Chapter XIII: Part 13

That self-sterility in all these different degrees is not a primary character of the species, but an adaptation to the advantages of cross-fertilization, is apparent--if indeed it seems doubtful to any one--especially from cases of heterostylism. I refer to the dimorphism and trimorphism which Darwin discovered in many flowers, and which shows itself in the fact that flowers otherwise almost exactly alike, as, for instance, primroses, may exhibit a long style in some individuals, and in others a short one (Fig. 128). At the same time, there is a difference in the position of the stamens, which are placed higher up in flowers with short styles, and much lower down in those with long styles. Experiments have proved that the dusting of the stigma has the best results if pollen from the short-styled reaches the stigma of the long-styled form, or if pollen from the long-styled form reaches the stigma of the short-styled. Thus we have again to deal with an arrangement for crossing, an adaptation to the advantages of cross-fertilization, and we can in this case see the reason why the pollen has a different effect upon the two stigmas; the pollen-grains of the flowers with short style are larger than those of the flowers with long style, and as the length of the pollen-tube that can be sent out must depend upon the mass of protoplasm within the pollen-grain, it follows that the smaller pollen-grains will send out too short a tube to reach through the long style to the embryo-sac. In addition to this there is a difference in the papillæ of the stigmas, and it is possible that these may form an obstacle to the penetrating of pollen from a similar type. The process of selection which gives rise to such arrangements as we find in Primulas may easily be imagined, as soon as we are able to assume that cross-fertilization is more advantageous than self-fertilization as regards progeny, that is, as regards the continuance of the species.

We have already seen that uninterrupted self-fertilization is unknown among animals, but that it is not even very rare among plants, and this emphatically corroborates our previous conclusion, that the reason for which amphimixis was introduced as a normal event in nature is not to be sought for in the necessity for a renewing of life, or 'rejuvenation.' It cannot be a necessity, but only an advantage, which can in certain circumstances be dispensed with.

Although it is obvious enough that continued inbreeding in its most extreme form, self-fertilization, does not imply an absolute abandonment of amphimixis, the adherents of the rejuvenescence theory have regarded the unfavourable consequences of pure inbreeding as a confirmation of their assumption, according to which amphimixis is indispensable to the continuance of the life of the species, and it is therefore an important fact, if it can be proved, that continued self-fertilization can occur persistently, among plants at least, and yet not cause any injurious results to the species.

But how can this fact be understood from our point of view? How does it happen that crossing is striven after in so many different ways and yet so often given up again, and continued self-fertilization resorted to?

To this it may be answered, in the first place, that it is not, as far as we can see, for _internal_ reasons that persistent self-fertilization becomes the rule; there is no peculiar condition of the germ-plasm which makes it disadvantageous or superfluous that the diversity of the id-combinations should be maintained; self-fertilization is due to _external_ influences which bring it about that the plant has only the alternative of producing no seeds at all or of producing them by self-fertilization. In this connexion Darwin's experiments with orchids are particularly noteworthy.

In this very diversified order of plants there are numerous species whose flowers are infertile with their own pollen, although it does not reach the stigma in natural conditions, and therefore there was no necessity--as far as we can see--for guarding against self-fertilization by 'self-sterility.' These flowers are thus doubly adapted, so to speak, for crossing by means of insects. But as regards many of these, as well as many other modern orchids, insect-visits are very rare, and in some cases do not occur at all, and therefore these species cannot produce seed or can do so only exceptionally.

This is true of most of the Epidendra of South America, and of _Coryanthus triloba_ of New Zealand, two hundred blossoms of which only yielded five seed-capsules, and also of our _Ophrys muscifera_ and _O. aranifera_, the latter of which yielded only a single seed-capsule from 3,000 flowers gathered in Liguria. We might expect that the species in question must have become very rare, but this is not always the case, since each of these capsules contains an enormous number of seeds, sometimes many thousands. As soon as the visits of insects cease altogether, the species must die out in the particular locality concerned, unless it can revert to self-pollination and self-fertility. There is a whole series of species in which the stigma of the flower is sensitive to its own pollen, and in many of these an adaptation to self-fertilization has actually been effected, for the pollinia detach themselves from their anthers at maturity and fall upon the stigma. I have already mentioned _Ophrys apifera_, which, according to Charles Darwin, is no longer visited by insects, although its flowers still possess the structure required for insect-fertilization. This species has saved itself from extinction by the normal occurrence of self-fertilization.

This seems to me noteworthy in two respects. In the first place, it shows that pure self-fertilization need not necessarily result in a weakening of the species, and secondly, it affords a clear instance of a species being transformed in one minute character only, all the other characters remaining unaltered. In this case it was only the pollinia that required to vary a little in their mode of attachment and maturation, in order to effect the transformation of the flower for self-fertilization, and in point of fact that is all that has varied. The case is not relevant to our investigation at this moment, but cases of the kind can so rarely be clearly demonstrated that I cannot lose the opportunity of calling attention to it. The germ-plasm of this _Ophrys_ must have varied at an earlier stage, for otherwise the detachment of the pollinia would not have become normal and hereditary, but it can only have varied to the extent that the structure of this one small part of the flower was affected by the variation; something must have varied in the germ-plasm that had no influence upon the other parts of the flower, that is, solely the _determinants_ of the pollinia.

