Chapter IX

CHARACTERS AS ADAPTIVE AND SPECIFIC (_continued_).

It must have appeared strange that hitherto I should have failed to distinguish between "true species" and merely "climatic varieties." But it will conduce to clearness of discussion if we consider our subject point by point. Therefore, having now given a fair statement of the facts of climatic variation, I propose to deal with their theoretical implications--especially as regards the distinction which naturalists are in the habit of drawing between them and so-called true species.

First of all, then, what is this distinction? Take, for example, the case of the Porto Santo rabbits. To almost every naturalist who reads what has been said touching these animals, it will have appeared that the connexion in which they are adduced is wholly irrelevant to the question in debate. For, it will be said that the very fact of the seemingly specific differentiation of these animals having proved to be illusory when some of them were restored to their ancestral conditions, is proof that their peculiar characters are not specific characters; but only what Mr. Wallace would term "individual characters," or variations that are not _inherited_. And the same remark applies to all the other cases which have been adduced to show the generality and extent of climatic variation, both in other animals and also in plants. Why, then, it will be asked, commit the absurdity of adducing such cases in the present discussion? Is it not self-evident that however general, or however considerable, such merely individual, or non-heritable, variations may be, they cannot possibly have ever had anything to do with the origin of _species_? Therefore, is it not simply preposterous to so much as mention them in relation to the question touching the utility of specific characters?

Well, whether or not it is absurd and preposterous to consider climatic variations in connexion with the origin of species, will depend, and depend exclusively, on what it is that we are to understand by a species. Hitherto I have assumed, for the sake of argument, that we all know what is meant by a species. But the time has now come for showing that such is far from being the case. And as it would be clearly absurd and preposterous to conclude anything with regard to specific characters before agreeing upon what we mean by a character as specific, I will begin by giving all the logically possible definitions of a species.

1. _A group of individuals descended by way of natural generation from an originally and specially created type._

This definition may be taken as virtually obsolete.

2. _A group of individuals which, while fully fertile_ inter se, _are sterile with all other individuals--or, at any rate, do not generate fully fertile hybrids._

This purely physiological definition is not nowadays entertained by any naturalist. Even though the physiological distinction be allowed to count for something in otherwise doubtful cases, no systematist would constitute a species on such grounds alone. Therefore we need not concern ourselves with this definition, further than to observe that it is often taken as more or less supplementary to each of the following definitions.

3. _A group of individuals which, however many characters they share with other individuals, agree in presenting one or more characters of a peculiar kind, with some certain degree of distinctness._

In this we have the definition which is practically followed by all naturalists at the present time. But, as we shall presently see more fully, it is an extremely lax definition. For it is impossible to determine, by any fixed and general rule, what degree of distinctness on the part of peculiar characters is to be taken as a uniform standard of specific separation. So long as naturalists believed in special creation, they could feel that by following this definition (3) they were at any rate doing their best to tabulate very real distinctions in nature--viz. between types as originally produced by a supernatural cause, and as subsequently more or less modified (i.e. within the limits imposed by the test of cross-fertility) by natural causes. But evolutionists are unable to hold any belief in such real distinctions, being confessedly aware that all distinctions between species and varieties are purely artificial. So to speak, they well know that it is they themselves who create species, by determining round what degrees of differentiation their diagnostic boundaries shall be drawn. And, seeing that these degrees of differentiation so frequently shade into one another by indistinguishable stages (or, rather, that they _always_ do so, unless intermediate varieties have perished), modern naturalists are well awake to the impossibility of securing any approach to a uniform standard of specific distinction. On this account many of them feel a pressing need for some firmer definition of a species than this one--which, in point of fact, scarcely deserves to be regarded as a definition at all, seeing that it does not formulate any definite criterion of specific distinctness, but leaves every man to follow his own standards of discrimination. Now, as far as I can see, there are only two definitions of a species which will yield to evolutionists the steady and uniform criterion required. These two definitions are as follows.

