[1] Considerable portions of this chapter have already appeared as an article in the Contemporary Review for May, 1890. My thanks are due to the editor for kindly allowing me to reproduce them here.

[2] In as far as these sundry theories of heredity are not more or less intermediate between those of Darwin and Weismann, the differences have reference either to points of comparative detail, or else to the introduction of ideas derived from chemistry and physics—whereby it is sought to show that the principles of chemical combination and of rhythmic vibration may have a more or less considerable share in the matter. For my own part I do not see that the introduction of such ideas has been of any avail in helping—even hypothetically—to explain the phenomena of heredity; and therefore I do not deem it worth our while to consider them.

[3] See Appendix.

[4] E.g., Variation, &c., vol. i. pp. 197, 398; vol. ii. pp. 237, 252.

[5] Since this chapter was written and sent as a contribution to the Contemporary Review, Professor Weismann has published in Nature (Feb. 6. 1890) an elaborate answer to a criticism of his theory by Professor Vines (Oct. 24, 1889). In the course of this answer Professor Weismann says that he does attribute the origin of sexual reproduction to natural selection. This directly contradicts what he says in his Essays; and, for the reasons given in the text, appears to me an illogical departure from his previously logical attitude. I herewith append quotations, in order to reveal the contradiction.

“But when I maintain that the meaning of sexual reproduction is to render possible the transformation of the higher organisms by means of natural selection, such a statement is not equivalent to the assertion that sexual reproduction originally came into existence in order to achieve this end. The effects which are now produced by sexual reproduction did not constitute the causes which led to its first appearance. Sexual reproduction came into existence before it could lead to hereditary individual variability [i.e., to the possibility of natural selection]. Its first appearance must, therefore, have had some other cause [than natural selection]; but the nature of this cause can hardly be determined with any degree of certainty or precision from the facts with which we are at present acquainted.”—Essay on the Significance of Sexual Reproduction in the Theory of Natural Selection. English Translation, pp. 281-282.

“I am still of opinion that the origin of sexual reproduction depends on the advantage which it affords to the operation of natural selection.... Sexual reproduction has arisen by and for natural selection as the sole means by which individual variations can be united and combined in every possible proportion.”—Nature, vol. xli. p. 322.

How such contradictory statements can be reconciled I do not perceive; but they furnish a good example of the extreme laxity with which the term “natural selection” is used by ultra-Darwinians.

[6] The meaning of this term, however, as originally used by NÄgeli, he so greatly changes to suit the requirements of his own theory, that I think it would have been better had he coined some new one.

[7] I think it is to be regretted that for this other kind of idio-plasm (i.e., idio-plasm-B) Weismann has not coined some distinctive name, or some distinctive prefix, such as that which he sometimes employs when speaking of the other kind (i.e., idio-plasm-A)—viz., “somatic-idioplasm.” Also, the interchangeable manner in which he uses his term “idio-plasm” with the term “ nucleo-plasm,” is somewhat confusing (e.g., pp. 217, 219, 220, 250, 251, &c.). I may add that the word “plasm” in all its combinations appears to me an unfortunate one, since it seems to betoken a substance that can be seen, instead of merely inferred. But, be this as it may, the following table of terms employed may be useful for ready reference:—

[8] See close of Appendix.

[9] See Part I, figs. 36, 37, and 38. The substance of this thread, in the various phases of its segmentation, is the “chromatin,” as there depicted, and so called because it takes a stain better than other parts of the nucleus—thus showing some distinctive character.

[10] For an account of the formation and expulsion of these bodies, see Part I, pp. 125-6. There is now no longer any doubt touching the statement there made as to the male-cell likewise parting with some of its nuclear substance prior to fertilizing the female.

[11] In the case of identical twins, both are probably always produced from the same ovum.

