The development of a half embryo from one of the first two blastomeres of the frog’s egg, in Roux’s experiment, seemed to support Roux’s hypothesis, but it was not long before it was seen that the presence of the other blastomere vitiated the evidence to such an extent as to render it worthless, so far as this hypothesis is concerned. Then followed the experiments with the isolated blastomeres of the sea-urchin, amphioxus, jelly-fish, teleost, ascidian, triton, etc., in which each blastomere, when completely separated, gives rise to a whole embryo. From these experiments Driesch and Hertwig drew the opposite conclusion, namely, that during the cleavage there is a quantitative division of the egg into blastomeres that are equivalent, or at least totipotent. Roux attempted to meet the results of these experiments in two ways. He pointed out that in several of these cases the isolated blastomere divides as a half or as a fourth of the egg, and that in the sea-urchin this leads to the formation of an open half-blastula. In the second place, Roux brought more to the front his subsidiary hypothesis of the reserve germ plasm. He supposed that along with the early qualitative division of the nucleus, by means of which each part receives its particular chromatic substance, there is also a quantitative division of a sort of reserve germ plasm contained in the nucleus. Each cell _may_ receive also a part of this material, and hence each cell may contain the potentialities of the whole egg. This reserve plasm may be awakened by any change that alters the normal development, as, for instance, when the blastomeres are separated. It may take some time for this reserve stuff to wake up, as shown by the half-development of the sea-urchin’s egg that goes on for some time after the separation of the blastomeres. This hypothesis cannot be objected to on purely formal grounds, but we are not so much concerned with a purely logical hypothesis as with a verifiable one.