4. _A group of individuals which, however many characters they share with other individuals, agree in presenting one or more characters of a peculiar and hereditary kind, with some certain degree of distinctness._

It will be observed that this definition is exactly the same as the last one, save in the addition of the words "and hereditary." But, it is needless to say, the addition of these words is of the highest importance, inasmuch as it supplies exactly that objective and rigid criterion of specific distinctness which the preceding definition lacks. It immediately gets rid of the otherwise hopeless wrangling over species as "good" and "bad," or "true" and "climatic," of which (as we have seen) Kerner's essay is such a remarkable outcome. Therefore evolutionists have more and more grown to lay stress on the hereditary character of such peculiarities as they select for diagnostic features of specific distinctness. Indeed it is not too much to say that, at the present time, evolutionists in general recognize this character as, theoretically, indispensable to the constitution of a species. But it is likewise not too much to say that, practically, no one of our systematic naturalists has hitherto concerned himself with this matter. At all events, I do not know of any who has ever taken the trouble to ascertain by experiment, with regard to any of the species which he has constituted, whether the peculiar characters on which his diagnoses have been founded are, or are not, hereditary. Doubtless the labour of constituting (or, still more, of _re_-constituting) species on such a basis of experimental inquiry would be insuperable; while, even if it could be accomplished, would prove undesirable, on account of the chaos it would produce in our specific nomenclature. But, all the same, we must remember that this nomenclature as we now have it--and, therefore, the partitioning of species as we have now made them--has no reference to the criterion of heredity. Our system of distinguishing between species and varieties is not based upon the definition which we are now considering, but upon that which we last considered--frequently coupled, to some undefinable extent, with No. 2.

5. There is, however, yet another and closer definition, which may be suggested by the ultra-Darwinian school, who maintain the doctrine of natural selection as the only possible cause of the origin of species, namely:--

_A group of individuals which, however many characters they share with other individuals, agree in presenting one or more characters of a peculiar, hereditary, and adaptive kind, with some certain degree of distinctness._

Of course this definition rests upon the dogma of utility as a necessary attribute of characters _qua_ specific--i.e. the dogma against which the whole of the present discussion is directed. Therefore all I need say with reference to it is, that at any rate it cannot be adduced in any argument where the validity of its basal dogma is in question. For it would be a mere begging of this question to argue that every species must present at least one peculiar and adaptive character, because, according to definition, unless an organic type does present at least one such character, it is not a specific type. Moreover, and quite apart from this, it is to be hoped that naturalists as a body will never consent to base their diagnostic work on what at best must always be a highly speculative extension of the Darwinian theory. While, lastly, if they were to do so with any sort of consistency, the precise adaptation which each peculiar character subserves, and which because of this adaptation is constituted a character of specific distinction, would have to be determined by actual observation. For no criterion of specific distinction could be more vague and mischievous than this one, if it were to be applied on grounds of mere inference that such and such a character, because seemingly constant, must "necessarily" be either useful, vestigial, or correlated.

Such then, as far as I can see, are all the definitions of a species that are logically possible[122]. Which of them is chosen by those who maintain the necessary usefulness of all specific characters? Observe, it is for those who maintain this doctrine to choose their definition: it is not for me to do so. My contention is, that the term does not admit of any definition sufficiently close and constant to serve as a basis for the doctrine in question--and this for the simple reason that species-makers have never agreed among themselves upon any criterion of specific distinction. My opponents, on the other hand, are clearly bound to take an opposite view, because, unless they suppose that there is some such definition of a species, they would be self-convicted of the absurdity of maintaining a great generalization on a confessedly untenable basis. For example, a few years ago I was allowed to raise a debate in the Biological Section of the British Association on the question to which the present chapters are devoted. But the debate ended as I had anticipated that it must end. No one of the naturalists present could give even the vaguest definition of what was meant by a species--or, consequently, of a character as specific. On this account the debate ended in as complete a destruction as was possible of the doctrine that all the distinctive characters of every species must necessarily be useful, vestigial, or correlated. For it became unquestionable that the same generalization admitted of being made, with the same degree of effect, touching all the distinctive characters of every "snark."