[12] We have no means of estimating exactly the proportional number of cases in which this is possible, either among the lower or the higher plants; but it is certainly much greater than Weismann supposes. “How is it that all plants cannot be reproduced in this way?” he asks, and then adds,—“No one has ever grown a tree from the leaf of a lime or an oak, or a flowering plant from a leaf of the tulip or the convolvulus.” But I am told by botanists that the only reason why the phenomenon thus appears to be a rare one, is because it is not worth anybody’s while to grow plants in this way at a necessarily unsuitable season of the year. Thus, the Rev. George Henslow writes me:—“The fact is that any plant will reproduce itself by its leaves, provided that the cells be ‘embryonic,’ (i.e., the leaf not too near its complete development), and that it be not too thin, so as to provide enough nutriment for the bud to form till it has roots.”

[13] Intracellulare Pangenesis, s. 55.

[14] I employ the term “particles,” instead of “molecules,” because although Weismann and his followers seem to prefer the latter term, I can scarcely imagine that they intend to use it in its original, or chemical, sense.

[15] This principle will be considered at some length in my next volume.

[16] Galton first published his theory in 1872 (Proc. R. S., No. 136), but presented it in a more complete form three years later (Contemporary Review, Dec. 1875, and Journl. Anthropol. Inst., 1875).

[17] Journ. Anthropol. Inst. 1875, p. 346.

[18] Essays, &c., 2nd ed., p. 105.

[19] Essays, &c., 2nd ed., p. 100.

[20] See for example, Essays, p. 229.

[21] On previous occasions, when inconsistencies have been brought to the notice of Professor Weismann by his critics, he has complained that sufficient allowance was not made for the fact of his having published his sundry essays at different times. This, of course, is a satisfactory answer in cases where criticism refers to a growing theory, the later additions to which supersede certain parts of the earlier construction. But clearly the answer is not available in cases where one set of statements, touching fundamental principles of the theory, are directly opposed to others. A logical contradiction is not affected by dates of publication, and where the contradictory statements have reference to the vital essence of a theory, it is equally impossible for the theory to comprise them whether they be presented simultaneously or successively.

[22] The possibility of any spermatozoa of the first impregnation surviving to take part in the second is excluded by the fact that the phenomenon occurs in mammals, and, apparently, may extend over two or three litters.

[23] Possibly the school of Weismann may simply refuse to accept the facts, which are confessedly rare, and, in many of the cases alleged, dubious. In other cases, however, the evidence is sufficient to have satisfied the cautious judgement of Darwin, who has discussed it in detail. Therefore, even if the Neo-Darwinians repudiate this evidence, at least they ought to state that such is the position which they adopt.

[24] Nature, Feb. 6th, 1890.

[25] Nature, vol. xl. p. 626.

[26] Ibid., vol. xli. p. 322.

[27] In his Essays (vol. i. p. 282) Weismann says:—“If it could be shown that a purely parthenogenetic species had become transformed into a new one, such an observation would prove the existence of some new force of transformation other than selective processes, for the new species could not have been produced by these latter.” But now it has been shown that a purely parthenogenetic species can be transformed into a new one, and therefore it seems desirable to note that the observation does not so much as tend to prove the existence of some new force of transformation other than selective processes. For this most singular statement can only stand on a prior acceptance of Weismann’s own assumption, as to amphigony being the only possible cause of individual hereditary variation. Only if we have already, and with absolute certainty, embraced the whole Weismannian creed, could we consent to affirm that “natural selection is an impossibility in a species propagated by a-sexual reproduction.”

[28] What he says is:—“It was only after a greater or less number of generations had elapsed that a variable proportion of double flowers appeared, sometimes accompanied by changes in the leaves and in the colours of the flowers. This fact admits of only one interpretation:—the changed conditions at first produced slight and ineffectual changes in the idio-plasm of the individual, which was transmitted to the following generation.... Now, the idio-plasm of the first ontogenetic grade (viz., germ-plasm) alone passes from one generation to another, and hence it is clear that the germ-plasm itself must have been gradually changed by the conditions of life, until the alteration became sufficient to produce changes in the soma, which appeared as visible characters in either flower or leaf.”—Essays, pp. 426-7; italics mine.

[29] Nature, Nov. 14, 1889, p. 41.

[30] Essays, 2nd Ed., pp. 331-2.

[31] Essays, p. 296.

