence of his theory of physiological selection is "that some amount of infertility characterizes the distinct varieties which are in process of differentiation into species," and that the occurrence of infertility among the members of an undifferentiated species is secondary and comparatively unimportant, I ask leave to quote one or two more of his original statements, in addition to the four emphatic passages quoted in my communication of November 27.

    (1) "When accidental variations of a non-useful kind occur in any of the other systems or parts of organisms, they are, as a rule, immediately extinguished by intercrossing. But whenever they happen to arise in the reproductive system in the way here suggested, they must inevitably tend to be preserved as new natural varieties, or incipient species. At first the difference would only be in respect of the reproductive system; but eventually, on account of independent variation, other differences would supervene, and the new variety would take rank as a new species"(Nature, vol. xxxiv. p. 316).

    The words I have italicized show clearly that variation in fertility only was what Dr. Romanes then claimed as essential to his theory. Again, after referring to variations in the season of flowering as a "well-known and frequently observed cause" of isolation, he adds:--

    (2) "But it is on what may be called spontaneous variability of the reproductive system itself that I mainly rely for evidence of physiological selection" (l.c., p. 337).

    The meaning of this is still further enforced by other passages. After discussing the supposed causes of infertility, he says:--

    (3) "Why should we suppose that, unlike all other such variations, it can never be independent, but must always be superinduced as a secondary result of changes taking place elsewhere? It appears to me that the only reason why evolutionists suppose this is because the particular variation in question happens to have as its result the origination of species" (l.c., p. 339).

    And again:--

    (4) "It appears to me much the more rational view that the primary specific distinction is likewise, as a rule, the primordial distinction; and that the cases where it has been superinduced by the secondary distinctions are comparatively few in number" (l.c.).

    Notwithstanding the passages I have now quoted, emphasizing eight times over, in different ways, that the theory is essentially one of variations as regards fertility and sterility alone, Dr. Romanes now says that, even if all this is wrong, "the principle of physiological selection, as I have stated it, is not thereby affected." If this is not an absolute change of front, words have no meaning; and it is further shown to be so by the fact that Dr. Romanes acknowledged that Mr. Catchpool had "very clearly put forward the theory of physiological selection." But Mr. Catchpool clearly distinguished between the old theory that species arise first by variation in form and structure, and only gradually become mutually infertile, and the new theory that they arise "by spontaneous variations in the generative elements, and are in this case originally mutually infertile, but only gradually become otherwise divergent" (l.c., vol. xxxi. p. 4).

    That this was the essential and original "physiological selection," that was claimed as supplying the missing link required to make the origin of species by natural selection a reality, is yet further shown by the repeated statements that physiological "selection" is a powerful preservative agent. Besides the statement already quoted, that variations in fertility "cannot escape the preserving agency of physiological selection," we have the assertion, quoted above, that such variations "must inevitably tend to be preserved as new natural varieties or incipient species," and the following still more emphatic assertion:--"Neither are we concerned with the degrees of sterility which the variation in question may in any particular case supply. For whether the degree of sterility with the parent form be originally great or small, the result of it will in the long run be the same: the only difference will be that in the latter case a greater number of generations would be required in order to separate the varietal from the parent form."

    Now my contention has always been, and still is, that there is no principle at work which can accumulate or even preserve the variations of infertility occurring in an otherwise undifferentiated species, and that the term physiological "selection" is therefore a misnomer, and altogether misleading. If Dr. Romanes will carefully work out numerically (as I have attempted to do) a few cases showing the preservative and accumulative agency of pure physiological selection within an otherwise undifferentiated species, he will do more for his theory than volumes of general disquisition or any number of assertions that it does possess this power.

    My next contention is, that this is the only new part of his theory--as he himself shows by his reference to the ordinary view, of sterility following other changes, as that which "evolutionists suppose." All the rest is to be found more or less fully discussed in Darwin's works; and I myself claim only to have carefully studied Darwin's facts, and his brief but most suggestive discussion of them in his chapter on "Hybridism" (vol. ii. of "Animals and Plants under Domestication"), and by arranging them more systematically to have shown that they do really give a fairly consistent and sufficient solution of the problem. The only part of my work I claim as a distinct addition to the theory is the proof that, under certain conditions that appear to me probable, natural selection is capable of increasing incipient infertility between distinct races or varieties; and the same view was submitted to Darwin twenty years ago.

