But is it true, as Fabre thinks, that "to be able to give to each larva the space and the nourishment which belong to it, whether it is male or female . . . the female should know the sex of the egg which she is going to lay"?

Let us first note that the solitary nest-builders must instinctively build or choose large cells and small ones, that they have acquired this instinct after long experience or through mutations, and that the species which manifest it from the first have been able to benefit largely from natural selection because this greatly reduced the architectural needs.

Given this instinct, it suffices, as Marchai observes, "from a simple psychic adaptation" of the female, "for the seminal receptacle to contract for one of the two kinds of cells and to remain inactive for the other." At the sight of a large cell the female will contract her receptacle in a reflex way, the spermatozoa will fertilize the egg situated in the oviduct, and the fertilized egg will produce a female ; with no reflex before the little cells there will be no fertilization and the egg will develop into a male.

This is the same explanation which we have adopted, following Marchai, for the social bees and wasps. It avoids the idea of the will of the female, but supposes that the male eggs are par-thenogenetic. But if it is true that this is the case with the social species, it is far from enabling us to affirm that the same rule is obeyed with the solitary species. This is, I believe, the sentiment of Cuénot (1899) and perhaps also that of Bugnion (1910), but upon this point research has been insufficient. I know of no other except that of Ver-hoefï (1891), who attributes a constant parthenogenetic origin to the males of two solitary wasps (Odynerus spinipes and 0. parietum), although those of 0. reni for mis result sometimes from a fertilization. We should not underestimate the importance of this gap, and it is to be hoped that serious study will soon make it disappear. If, as we may believe, research establishes the fact that the males of the solitary nesting bees have a par-thenogenetic origin, it will completely justify the hypothesis proposed by Marchai. In a contrary case, the explanation of the mechanism of the division of the sexes will be very delicate. The fact is established that with non-parthenogenetic species the sex of the egg is determined by the spermatozoa, which are of two kinds from the point of view of their power and, sometimes, from the point of view of structure. It is necessary, then, to admit that the psychic reflex of origin provokes the issuance of spermatozoa of the one or the other kind. We do not know of any such mechanism, and we must pass over this difficulty, which will, without doubt, vanish after badly needed studies have been made, as we have suggested above.