[32] In this connexion it ought to be observed that Darwin believed the causes of variation to be internal as well as external—or arising from “the nature of the organism” no less—or even more—than from “changed conditions of life.” But although he appears to have entertained the admixture of hereditary endowments in sexual unions as one of the causes of variation belonging to the former category, he expressly says that he did not regard it as the only, or even the main, cause. (See Variation, &c., vol. i, pp. 197, 398; vol. ii, pp. 237, 252.)

[33] See above, p. 54, note.

[34] See Darwin and after Darwin, Part I, p. 129.

[35] It must always be remembered that the view adopted by Weismann touching the nucleus (and more especially the chromosomes) of a germ-cell being the sole seat of heredity, is still far from having been established.

[36] Essays, vol. ii. p. 122.

[37] The Germ-plasm, p. 342.

[38] The Germ-plasm, p. 342.

[39] See, however, p. 83, note.

[40] Essays, vol. i. p. 101. Italics mine.

[41] The Germ-plasm, p. 406.

[42] Galton, loc. cit., pp. 343-344.

[43] Professor Weismann still maintains that there is a further important distinction between the theories of pangenesis and germ-plasm, in that the one is pre-formative while the other is epigenetic. But I am still unable to perceive that such is the case. He argues, indeed, that his new doctrine of determinants emphasizes this distinction: the argument, however, appears to me radically unsound. For instance, he says, “The hereditary continuation in each part is pre-determined in each part from the germ onwards. The right and left ears could not possibly resemble each other, if the relative strength of the hereditary tendencies on both sides were not pre-determined for all parts of the child by the nature of the paternal and maternal idants.” Very well. But, if so, the theory of determinants is just as much pre-formative as is that of gemmules. Or, conversely, the latter is quite as epigenetic as the former. Both are alike determinative, while neither supposes that the determination is due to a pre-formed miniature of the future child in the fertilized egg of its mother; but to a particulate representation in the latter of every heritable part of the former.

[44] By “germ” Galton means a carrier of heredity, which is capable of self-multiplication. In these fundamental respects, therefore, it is equivalent to a “gemmule” on the one hand and a “determinant” on the other. The three terms are so far synonymous.

[45] Loc. cit., p. 338.

[46] Loc. cit., p. 339.

[47] The Germ-plasm, pp. 199, 220.

[48] pp. 72-4.

[49] The Germ-plasm, pp. 383-386.

[50] Quoted from above, p. 78.

[51] Morph. Journal, vol. ii.

[52] See Appendix II.

[53] Essays on Heredity, vol. ii. pp. 193-4.

[54] See above, pp. 63-67.

[55] Nature, vol. xli. p. 322.

[56] The Germ-plasm, pp. 414-415. Italics Weismann’s.

[57] Essays, vol. i. p. 284.

[58] The Germ-plasm, p. 431.

[59] Variation &c., vol. i. p. 398.

[60] Ibid., vol. ii. p. 242.

[61] Nature, May 11, pp. 28-29.—In 1891-2 Professor Hartog furnished a criticism of Weismann’s theory of Heredity (Nature, vol. 44. p. 613, and Contemporary Review, July, 1892). Although disputed at the time by some of Weismann’s followers in England, this criticism was one of unquestionable cogency, and has now been recognized as such by Weismann himself (The Germ-plasm, pp. 434-5). The main point of the criticism had been missed by previous critics of Weismann, and consisted in revealing an important “difficulty” inherent in the structure of the theory itself. How far this criticism had the effect of causing Professor Weismann to abandon his theory of variation being exclusively due to amphimixis, as Professor Hartog appears to think (Nature, May 11, 1893, p. 28), is immaterial. But it must be observed that as far back as February, 1890, Professor Weismann in his answer to Professor Vines’ criticism wrote the passage already quoted on page 152.

