The “one or two” run to three, viz. Stocks (hoariness and colour); Datura (character of fruits and colour of flowers); and lastly colours of Rats and Mice. Each of these subjects, as it happens, has been referred to in the forthcoming paper by Miss Saunders and myself. Datura and Matthiola have been subjected to several years’ experiment and I venture to refer the reader who desires to see whether the facts are or are not in accord with Mendel’s expectation and how far there is “fundamental inconsistency” amongst them to a perusal of our work.

But as Professor Weldon refers to some points that have not been explicitly dealt with there, it will be safer to make each clear as we proceed.

1. Stocks (Matthiola). Professor Weldon quotes Correns’ observation that glabrous Stocks crossed with hoary gave offspring all hoary, while Trevor Clarke thus obtained some hoary and some glabrous. As there are some twenty different sorts of Stocks128 it is not surprising that different observers should have chanced on different materials and obtained different results. Miss Saunders has investigated laws of heredity in Stocks on a large scale and an account of her results is included in our forthcoming Report. Here it must suffice to say that the cross hoary ♀ × glabrous ♂ always gave offspring all hoary except once: that the cross glabrous ♀ × hoary ♂ of several types gave all hoary; but the same cross using other hoary types did frequently give a mixture, some of the offspring being hoary, others glabrous. Professor Weldon might immediately decide that here was the hoped for phenomenon of “reversed” dominance, due to ancestry, but here again that hypothesis is excluded. For the glabrous (recessive) cross-breds were pure, and produced on self-fertilisation glabrous plants only, being in fact, almost beyond question, “false hybrids” (see p. 34), a specific phenomenon which has nothing to do with the question of dominance.

Professor Weldon next suggests that there is discrepancy between the observations as to flower-colour. He tells us that Correns found violet Stocks crossed with “yellowish white” gave violet or shades of violet flaked together. According to Professor Weldon

On comparing Professor Weldon’s version with the originals we find the missing explanations. Having served some apprenticeship to the breeding of Stocks, we, here, are perhaps in a better position to take the points, but it is to me perfectly inexplicable how in such a simple matter as this he can have gone wrong.

Note then

(1) That Nobbe does not specify which colours he crossed together, beyond the fact that white was crossed with each fertile form. The crimson form (Karmoisinfarbe), being double to the point of sterility, was not used. There remain then, white, carmine, and two purples (violet, “dark blue”). When white was crossed with either of these, Nobbe says the colour becomes paler, whichever sort gave the pollen. Nobbe does not state that he crossed carmine with the purples.

(2) Professor Weldon gives no qualification in his version. Nobbe however states that he found it very difficult to distinguish the result of crossing carmine with white from that obtained by crossing dark blue or violet with white130‍, thereby nullifying Professor Weldon’s statement that in every case the cross was a simple mixture of the parental colours—a proposition sufficiently disproved by Miss Saunders’ elaborate experiments.

(3) Lately the champion of the “importance of small variations,” Professor Weldon now prefers to treat the distinctions between established varieties as negligible fluctuations instead of specific phenomena131. Therefore when Correns using “yellowish white” obtained one result and Nobbe using “white” obtained another, Professor Weldon hurries to the conclusion that the results are comparable and therefore contradictory. Correns however though calling his flowers gelblich-weiss is careful to state that they are described by Haage and Schmidt (the seed-men) as “schwefel-gelb” or sulphur-yellow. The topics Professor Weldon treats are so numerous that we cannot fairly expect him to be personally acquainted with all; still had he looked at Stocks before writing, or even at the literature relating to them, he would have easily seen that these yellow Stocks are a thoroughly distinct form132; and in accordance with this fact it would be surprising if they had not a distinctive behaviour in their crosses. To use our own terminology their colour character depends almost certainly on a compound allelomorph. Consequently there is no evidence of contradiction in the results, and appeal to ancestry is as unnecessary as futile.

2. Datura. As for the evidence on Datura, I must refer the reader again to the experiments set forth in our Report.

The phenomena obey the ordinary Mendelian rules with accuracy. There are (as almost always where discontinuous variation is concerned) occasional cases of “mosaics,” a phenomenon which has nothing to do with “ancestry.”

3. Colours of Rats and Mice. Professor Weldon reserves his collection of evidence on this subject for the last. In it we reach an indisputable contribution to the discussion—a reference to Crampe’s papers, which together constitute without doubt the best evidence yet published, respecting colour-heredity in an animal. So far as I have discovered, the only previous reference to these memoirs is that of Ritzema Bos133, who alludes to them in a consideration of the alleged deterioration due to in-breeding.

Now Crampe through a long period of years made an exhaustive study of the peculiarities of the colour-forms of Rats, white, black, grey and their piebalds, as exhibited in Heredity.

Till the appearance of Professor Weldon’s article Crampe’s work was unknown to me, and all students of Heredity owe him a debt for putting it into general circulation. My attention had however been called by Dr Correns to the interesting results obtained by von Guaita, experimenting with crosses originally made between albino mice and piebald Japanese waltzing mice. This paper also gives full details of an elaborate investigation admirably carried out and recorded.