Let us return after this digression to our previous train of thought; we have to inquire how we can interpret the fact of continued self-fertilization without any visible injurious results to the species. If cross-fertilization be a material advantage as regards the continuance of the species, how can it be transformed into its opposite without evil effects? And there are no visible evil effects in _Ophrys apifera_. It is indeed not so abundant as _Ophrys muscifera_, or other allied species, but it certainly does not follow from that that it is on the way to extinction; certainly no decrease either of vigour of growth or of fertility can be observed.

If we inquire from the standpoint of our theory, how the composition of the germ-plasm must have altered through continual inbreeding, we have already found the answer--that through the reduction of the number of ids at the maturation of every germ-cell the diversity of the germ-plasm would gradually be lessened, that the number of different ids would thereby be lessened possibly even to the identity of the whole of the ids.

The consequences of such extreme uniformity of the germ-plasm would not, according to our theory, necessarily be that the species would be incapable of continued existence, but it would be that the species would become incapable of adaptations in many directions. Adaptations in one direction, such, for instance, as the variation in the mode of attachment and detachment of the pollinia of an Orchid, would still be possible. Thus a species which has long been perfectly adapted will be able to make the transition to inbreeding without injury to its chances of continued existence, if it be compelled by circumstances to do so. Species, on the other hand, which are still undergoing considerable transformations in many directions must be exposed by these to the danger of degeneration, just as happens in the artificial experiments with domesticated animals, whose secret weaknesses are greatly exaggerated by inbreeding.

We might be inclined to regard the effects of inbreeding as similar to those of parthenogenesis; they are certainly analogous, for both modes of reproduction must lead to a certain degree of uniformity in the germ-plasm. But there seems to me to be a difference and one which is not without importance.

In parthenogenesis no amphimixis occurs, but neither does any reduction of the number of the ids to one-half; all the ids present at the beginning of parthenogenesis are retained; they are only no longer mingled with strange ids. In inbreeding both amphimixis and reduction take place, but the former soon ceases to convey any really strange ids to the germ-plasm, but only the same as those which it already contains, so that a rapidly increasing monotony of the germ-plasm must result. To this must be added the possibility that among the few ids which now--many times repeated--form the germ-plasm, some must occur which exhibit unfavourable variational tendencies in one or many determinants, and then the same thing will occur which usually occurs in experimental inbreeding of domesticated animals, namely, _degeneration of the progeny_. In parthenogenesis the case is otherwise; unfavourable variational tendencies, as soon as they attain selection-value, are, so to speak, eliminated root and branch, because the individuals which exhibit them, and their whole lineage, are exterminated, without their having any effect upon the other collateral lines of descent. A purely parthenogenetic species will, therefore, not degenerate as long as individuals of normal constitution are present, for these reproduce with perfect purity. But if in later generations unfavourable variational tendencies crop up in the germ-plasm through germinal selection, the process of personal selection will be reinforced on these or on their descendants, and it is conceivable, and even probable, that in perfectly adapted species parthenogenesis may last for a very long time without doing any injury to the constitution of the species.

The same is true of purely asexual reproduction, to the investigation of which we shall now turn.

Let us leave out of account the simplest animals (Monera) without amphimixis, which we have already discussed. In simple animals reproduction by budding or by fission is frequent, or it occurs in alternation with sexual reproduction; in higher animals, Arthropods, Mollusca, and Vertebrates, asexual reproduction is wholly absent. In plants it plays an enormously greater part, and what is called 'vegetative reproduction,' which is purely asexual without any amphimixis, is to be found in all groups of plants, especially in the form of budding and spore-formation, besides which there is multiplication by runners, rhizomes, tubers, bulbs, and bulbils. In most cases there is, in addition to the purely asexual reproduction, so-called sexual reproduction associated with amphimixis, and often the sexual and asexual generations alternate with each other, so that 'alternation of generations' occurs, as is common in lower animals, especially polyps, medusæ, and worms.

But it sometimes happens among plants that the sexual reproduction is absent, and that a species reproduces by the asexual mode only, and this is the case which we must now consider more closely.

Let us first of all seek to gain clearness as to the composition of the germ-plasm in the case of purely asexual multiplication, and what conclusions may be drawn from this, and then let us compare these with the known observational data, and it will be apparent that in individuals which have arisen by budding the complete germ-plasm of the species must be contained; the number of ids will not only remain the same in the bud as it was in the mother plant, but the number of _different ids_ will not be diminished. The case is analogous to that of pure parthenogenesis, in which the absence of the second maturation-division of the ovum allows the germ-plasm to retain the full complement of ids. Charles Darwin held that purely asexual multiplication was 'closely analogous to long-continued self-fertilization,' yet, as we have seen, according to our theory there must be a not inconsiderable difference between the two processes, depending on the fact that in exclusive self-fertilization the number of different ids is continually decreasing, while in purely asexual reproduction the germ-plasm loses nothing of the diversity of its ids. If, therefore, the germ-plasm in purely asexual reproduction no longer receives fresh ids through amphimixis, it at least loses none of those it formerly possessed. Although we cannot consider it adapted for entering upon new adaptations in many directions, yet we may expect that the species will continue to reproduce unchanged for longer than in the case of exclusive self-fertilization, the more so since all unfavourable variational tendencies which crop up are eliminated as soon as they attain to selection-value, and, as in the case of parthenogenesis, they are eliminated without being mingled with other lines of descent.