[62] It is almost needless to say that no fault is to be found with Weismann for having thus reversed his opinion touching one of his fundamental postulates. Consistency is no merit in a man of science; and least of all where matters of such high speculation are concerned. I think, however, that it is open to question whether an author of any kind should suffer an elaborate system of theories to be published and translated, at the very time when he is himself engaged in producing another work showing the untenable character of their basal premises. At any rate, it would have saved his English readers no small trouble and confusion, if Weismann had added notes to the translations of his essays on Polar Bodies, on The Significance of Sexual Reproduction, and on Amphimixis, to the effect that he had abandoned some of their most distinctive features before the translations had gone to press.

[63] See especially pp. 86-89. All that is there said about the unicellular organisms is not, in the present connexion, affected by Weismann’s change of view with regard to them. We have only to substitute “primordial” or “protoplasmic” for “unicellular,” and nearly all the points of the criticism remain.

[64] Professor Weismann has now considered more fully than heretofore the phenomena of bud-variation (The Germ-plasm, pp. 439-442); but as he continues (though with diffidence) to take substantially the same view of them as that which I have already quoted on pp. 95-96, it is needless for me to re-discuss the matter here.

[65] “Rejuvenescence” means the renewal of vital energies which is supposed to result from a fusion of the contents of two cells. For an excellent discussion of this and the other theories on the object of sexual propagation, see a brief article by Professor Marcus Hartog, in the Contemporary Review for July, 1892. Since then Weismann has published The Germ-plasm, and here his main argument against this theory is that tens, or even hundreds of generations of unicellular organisms have been observed to succeed one another before any act of conjugation takes place. But I cannot see that it signifies how many generations may in different species be proved capable of resulting from a single act of conjugation. Weismann himself now accepts the analogy between cell-proliferation as resulting from conjugation in unicellular organisms, and from fertilization in multicellular. But even three hundred generations of the former can scarcely be regarded as equal to all the “ontogenetic stages” of the latter.

[66] This view of the function of sexual propagation is now universally ascribed to Strasburger, and it is quite true that he has independently adduced it. But as this was not done until about ten years after it had been published by Galton, I have designedly associated the idea with Galton’s name. The following are the words in which it was announced by him:—

“The necessity of a system of double parentage in complex organisations is the immediate consequence of a theory of organic units and germs, as we shall see if we fix our attention upon any one definite series of unisexual descents, and follow out its history. Suppose we select, cut off, and plant the second bud, then after it has grown to maturity we similarly take the second of its buds, and so on consecutively. At each successive stage there is always a chance of some one or more of the various species of germs in the stirp dying out, or being omitted; and of course when they are gone they are lost for ever, and are irreplaceable by others. From time to time this chance must fall unfavourably, and will cause a deficiency in some of the structural elements, and a consequent deterioration of the race. If the loss be vital, this particular line of descent will of course be extinguished at once; but on the more favourable supposition, the race will linger on, submitting to successive decrements in its constituent elements, until the accumulation of small losses becomes fatal.”—loc. cit., p. 333.

Galton also points out a further advantage that is secured by “amphimixis,” and one which shows the non-necessity of what remains of Weismann’s theory of polar bodies, thus:—

“There is yet another advantage in double parentage, namely, that as the stirp whence the child sprang can only be half the size of the combined stirps of his two parents, it follows that one half of his possible heritage must have been suppressed. This implies a sharp struggle for place among the competing germs, and the success, as we may infer, of the fitter half of their numerous varieties.”—loc. cit., p. 334.

[67] In fact, it seems to me that this is the sole supposition whereby it can be held that sexual propagation has been developed both “by” and “for” natural selection, in order to supply variations as material for the action of this principle. Natural selection cannot thus supply the conditions to its own activity, if, as Weismann supposes, there is but one purpose for it to subserve (see above, pp. 13-15). But, if it is acting for more than one purpose, the “by” and the “for” argument may hold.

[68] I find that a passage explaining the sense in which I use these terms has been accidentally omitted from Chapter III, where they are first introduced; and, as the sheets of that chapter have been already printed off, I here supply the omission. The terms in italics are not Weismann’s, and I have employed them merely for the purpose of giving precision to his views. By “absolute stability of germ-plasm” I mean to indicate that degree of stability which he has hitherto postulated as the necessary basis for his doctrine of heritable variations being solely due to admixtures of germ-plasm in sexual unions. By “perpetual continuity of germ-plasm” I intend to denote that amount of continuity which he still postulates as the necessary basis for his correlative doctrine touching the non-inheritance of acquired characters.