    Lastly, I totally and emphatically deny that any portion of my facts or conclusions on the subject were derived from Dr. Romanes's writings on "physiological selection." The only two sentences he has quoted from my book to prove that I have done so merely express what he himself has declared to be the common opinion of evolutionists, and which is also the direct outcome of the facts collected by Darwin. If this is "the whole essence of physiological selection," then physiological selection is but a re-statement and amplification of Darwin's own views, since he certainly assumed, and proved, that "some amount of infertility" characterized "some varieties" of animals and plants, and that this infertility, when it occurs, is of some use in preventing the swamping effects of intercrossing. I feel sure that if this had been stated, at the outset, to be what was termed "physiological selection," no discussion would have arisen as to the principle involved, but only as to its novelty and as to the appropriateness of the name given to it.

    If now, notwithstanding his repeated and emphatic statements that variation as regards fertility in otherwise undifferentiated species was what constituted the basis of his theory of physiological selection, Dr. Romanes continues to assert that I have adopted that theory "purely and simply, without any modification whatever," it will show that our respective standards of scientific reasoning and literary consistency are so entirely different as to render any further discussion of the subject on my part unnecessary as regards myself and useless as regards Dr. Romanes.

*                *                *                *                *

Comment by Prof. Donald R. Forsdyke, Queen's University, Canada (pers. commun. 2/04):

To savour fully the fascinating debate between Wallace and Romanes on the pages of journals such as Nature and Fortnightly Review, the reader should consult the original texts of Romanes in addition to the Wallace texts displayed here. I would like to draw attention to the important term "physiological complements" which Wallace introduces to characterize individuals (in say group A) that have diverged from most other species members (the "bulk species"). They have diverged in some unspecified (abstract) aspect of their reproductive systems to the extent that, when crossed with members of the bulk species, the offspring, although otherwise healthy, would tend to be sterile. Thus, even though the offspring were healthy, the parents of such sterile offspring would tend to be reproductively isolated from each other (and the line could not continue). This identifies individuals A as potential members of an incipient new species. Romanes postulated that, although diverged from the bulk species, A individuals might not be reproductively isolated from each other to the same extent as they were reproductively isolated from members of the bulk species. So A x A crosses could result in some healthy non-sterile offspring, thus allowing the deviant A line to continue. A with A would be physiological complements, but A with a member of the bulk species would not be physiological complements.

Wallace acknowledges that a (say) male individual of genetic constitution A might produce fertile offspring if crossed with females of genetic constitutions a, b or c, where b and c might be different in some (abstract) aspect of their reproductive systems (other than being of the opposite sex to A):

"In a given species there is born an individual, A, which is infertile with the bulk of the species, but fertile with some few individuals of the opposite sex, a, b, c".

The latter are not bulk species members. However, he seems to have seen fertility in black-and-white terms:

"Let there be a second individual, E, born from other parents in another part of the area occupied by the species, and fertile only with e, f, g. Other individuals, K, P, R, &c., may have similar relations, each infertile with the bulk of the species, fertile only with a few individuals which may be termed their physiological complements. Now each of these, separately, is a physiological variety, but the whole set, A, E, K, P, R, do not form one, but five distinct varieties. To form one variety all of them must be fertile with the same identical set of individuals of the opposite sex, and this seems to me to be so highly improbable that it must not be assumed till rigidly proved."

What is missing is that Wallace does not acknowledge the possibility of degrees of reproductive incompatibility. At the outset, individuals of group A might be more likely to produce fertile offspring when crossed with another group A member than when crossed with a member of sets E, K, P and R. But some of the offspring of the latter crosses might be fertile. On the other hand, individuals of group A (or E, K, P and R) do not (by Wallace's definition) produce fertile offspring when crossed with average members of the bulk species. The possibility of some fertile offspring is absent. In this way an initially very small reproductive advantage (affecting members of one or more of groups A, E, K, P and R) might allow the preservation of a collective entity whose members could interbreed. The effects of natural selection (should any potential beneficial phenotypic differences from the bulk species emerge) would then not be countermanded by blending effects (see http://post.queensu.ca/~forsdyke/evolutio.htm).