In the light of modern knowledge both these two researches furnish material of the most convincing character demonstrating the Mendelian principles. It would be a useful task to go over the evidence they contain and rearrange it in illustration of the laws now perceived. To do this here is manifestly impossible, and it must suffice to point out that the albino is a simple recessive in both cases (the waltzing character in mice being also a recessive), and that the “wild grey” form is one of the commonest heterozygotes—there appearing, like the yellow cotyledon-colour of peas, in either of two capacities, i.e. as a pure form, or as the heterozygote form of one or more combinations134.

Professor Weldon refers to both Crampe and von Guaita, whose results show an essential harmony in the fact that both found albino an obvious recessive, pure almost without exception, while the coloured forms show various phenomena of dominance. Both found heterozygous colour-types. He then searches for something that looks like a contradiction. Of this there is no lack in the works of Johann von Fischer (11)—an authority of a very different character—whom he quotes in the following few words:

But this is doing small justice to the completeness of Johann von Fischer’s statement, which is indeed a proposition of much more amazing import.

That investigator in fact began by a study of the cross between the albino Ferret and the Polecat, as a means of testing whether they were two species or merely varieties. The cross, he found, was in colour and form a blend of the parental types. Therefore, he declares, the Ferret and the Polecat are two distinct species, because, “as everybody ought to know,”

whereas in crosses (between species) this is not the case.

And again, after reciting that the Ferret-Polecat crosses gave intermediates, he states:

These are admirable illustrations of what is meant by a “universal” proposition. But von Fischer doesn’t stop here. He proceeds to give a collection of evidence in proof of this truth which he says “ought to be known to everyone.” He has observed the fact in regard to albino mole, albino shrew (Sorex araneus), melanic squirrel (Sciurus vulgaris), albino ground-squirrel (Hypudaeus terrestris), albino hamster, albino rats, albino mice, piebald (grey-and-white or black-and-white) mice and rats, partially albino sparrow, and we are even presented with two cases in Man. No single exception was known to von Fischer137.

In his subsequent paper von Fischer declares that from matings of rats in which the mothers were grey and the fathers albino he bred 2017 pure albinos; and from albino mothers and grey fathers 3830 normal greys. “Not a single individual varied in any respect, or was in any way intermediate.”

With piebalds the same result is asserted, save that certain melanic forms appeared. Finally von Fischer repeats his laws already reached, giving them now in this form: that if the offspring of a cross show only the colour of the father, then the parents are varieties of one species; but if the colour of the offspring be intermediate or different from that of the father, then the parents belong to distinct species.

The reader may have already gathered that we have here that bane of the advocate—the witness who proves too much. But why does Professor Weldon confine von Fischer to the few modest words recited above? That author has—so far as colour is concerned—a complete law of heredity supported by copious “observations.” Why go further?

Professor Weldon “brings forth these strong reasons” of the rats and mice with the introductory sentence:

So once again Professor Weldon suggests that his laws of ancestry will explain even the discrepancies between von Fischer on the one hand and Crampe and von Guaita on the other but he does not tell us how he proposes to apply them.

In the cross between the albino and the grey von Fischer tells us that both colours appear in the offspring, but always, without exception or variation, that of the father only, in 5847 individuals.

Surely, the law of ancestry, if he had a moment’s confidence in it, might rather have warned Professor Weldon that von Fischer’s results were wrong somewhere, of which there cannot be any serious doubt. The precise source of error is not easy to specify, but probably carelessness and strong preconception of the expected result were largely responsible, though von Fischer says he did all the recording most carefully himself.

Such then is the evidence resting “on excellent authority”: may we some day be privileged to see the “examples which may be doubted”?

The case of mice, invoked by Professor Weldon, has also been referred to in our Report. Its extraordinary value as illustrating Mendel’s principles and the beautiful way in which that case may lead on to extensions of those principles are also there set forth (see the present Introduction, p. 25). Most if not all of such “conflicting” evidence can be reconciled by the steady application of the Mendelian principle that the progeny will be constant when—and only when138—similar gametes meet in fertilisation, apart from any question of the characters of the parent which produces those gametes.

V. Professor Weldon’s quotations from Laxton.

In support of his conclusions Professor Weldon adduces two passages from Laxton, some of whose testimony we have just considered. This further evidence of Laxton is so important that I reproduce it in full. The first passage, published in 1866, is as follows:—

Here Professor Weldon’s quotation ceases. It is unfortunate he did not read on into the very next sentence with which the paragraph concludes:—

Now if Professor Weldon instead of leaving off on the word “ancestry” had noticed this passage, I think his article would never have been written.

Laxton proceeds:—

The significance of this latter testimony I will presently discuss.