Let us take, for instance, the purely asexual reproduction which obtains in Algæ of the genus _Laminaria_, in regard to which it is stated that it multiplies only through asexual swarm-spores. There are quite a number of species of this large tangle, and if it should be established that in all these the spore-cells really do not conjugate, then the case would prove that the species of a genus can maintain a well-defined existence for a long time after amphimixis has been given up. But this would not be a proof of the possibility of _species-formation_, for that the ancestral forms of the Laminarians must have possessed amphigony may be assumed, since their nearest relatives exhibit it still. It cannot be proved, but there seems nothing against the assumption that these tangles have existed for a long time under uniform conditions, and have become adapted to these with a high degree of constancy.

The conditions are similar in the marine Algæ of the genus _Caulerpa_, the nearest relatives of which reproduce sexually, though they themselves, as far as is known, reproduce only by spores.

In the Lichens, which represent, as we have already seen, a life-partnership between Fungi and Algæ, amphimixis appears not to occur at all; the unicellular Alga reproduces by cell-division, the Fungus by producing a great number of swarm-spores, which do not conjugate with one another. As far as the Alga is concerned we might perhaps suppose that the simplicity of its structure makes it possible for it to dispense with a constant recombination of its few characters to bring about the most favourable composition in its idioplasm; in support of this we may note that even the life-long combination with the Fungus has caused no visible variation in the Alga, as we must conclude from the fact that these Algæ can also live independently, and that the same species of Alga may combine with several different Fungi to form different species of lichen, just as the same Fungus may also form part of several species of lichen. We might also imagine that we have here no more than a direct influence of the Alga and Fungus upon one another, and that there is no adaptation to the new conditions of life at all, yet that can hardly be seriously maintained in regard to species which live under such definite and diverse conditions. It now seems to be established--contrary to the older statements--that the lichen-fungus only reproduces asexually, and in face of this it seems to me that nothing remains except to make the assumption that lichens formerly possessed sexual reproduction, but that they have lost it, though whether all have done so is, perhaps, not yet quite certain.

The same assumption must be made in regard to the Basidiomycetes among the Fungi, and for most of the Ascomycetes, for in these groups of Fungi sexual reproduction has only been demonstrated 'with certainty in a few genera.' That in these cases also there has been a degeneration of amphigony, until it has completely disappeared, seems probable from the two other groups of Fungi, the Zygomycetes and Oomycetes, since in these 'a reduction of sexuality amounting in some cases to complete disappearance' can be demonstrated even in existing forms. But whether it may be assumed that the Fungi which are now asexual are no longer capable of new adaptations, and whether their parasitic habit may be regarded as making up in some way for the lack of the remingling of the germ-plasm, as the botanist Möbius supposes, I am not able to decide. It is obvious that data in regard to amphimixis among the Fungi are still incomplete, and recent investigations lead us to suspect that sexual mingling may not be absent, but only disguised. Dangeard, Harold Wager, and others have observed that a fusion of nuclei precedes the formation of spores, and this may be regarded as amphimixis, although the conjugating nuclei belong to cells of the same plant and sometimes even to the same cell. But although we are here dealing with a set of facts which cannot yet be satisfactorily formulated in terms of our theory, it is nevertheless not contradictory to it that amphimixis should be wholly absent in the higher Fungi. But the fact would be contradictory to the unadulterated rejuvenescence-theory, for if amphimixis were really a condition of the continuance of life, no species--as we have already said--could continue to exist without it for countless generations.

The same argument holds true for the higher plants, which have become purely asexual under the influence of cultivation. I refer to many of the well-marked varieties of our cultivated plants which multiply exclusively, or almost exclusively, by means of tubers and slips, as is the case with the potato, the manioc, the sugar-cane, the arrowroot-plant (_Maranta arundinacea_), and others. All these facts can easily be reconciled with our interpretation of the meaning of amphimixis, although the attempt to range them as evidence against our theory has more than once been made. We have thus arrived at the conclusion that while many-sided adaptations, that is, variations which transform the plant in accordance with the indirect influences of new conditions of life, cannot be brought about without a persistent mingling of germ-plasms, simple modifications may readily appear although amphimixis is altogether absent. If a wild plant be permanently transferred to a well-manured culture-bed, it is probable that certain changes will occur in it, either gradually or at once. But these are not adaptations; they are, so to speak, direct reactions of the organism which do not even require selection to make them increase, but depend upon the influencing of certain determinants of the germ-plasm, and which, like all germinal variations, will follow their course steadily until a halt is called either by germinal or by personal selection. When a given plant is exposed to these new and artificial conditions, the changes in question make their appearance sooner or later, and follow their course, and go on increasing as long as that is compatible with the harmony of the structure and functioning of the plant, this depending, as in all individual development, on the struggle between the parts, that is to say, on histonal selection. Only in this respect is the utility or injuriousness of the change of importance, for personal selection, the struggle between individuals, does not affect plants which are under cultivation.