[122] It is almost needless to say that by a definition as "logical"
is meant one which, while including all the differentiae of
the thing defined, excludes any qualities which that thing may
share in common with any other thing. But by definitions as
"logically possible" I mean the number of separate definitions
which admit of being correctly given of the same thing from
different points of view. Thus, for instance, in the present
case, since the above has been in type the late M.
Quatrefages' posthumous work on _Darwin et ses Precurseurs
Francais_ has been published, and gives a long list of
definitions of the term "species" which from time to time have
been enunciated by as many naturalists of the highest standing
as such (pp. 186-187). But while none of these twenty or more
definitions is logical in the sense just defined, they all
present one or other of the differentiae given by those in the
text.

Probably, however, it will be thought unfair to have thus sprung a difficult question of definition in oral debate. Therefore I allude to this fiasco at the British Association, merely for the purpose of emphasizing the necessity of agreeing upon some definition of a species, before we can conclude anything with regard to the generalization of specific characters as necessarily due to natural selection. But when a naturalist has had full time to consider this fundamental matter of definition, and to decide on what his own shall be, he cannot complain of unfairness on the part of any one else who holds him to what he thus says he means by a species. Now Mr. Wallace, in his last work, has given a matured statement of what it is that he means by a species. This, therefore, I will take as the avowed basis of his doctrine touching the necessary origin and maintenance of all specific characters by natural selection. His definition is as follows:--

"An assemblage of individuals which have become somewhat modified
in structure, form, and constitution, _so as to adapt them to
slightly different conditions of life_; which can be differentiated
from allied assemblages; which reproduce their like; which usually
breed together; and, perhaps, when crossed with their near allies,
always produce offspring which are more or less sterile _inter
se_[123]."

[123] Darwinism, p. 167.

From this definition the portion which I have italicized must be omitted in the present discussion, for the reasons already given while considering definition No. 5. What remains is a combination of Nos. 2 and 4. According to Mr. Wallace, therefore, our criterion of a species is to be the heredity of peculiar characters, combined, perhaps, with a more or less exclusive fertility of the component individuals _inter se_. This is the basis on which his generalization of the utility of specific characters as necessary and universal is reared. Here, then, we have something definite to go upon, at all events as far as Mr. Wallace is concerned. Let us see how far such a basis of definition is competent to sustain his generalization.

First of all it must be remarked that, as species have actually been constituted by systematists, the test of exclusive fertility does not apply. For my own part I think this is to be regretted, because I believe that such is the only natural--and therefore the only firm--basis on which specific distinctions can be reared. But, as previously observed, this is not the view which has been taken by our species-makers. At most they regard the physiological criterion as but lending some additional weight to their judgement upon morphological features, in cases where it is doubtful whether the latter alone are of sufficient distinctness to justify a recognition of specific value. Or, conversely, if the morphological features are clearly sufficient to justify such a recognition, yet if it happens to be known that there is full fertility between the form presenting them and other forms which do not, then the latter fact will usually prevent naturalists from constituting the well differentiated form a species on grounds of its morphological features alone--as, for instance, in the case of our domesticated varieties. In short, the physiological criterion has not been employed with sufficient closeness to admit of its being now comprised within any practical definition of the term "species"--if by this term we are to understand, not what any one may think species _ought to be_, but what species actually _are_, as they have been constituted for us by their makers.

