Quite recently, Cuénot, Beard, Castle, and Lenhossek have contended that the differentiation of sex is the outcome of internal factors. They think that the view that sex is determined by external agents is fundamentally erroneous. The fallacies that have given rise to this conception, Castle points out, are, first, that in animals that reproduce sometimes by parthenogenesis and sometimes by fertilized eggs, the former process is favored by good nutrition and the latter by poor nutrition. This only means, in reality, Castle thinks, that parthenogenetic reproduction is favored by external conditions, and this kind of reproduction, he thinks, is a thing _sui generis_, and not to be compared to the formation of more females in the sexual forms of reproduction. There is no proof, however, that this is anything more than a superficial distinction, and it ignores the fact that in ordinary cases the females sometimes lay parthenogenetic eggs which differ, as far as we can see, from eggs that are destined to be fertilized in no important respect. More significant, it seems to me, is the fact that only parthenogenetic females develop the following spring from the fertilized eggs of the last generation of the autumn series, whose origin is described to be due to lack of food. We find, in the case of aphids, that unfertilized parthenogenetic eggs and also fertilized eggs give rise to females only, while a change in the amount of food causes the parthenogenetic eggs to give rise both to males and to females. This point is not, I think, fully met by Castle, for even if the change in food does not, as he claims, cause only one sex to appear, yet lack of food does seem to account for the appearance of the males at least. The other fallacy, mentioned by Cuénot, is that the excess of males that has been observed when the food supply is limited is due to the early death of a larger percentage of females, which require more food, but this still fails to account for the excess of females when more food is given, provided Yung’s experiments on tadpoles are correct. It may be, however, in the light of Pflüger’s results, that there has been some mistake in the experiments themselves. We may now proceed to examine Castle’s argument, attempting to show in what way sex is predetermined in the embryo. His hypothesis rests on the three following premises: “(1) the idea of Darwin, that in animals and plants of either sex the characters of the opposite sex are latent; (2) the idea of Mendel, that in the formation of the gametes [germ-cells] of hybrids a segregation of the parental characters takes place, and when in fertilization different segregated characters meet, one will dominate, the other become latent or recessive; (3) the idea of Weismann, that in the maturation of egg and spermatozoon a segregation is attended by a visible reduction in the number of chromosomes in the germinal nuclei.”