Professor Weldon next appeals to a later paper of Laxton’s published in 1890. From it he quotes this passage:

Here again the reader would have gained had Professor Weldon, instead of leaving off at the comma, gone on to the end of the paragraph, which proceeds thus:—

Professor Weldon declares that Laxton’s “experience was altogether different from that of Mendel.” The reader will bear in mind that when Laxton speaks of fixing a variety he is not thinking particularly of seed-characters, but of all the complex characters, fertility, size, flavour, season of maturity, hardiness, etc., which go to make a serviceable pea. Considered carefully, Laxton’s testimony is so closely in accord with Mendelian expectation that I can imagine no chance description in non-Mendelian language more accurately stating the phenomena.

Here we are told in unmistakable terms the breaking up of the original combination of characters on crossing, their re-arrangement, that at the fourth or fifth generation the possibilities of sporting [sub-division of compound allelomorphs and re-combinations of them?] are exhausted, that there are then definite forms which if selected are thenceforth fixed [produced by union of similar gametes?] that it takes longer to select some forms [dominants?] than others [recessives?], that there may be “mule” forms144 or forms which cannot be fixed at all145 [produced by union of dissimilar gametes?].

But Laxton tells us more than this. He shows us that numbers of varieties may be obtained—hundreds—“incalculable numbers.” Here too if Professor Weldon had followed Mendel with even moderate care he would have found the secret. For in dealing with the crosses of Phaseolus Mendel clearly forecasts the conception of compound characters themselves again consisting of definite units, all of which may be separated and re-combined in the possible combinations, laying for us the foundation of the new science of Analytical Biology.

How did Professor Weldon, after reading Mendel, fail to perceive these principles permeating Laxton’s facts? Laxton must have seen the very things that Mendel saw, and had he with his other gifts combined that penetration which detects a great principle hidden in the thin mist of “exceptions,” we should have been able to claim for him that honour which must ever be Mendel’s in the history of discovery.

When Laxton speaks of selection and the need for it, he means, what the raiser of new varieties almost always means, the selection of definite forms, not impalpable fluctuations. When he says that without selection there will be utter confusion, he means—to use Mendelian terms—that the plant which shows the desired combination of characters must be chosen and bred from, and that if this be not done the grower will have endless combinations mixed together in his stock. If however such a selection be made in the fourth or fifth generation the breeder may very possibly have got a fixed form—namely, one that will breed true146. On the other hand he may light on one that does not breed true, and in the latter case it may be that the particular type he has chosen is not represented in the gametes and will never breed true, though selected to the end of time. Of all this Mendel has given us the simple and final account.

At Messrs Sutton and Sons, to whom I am most grateful for unlimited opportunities of study, I have seen exactly such a case as this. For many years Messrs Sutton have been engaged in developing new strains of the Chinese Primrose (Primula sinensis, hort.). Some thirty thoroughly distinct and striking varieties (not counting the Stellata or “Star” section) have already been produced which breed true or very nearly so. In 1899 Messrs Sutton called my attention to a strain known as “Giant Lavender,” a particularly fine form with pale magenta or lavender flowers, telling me that it had never become fixed. On examination it appeared that self-fertilised seed saved from this variety gave some magenta-reds, some lavenders, and some which are white on opening but tinge with very faint pink as the flower matures.

On counting these three forms in two successive years the following figures appeared. Two separately bred batches raised from “Giant Lavender” were counted in each year.

The numbers 54 : 96 : 45 approach the ratio 1 : 2 : 1 so nearly that there can be no doubt we have here a simple case of Mendelian laws, operating without definite dominance, but rather with blending.

When Laxton speaks of the “remarkably fine but unfixable pea Evolution” we now know for the first time exactly what the phenomenon meant. It, like the “Giant Lavender,” was a “mule” form, not represented by germ-cells, and in each year arose by “self-crossing.”

This is only one case among many similar ones seen in the Chinese Primrose. In others there is no doubt that more complex factors are at work, the subdivision of compound characters, and so on. The history of the “Giant Lavender” goes back many years and is not known with sufficient precision for our purposes, but like all these forms it originated from crossings among the old simple colour varieties of sinensis.

VI. The Argument Built on Exceptions.

So much for the enormous advance that the Mendelian principles already permit us to make. But what does Professor Weldon offer to substitute for all this? Nothing.

Professor Weldon suggests that a study of ancestry will help us. Having recited Tschermak’s exceptions and the great irregularities seen in the Telephone group, he writes:

Premising that when pure types are used the exceptions form but a small part of the whole, and that any supposed absence of “segregation” may have been variation, this statement is perfectly sound. He proceeds:—

In this passage the Mendelian view is none too precisely represented. I should rather have said that it was from Mendel, first of all men, that we have learnt not to regard the effects produced on offspring “as due to the existence in the parent of particular structural characters.” We have come rather to disregard the particular structure of the parent except in so far as it may give us a guide as to the nature of its gametes.

This indication, if taken in the positive sense—as was sufficiently shown in considering the significance of the “mule” form or “hybrid-character”—we now know may be absolutely worthless, and in any unfamiliar case is very likely to be so. Mendel has proved that the inheritance from individuals of identical ancestry may be entirely different: that from identical ancestry, without new variation, may be produced three kinds of individuals (in respect of each pair of characters), namely, individuals capable of transmitting one type, or another type, or both: moreover that the statistical relations of these three classes of individuals to each other will in a great number of cases be a definite one: and of all this he shows a complete account.