From all this it follows that the definition of the term "species" on which Mr. Wallace relies for his deduction with respect to specific characters, is the definition No. 4. In other words, omitting his _petitio principii_ and his allusion to the test of fertility, the great criterion in his view is the criterion of Heredity. And in this all other evolutionists, of whatever school, will doubtless agree with him. They will recognize that it is really the distinguishing test between "climatic varieties" and "true species," so that however widely or however constantly the former may diverge from one another in regard to their peculiar characters, they are not to be classed among the latter unless their peculiar characters are likewise hereditary characters.

Now, if we are all agreed so far, the only question that remains is whether or not this criterion of Heredity is capable of supplying a basis for the generalization, that all characters which have been ranked as of specific value must necessarily be regarded as presenting also an adaptive, or life-serving, value? I will now endeavour to show that there are certain very good reasons for answering this question in the negative.

(A.)

In the first place, even if the modifications induced by the direct action of a changed environment are not hereditary, who is to know that they are not? Assuredly not the botanist or zoologist who in a particular area finds what he is fully entitled to regard as a well-marked specific type. Only by experiments in transposition could it be proved that the modifications have been produced by local conditions; and although the researches of many experimentalists have shown how considerable and how constant such modifications may be, where is the systematic botanist who would ever think of transplanting an apparently new species from one distant area to another before he concludes that it is a new species? Or where is the systematic zoologist who would take the trouble to transport what appears to be an obviously endemic species of animal from one country to another before venturing to give it a new specific name? No doubt, both in the case of plants and animals, it is tacitly assumed that constant differences, if sufficient in amount to be regarded as specific differences are hereditary; but there is not one case in a hundred where the validity of this assumption has ever been tested by experiments in transposition. Therefore naturalists are apt to regard it as remarkable when the few experiments which have been made in this direction are found to negative their assumption--for example, that a diagnostic character in species of the genus _Hieratium_ is found by transplantation not to be hereditary, or that the several named species of British trout are similarly proved to be all "local varieties" of one another. But, in point of fact, there ought to be nothing to surprise us in such results--unless, indeed, it is the unwarrantable nature of the assumption that any given differences of size, form, colour, &c., which naturalists may have regarded as of specific value, are, on this account, hereditary. Indeed, so surprising is this assumption in the face of what we know touching both the extent and the constancy of climatic variation, that it seems to me such a naturalist as Kerner, who never considers the criterion of heredity at all, is less assailable than those who profess to constitute this their chief criterion of specific distinction. For it is certain that whatever their professions may have nowadays become, systematic naturalists have never been in the habit of really following this criterion. In theory they have of late years attached more and more weight to definition No. 4; but in practice they have always adopted definition No. 3. The consequence is, that in literally numberless cases (particularly in the vegetable kingdom) "specific characters" are assumed to be hereditary characters merely because systematic naturalists have bestowed a specific name on the form which presents them. Nor is this all. For, conversely, even when it is known that constant morphological characters are unquestionably hereditary characters, if they happen to present but small degrees of divergence from those of allied forms, then the form which presents them is not ranked as a species, but as a constant variety. In other words, when definitions 3 and 4 are found to clash, it is not 4, but 3, that is followed. In short, even up to the present time, systematic naturalists play fast and loose with the criterion of Heredity to such an extent, that, as above observed, it has been rendered wellnigh worthless in fact, whatever may be thought of it in theory.

Now, unless all this can be denied, what is the use of representing that a species is distinguished from a variety--"climatic" or otherwise--by the fact that its constituent individuals "reproduce their like"? We are not here engaged on any abstract question of what might have been the best principles of specific distinction for naturalists to have adopted. We are engaged on the practical question of the principles which they actually have adopted. And of these principles the reproduction of like by like, under all circumstances of environment, has been virtually ignored.

(B.)

In the second place, supposing that the criterion of Heredity had been as universally and as rigidly employed by our systematists in their work of constructing species as it has been but occasionally and loosely employed, could it be said that even then a basis would have been furnished for the doctrine that all specific characters must necessarily be useful characters? Obviously not, and for the following reasons.