That such modifications may increase and may persist through many generations, even with asexual multiplication, depends upon the fact that the budding cells contain germ-plasm, as well as the germ-cells, and if particular determinants of the germ-plasm in general are caused to vary by these new influences, the variation may be transmitted from bud to bud, from shoot to shoot, and so go on increasing as long as the new conditions persist, as well as in amphigonic (bisexual) reproduction, where they are transmitted from germ-cell to germ-cell. It is not inconceivable that an individual adaptation, that is to say a useful adjustment, might be effected in the course of asexual reproduction, although it is improbable that direct influences would give rise to just those changes which would be useful under the new conditions. But there are a number of cases which have been interpreted in this way. In several of the cultivated plants named, the reproductive organs have themselves degenerated, either only the male, or only the female, or both at the same time; and some observers, accepting the hypothesis of an inheritance of functional modifications, have regarded this as the direct result of disuse during the long period of asexual reproduction.

Leaving out of account this erroneous presupposition, we may ask how asexual reproduction, such as that of the potato by tubers instead of by seed, which has gone on exclusively for several centuries, could exercise any influence upon the flowers and seed-forming of this species? In point of fact it has exercised none in most potatoes, for the flowers and seeds are just as fertile now as they were when the potato was first discovered.

Whether the pollen of a flower is utilized in one or other of its thousands of pollen-grains by reaching the stigma of another plant of the same species, or whether all the pollen-grains are uselessly scattered abroad, cannot possibly affect the flower so as to cause degeneration; the theory of disuse cannot be applied in this case. What is true of the potato holds good also of the manioc (_Manihot utilissima_), but, on the other hand, many of the best varieties of common fruits--pears, figs, grapes, pine-apples, and bananas--are seedless. In _Maranta arundinacea_ 'the whole wonderful structure of the flower has persisted, but the pollen-grains, that is the germ-cells, are wanting.' Whether this implies a permanent degeneration of the sexual organs, that is to say, one that is embodied in the primary constituents of the species, or whether it is only the result of over-abundant nourishment, or of other causes in the circumstances affecting the particular plant, can only be decided by experiment. Probably both occur. The common ivy, for instance, does not now blossom in the northern parts of Sweden and Russia, but it does so still in the southern provinces. If plants were brought to us from the most northerly zone of distribution, they would in all probability flower and bear fruit with us, and in that case the absence of bloom in these plants must have been a direct effect of the cold climate. But it is quite conceivable that cultivated plants have in many cases become hereditarily infertile, when they are constantly propagated only by means of buds, layering, and so on, not however because of any direct effect of this mode of propagation, but through chance germinal variations. For in regard to many of them man has lost all interest in the flowers and fruit, as, for instance, in the case of the potato; in other cases he is even interested in procuring seedless fruits.

In the first case he will quite readily make use of plants with imperfect flowers for propagating, if they are otherwise fit and exhibit what he wants in other respects; in the second case, he will give a preference to individuals with seedless fruits, and thus increase and strengthen the tendency to degeneration of the seeds in the race concerned.

All these cases are quite in harmony with our conception of amphimixis, which, now that we have investigated the facts throughout the animate kingdom, we may sum up in the following propositions. In the whole organic world, from unicellular organisms up to the highest plants and animals, amphimixis now means an augmentation of the organism's power of adaptation to the conditions of its life, since it is only through amphimixis that simultaneous harmonious adaptation of many parts becomes possible. It effects this by the mingling and constant recombination of the germ-plasm ids of different individuals, and thus gives the selection-processes the chance of favouring advantageous variational tendencies and eliminating those which are unfavourable, as well as of collecting and combining all the variations which are necessary for the further evolution of the species. This indirect influence of amphimixis on the capacity of organisms for surviving and being transformed is the fundamental reason for its general introduction and for its persistence through the whole known realm of organisms from unicellulars upwards.

The reason for its _first_ introduction among the lower forms of life must have been a direct effect which had a favourable influence on the metabolism, and this is so far coincident with the subsequent import of amphimixis, inasmuch as it may be regarded not only as a heightening of the power of adaptation, but as an immediate and direct increase and extension of the power of assimilation. In any case, amphimixis is not necessary to the actual preservation of life itself, but it does bring about a wealth and diversity of organic architecture which without it would have been unattainable.

If amphimixis has been abandoned in the course of phylogeny by isolated groups of organisms, this has happened because other advantages accrued to them in consequence, which gave them greater security in the struggle for existence; but it must be admitted that they thereby lost their perfect power of adaptation, and that they have thus bartered their future for the temporary securing of their existence.

In addition to this variational influence, amphimixis has also played a part in the evolution of sharply defined organic types, especially of specific types; but of this we shall have more to say later on.

