When we consider the bearing of recent discoveries on those comprehensive schemes of evolution with which we were formerly satisfied, we find that certain details of the process are more easy to imagine. We readily now understand how varieties once formed, can persist, but at the same time difficulties hitherto faced with complacency become formidable in the light of the new knowledge. So generally is this admitted by those familiar with modern genetic research that most are rightly inclined to postpone the discussion. The premisses, indeed, on which such a discussion must be based are almost wholly wanting.

The difficulties to which I chiefly refer are not those created by the phenomena of adaptation, though they are serious enough. In treating of that subject I have felt obliged to express scepticism as to the validity of nearly all the new evidence for the transmission of acquired characters. At the present time the utmost we are bound to accept is the proof that (1) in some parthenogenetic forms variations, or perhaps we may say malformations, produced in response to special conditions, recur in one or perhaps two generations asexually produced after removal to other conditions. (2) That violent maltreatment may in rare instances so affect the germ-cells contained in the parents as to cause the individuals resulting from the fertilisation of those cells to exhibit an arrest of development similar to that which their parents underwent.

I do not doubt that evidence of this type will be greatly extended. As a contribution to genetic physiology these facts are very important and interesting, but I cannot think that any one, on reflexion, will feel encouraged by such indications to revive old beliefs in the direct origin of adaptations.

In these respects we are simply left where we were. The force of objections based upon the existence of adaptative mechanisms is no greater than it has always been. On the contrary the fact that variations can now so generally be recognized as definite is some alleviation of the difficulty. We can moreover disabuse ourselves of the notion that for all characters which are definite or fixed, some utilitarian rationale may be presumed. Upon that point the study of variation has provided a perfectly clear answer.

In frankly recognizing that the fixity of characters in general need not connote usefulness to their possessors we deliver ourselves of a distracting pre-occupation and prepare our minds for an investigation of the properties of living organisms in the same spirit as that in which the chemist and the physicist examine the properties of unorganized materials. The creature persists not merely by virtue of its characteristics but in spite of them, and the fact of its persistence proves no more than that on the whole the balance of its properties leaves something in its favour.

It may be noted by the way that the fact that the structures of living things are on the whole adaptative was not always obvious. Though to naturalists of this generation it is a truism, we have only to turn to Buffon to find that in his philosophy of nature it played no essential part. The passage in which Buffon describes what he regards as the forlorn and degraded condition of the Woodpecker is well known. We have come to think of the Woodpecker as a capital example of adaptation to the mode of life; but Buffon after enumerating the hard features of the bird's existence, forced to earn its living by piercing the bark of trees in an attitude of perpetual constraint, remarks[1] "Tel est l'instinct Étroit et grossier d'un oiseau bornÉ a une vie triste et chÉtive. Il a reÇu de la Nature des organes et des instrumens appropriÉs a cette destinÉe ou plutÔt il tient cette destinÉe mÊme des organes avec lesquels il est nÉ" (my italics). His reflexions on the Stilt (Himantopus) read even more strangely to us, accustomed as we are to see in the prodigious length and thinness of the shanks and in the other features of its organisation palpable adaptations to a wading life. For Buffon, however, this curious bird seemed a poor, neglected production, extravagant in its disproportions, one of the misfits of creation, left as a shadow in the picture composed of nature's more successful efforts.[2] This theme he develops at some length, being evidently well pleased with the idea.

Our way of regarding these things is doubtless sounder and more fruitful than Buffon's, but it is well to remember that what seems so obvious to us looked quite differently to other excellent observers; and stupid as it may have been to have overlooked plain examples of adaptation, it is a far worse mistake to see adaptation everywhere. I do not seek to minimise the real and permanent difficulty which the existence of adaptations creates, but by the suggestion that all normal specific differences are adaptational that difficulty was quite gratuitously increased.

In these respects it may be claimed that progress has been made, even if that progress seem outwardly of small account.

