If we refer to "Weismannism up to date," as Mr. Romanes so happily indicates the latest teaching of this great savant, we shall find that he has arrived now at a definite conclusion.

"Many climatic varieties of plants may also be due wholly or in part to the simultaneous variation of corresponding determinants in some part of the soma and in the germ plasm of the reproductive cells, and these variations must necessarily be hereditary." -(The Germ Plasm. p. 406.)

Speaking of amphimixis, he says:

"It is not the primary cause of hereditary variation. By its means those specific variations which already exist in a species may continually be blended in a fresh manner, but it is incapable of giving rise to new variations even though it often appears to do so. The cause of hereditary variation must lie deeper than this. It must be due to the direct effect of external influences on the biophors and determinants."-(The Germ Plasm. pp. 414-15.)

Now the effect of this modification of the theory is enormous; for whereas the denial of any change of the continuous germ plasm, due to the effect of changed conditions, emphasised the importance of sexual reproduction as the sole means of modification, and variations thus produced could only be fixed in a race living in a state of nature by the principle of Natural Selection, the admission that external influences produce a direct effect upon the germ plasm and are the sole source of new variations establishes the influence of the principle of the action of external conditions, so as to render Natural Selection no longer necessary. The effect of all this is to make the theory of the continuity of the germ plasm the ally of Lamarckism rather than of Darwinism.

And this result is not affected by some of the modifications of the theory. The essential point is that the

be

germ plasm should be modified by the action of external circumstances; that the germ plasm of all the individuals subjected to the same outward conditions should modified in the same way; these similar variants will intermarry apart from selection. Let this essential change take place, and it matters little for our argument under what especial conditions and limitations it takes place. It is asserted that sexual reproduction is necessary to develop this influence. We cannot see why this should be necessary any more than it is necessary to mix the contents of two wine vats of the same vineyard, in order to bring out the characteristic aroma of this particular growth. But grant that "amphimixis" is necessary; that does not vitiate but develop the change that has already taken place. The preparatory modification has taken place in all and produces its undoubted result after sexual reproduction. (How could it produce any result on offspring before?) And it is obvious that the germ plasm of all being transformed in the same way, there is no need, no room for the selection of some which have been

modified in a special way. Nor is this principle affected by the assertion as to the time when this modification of the germ plasm takes place to sufficiently modify the organism itself. If it be asserted that the result is only visible after several generations, this is only in accordance with the fact that some species do not vary so readily as others.

It would require a volume to deal properly with the theory of Dr. Weismann, regarding it as a statement of some of the most difficult phenomena of nature. I have endeavoured to look at it simply from the point of view of the theory of Natural Selection; and I trust that nothing has been omitted which is essential to the argument, although I have avoided as much as possible the

technical language in which the theory is stated by its author.

Before leaving the theory of Dr. Weismann, it may be well to indicate a certain ambiguity in the use of the word "acquired." It is generally used to indicate variations acquired by the body cells of an organism; and unless I am mistaken it is tacitly assumed that no acquired modification whatever can be inherited, and that therefore the transmutation of species in nature is due to the Natural Selection of variations necessarily associated with reproduction. But surely if external conditions can modify the sexual elements, these modifications have been acquired by the sexual elements; and, as we have seen, the necessity for Natural Selection at once disappears, and something very like a Lamarckian principle is at once established.

CHAPTER VI.

PURE DARWINISM (continued).

(b) MORE METHODS OF SELECTION THAN ONE.

"Il y a fagots et fagots."-MOLIÈRE.

THERE are, to say the least, three conceivable ways in which similar variations may be isolated for breeding purposes in the world of nature, apart from Natural Selection. Physical separation, which is so effectual in artificial selection, might also operate in nature. Similar variants might segregate themselves with results similar to those of artificial selection; similar variants might be physiologically differentiated so that either their period of sexual maturity differed from that of the rest of the race, or they might be fertile inter se and infertile with the rest of the race. And I see no reason whatever why these three principles should not prove as efficient in nature as the principle of Natural Selection is supposed to be, on the hypothesis that the tacit assumptions of that theory are to be relied on. Let us consider, for a moment, what these assumptions are. Those who believe that Natural Selection is the sole cause of the transmutation of species, argue on the ground of the necessity of" selecting "—that is, of isolating for breeding purposes-those similar favourable variations which arise in connection with variations necessarily associated with sexual reproduction. They assert that the agent of

this selection is found in the struggle for existence arising from the tendency of all organisms to increase in a geometrical ratio.

But when this assertion is made, it is tacitly assumed that sexual reproduction will necessarily cause the output of variations which are unmodified by any other influence; that these variations will afford the necessary material for selection; and that Natural Selection-selection by life and death is the only isolation for breeding purposes possible in nature. But if the first two assumptions are taken for granted, it does not necessarily follow that Natural Selection is the sole agent in nature for isolating similar variations. And, first, it may be observed that, on the principles of Natural Selection, isolation by life and death cannot be the only possible means of isolating similar variants for breeding purposes; for in artificial selection this is generally effected by physical separation; and hence we may presume that there may be two or more methods of isolation in nature. The next point to be observed is that it is the fact of isolation which is important, and that the same result may be attained by many different methods. As Mr. Romanes says:

"All that the causation of the case requires is that there should be exclusive breeding between the similarly modified individuals. Whether this exclusive breeding is secured by killing off all other individuals, or by fencing them all out of a field, is plainly immaterial.”—(Contemporary Review. vol. liii., p. 846.)

The same remarks will apply with equal force to social segregation and to physiological differentiation.

This is true as to the immediate result of the isolation of similar variants, but it is not immaterial what methods are suggested; there may be methods practicable to art which would be impossible in nature; methods appearing like a splendid vision in the realm of theory and yet not