It is admitted that climatic characters are not necessarily--or even generally--useful characters. Consequently, if there be any reason for believing that climatic characters may become in time hereditary characters, the doctrine in question would collapse, even supposing that all specific types were to be re-constituted on a basis of experimental inquiry, for the purpose of ascertaining which of them conform to the test of Heredity. Now there are very good reasons for believing that climatic characters not unfrequently do become hereditary characters; and it was mainly in view of those reasons that I deemed it worth while to devote so much space in the preceding chapter to the facts of climatic variation. I will now state the reasons in question under two different lines of argument.

We are not as yet entitled to conclude definitely against the possible inheritance of acquired characters. Consequently, we are not as yet entitled to assume that climatic characters--i. e. characters acquired by converse with a new environment, continued, say, since the last glacial period--can never have become congenital characters. But, if they ever have become congenital characters, they will have become, at all events as a general rule, congenital characters that are useless; for it is conceded that, _qua_ climatic characters, they have not been due to natural selection.

Doubtless the followers of Weismann will repudiate this line of argument, if not as entirely worthless, at all events as too questionable to be of much practical worth. But even to the followers of Weismann it may be pointed out, that the Wallacean doctrine of the origin of all specific characters by means of natural selection was propounded many years before either Galton or Weismann had questioned the transmission of acquired characters. However. I allow that this line of argument has now become--for the time being at all events--a dubious line, and will therefore at once pass on to the second line, which is not open to doubt from any quarter.

Whether or not we accept Weismann's views, it will here be convenient to employ his terminology, since this will serve to convey the somewhat important distinctions which it is now my object to express.

In the foregoing paragraphs, under heading (A), we have seen that there must be "literally numberless forms" which have been ranked as true species, whose diagnostic characters are nevertheless not congenital. In the case of plants especially, we know that there must be large numbers of named species which do not conform to the criterion of Heredity, although we do not know which species they are. For present purposes, however, it is enough for us to know that there are many such named species, where some change of environment has acted directly and similarly on all the individual "somas" exposed to it, without affecting their "germ-plasms," or the material bases of their hereditary qualities. For named species of this kind we may employ the term _somatogenetic species_.

But now, if there are any cases where a change of environment does act on the germ-plasms exposed to it, the result would be what we may call _blastogenetic species_--i.e. species which conform to the criterion of Heredity, and would therefore be ranked by all naturalists as "true species." It would not signify in such a case whether the changed conditions of life first affected the soma, and then, through changed nutrition, the germ-plasm; or whether from the first it directly affected the germ-plasm itself. For in either case the result would be a "species," which would continue to reproduce its peculiar features by heredity.

Now, the supposition that changed conditions of life may thus affect the congenital endowments of germ-plasm is not a gratuitous one. The sundry facts already given in previous chapters are enough to show that the origin of a blastogenetic species by the direct action on germ-plasm of changed conditions of life is, at all events, a possibility. And a little further thought is enough to show that this possibility becomes a probability--if not a virtual certainty. Even Weismann--notwithstanding his desire to maintain, as far as he possibly can, the "stability" of germ-plasm--is obliged to allow that external conditions acting on the organism may in some cases modify the hereditary qualities of its germ-plasm, and so, as he says, "determine the phyletic development of its descendants." Again, we have seen that he is compelled to interpret the results of his own experiments on the climatic varieties of certain butterflies by saying, "I cannot explain the facts otherwise than by supposing the passive acquisition of characters produced by direct influences of climate"; by which he means that in this case the influence of climate acts directly on the hereditary qualities of germ-plasm. Lastly, and more generally, he says:--

"But although I hold it improbable that individual variability can
depend on a direct action of external influences upon the
germ-cells and their contained germ-plasm, because--as follows from
sundry facts--the molecular structure of the germ-plasm must be
very difficult to change, yet it is by no means to be implied that
this structure may not possibly be altered by influences of the
same kind continuing for a very long time. Thus it seems to me the
possibility is not to be rejected, that influences continued for a
long time, that is, for generations, such as temperature, kind of
nourishment, &c., which may affect the germ-cells as well as any
other part of the organism, may produce a change in the
constitution of the germ-plasm. But such influences would not then
produce individual variation, but would necessarily modify in the
same way all the individuals of a species living in a certain
district. It is possible, though it cannot be proved, that many
climatic varieties have arisen in this manner."