But all constructive theories of evolution have been built on the understanding that what we know of the relation of varieties to species justifies the assumption that the one phenomenon is a phase of the other, and that each species arises or has arisen from another species either by one or several genetic steps. In the varieties we have accustomed ourselves to think that we see those steps. We still know little enough of the mode of occurrence of variation, but we do begin to know something, and if we ask ourselves whether our knowledge, such as it is, conforms at all readily with our former expectations, we cannot with any confidence assert that it does. Among the plants and animals genetically investigated are many illustrations of very striking and distinct varieties. Many of these might readily enough be accepted as species by even the most exacting systematists, and not a few have been so treated in classification; but when we have examined their relationship to each other we feel not merely that they are not species in any strict sense but that the distinctions they present cannot be regarded as stages in the direction of specific difference. Complete fertility of the results of inter-crossing is and I think must rightly be regarded as inconsistent with actual specific difference; and of variations leading to that consequence no clear indication has yet been found. As an example of possible exceptions mention should perhaps be made of the case of a giant form of Primula sinensis investigated by Keeble.[3] It arose from a "Star" Primula of normal size, and though fertile with its own pollen all attempts to fertilise it with the pollen of other forms failed. Miss Pellew, who did these fertilisations, tells me that very extensive trials were made, and repeated in several seasons. Ultimately two plants were raised from it fertilised with a plant of the strain from which it sprang, and these proved sterile. In the light of modern experience the significance of such isolated instances is doubtful.

All the strains known as "Giants" are, as Messrs. Sutton have always found, more or less sterile, and their sterility is presumably due to some negative defect.

In regard to the fertility of Primula species there are several paradoxes. For example the long-styled varieties, apart from giants, are fertile with their own pollen, and for many years short-styled plants have not been used in most strains. Auriculas and Polyanthuses, on the contrary, are generally if not always bred from short-styled plants, as the florists have decided that the long-styled are inadmissible. Mr. R. P. Gregory tells me that, though most strains of P. sinensis give seed enough when only long-styled plants are used, he finds nevertheless that when a "legitimate" union is made the amount of seed usually increases much as Darwin observed. Darwin's statement that plants of "illegitimate" origin are less fertile than the "legitimately" raised plants is also in general confirmed by his experience. To this rule there were some marked exceptions in strains derived from long-styled plants, which though illegitimate showed a high degree of fertility, but illegitimate unions between short-styled plants always produced comparatively sterile offspring. I have no records of the behavior of Auriculas and Polyanthuses. It would be interesting to know whether among them pure strains of short-styled plants (dominants) have appeared, and, if so, how their fertility is affected. Without much more critical data I suppose no one would nowadays be inclined to follow Darwin in instituting a comparison between the sterility of hybrids and that of illegitimately raised plants of heterostyle species.[4] It is even difficult to imagine any essential resemblance between these two phenomena, nor has evidence ever been produced to show that illegitimately raised plants have bad pollen grains, which is the usual symptom of sterility in hybrid plants and the consequence, as we believe, of failure of some essential division in the process of maturation.

The difficulty that we have no knowledge of the contemporary origin of forms, from a common stock, which when crossed together give a sterile product, is one of the objections constantly and prominently adduced from the time of the first promulgation of evolutionary ideas. In the light of recent work the objection has gathered strength. Why, if we are able to produce instances of variation colourably simulating specific difference in almost all other respects, do we never find an original appearance of this most widely spread of all specific characteristics? No doubt all breeders know that sterile animals and plants occasionally appear in their cultures, but it is more in accordance with probability that the sterility in these sporadic instances should be regarded as due to defect than that it should be thought comparable with that of the sterile hybrids. For their sterility must, by all analogy with results elsewhere seen, be attributed not to the absence of something, but to the presence and operation of complementary factors leading to the production of inhibition of division; and consistently with that interpretation, we find that when from a partially sterile hybrid comparatively fertile offspring can be raised, their comparative fertility continues in the posterity generally if not always without diminution. The distinction between these several kinds of sterility was of course not understood in Darwin's time. The comparison, for example, which he instituted[5] between the sterility of "contabescent" anthers and that of hybrids no longer holds, for at least in those cases in which the nature of contabescent anthers have been genetically investigated (Sweet Pea, Tropaeolum) they proved to be a simple recessive character. Nor can we now easily suppose that the attempt there made by Darwin to suggest resemblance between the sterility produced by unnatural conditions and that of hybrids has any physiological justification.