Professor Weldon cannot deal with any part of this phenomenon. He does little more than allude to it in passing and point out exceptional cases. These he suggests a study of ancestry will explain.

As a matter of fact a study of ancestry will give little guide—perhaps none—even as to the probability of the phenomenon of dominance of a character, none as to the probability of normal “purity” of germ-cells. Still less will it help to account for fluctuations in dominance, or irregularities in “purity.”

Ancestry and Dominance.

In a series of astonishing paragraphs (pp. 241–2) Professor Weldon rises by gradual steps, from the exceptional facts regarding occasional dominance of green colour in Telephone to suggest that the whole phenomenon of dominance may be attributable to ancestry, and that in fact one character has no natural dominance over another, apart from what has been created by selection of ancestry. This piece of reasoning, one of the most remarkable examples of special pleading to be met with in scientific literature, must be read as a whole. I reproduce it entire, that the reader may appreciate this curious effort. The remarks between round parenthetical marks are Professor Weldon’s, those between crotchets are mine.

Let us pause a moment before proceeding to the climax. Let the reader note we have been told of two groups of cases in which dominance of yellow failed or was irregular. (Why are not Gärtner’s and Seton’s “exceptions” referred to here?) In one of these groups Couturier was always one parent, either father or mother, and were it not for Tschermak’s own obvious hesitation in regard to his own exceptions (see p. 148), I would gladly believe that Couturier—a form I do not know—may be an exceptional variety. How Professor Weldon proposes to explain its peculiarities by reference to ancestry he omits to tell us. The Buchsbaum case is already disposed of, for on Tschermak’s showing, it is an unstable form.

Happily, thanks to Professor Weldon, we know rather more of the third case, that of Telephone, which, whether as father or mother, was frequently found by Tschermak to give either green, greenish, or patchwork-seeds when crossed with yellow varieties. It behaves, in short, “like a green-seeded pea of exceptional dominance,” as we are now told. For this dominant quality of Telephone’s greenness we are asked to account by appeal to its ancestry. May we not expect, then, this Telephone to be—if not a pure-bred green pea from time immemorial—at least as pure-bred as other green peas which do not exhibit dominance of green at all? Now, what is Telephone? Do not let us ask too much. Ancestry takes a lot of proving. We would not reject him “parce qu’il n’avait que soixante & onze quartiers, & que le reste de son arbre généalogique avait été perdu par l’injure du tems.”

But with stupefaction we learn from Professor Weldon himself that Telephone is the very variety which he takes as his type of a permanent and incorrigible mongrel, a character it thoroughly deserves.

From Telephone he made his colour scale! Tschermak declares the cotyledons to be “yellowish or whitish green, often entirely bright yellow151.” So little is it a thorough-bred green pea, that it cannot always keep its own self-fertilised offspring green. Not only is this pea a parti-coloured mongrel, but Professor Weldon himself quotes Culverwell that as late as 1882 both Telegraph and Telephone “will always come from one sort, more especially from the green variety”; and again regarding a supposed good sample of Telegraph that “Strange to say, although the peas were taken from one lot, those sown in January produced a great proportion of the light variety known as Telephone. These were of every shade of light green up to white, and could have been shown for either variety,” Gard. Chron. 1882 (2), p. 150. This is the variety whose green, it is suggested, partially “dominates” over the yellow of Pois d’Auvergne, a yellow variety which has a clear lineage of about a century, and probably more. If, therefore, the facts regarding Telephone have any bearing on the significance of ancestry, they point the opposite way from that in which Professor Weldon desires to proceed.

In view of the evidence, the conclusion is forced upon me that the suggestion that “ancestry” may explain the facts regarding Telephone has no meaning behind it, but is merely a verbal obstacle. Two words more on Telephone. On p. 147 I ventured to hint that if we try to understand the nature of the appearance of green in the offspring of Telephone bred with yellow varieties, we are more likely to do so by comparing the facts with those of false hybridisation than with fluctuations in dominance. In this connection I would call the reader’s attention to a point Professor Weldon misses, that Tschermak also got yellowish-green seeds from Fillbasket (green) crossed with Telephone. I suggest therefore that Telephone’s allelomorphs may be in part transmitted to its offspring in a state which needs no union with any corresponding allelomorph of the other gamete, just as may the allelomorphs of “false hybrids.” It would be quite out of place here to pursue this reasoning, but the reader acquainted with special phenomena of heredity will probably be able fruitfully to extend it. It will be remembered that we have already seen the further fact that the behaviour of Telephone in respect to seed-shape was also peculiar (see p. 152).

Whatever the future may decide on this interesting question it is evident that with Telephone (and possibly Buchsbaum) we are encountering a specific phenomenon, which calls for specific elucidation and not a case simply comparable with or contradicting the evidence of dominance in general.