LECTURE XXXI

THE INFLUENCES OF ENVIRONMENT

Different modes and grades of selection--Changes due to the influences of environment--Superfluity and lack of food--The horses and cattle of the Falkland Islands--Angora animals--Protection against cold in Arctic and marine mammals--Plant-galls--Nägeli's _Hieracium_ experiments--Experiments with _Polyommatus phlæas_--Artificially produced _Vanessa_-aberrations--Vöchting's experiments on the influence of light in the production of flower-forms--Heliotropism and other tropisms--Primary and secondary reactions of organisms--Herbst's 'lithium larvæ'--Schmankewitsch's experiments with _Artemia_--Poulton's caterpillars with facultative colour adaptation--Colour-change in fishes, chamæleon, &c.--Actual scope of those influences which directly produce organic changes.

Through a long series of lectures we have devoted our attention to those phenomena which bear some relation to the processes of selection; we have attempted to gain clearness in regard to the modes and stages of these, and we reached the result that all variations which have taken place in organisms since the first appearance of living matter are directed by processes of selection, that is, their direction and duration are determined by these processes, although they may have their roots in external influences. But it is not to be supposed that this guidance is due solely to that one kind of selection which, with Darwin and Wallace, we designate 'natural selection'; on the contrary, we must regard this as only one of the different modes of the processes of selection, necessarily occurring between all living units which are equivalent to one another, and which, therefore, must maintain a continual struggle with one another for space and food. If the expression 'natural selection' were not already so firmly fixed in its meaning, I should propose that it should be employed in the most general sense for all the processes of selection collectively, but we must keep to its original meaning and use it only for personal selection.

We have seen that processes of selection take place even between the elements of the germ-plasm in all organisms which possess a germ-plasm as distinguished from the mass of the body, and that through these processes there arise those hereditary individual variations which, under some circumstances, form the basis of transformations in the species.

Obviously this may come about in a twofold manner: firstly, a variation movement originating in the germ-plasm may go on increasing till it attains to selection-value, and then 'personal selection' steps in, and seeks to make it the common property of the species. But it is obviously also conceivable that variational tendencies arising in the germ-plasm may never attain to selection-value at all, and then in most cases they will only continue to exist through a longer or shorter series of generations as individual distinguishing characters, without being transmitted to a larger number of individuals or becoming a constant character of the species. Their persistence will depend essentially on the chance of mingling with other individuals, and on the halving of the germ-plasm which precedes sexual reproduction. Sooner or later these individual peculiarities disappear again, as may often be observed in the case of abnormalities or morbid tendencies in man, in as far as these do not weaken vitality. In the latter case they attain selection-value, though only negatively.

But even quite indifferent germinal variations, which neither raise nor lower the individual's power of survival, may, under some circumstances, increase and lead to permanent variations of all the individuals of a species, and this happens when they are conditioned by external influences which affect all the individuals of a species, or of the particular colony concerned, and it is this kind of organismal change which we shall now study for a little in detail.

The ordinary never-ceasing, always active germinal selection depends, we must assume, upon intra-germinal fluctuations of nutrition, or inequalities in the nutritive stream which circulates within the germ-plasm. The variations which it produces may, therefore, be different in each individual, since these fluctuations are a matter of chance and may affect the determinants _A_ in one individual and the determinants _B_, _C_, or _X_ in another, or alternating groups of these. Or it may be that the homologous determinants _A_ may vary in a plus direction in one individual, and in a minus direction in another, while in a third they may remain unchanged, and although the same direction of variation of a determinant _N_ may occur in many individuals, it will certainly not do so in all, and still less will it occur in all along with the same combination of fluctuations in the rest of the determinants. It is only if this occurs that the variation can become a specific character.

We might expect on _a priori_ grounds that not only the chance fluctuations of nutrition within the germ-plasm would cause its elements to vary in this or that direction, but that there would also be influences of a more general kind, especially those of nutrition and climate, which would in the first place affect the body as a whole, but with it also the germ-plasm, and which would therefore bring about variations, either in all or only in certain determinants. In this case all the individuals would vary in the same way, because all would be similarly affected by the same causes of change.

This is actually the case; it is indubitable that external influences, such as those emanating from the environment or media in which species live, are able to cause direct variation of the germ-plasm, that is, permanent, because hereditary variations. We have already referred to this process and called it 'induced germinal selection.'

That such influences of environment may bring about changes in _individual_ organisms is obvious enough; that, for instance, good nutrition makes the body strong and vigorous, that too abundant food makes it fat and causes degeneration, that insufficient food lessens its stamina and vigour, are well-known facts. We have to inquire, on the one hand, to what extent such influences are able to cause changes in the individual body in the course of a lifetime, and, on the other hand, more particularly, how far such changes or modifications of the soma can call forth corresponding variations in the determinant system of the germ-cells, and whether and under what circumstances they may be transmitted; for where this is not the case there can be no permanent hereditary variation of the whole species, and the variation will only persist as long as the conditions which gave rise to it endure, and will disappear again with these.

The influence of nutrition as a cause of variation has often been over-estimated. The old statement which has gone the round of the textbooks since the time of John Hunter, that the stomach of carnivores may be transformed by vegetable diet into a herbivore stomach, is absolutely unproved. Brandes at least, who not only subjected all the statements in the literature on this point to a critical investigation, but also instituted experiments of his own, regards the statement as altogether unfounded. All the 'cases' cited, in which the stomach of a gull or of an owl fed on grain became transformed into an organ with stronger muscles and covered with horny plates, depend, according to Brandes, upon inexact observation. There can therefore be no question of any inheritance of this fictitious stomach-transformation, and the idea that such a fundamental histological adaptation as the alleged transformation of the stomach of the grain-eating bird should arise as a direct effect of the food is wholly without foundation.