So far, then, we have testimony to this point, as it were, from a reluctant witness. But if we have no theory involving the "stability of germ-plasm" to maintain, we can scarcely fail to see how susceptible the germ-plasm is likely to prove to changed conditions of life. For we know how eminently susceptible it is in this respect when gauged by the practical test of fertility; and as this is but an expression of its extraordinarily complex character, it would indeed be surprising if it were to enjoy any immunity against modification by changed conditions of life. We have seen in the foregoing chapter how frequently and how considerably somatogenetic changes are thus caused, so as to produce "somatogenetic species"--or, where we happen to know that the changes are not hereditary, "climatic varieties." But the constitution of germ-plasm is much more complex than that of any of the structures which are developed therefrom. Consequently, the only wonder is that hitherto experimentalists have not been more successful in producing "blastogenetic species" by artificial changes of environment. Or, as Ray Lankester has well stated this consideration, "It is not difficult to suggest possible ways in which the changed conditions, shown to be important by Darwin, could act through the parental body upon the nuclear matter of the egg-cell and sperm-cell, with its immensely complex and therefore unstable constitution.... The wonder is, not that [blastogenetic] variation occurs, but that it is not excessive and monstrous in every product of fertilization[124]."

[124] _Nature_, Dec. 12, 1889, p. 129.

If to this it should be objected that, as a matter of fact, experimentalists have not been nearly so successful in producing congenital modifications of type by changed conditions of life as they have been in thus producing merely somatic modifications; or if it should be further objected that we have no evidence at all in nature of a "blastogenetic species" having been formed by means of climatic influences alone,--if these objections were to be raised, they would admit of the following answer.

With regard to experiments, so few have thus far been made upon the subject, that objections founded on their negative results do not carry much weight--especially when we remember that these results have not been uniformly negative, but sometimes positive, as shown in Chapter VI. With regard to plants and animals in a state of nature, the objection is wholly futile, for the simple reason that in as many cases as changed conditions of life may have caused an hereditary change of specific type, there is now no means of obtaining "evidence" upon the subject. But we are not on this account entitled to conclude against the probability of such changes of specific type having been more or less frequently thus produced. And still less can we be on this account entitled to conclude against the _possibility_ of such a change having ever occurred in any single instance. Yet this is what must be concluded by any one who maintains that the origin of all species--and, _a fortiori_, of all specific characters--must _necessarily_ have been due to natural selection.

Now, if all this be admitted--and I do not see how it can be reasonably questioned--consider how important its bearing becomes on the issue before us. If germ-plasm (using this term for whatever it is that constitutes the material basis of heredity) is ever capable of having its congenital endowments altered by the direct action of external conditions, the resulting change of hereditary characters, whatever else it may be, need not be an adaptive change. Indeed, according to Weismann's theory of germ-plasm, the chances must be infinitely against the change being an adaptive one. On the theory of pangenesis--that is to say, on the so-called Lamarckian principles--there would be much more reason for entertaining the possibly adaptive character of hereditary change due to the direct action of the environment. Therefore we arrive at this curious result. The more that we are disposed to accept Weismann's theory of heredity, and with it the corollary that natural selection is the sole cause of adaptive modification in species the less are we entitled to assume that all specific characters must necessarily be adaptive. Seeing that in nature there are presumably many cases like those of Hoffmann's plants, Weismann's butterflies, &c., where the hereditary qualities of germ-plasm have (on his hypothesis) been modified by changed conditions of life, we are bound to believe that, in all cases where such changes do not happen to be actively deleterious, they will persist. And inasmuch as characters which are only of "specific" value must be the characters most easily--and therefore most frequently--induced by any slight changes in the constitution of germ-plasm, while, for the same reason (namely, that of their trivial nature) they are least likely to prove injurious, it follows that the less we believe in the functionally-produced adaptations of Lamarck, the more ought we to resist the assumption that all specific characters must necessarily be adaptive characters.