In regarding the power to produce a sterile or partially sterile hybrid as a distinction in kind, of a nature other than those which we perceive among our varieties, I am aware that I am laying stress on an impression which may hereafter prove false. The distinction nevertheless is so striking and so continually before the eyes of a practical breeder that he can scarcely avoid the inference that when he meets a considerable degree of sterility in a cross-bred he is dealing with something belonging to a distinct category, and not merely a varietal feature of an exceptional kind.

Besides the sterility of hybrids appeal has often been made to the phenomenon of incompatibility, in its several stages of completeness, as distinguishing species. No one doubts that incompatibility may arise from a variety of causes of most diverse degrees of importance, but though sometimes referred to as an extreme case of interspecific sterility, it is really a very different matter. In regard to one phase of this incompatibility, that associated with self-sterility, some progress has been made, and we are not wholly without experimental evidence of its being within the range of contemporary variation.

Given the outline of Mendelian teaching as to gametic differentiation and the classification of individuals in a mixed population, it seemed highly probable that what we call self-sterility must mean that the species really consisted of classes, some of which are capable of interbreeding with others while others are not. According to the received account every individual, though incapable of fertilising itself, was supposed to be able both to fertilise and to be fertilised by any other individual. This notion has always seemed to me a self-evident absurdity, for it would imply that there can be as many categories as individuals. Such experiments, however, as I made did certainly give results consistent with that belief. I first tried Cinerarias, which are usually self-sterile, but I found no incompatible pairs of plants. Whether I was deceived by the consequences of apogamy, or whether the pollen of certain plants may belong to more than one class I do not know. The results were confused in various ways. Usually the self-fertilised plants set little or nothing, and cross-fertilised they set fully with such uniformity that the few failures could plausibly be attributed to mistakes in manipulation or to other extraneous causes. Later de Vries announced[6] (without giving particulars) that he had proved the existence of such classes in Linaria vulgaris; but on making experiments with that species I again got no positive results, and I came to the conclusion that in spite of inherent improbability the conventional belief must be substantially true. At last, however, the work of Correns, lately published,[7] does definitely show that in one species, Cardamine pratensis, classes of individuals exist such that individuals of the same class are incapable of fertilising themselves or each other, but fertilisation made between the classes is usually completely effective. Many complications were encountered and some contradictory evidence is recorded, but the general bearing of the results was positive and indubitable.

We know far too little of this phenomenon as yet to be able to understand its significance, but I suppose we may anticipate with some confidence that it will be found to be a manifestation of dissimilarity between the male and female gametes of the same individual, comparable with that first seen in the Stocks (Matthiola) which throw doubles—a state of things in all likelihood to be found widely spread among hermaphrodite organisms. Whether the incompatibility between species is to be associated with that of the self-steriles also cannot be positively asserted, though it seems not unreasonable to expect that such an association will be discovered.

The case of the apple and the pear is an impressive illustration of this possibility. The two species are of course exceedingly alike in all outward respects, but nevertheless the pollen of each is entirely without effect on the other. Presumably we should interpret this fact as meaning not so much that the apple and the pear are in reality very wide apart, but rather that either, each is lacking in one of two complementary elements, or that each possesses a factor with an inhibitory effect. Their incompatibility may well be of the same nature as that of the classes in Cardamine pratensis.