But it is quite otherwise with purely quantitative differences in nutrition. That meagre diet influences individuals unfavourably is indubitable, and we are certainly justified in considering whether this may not have an effect on the germ-cells, and one which will correspond to the changes induced on the body, so that if the poor nutrition should last through many generations an hereditary degeneration of the species would occur, which would not at once disappear though the animals were transferred to more favourable conditions.

We certainly know nothing of how far the minuteness of the determinants of the germ-plasm, the whole quantity of the germ-plasm, or the reduced size of the germ-cell, may bear an internal relation to the smallness of the animal which develops therefrom, but it surely cannot be regarded as absurd to suppose that there is some such relation. There are no experiments known to me which prove that meagre diet brings about a progressive decrease in the size of the body. Carl von Voit has observed that dogs of the same litter grew to very different sizes of body according as they received abundant or scanty food, but it would be difficult to make animals small through scantiness of food and at the same time to keep them capable of reproduction, and thus proofs of the inheritance of the dwarfing are lacking. Moreover, the experiments which Nature herself has made are never quite convincing, because we never can definitely exclude the indirect effect of altered circumstances. The case of the feral horses of the Falkland Islands, so often cited since the time of Darwin, which have become small 'through the damp climate and scanty food,' seems to me, of all known cases of the kind, the one we should most readily attribute to the direct effect of continued scanty diet; but even here we cannot altogether exclude the possibility of the co-operation of adaptations of some kind to the very peculiar conditions of life in these islands, as far as the feral horses are concerned. I have not been able to find any record of more modern exact investigations either regarding these feral horses, or in regard to the others which are reared in the Falklands under conditions of domestication. Darwin himself, however, in the Journal of his famous voyage tells us much that is interesting in regard to the mammals of the Falkland Islands. Cattle and horses were brought there in 1764 by the French, and have increased greatly in numbers since that time; they roam about wild in large herds, and the cattle are strikingly large and strong, while the horses both wild and tame are rather small, and have lost so much of their original strength that they cannot be used for catching wild cattle with the lasso, and horses have to be imported from La Plata for this purpose. From this contrast between the horses and the cattle we may at least conclude that it cannot be 'scanty food' alone which causes the horses to become smaller, but that the climatic conditions as a whole are concerned in the matter. Whether the total amount of variation which has taken place in the horses which have lived wild there for a hundred years would take place in the course of a single life, or whether it is a cumulative phenomenon, has still to be decided.

Similar statements, for the most part still more uncertain, are made in regard to changes in the hair of goats, sheep, cattle, cats, and sheep-dogs, which are referred to climatic influence. The raw climate of many highlands, like Tibet and Angora, is said to have directly produced the long and fine-haired breeds. But there is a lack of proof that adaptation or artificial selection did not also play a part, and the fact that similar long-haired breeds have arisen among rabbits and guinea-pigs in quite different places and under quite different climatic conditions, but under the directing care of man, speaks in favour of our supposition. But, on the other hand, it does not seem impossible that the climate may have a variational influence upon certain determinants of the germ-plasm, for we have already seen that the influence of cultivation may incite plants and animals to hereditary variations, and that slowly increasing disturbances in the equilibrium of the determinant system may thereby be produced, which may suddenly find marked expression as 'mutations.' But there is little probability that _adaptations_, that is, transformations corresponding to the altered climate, can arise in this way. The thick fur of the Arctic mammals is assuredly not a direct effect of the cold, although it has developed in all Arctic animals, not only in the modern polar bears, foxes, and hares of the polar regions, but also in the shaggy-haired mammoth of diluvial Siberia, whose tropical relatives of to-day, the elephants, have an almost naked skin. Another interesting case, recently brought to light, shows that a group of animals which, in correspondence with their otherwise exclusively tropical distribution, have only a moderately developed coat of hair, may, on migrating to a cold country, grow as good a fur as the members of other families. I refer to one of the higher apes, _Rhinopithecus roxellanæ_, which live in companies in the forest on the high mountains of Tibet, notwithstanding that the snow lies there for six months[26].

[26] See Milne-Edwards, _Recherches pour servir à l'histoire nat. d. mammifères_, Paris, 1868-74.

