With the social wasps as with the higher bees, the egg is laid singly in the still empty cell, and the food of the larvae is given them day by day. But this food is composed of fragments of fresh prey more or less masticated and mixed with honey, and, except for their reproductive organs, the queens differ very little from the workers.

With H. von Ihering and Ferton, we had thought that the social wasps were near the solitary sphegids, more or less closely related to Bembex, because Monedula punctata studied by Hudson, Bembex mediterraneus, Stizus tridens, and S. errans, studied by Ferton, lay their egg in an empty cell and nourish their larvae from day to day with killed or partially paralyzed prey. But the recent studies of Roubaud1 have shown that, aside from this primitive method, the sphegids have evolved in the sense of causing more or less perfect paralysis, which allows them to accumulate in their cells all the prey necessary for the food of their progeny; also, that they have not passed this stage, and that the social wasps are directly related to the solitary wasps of the family Eumenidœ.

1 E. Roubaud, Recherches biologiques sur les Guêpes sociales et solitaires d'Afrique. {Ann. Sci. Nat. Zool., ser. 10 v. 1. pp. 1-160, 1916).

As a matter of fact the Eumenidœ have become detached from the sphegid group, but their biological evolution has been brought about in a different way. The Odynerus and the Eumenes of our country are excellent paralyzers which lay their egg and afterward make a provisioning in mass like the sphegids. The African species studied by Roubaud, Synagris calida, S. sicheliana, Eumenes tinctor, Odynerus bellatulus, do the same in the wet season, since then there is an abundance of caterpillars, but in the dry season, when this food is much rarer, they await the hatching of the egg, furnish the larva "with more than enough food to last through the day," and before closing the cell complete the provisioning "with a sufficiently large supplementary supply of caterpillars." It is the delayed provisioning in bulk which leads us to the daily provisioning practised by Synagris cornuta. Roubaud has stated, indeed, that this species feeds its larvae from day to day with fragments of caterpillars. After a detour which leads them to perfect paralysis and provisioning in bulk, Synagris returns to the brutal habit of the Bem-becids and to the mode of provisioning which is its consequence. This method is followed by all social wasps, notably by one of the African species, Belonogaster junceus, studied by Roubaud. With this wasp the habits resemble those of Synagris cornuta, but the young live in the nest by the side of the parents which built it, and the work which they do retards somewhat the maturation of their ovarian glands. In fact, these young become sooner or later fertile, and as the nests may be founded by several queens, we see that the societies of this species are polygenous at an approximate time. Having no workers, these very primitive societies dissolve by emigration when bad weather comes. Then it is not rare to see a queen raise a single larva at a time and behave exactly like Synagris cornuta. Without doubt Belonogaster builds paper nests like most of the social wasps, but certain of the latter also know how to build masonry nests, like Eumenes and Synagris, or to dig readily in the soil like Odynerus.

From Belonogaster we pass to Icaria and to Polistes marginalis, social African wasps in which Roubaud has noted the presence of numerous queens and of a lesser number of true workers. Polygeny is also frequent with the American social wasps, among which it was discovered by Hermann von Ihering (1896) with Polybia scutellaris. In this species the number of queens may reach fifteen per cent, of the population, which is principally composed of neuter females. Extending the researches of his father on other species, Rodolphe von Ihering (1905) has divided the social wasps into polygamous forms (polygynes) and monogamous forms (monogynes) according as their societies include several queens or one only. This classification has been systematized by Ducke in a recent work,1 but it would hardly seem natural when we consider that certain genera, such as Polybia and Polistes, include at the same time monogamous and polygamous species. However this may be, we cannot doubt that Cha-tergus, Nectarina, and the most of the Polybias are polygamous; while the Vespidae, the true wasps, are monogamous in all countries. It is to be noted that the polygamous American societies are perennial and that they multiply by swarming, that many store up honey, and that their cells are all alike, which is a primitive character. According to H. von Ihering, Polybia scutellaris gives its larvae mutilated prey, and not bits of flesh.

Our Polistes gallicus is an annual species whose pedunculate nest is made of cells of the same size. It forms monogamous societies, but Ferton has seen females of this species (sisters, to all appearances) living in perfect amity and associating together in the work of the nest. As to the true wasps of the genus Vespa, they also are monogamous, but their nests are composed of many shelves in which are found cells of two kinds, little ones for the workers, larger ones for the true females, the eggs of the males being placed in those of either size, but preferably in the smaller ones. This is the maximum of the division of work and of castes in the group of social wasps. This division is not pushed as far as with the bees. When she is definitely held in the nest by the need of laying, the original queen, says Robert du Buys-son,1 "still continues to work; she gives slight retouches here and there to the covering and to the cells; she distributes to the larvae morsels which are furnished to her by the workers," and this recalls a little the very primitive habits of Belonogaster.

1 A. Ducke, Revision des Guêpes sociales polygames d' Amérique: 1910. (Ann. Mus. Nat. Hungar. Budapest, v. pp. 499-544).