Returning now to the problem of inter-specific sterility; we note, as I have said, the absence of contemporary evidence that variation can confer on a variety the power to form a sterile hybrid with the parent species. The considerations based on this want of evidence have for a long while been familiar to all who have discussed evolutionary theories, and it is worth observing the exact reason why the difficulty strikes us now with a new and special force. In pre-Mendelian times all that was known was that some forms could freely interbreed without diminution of fertility in the product, while others could not. But now we find that, by virtue of segregation, from one and the same pair of parents, or even, in the case of hermaphrodites, from one and the same individual, offspring commonly arises showing among themselves exactly such differences as distinguish species—and very good species too. This we see happening again and again. But to forms capable of arising as brethren in one family the title species has never been meant to apply, and if we are going to use the term in application to fraternal groups we must definitely recognise that by "specific" difference is to be understood simply difference, without any immediate or even ulterior physiological limitation whatever. Naturally, therefore, we begin to think of the appearance of sterility in crosses as something apart, and as a manifestation which distinguishes certain kinds of unions in a very special way.

I am perfectly aware that there are gradations in the sterility of hybrids as in every other characteristic upon which it has been proposed to base specific definitions; but, as also so often happens in the matter of defining intergrading categories, the difficulty in practice is not often such as to lead to actual ambiguity. I am speaking of course of those examples which are amenable to genetic experiment. As to the rest there is complete and permanent uncertainty. But the experience of the practical breeder does, I think, on the whole, support the contention to which systematists have so steadily clung under all the assaults of evolutionary philosophers, that, though we cannot strictly define species, they yet have properties which varieties have not, and that the distinction is not merely a matter of degree.

The first step is to discover the nature of the factors which by their complementary action inhibit the critical divisions and so cause the sterility of the hybrid. Thus expressed, we see the problem of inter-specific sterility in its right place; and the question why we do not now find contemporary instances of varieties lately arisen in domestication, which when crossed back with their parents, or with their coderivatives, can produce sterile products, is perceived to be only a special case of a problem which in its more general form is that of the origin of new and additional factors.

For the requisite evidence no comprehensive search has been made, but perhaps it will yet be found. All that we can say at the present time is that the incidence both of hybrid sterility, and of incompatibility also, is most capricious; and provided that two forms have such features in common that a cross between them seems not altogether out of the question, no one can predict without experiment whether such a cross is feasible, and if feasible whether the product will be fertile, or sterile more or less completely. For instance, though probably all the British and some Foreign Finches (Fringillidae) have been crossed together, and some of these crosses, as for instance, the various Canary-mules have been made in thousands, I believe no quite clear example of a fertile hybrid can be produced. Many species of Anatidae cross readily and produce fertile hybrids: others give results uniformly sterile. Though most of the Equidae can be crossed and some of the hybrids are among the commonest of domesticated animals there is no certain record of a fertile mule. Among the Canidae the dogs, wolves and jackals all give fertile hybrids, but there is no clearly authenticated instance of a cross between any of these forms and the European fox. In spite of their close anatomical resemblance it is doubtful if the rabbit and the hare have ever interbred. Many of the wild species of Bos have been crossed and recrossed both with each other and with many domesticated races, but I understand that no cross with the Indian buffalo (Bos bubalus) has yet been successful even in producing a live calf.[8] In the genus Primula many hybrids are known and several of them occur in nature, but hitherto no certain hybrid between P. sinensis and any other species has been made, in spite of repeated attempts.

In Nicotiana many—doubtless all—the various forms of N. tabacum can be crossed together without diminution of fertility, though some are very distinct in appearance, but crosses between tabacum and sylvestris are highly sterile (in my experience totally sterile[9]), though the distinctions between them are not to outward observation nearly so great as those which can be found between the various races of Primula sinensis.