But we should assuredly make a mistake if we were to regard the thick fur of these apes as a direct reaction of the organism to the cold. We see at once that this cannot be the case if we compare them with marine mammals, which differ just as much from one another in this respect and yet are exposed to the same low temperature. The whale and the dolphin are quite naked, absolutely hairless, but the seals possess a thick hairy coat. This striking difference is obviously connected with the mode of life; the whales remain always in the water, the seals leave it often and therefore require the hairy coat, especially in colder climates, since otherwise they would be too rapidly cooled by the evaporation of the water from their bodies. For the whales, on the other hand, even a very thick hairy coat would not have sufficed as a protection against cold, since water is a much better conductor of heat than air, and so it was necessary for them to become enveloped with the well-known thick layer of blubber, a deposit of fat lying under the skin, and this--after it was once developed--made the hairy coat superfluous, so that it disappeared. The seals certainly also possess a layer of fat under the skin, but it is only in the largest of them that it affords sufficient protection against the cooling effect of evaporation when they go upon land or on the ice, and it is therefore only in these larger ones that the hairy coat has markedly degenerated, as, for instance, in the walrus and the sea-lion; in all the smaller seals, in which the mass of the body is much less, the hairy coat is necessarily very thick and protected from soaking by being very oily, because the layer of fat under the skin would not be sufficient to prevent excessive cooling when on land. But the thick coat of hair is no more _produced_ by the cold than is the layer of fat. As Kükenthal has shown, all these characters are adaptations, and may depend here as elsewhere upon natural selection and upon the 'fluctuating' variations of the germ-plasm upon which that process is based. They are directed by personal selection because there is the need for them, and they are produced and augmented by germinal selection.

In all these cases the _direct_ effect of external influences has nothing to do with the matter, but in other cases that alone brings about the whole change, which is then limited to the individual and does not affect the species as a whole at all.

Plant-galls afford striking illustration of the extraordinary changes that may be brought about in an organism or in its parts by external influence in the course of the individual life. All possibility of adaptation on the part of the plant is excluded in this case. The gall can only depend upon the direct influence of a stimulus, which is exercised by the young animal, the larva, upon the cells which surround it; and yet these cells vary to a considerable extent, become filled with starch or form a woody layer, secrete special substances, such as tannic acid, in large quantities, or develop hairs, moss-like growths, pigments, and so on, which do not otherwise occur in that particular part of the plant. Since Adler and Beyerinck have proved that it is not a poison conveyed by the mother animal into the leaf or bud when laying the eggs, which gives rise to the gall-formation, the matter has become rather clearer. We can now understand that different stimuli in succession affect the cells which enclose the larva, and that the ordered succession of these and the exactly graded stimulation incite the cells to activity in various ways, whether to mere growth and multiplication in a given direction, or to the secretion of tannic acid, or to the formation of wood, or to the deposition of reserve material, and so on. Even the feeble movements of the young larva may form a stimulus that increases with its growth; then the movements made by the larva in feeding, and not least the different secretions emanating from the salivary glands of the animal, which must contain some substances capable of acting as stimuli and probably changing in character as time goes on. All these factors must act as specific stimuli to the plant-cells, influencing and modifying their processes of growth and metabolism in one direction or another. In principle at least, if not in detail, we understand the possibility that through the ordered succession and exact balancing of these different cell-stimuli the really marvellous structure of the gall may be brought about as the product of the direct influence, exercised only once, of the gall-insect upon the plant's parts. But the animal's power of exercising such a succession of finely graded stimuli upon the plant must be referred to long-continued processes of selection, and the structure of the gall, which is adapted to its purpose down to the minutest details, can thus be understood. The assumption of substances which can act even in minute quantities as specific cell-stimuli, which we require to make in this attempt to explain galls, is no longer without corroboration since we find analogies in the Iodothyrin of Baumann, the specific secretions of the thymus and the supra-renal bodies in the higher animals, not to speak of the 'anti-toxins' of the pathogenic bacteria, which are only known by their effects.

The case of plant-galls is thus of great theoretical interest because we can exclude all preparation of the plant-cells for the stimuli exercised by the animal, since the gall is quite useless for the plant, though many have endeavoured to discover some utility. We have therefore here a clear case of modification due to the effect, exercised once only, of external influences, an adaptation of the animal to the mode of reaction of particular plant-tissues.

It might be supposed that if any inheritance of somatogenic modifications, any transmission of the acquirements of the personal part to the germinal part, were possible at all, it would occur in this case, for many species of gall-insects attack plants, particularly oaks, in great numbers every year. It has actually been maintained that galls may arise spontaneously, that is without the presence of a gall-insect. But no proof of this has ever been found, and the fact that no one has paid any attention to the assertion probably implies an unconscious condemnation of the hypothesis of the transmissibility of acquired characters.

It has been proved by Nägeli's often discussed experiments on hawkweeds (_Hieracium_) that much less specialized external influences can give rise to changes which are not hereditary. The Alpine species of hawkweed varied considerably in their whole habit in the rich soil of the Botanic Gardens at Munich, but their descendants, when transferred to a poor flinty soil, returned to the habit of the Alpine species. The changes which occurred in garden soil were therefore somatic and, as I have called them, 'transient,' and they did not depend upon variations of the germ-plasm. It may be objected in regard to these experiments that they were not continued long enough to prove that hereditary variations would not also have cropped up in consequence of the altered conditions. But in any case they prove that marked changes in the whole body of the plant may occur without any obvious variation of the germ-plasm. This does not mean, however, that the possibility of variations of the germ-plasm through such direct external influences is disputed. We must assume the occurrence of these on _a priori_ grounds, if we refer--as we have done--individual hereditary variation to fluctuations in the nutrition of the individual determinants of the germ-plasm. It is probable that many general nutritive variations or climatic factors affect the germ-plasm as well as the soma, and it is by no means inconceivable that it is not all, but only certain definite determinants that are caused to vary.