[69] Essays, pp. 76-77, from which the following quotations are likewise taken seriatim.

[70] “Or, more precisely, they must give up as many molecules as would correspond to the number of the kind of cell in question found in the mature organism.” Of course by “molecules” Weismann means what Darwin does by “gemmules.”

[71] If there are such things as gemmules, it appears to me to follow that the only physiological distinction between the reproductive glands and glands in general is, that the former discharge their products in the form of living cells. Even here, however, there appears to be one analogous case in those salivary glands which discharge the so-called salivary corpuscles—i.e., nucleated cells, undergoing amoeboid changes of form, and exhibiting the movements of living protoplasm in their interior.

[72] Variation, &c., 2nd ed., vol. ii. pp. 374-6.

[73] Nature, vol. xl. p. 624. Weismann’s answer to this and other parts of Professor Vines’ criticism where the term “somato-plasm” occurs, will be considered later on.

[74] Weismann speaks disparagingly of Darwin’s theory as a “theory of preformation” (p. 316). “We must assume,” he adds by way of explanation, “that each single part of the body at each developmental stage is, from the first, represented in the germ-cell as distinct particles of matter, which will reproduce each part of the body at its appropriate stage as their turn for development arrives.” But must we not likewise “assume” exactly the same thing in the case of Weismann’s own theory? To me, at any rate, it appears that the description is quite as appropriate to germ-plasm as it is to gemmules. Nor can I see any distinction, even where he seeks to draw it more expressly, as for instance—“Every detail in the whole organism must be represented in the germ-plasm by its own special and peculiar arrangement of the groups of molecules, ... not indeed as the pre-formed germs of structure (the gemmules of pangenesis), but as variations in its molecular constitution.” [Essays, p. 194.] Again, on page 325 he gives a foot-note explaining the distinction by alluding to the controversy between the preformationists and epigenesists. But the theory of pangenesis does not suppose the future organism to exist in the egg-cell as a miniature: it supposes merely that every part of the future organism is represented in the egg-cell by corresponding material particles. And this, as far as I can understand, is exactly what the theory of germ-plasm supposes; only it calls the particles “molecules,” and seemingly attaches more importance to the matter of variations in their arrangement or “constitution,” whatever these vague expressions may be intended to signify.

[75] ‘Philosophical Transactions of the Royal Society for the Year 1821,’ Part I. pp, 20-24.

[76] Readers who may happen to be acquainted with De Vries’ important essay on heredity will perceive how well this suggestion fits in with his modification of Pangenesis.

[77] As already indicated, I cannot gather from his remarks on the subject which, if any, of the alternative interpretations of the phenomena that we are considering Mr. Spencer adopts. From the following sentences it would appear that he assigns yet a third interpretation, and this as the only possible one. For he says of these phenomena: “They prove that while the reproductive cells multiply and arrange themselves during the evolution of the embryo, some of their germ-plasm passes into the mass of somatic cells constituting the parental body, and becomes a permanent component of it. Further, they necessitate the inference that this introduced germ-plasm, everywhere diffused, is some of it included in the reproductive cells subsequently formed” (Contemporary Review, March, p. 452). This appears to mean that the influence of a previous sire can only be explained by supposing that the developing embryo inoculates the somatic tissues of its mother with hereditary material derived from its father, and that the maternal tissue afterwards reflect some of this material (or its influence) to the still unripe ovarian ova. If this be the hypothesis intended, it seems to me more complex than any of the three which I have suggested. But, be this as it may, we certainly cannot agree that such an hypothesis is “proved” by the facts, or that the latter “necessitate” the inference as to its being some of the embryo’s germinal matter which enters the unripe ova.

[78] “A Text Book of Human Physiology.” By Austin Flint, M.D., LL. D. Fourth edition. New York: D. Appleton & Co. 1888. Page 797.