Castle assumes, in fact, that in animals like daphnids and rotifers one polar body only is extruded, and the other (the second) is retained in the egg, and hence the potentiality of producing males is present. In the honey-bee, on the contrary, Castle assumes that both polar bodies are extruded in the unfertilized egg (and there are some observations that support this idea), and since only males are produced from these, he believes it is the female element that has been sent out into the second polar body. This hypothesis is necessary, because Castle assumes that when both elements are present in the bee’s eggs, the female element dominates. “Hence, if the egg which has formed two polar cells develops without fertilization, it must develop into a male. But if such an egg is fertilized, it invariably forms a parthenogenetic female ♀ (♂), that is, an individual in which the male character is recessive. Accordingly the _functional_ spermatozoon must in such cases invariably bear the _female character_, and this is invariably dominant over the male character when the two meet in fertilization.” If it should prove generally true that the size of the egg is one of the factors determining the sex, we have still the further question to consider as to whether some eggs are bigger because they are already female, or whether all eggs that go beyond a certain size are females, and all those that fail to reach this are males. If this is the case, an animal might produce more females if the external conditions were favorable to the growth of the eggs, and if in some cases these large eggs were capable of developing, parthenogenetic races might become established. Should, however, the conditions for nutrition become less favorable, some of the eggs might fall below the former size and produce males. It is not apparent, however, why all the fertilized autumn eggs of the aphids should give rise to females, for although these eggs are known to be larger than the summer eggs, yet they are produced under unfavorable conditions. The preceding discussion will show how far we still are from knowing what factors determine sex. Castle’s argument well illustrates how many assumptions must be made in order to make possible the view that sex is a predetermined quality of each germ-cell. Even if these assumptions were admissible, we still return to the old idea that the _fertilized egg_ has both possibilities, and something determines which shall dominate. Until we have ascertained definitely by experimental work whether the sex in some forms can be determined by external conditions, it is almost worthless to speculate further. Whatever decision is reached, the conclusion will have an immediate bearing on the question to be next discussed. Meanwhile, we can at least examine some of the theories that have been advanced as to what advantage, if any, has been gained by having the individuals of many classes divided into two kinds, male and female.