In this excursion we have seen something more of the “exceptions.” Many have fallen, but some still stand, though even as to part of the remainder Tschermak entertains some doubts, and, it will be remembered, cautions his reader that of his exceptions some may be self-fertilisations, and some did not germinate152. Truly a slender basis to carry the coming structure!

But Professor Weldon cannot be warned. He told us the “law of dominance conspicuously fails for crosses between certain races.” Thence the start. I venture to give the steps in this impetuous argument. There are exceptions153—a fair number if we count the bad ones—there may be more—must be more—are more—no doubt many more: so to the brink. Then the bold leap: may there not be as many cases one way as the other? We have not tried half the sorts of Peas yet. There is still hope. True we know dominance of many characters in some hundreds of crosses, using some twenty varieties—not to speak of other plants and animals—but we do know some exceptions, of which a few are still good. So dominance may yet be all a myth, built up out of the petty facts those purblind experimenters chanced to gather. Let us take wider views. Let us look at fields more propitious—more what we would have them be! Let us turn to eye-colour: at least there is no dominance in that. Thus Professor Weldon, telling us that Mendel “had no proper basis for generalisation about yellow and green peas of any ancestry,” proceeds to this lamentable passage:—

The suggestion amounts to this: that because there are exceptions to dominance in peas; and because by some stupendous coincidence, or still more amazing incompetence, a bungler might have thought he found dominance of one eye-colour whereas really there was none154; therefore Professor Weldon is at liberty to suggest there is a fair chance that Mendel and all who have followed him have either been the victims of this preposterous coincidence not once, but again and again; or else persisted in the same egregious and perfectly gratuitous blunder. Professor Weldon is skilled in the Calculus of Chance: will he compute the probabilities in favour of his hypothesis?

Ancestry and purity of germ-cells.

To what extent ancestry is likely to elucidate dominance we have now seen. We will briefly consider how laws derived from ancestry stand in regard to segregation of characters among the gametes.

For Professor Weldon suggests that his view of ancestry will explain the facts not only in regard to dominance and its fluctuations but in regard to the purity of the germ-cells. He does not apply this suggestion in detail, for its error would be immediately exposed. In every strictly Mendelian case the ancestry of the pure extracted recessives or dominants, arising from the breeding of first crosses, is identical with that of the impure dominants [or impure recessives in cases where they exist]. Yet the posterity of each is wholly different. The pure extracted forms, in these simplest cases, are no more likely to produce the form with which they have been crossed than was their pure grandparent; while the impure forms break up again into both grand-parental forms.

Ancestry does not touch these facts in the least. They and others like them have been a stumbling-block to all naturalists. Of such paradoxical phenomena Mendel now gives us the complete and final account. Will Professor Weldon indicate how he proposes to regard them?

Let me here call the reader’s particular attention to that section of Mendel’s experiments to which Professor Weldon does not so much as allude. Not only did Mendel study the results of allowing his cross-breds (DR’s) to fertilise themselves, giving the memorable ratio

1 DD : 2 DR : 1 RR,

but he fertilised those cross-breds (DR’s) both with the pure dominant (D) and with the pure recessive (R) varieties reciprocally, obtaining in the former case the ratio

1 DD : 1 DR

and in the latter the ratio

1 DR : 1 RR.

The DD group and the RR group thus produced giving on self-fertilisation pure D offspring and pure R offspring respectively, while the DR groups gave again

1 DD : 2 DR : 1 RR.

How does Professor Weldon propose to deal with these results, and by what reasoning can he suggest that considerations of ancestry are to be applied to them? If I may venture to suggest what was in Mendel’s mind when he applied this further test to his principles it was perhaps some such considerations as the following. Knowing that the cross-breds on self-fertilisation give

1 DD : 2 DR : 1 RR

three explanations are possible:

(a) These cross-breds may produce pure D germs of both sexes and pure R germs of both sexes on an average in equal numbers.

(b) Either the female, or the male, gametes may be alone differentiated according to the allelomorphs, into pure D’s, pure R’s, and crosses DR or RD, the gametes of the other sex being homogeneous and neutral in regard to those allelomorphs.

(c) There may be some neutralisation or cancelling between characters in fertilisation occurring in such a way that the well-known ratios resulted. The absence of and inability to transmit the D character in the RR’s, for instance, might have been due not to the original purity of the germs constituting them, but to some condition incidental to or connected with fertilisation.

It is clear that Mendel realized (b) as a possibility, for he says DR was fertilised with the pure forms to test the composition of its egg-cells, but the reciprocal crosses were made to test the composition of the pollen of the hybrids. Readers familiar with the literature will know that both Gärtner and Wichura had in many instances shown that the offspring of crosses in the form (a × b) ♀ × c ♂ were less variable than those of crosses in the form a ♀ × (b × c) ♂, &c. This important fact in many cases is observed, and points to differentiation of characters occurring frequently among the male gametes when it does not occur or is much less marked among the maternal gametes. Mendel of course knew this, and proceeded to test for such a possibility, finding by the result that differentiation was the same in the gametes of both sexes155.

Of hypotheses (b) and (c) the results of recrossing with the two pure forms dispose; and we can suggest no hypothesis but (a) which gives an acceptable account of the facts.