* * * * *

Upon the whole, then, and with regard to the direct action of external conditions, I conclude--not only from general considerations, but also from special facts or instances quite sufficient for the purpose--that these must certainly give rise to immense numbers of somatogenetic species on the one hand, and probably to considerable numbers of blastogenetic species on the other; that in neither case is there any reason for supposing the distinctively "specific characters" to be other than "neutral" or "indifferent"; while there are the best of reasons for concluding the contrary. So that, under this division of our subject alone (B), there appears to be ample justification for the statement that "a large proportional number of specific characters" are in reality, as they are in appearance, destitute of significance from a utilitarian point of view.

(C.)

Thus far in the present chapter we have been dealing exclusively with the case of "climatic variation," or change of specific type due to changes in the external conditions of life. But it will be remembered that, in the preceding chapter, allusion was likewise made to changes of specific type due to internal causes, or to what Darwin has called "the nature of the organism." Under this division of our subject I mentioned especially Sexual Selection, which is supposed to arise in the aesthetic taste of animals themselves; Isolation, which is supposed to originate new types by allowing the average characters of an isolated section of an old type to develop a new history of varietal change, as we shall see more fully in the ensuing part of this treatise; and the Laws of Growth, which is a general term for the operation of unknown causes of change incidental to the living processes of organisms which present the change.

Now, under none of these divisions of our subject can there be any question touching the criterion of Heredity. For if new species--or even single specific characters of new species--are ever produced by any of these causes, they must certainly all "reproduce their like." Therefore the only question which can here obtain is as to whether or not such causes ever do originate new species, or even so much as new specific characters. Mr. Wallace, though not always consistently, answers this question in the negative; but the great majority of naturalists follow Darwin by answering it in the affirmative. And this is enough to show the only point which we need at present concern ourselves with showing--viz. that the question is, at the least, an open one. For as long as this question is an open one among believers in the theory of natural selection, it must clearly be an unwarrantable deduction from that theory, that all species, and _a fortiori_ all specific characters, are necessarily due to natural selection. The deduction cannot be legitimately drawn until the possibility of any other cause of specific modification has been excluded. But the bare fact of the question as just stated being still and at the least an open question, is enough to prove that this possibility has not been excluded. Therefore the deduction must be, again on this ground alone (C), unwarrantable.

* * * * *

Such are my several reasons--and it is to be observed that they are all _independent_ reasons--for concluding that it makes no practical difference to the present discussion whether or not we entertain Heredity as a criterion of specific distinction. Seeing that our species-makers have paid so little regard to this criterion, it is neither absurd nor preposterous to have adduced, in the preceding chapter, the facts of climatic variation. On the contrary, as the definition of "species" which has been practically followed by our species-makers in No. 3, and not No. 4, these facts form part and parcel of our subject. It is perfectly certain that, in the vegetable kingdom at all events, "a large proportional number" of specifically diagnostic characters would be proved by experiment to be "somatogenetic"; while there are numerous constant characters classed as varietal, although it is well known that they are "blastogenetic." Moreover, we can scarcely doubt that many specific characters which are also hereditary characters owe their existence, not to natural selection, but to the direct action of external causes on the hereditary structure of "germ-plasm"; while, even apart from this consideration, there are at least three distinct and highly general principles of specific change, which are accepted by the great majority of Darwinists, and the only common peculiarity of which is that they produce hereditary changes of specific types without any reference to the principle of utility.