Recently some remarkable experiments bearing closely on these questions have been published by F. Rosen.[10] They concern the forms of Erophila (Draba) verna, celebrated in the history of evolutionary theory as the plants especially chosen by Alexis Jordan for the exposition of his views on these subjects.

The "species" contains a profusion of forms dissimilar in many structural characters, such as the size and shape of leaves, flowers, fruits, etc. Of these forms many grow in association. Jordan found, on experiment, that each, to the number of some two hundred, bred true, and that therefore, the conventional assumption that polymorphism of this kind must mean great contemporary variability had no foundation in fact. So far indeed is the evidence from favouring the belief that such forms are in any way transitional or indeterminate, that, as is well known, Jordan used it with every plausibility to support the doctrine of the fixity of species. To certain aspects of Jordan's work we will return later in this chapter, but the matter is in the present connection of especial interest for the reason that Rosen has lately found by experiment that some of these presumably very closely allied forms, crossed together, gave hybrids more or less sterile. In the case of the offspring of one pair of forms only (E. cochleata and stricta) was the fertility undiminished, and the various degrees of sterility found in the other crosses ranged up to the extreme infertility of the hybrids between E. stricta × elata. From this cross ten plants were bred. Of these the four strongest were chosen to breed from, but two of the four proved totally sterile; one had only bad seeds; and from the fourth a single seedling was raised which in its turn proved to be sterile. From the less sterile hybrids F₂ families were raised, with the usual experience that in this and subsequent generations the sterility diminished among extracted forms, new and true-breeding types with complete fertility being thus derived from the original cross.[11]

The production of sterility as a consequence of crossing plants so nearly approaching each other as these Erophila "species" do is not a little interesting, and the fact well exemplifies the futility of the various attempts to frame general expressions as to specific properties or behaviour. Commenting on his results Rosen argues that the polymorphic group commonly called by systematists Erophila (Draba) verna may now be regarded as having arisen by crossing, as did his own types mentioned above. The question, however, what species were the original progenitors of the group cannot be answered. Rosen considers that no form which he knows satisfies the requirements, and that it or they must be supposed to be lost. This conclusion will recall the similar problem raised by the Oenothera mutants (Chap. V); and unsatisfactory as it may be to have recourse to such hypotheses we must remember the possibility that as a consequence of hybridisation, subsequent segregation and recombination of factors, species may have thus actually, as we may say, exploded, and left nothing but a polymorphic group of miscellaneous types to represent them in posterity. If this way of regarding the phenomena be a true one, the sterility now seen when some of the group are re-crossed, becomes analogous to that "reversion or crossing" which we now so well understand to be a consequence of the recombination of characters separated at some previous point in the history of descent. In the partial sterility of the contemporary hybrid we see this character reappearing, formed now as it was on the occasion of the original cross, by the meeting of complementary factors.

Another case that may be mentioned in this connection is that of the crosses between various culinary peas (Pisum sativum) and a peculiar form found by Mr. Arthur Sutton growing ostensibly in a wild state in Palestine. This Palestine Pea is low growing, rarely reaching 18 inches. It is in general appearance like a small and poorly grown field pea. The stems are thin and rather hard. The most obvious differences which distinguish this from other field peas are the marked serration of the stipules, and the development of pith in the pods. Such pith is often present in the pods of peas more or less, but in the Palestines it is so strongly developed as almost to form a lomentum. Curiously enough, though the flowers are purple much as those of ordinary field peas, there is no coloured spot in the axils. On the other hand, the stems have coloured stripes running up from the axils. Though this plant differs so little from domesticated peas, all crosses with them either failed, or produced hybrids quite or almost quite sterile. This was Mr. Sutton's experience, and on repeating the experiments with material kindly given by him I found the same result.[12]

In a large series of crosses some seeds died or gave rise to feeble plants. Of the plants which lived, few gave any seed. The seed, however, that was obtained from F₁ plants grew well enough, and the F₂ plants proved, as often in such cases, fertile. In these, indeed, no sign of sterility was noticeable. The experiment is being repeated in various ways, for, as the genetic behaviour of peas is comparatively well known, the subject is an exceptionally favourable one for these investigations.