It is the purity of the “extracted” recessives and the “extracted” dominants—primarily the former, as being easier to recognize—that constitutes the real proof of the validity of Mendel’s principle.

Using this principle we reach immediately results of the most far-reaching character. These theoretical deductions cannot be further treated here—but of the practical use of the principle a word may be said. Where-ever there is marked dominance of one character the breeder can at once get an indication of the amount of trouble he will have in getting his cross-bred true to either dominant or recessive character. He can only thus forecast the future of the race in regard to each such pair of characters taken severally, but this is an immeasurable advance on anything we knew before. More than this, it is certain that in some cases he will be able to detect the “mule” or heterozygous forms by the statistical frequency of their occurrence or by their structure, especially when dominance is absent, and sometimes even in cases where there is distinct dominance. With peas, the practical seedsman cares, as it happens, little or nothing for those simple characters of seed-structure, &c. that Mendel dealt with. He is concerned with size, fertility, flavour, and numerous similar characters. It is to these that Laxton (invoked by Professor Weldon) primarily refers, when he speaks of the elaborate selections which are needed to fix his novelties.

We may now point tentatively to the way in which some even of these complex cases may be elucidated by an extension of Mendel’s principle, though we cannot forget that there are other undetected factors at work.

The value of the appeal to Ancestry.

But it may be said that Professor Weldon’s appeal to ancestry calls for more specific treatment. When he suggests ancestry as “one great reason” for the different properties displayed by different races or individuals, and as providing an account of other special phenomena of heredity, he is perhaps not to be taken to mean any definite ancestry, known or hypothetical. He may, in fact, be using the term “ancestry” merely as a brief equivalent signifying the previous history of the race or individual in question. But if such a plea be put forward, the real utility and value of the appeal to ancestry is even less evident than before.

Ancestry, as used in the method of Galton and Pearson, means a definite thing. The whole merit of that method lies in the fact that by it a definite accord could be proved to exist between the observed characters and behaviour of specified descendants and the ascertained composition of their pedigree. Professor Weldon in now attributing the observed peculiarities of Telephone &c. to conjectural peculiarities of pedigree—if this be his meaning—renounces all that had positive value in the reference to ancestry. His is simply an appeal to ignorance. The introduction of the word “ancestry” in this sense contributes nothing. The suggestion that ancestry might explain peculiarities means no more than “we do not know how peculiarities are to be explained.” So Professor Weldon’s phrase “peas of probably similar ancestral history156” means “peas probably similar”; when he speaks of Mendel having obtained his results with “a few pairs of plants of known ancestry157,” he means “a few pairs of known plants” and no more; when he writes that “the law of segregation, like the law of dominance appears to hold only for races of particular ancestry158,” the statement loses nothing if we write simply “for particular races.” We all know—the Mendelian, best of all—that particular races and particular individuals may, even though indistinguishable by any other test, exhibit peculiarities in heredity.

But though on analysis those introductions of the word “ancestry” are found to add nothing, yet we can feel that as used by Professor Weldon they are intended to mean a great deal. Though the appeal may be confessedly to ignorance, the suggestion is implied that if we did know the pedigrees of these various forms we should then have some real light on their present structure or their present behaviour in breeding. Unfortunately there is not the smallest ground for even this hope.

As Professor Weldon himself tells us159, conclusions from pedigree must be based on the conditions of the several ancestors; and even more categorically (p. 244), “The degree to which a parental character affects offspring depends not only upon its development in the individual parent, but on its degree of development in the ancestors of that parent.” [My italics.] Having rehearsed this profession of an older faith Professor Weldon proceeds to stultify it in his very next paragraph. For there he once again reminds us that Telephone, the mongrel pea of recent origin, which does not breed true to seed characters, has yet manifested the peculiar power of stamping the recessive characters on its cross-bred offspring, though pure and stable varieties that have exhibited the same characters in a high degree for generations have not that power. As we now know, the presence or absence of a character in a progenitor may be no indication whatever as to the probable presence of the character in the offspring; for the characters of the latter depend on gametic and not on zygotic differentiation.

The problem is of a different order of complexity from that which Professor Weldon suggests, and facts like these justify the affirmation that if we could at this moment bring together the whole series of individuals forming the pedigree of Telephone, or of any other plant or animal known to be aberrant as regards heredity, we should have no more knowledge of the nature of these aberrations; no more prescience of the moment at which they would begin, or of their probable modes of manifestation; no more criterion in fact as to the behaviour such an individual would exhibit in crossing160, or solid ground from which to forecast its posterity, than we have already. We should learn then—what we know already—that at some particular point of time its peculiar constitution was created, and that its peculiar properties then manifested themselves: how or why this came about, we should no more comprehend with the full ancestral series before us, than we can in ignorance of the ancestry. Some cross-breds follow Mendelian segregation; others do not. In some, palpable dominance appears; in others it is absent.

