Furthermore, according to Professor Weismann, natural selection is really working, not on the organism at large, but on the germ-plasm which produces it; and it is conceivable that the variation of one or more of the few cells in early embryonic life may introduce a great number of variations in the numerous derivative cells. In explanation of my meaning, I will quote a paragraph from a paper of Mr. E. B. Poulton's on "Theories of Heredity."[DJ] "It appears," he says, "that, in some animals, the great groups of cells are determined by the first division [of the ovum in the process of cleavage[DK]]; in others, the right and left sides, or front and hind ends of the body; while the cells giving rise to the chief groups on each side would then be separated at some later division. This is not theory, but fact; for Roux has recently shown that, if one of the products of the first division of the egg of a frog be destroyed with a hot needle, development is not necessarily arrested, but, when it proceeds, leads to the formation of an embryo from which either the right or the left side is absent. When the first division takes place in another direction, either the hind or the front half was absent from the embryo which was afterwards produced. After the next division, when four cells were present, destruction of one produced an embryo in which one-fourth was absent." Now, it is conceivable that a single modification or variation of the primitive germ might give rise to many correlated modifications or variations of the numerous cells into which it develops; just as an apparently trivial incident in childhood or youth may modify the whole course of a man's subsequent life. It is difficult, indeed, to see how this could be effected; to understand what could be the nature of a modification of the germ which could lead simultaneously to many favourable variations of bones, muscles, blood-vessels, and nerves in different parts of the body. This, however, is a question of the origin of variations; and it is, at any rate, conceivable that, just as by the extirpation with a hot needle of one cell of the cleaved frog's ovum all the anterior part of the body should be absent in development, so by the appropriate modification of this one cell, or the germinal matter which produced it, all the anterior part of the body should be appropriately modified. These considerations, perhaps, somewhat weaken the force of Mr. Spencer's argument, which is not quite so strong now as it was when the "Principles of Biology" was published.