Such an example shows the confusion produced the moment we attempt to harmonize conceptions of specific difference with results attained by experimental methods. It has been usual to regard the field pea (P. arvense) as a species distinct from the edible pea (P. sativum). De Candolle and others regard the field pea as derived from a form wild in Italy, but the origin of the edible pea is considered to be unknown. From breeding experiments we find no sterility whatever in the crosses between the various arvense and sativum types, nor in the crosses made between them and several other peculiar types from various countries; whereas this Palestine Pea, which only differs from a small arvense in what might have been thought trivial characters,[13] either fails to cross altogether or gives a sterile product, whatever type be chosen as the other parent.

Examples of this kind have at least the merit that they lead to more precise delimitations of the problem. We are confronted with two distinct alternatives.

1. We may apply the term Species promiscuously to all distinct forms. If we do so it must be clearly understood that we cannot even rule out the several combinations of "presences and absences" represented by the various types whether wild or domesticated. For we may feel perfectly assured that at least all the arvense and all the sativum types yet subjected to experimental tests are on precisely the same level in this respect. There is no distinction, logical or physiological, to be drawn between them. Some contain more factors, and others contain fewer. In some the re-combinations have been brought about by natural variation or crossing, while the same consequences in the others have resulted from man's interference.

2. We may follow the conventions of systematists and distinguish the outstanding or conspicuous forms such as arvense, quadratum, sativum and perhaps a few more as species, and leave the rest unheeded. If this course is followed it must be clearly understood and permitted as a piece of pure pragmatism, deliberately adopted for the convenience of cataloguers and collectors, without regard to any natural fact or system whatsoever.

But while following either the one plan or the other we shall be still awaiting the answer, which only genetic experiment can provide, to the question whether among the various types there are some which differ from the rest in a peculiar way: whether by having groups of characters linked together in especially durable combinations, or by possessing ingredients which cause greater or less disturbance in the processes of cell-division, and especially in the processes of gametic maturation, when they are united by fertilisation with complementary ingredients.

Before any but the vaguest ideas regarding the nature and significance of inter-specific sterility can be formed, a vast amount of detailed work must be done. Sterility as a result of crossing, as well as that which is alleged sometimes to arise in consequence of changed conditions, is at best a negative characteristic, and there are endless opportunities for mistake and misinterpretation in studying features of this kind. No one, I suppose, would now feel any great confidence in most of the data which from time to time are resuscitated for the purpose of such discussions. Even the best collections of evidence, such as those given by Darwin in Forms of Flowers, cannot be regarded as critical when judged by present-day standards. Nothing short of the most familiar acquaintance with the habitual behaviour of individuals, and of strains kept under constant scrutiny for several years would enable the experimenter to form reliable judgments as to the value to be attached to observations of this class.

The admission must, however, be faced that nothing in recent work materially tends to diminish the surprise which has always been felt at the absence of sterility in the crosses between co-derivatives. We should expect such groups of forms to behave like the Erophila types, and frequently to produce sterile products on crossing. Whatever be the explanation, the fact remains that such evidence is wanting almost completely. In spite of all that we know of variability nothing readily comparable with the power to produce a sterile hybrid on crossing with a near ally, has yet been observed spontaneously arising, though that characteristic of specificity is one of the most widely distributed in nature. It may be that the lacuna in our evidence is due merely to want of attention to this special aspect of genetic inquiry, and on the whole that is the most acceptable view which can be proposed. But seeing that naturalists are more and more driven to believe the domesticated animals and plants to be poly-phyletic in origin—the descendants, that is to say, of several wild forms—the difficulty is proportionately greater than it was formerly, when variation spontaneously occurring was regarded as a sufficient account of their diversity.