If there were no ancestry, there would be no posterity. But to answer the question why certain of the posterity depart from the rule which others follow, we must know, not the ancestry, but how it came about either that at a certain moment a certain gamete divided from its fellows in a special and unwonted fashion; or, though the words are in part tautological, the reason why the union of two particular gametes in fertilisation took place in such a way that gametes having new specific properties resulted161. No one yet knows how to use the facts of ancestry for the elucidation of these questions, or how to get from them a truth more precise than that contained in the statement that a diversity of specific consequences (in heredity) may follow an apparently single specific disturbance. Rarely even can we see so much. The appeal to ancestry, as introduced by Professor Weldon, masks the difficulty he dare not face.

In other words, it is the cause of variation we are here seeking. To attack that problem no one has yet shown the way. Knowledge of a different order is wanted for that task; and a compilation of ancestry, valuable as the exercise may be, does not provide that particular kind of knowledge.

Of course when once we have discovered by experiment that—say, Telephone—manifests a peculiar behaviour in heredity, we can perhaps make certain forecasts regarding it with fair correctness; but that any given race or individual will behave in such a way, is a fact not deducible from its ancestry, for the simple reason that organisms of identical ancestry may behave in wholly distinct, though often definite, ways.

It is from this hitherto hopeless paradox that Mendel has begun at last to deliver us. The appeal to ancestry is a substitution of darkness for light.

VII. The question of absolute purity of germ-cells.

But let us go back to the cases of defective “purity” and consider how the laws of ancestry stand in regard to them. It appears from the facts almost certain that purity may sometimes be wanting in a character which elsewhere usually manifests it.

Here we approach a question of greater theoretical consequence to the right apprehension of the part borne by Mendelian principles in the physiology of heredity. We have to consider the question whether the purity of the gametes in respect of one or other antagonistic character is or is likely to be in case of any given character a universal truth? The answer is unquestionably—No—but for reasons in which “ancestry” plays no part162.

Hoping to interest English men of science in the Mendelian discoveries I offered in November 1900 a paper on this subject to “Nature.” The article was of some length and exceeded the space that the Editor could grant without delay. I did not see my way to reduce it without injury to clearness, and consequently it was returned to me. At the time our own experiments were not ready for publication and it seemed that all I had to say would probably be common knowledge in the next few weeks, so no further attempt at publication was made.

In that article I discussed this particular question of the absolute purity of the germ-cells, showing how, on the analogy of other bud-variations, it is almost certain that the germ-cells, even in respect to characters normally Mendelian, may on occasion present the same mixture of characters, whether apparently blended or mosaic, which we know so well elsewhere. Such a fact would in nowise diminish the importance of Mendel’s discovery. The fact that mosaic peach-nectarines occur is no refutation of the fact that the total variation is common. Just as there may be trees with several such mosaic fruits, so there may be units, whether varieties, individual plants, flowers or gonads, or other structural units, bearing mosaic egg-cells or pollen grains. Nothing is more likely or more in accordance with analogy than that by selecting an individual producing germs of blended or mosaic character, a race could be established continuing to produce such germs. Persistence of such blends or mosaics in asexual reproduction is well-known to horticulturists; for example “bizarre” carnations, oranges streaked with “blood”-orange character, and many more. In the famous paper of Naudin, who came nearer to the discovery of the Mendelian principle than any other observer, a paper quoted by Professor Weldon, other examples are given. These forms, once obtained, can be multiplied by division; and there is no reason why a zygote formed by the union of mosaic or blended germs, once arisen, should not in the cell-divisions by which its gametes are formed, continue to divide in a similar manner and produce germs like those which united to form that zygote. The irregularity, once begun, may continue for an indefinite number of divisions.

I am quite willing to suppose, with Professor Weldon (p. 248), that the pea Stratagem may, as he suggests, be such a case. I am even willing to accept provisionally as probable that when two gametes, themselves of mosaic or blended character, meet together in fertilisation, they are more likely to produce gametes of mosaic or blended character than of simply discontinuous character. Among Messrs Sutton’s Primulas there are at least two striking cases of “flaked” or “bizarre” unions of bright colours and white which reproduce themselves by seed with fair constancy, though Mendelian purity in respect of these colours is elsewhere common in the varieties (I suspect mosaics of “false hybridism” among allelomorphs in some of these cases). Similarly Galton has shown that though children having one light-eyed and one dark-eyed parent generally have eyes either light or dark, the comparatively rare medium eye-coloured persons when they mate together frequently produce children with medium eye-colour.

In this connection it may be worth while to allude to a point of some practical consequence. We know that when pure dominant—say yellow—is crossed with pure recessive—say green—the dominance of yellow is seen; and we have every reason to believe this rule generally (not universally) true for pure varieties of peas. But we notice that in the case of a form like the pea, depending on human selection for its existence, it might be possible in a few years for the races with pure seed characters to be practically supplanted by the “mosaicized” races like the Telephone group, if the market found in these latter some specially serviceable quality. In the maincrop peas I suspect this very process is taking place163. After such a revolution it might be possible for a future experimenter to conclude that Pisum sativum was by nature a “mosaicized” species in these respects, though the mosaic character may have arisen once in a seed or two as an exceptional phenomenon. When the same reasoning is extended to wild forms depending on other agencies for selection, some interesting conclusions may be reached.