A proof of this may be found in the results of experiments made upon the little red-gold fire-butterfly (_Polyommatus phlæas_), to which I have briefly referred in a former lecture. This little diurnal butterfly of the family Lycænidæ has a wide distribution and occurs in two climatic varieties. In the far north and also in the whole of Germany the upper surface is red-gold with a narrow black outer margin, but in the south of Europe the red-gold has been almost crowded out by the black. I reared caterpillars in Germany from eggs of _P. phlæas_ found at Naples and exposed them directly after they had entered on the pupa-stage to a relatively low temperature (10° C.). Butterflies emerged which were not quite so black as those of Naples, but considerably darker than the German form. Conversely, German pupæ were exposed to greater warmth (38° C.), and these gave rise to butterflies which were rather less fiery gold and considerably blacker than the ordinary German form. If I had to repeat these experiments I should use a much lower temperature in the case of the cold experiments, because we now know from the experiments of Standfuss, E. Fischer, and Bachmetjeff, that most of the pupæ of diurnal butterflies can stand a temperature below zero for a considerable time; probably the results would be even more marked then.

But even from the results of my former experiments we are justified in concluding that the blackening of the upper surface of the wing is really the direct result of the increased temperature during pupahood, and that the pure red-gold results from the lowered temperature. Similar experiments made by Merrifield with English _Phlæas_ pupæ agree exactly with mine. But we may conclude further from these experiments that both warmth and cold only give rise to slight variations in the individual pupæ, and that the pure red-gold of the northern form and the black of the southern are the result of a long process of inheritance and accumulation, in which the germ-plasm has been caused to vary in as far as the relevant determinants are concerned, so that these yield the respective northern and southern forms even in less extreme temperatures.

As it is to be assumed that these determinants are present not only in the primordium of the wing in the pupa, but also in the germ-cells, both must be affected by the varying temperature, and, in accordance with the continuity of the germ-plasm, each variation of these determinants, however slight, would be continued in the next generation. It is thus intelligible that somatic variations like the blackening of the wings through warmth appear to be directly inherited and accumulate in the course of generations; in reality, however, it is not the somatic change itself which is transmitted, but the corresponding variation evoked by the same external influence in the relevant determinants of the germ-plasm within the germ-cells, in other words, in the determinants of the following generation.

This interpretation of these experiments, which I offered some years ago, has been confirmed in several ways in regard to various other diurnal Lepidoptera. By employing a temperature as low as 8° C. in the case of fresh pupæ of various species of _Vanessa_ Standfuss and Merrifield, and especially E. Fischer, succeeded in getting great deviations in the marking and colour of the full-grown insects,--so-called aberrations, such as had previously been found only very rarely and singly under natural conditions. The deviations from the normal must undoubtedly be ascribed to the effect of cold, but it does not follow that they are new forms which have suddenly sprung into existence, as many have assumed without further experiment. Dixey, on the other hand, has attempted to establish, by a comparison of the different species of _Vanessa_, the phyletic development of their markings, and has found that these aberrations due to cold are more or less complete reversions to earlier phyletic stages. As regards the common small painted lady (_Vanessa cardui_), the small tortoise-shell butterfly (_Vanessa urticæ_), the 'Admiral' (_Vanessa atalanta_), the peacock (_Vanessa io_), and the large tortoise-shell (_Vanessa polychioros_), I can agree with this interpretation, and I do so the more readily because some years ago I suggested that the alternation of differently coloured generations of seasonally dimorphic Lepidoptera might be considered as a reversion. But this by no means excludes the possibility that other than atavistic aberrations may be produced by cold or heat. There is nothing against this theoretically. Yet we must not, without due consideration, compare these abruptly occurring variations to the sport-varieties of plants which we have already discussed; there is an important difference between the two sets of cases. In the Lepidoptera a single interference, lasting only for a short time, modifies the wing-marking, but in the plant varieties the visible appearance of the variation is preceded by a long period of preparatory change within the germ-plasm. This period required for the external influences to take effect was already recognized by Darwin, and it has recently been named by De Vries the 'premutation period.'

We may explain these remarkable aberrations theoretically in the following way: The determinants of the wing-scales in the wing-primordium of the young pupa are influenced by the cold in different ways, some kinds of determinants being strengthened by it, others markedly weakened, even crippled so to speak, and in this way one colour-area spreads itself out more than is normal on the surface of the wing, and another less, while a third is suppressed altogether. That this disturbance of the equilibrium between the determinants leads usually to the development of a phyletically older marking pattern leads us to the conclusion that in the germ-plasm of the modern species of _Vanessa_ a certain number of determinants of the ancestors must be contained in addition to the modern ones. We might even inquire whether these were not better able to endure cold than their modern descendants, since their original possessors, the old species of the Ice age, were accustomed to greater cold, but this idea is contradicted by the experiments of E. Fischer, which go to show that the same aberrations are evoked by abnormally high temperature. That the old ancestral determinants are present in _different_ numbers in the germ-plasm of the modern species, I am inclined to infer from the fact that among a large number of experiments made by me in the course of several years the aberrations have always occurred in very different numbers in the different broods, although the greatest care was taken to have the conditions as nearly alike as possible; absolutely alike, of course, they never can be.