But in Mendelian cases we are concerned primarily not with the product of gametes of blended character, but with the consequences of the union of gametes already discontinuously dissimilar. The existence of pure Mendelian gametes for given characters is perfectly compatible with the existence of blended or mosaic gametes for similar characters elsewhere, but this principle enables us to form a comprehensive and fruitful conception of the relation of the two phenomena to each other. As I also pointed out, through the imperfection of our method which does not yet permit us to see the differentiation among the gametes though we know it exists, we cannot yet as a rule obtain certain proof of the impurity of the gametes (except perhaps in the case of mosaics) as distinct from evidence of imperfect dominance. If however the case be one of a “mule” form, distinct from either parent, and not merely of dominance, there is no a priori reason why even this may not be possible; for we should be able to distinguish the results of breeding first crosses together into four classes: two pure forms, one or more blend or mosaic forms, and “mule” forms. Such a study could as yet only be attempted in simplest cases: for where we are concerned with a compound allelomorph capable of resolution, the combinations of the integral components become so numerous as to make this finer classification practically inapplicable.

But in many cases—perhaps a majority—though by Mendel’s statistical method we can perceive the fluctuations in the numbers of the several products of fertilisation, we shall not know whether abnormalities in the distribution of those products are due to a decline in dominance, or to actual impurity of the gametes. We shall have further to consider, as affecting the arithmetical results, the possibility of departure from the rule that each kind of gamete is produced in equal numbers164; also that there may be the familiar difficulties in regard to possible selection and assortative matings among the gametes.

I have now shown how the mosaic and blend-forms are to be regarded in the light of the Mendelian principle. What has Professor Weldon to say in reference to them? His suggestion is definite enough—that a study of ancestry will explain the facts: how, we are not told.

In speaking of the need of study of the characters of the race he is much nearer the mark, but when he adds “that is their ancestry,” he goes wide again. When Telephone does not truly divide the antagonistic characters among its germ-cells this fact is in nowise simply traceable to its having originated in a cross—a history it shares with almost all the peas in the market—but to its own peculiar nature. In such a case imperfect dominance need not surprise us.

What we need in all these phenomena is a knowledge of the properties of each race, or variety, as we call it in peas. We must, as I have often pleaded, study the properties of each form no otherwise than the chemist does the properties of his substances, and thus only can we hope to work our way through these phenomena. Ancestry holds no key to these facts; for the same ancestry is common to own brothers and sisters endowed with dissimilar properties and producing dissimilar posterity. To the knowledge of the properties of each form and the laws which it obeys there are no short cuts. We have no periodic law to guide us. Each case must as yet be separately worked out.

We can scarcely avoid mention of a further category of phenomena that are certain to be adduced in opposition to the general truth of the purity of the extracted forms. It is a fact well known to breeders that a highly-bred stock may, unless selections be continued, “degenerate.” This has often been insisted on in regard to peas. I have been told of specific cases by Messrs Sutton and Sons, instances which could be multiplied. Surely, will reply the supporters of the theory of Ancestry, this is simply impurity in the extracted stocks manifesting itself at last. Such a conclusion by no means follows, and the proof that it is inapplicable is obtained from the fact that the “degeneration,” or variation as we should rather call it, need not lead to the production of any proximate ancestor of the selected stock at all, but immediately to a new form, or to one much more remote—in the case of some high class peas, e.g., to the form which Mr Sutton describes as “vetch-like,” with short pods, and a very few small round seeds, two or three in a pod. Such plants are recognized by their appearance and are rigorously hoed out every year before seeding.

To appreciate the meaning of these facts we must go back to what was said above on the nature of compound characters. We can perceive that, as Mendel showed, the integral characters of the varieties can be dissociated and re-combined in any combination. More than that; certain integral characters can be resolved into further integral components, by analytical variations. What is taking place in this process of resolution we cannot surmise, but we may liken the consequences of that process to various phenomena of analysis seen elsewhere. To continue the metaphor we may speak of return to the vetch-like type as a synthetical variation: well remembering that we know nothing of any substance being subtracted in the former case or added in the latter, and that the phenomenon is more likely to be primarily one of alteration in arrangement than in substance.

A final proof that nothing is to be looked for from an appeal to ancestry is provided by the fact—of which the literature of variation contains numerous illustrations—that such newly synthesised forms, instead of themselves producing a large proportion of the high class variety which may have been their ancestor for a hundred generations, may produce almost nothing but individuals like themselves. A subject fraught with extraordinary interest will be the determination whether by crossing these newly synthesised forms with their parent, or another pure form, we may not succeed in reproducing a great part of the known series of components afresh. The pure parental form, produced, or extracted, by “analytical” breeding, would not in ordinary circumstances be capable of producing the other components from which it has been separated; but by crossing it with the “synthesised” variety it is not impossible that these components would again reappear. If this can be shown to be possible we shall have entirely new light on the nature of variation and stability.