It seems that an abyss, from the social point of view, separates the solitary bees from our domestic honey-bee, with which the division of work is so perfect ; but comparative biology permits us to bridge that abyss.

We note first that certain solitary bees capable of living in small colonies are also capable of work in common. The females of Halictus frequently have the habit of digging a burrow, in which they hide away for the winter, but Verhoeff has found hibernating under a stone sixteen females of Halictus morio distributed here and there in branching galleries. We believe with Buttel-Reepen that a first female had established herself under the stone and that others, finding the place favorable, had continued the work of excavation. Certain Halicti studied by Nicolas (1887) offer the curious particularity "that a single door gave access to several habitations ; one door opened upon several galleries where certain females were isolated. Such a disposition demands a special surveillance" ; at the entry of the common gallery an individual was placed who acted as porter, chased away the intruders, and took care in the vestibule to keep a passage for the members of the colony. Aurivillius (1896) observed the same habit with the Hungarian species, Halictus longulus. The Halicti conduct us a little further on the way to communism. Most bees of this genus establish their cells isolated on the sides of the gallery, but Verhoeff (1891) has observed that the cells are a little nearer together in H. sexcinctus and that in H. quadrimaculatus their ensemble forms a sort of a ray everywhere isolated from the neighboring earth except in the lower part. There were found in these nests empty cells, without provisions, while others were inhabited by nymphs on the point of issuing, so that it is possible that the first nests were always in contact with their mother; so it is, says Verhofr, "colonization is ready," and will be realized if the children aid their mother in her work.

1 A. Forel, Les Fourmis de la Suisse, 1874. (Soc. helvet. des sci. nat. Nouv. Mém. 26).

This realization does not seem to be attained with Halictus, but it becomes almost absolutely the rule with the bumblebees, large, velvety bees, concerning which Buttel-Reepen (1903) has conducted a study rich in achievement. When she starts her nest in the first warm days, the queen provisions it before laying, like the solitary bees, but the food is insufficient, for the insect lays several eggs in the rearing-cell and she must open and close this cell several times to give her larvae supplementary food, and that is the method followed by the bees. Tins mode is employed from the beginning until the first females have issued and replace the queen in her work. Poorly nourished by their mother, who assumed all the work, these females are small, undersized; but as more and more are born, their vivacity increases and also the size of the new-born, which in the end reach the dimensions of the queen. These female workers do not lay, but they may not be regarded as neuters, for they have the same structure as the queen, and may be distinguished only by the feeble development of their ovaries. At the end of the season, when the males appear, the largest are fertilized and hunt refuge for the winter. All the other individuals perish. It is not the same in the tropics, at least in Brazil, where Rodolphe von Ihering (1904) has found that the colonies of Bom-bus are perennial and multiply by swarming. Then the fertile females remain at home, and the nest can count numerous queens in addition to the workers. On the contrary, in the Northern countries, where the summer is short, certain species of bumblebees return to isolated work; Friese (1901) reports, in fact, that Spar re Schneider, in the course of long studies carried on at Tromso, has never found worker females in the nests of Bombus kirbyellus, and that their rarity is extreme in those of B. hyperboreus. I may add that by the ovoid form and the irregular arrangement of their cells, by the employment of little twigs and moss in making the nest, the bumblebees recall certain solitary bees of the genus Osmia; without doubt they store up honey and secrete wax, but it is hardly possible that B. kirbyellus can accumulate provisions, and we know that certain solitary bees (Centris, some Anthophoras) already produce a waxy secretion.

The tropical bees of the family Meliponidœ, the American Meliponas and Trigonas of the two continents, build with wax ovoid cells similar to those of bumblebees, but these cells are used only for provisions and are included in the nest which they build of clay and resin. What especially distinguishes the constructions of these bees is the presence of horizontal or oblique rearing floors formed by a single row of hexagonal cells which open upward, and are provisioned before the laying, like those of the solitary bees. The societies of these bees are formed of a queen and a number of neuter females which possess only collecting organs. We know from Hermann von Ihering1 that the young queens are of almost the same size as the workers, that those of Melipona are born in the ordinary cells and remain some time in the nest until their organs harden, while those of the Trigonas grow in large royal cells and are fully hardened before their issuance. In the two cases, however, the societies are perennial and multiply by swarming; and since the queen mothers become very obese and incapable of taking flight, it must be admitted that the swarms are formed by the young queens and by a crowd of workers.

The societies of Apis, or true bees, so-called, are also composed of a single queen, of neuter females, and males; they are also perennial and multiply by swarming, but they occupy a much higher place in the social hierarchy. Their combs are always vertical and formed of two layers of hexagonal cellules, which are connected at the bottom; these cells serve at the same time for provisions and for rearing, while the larvaa which they enclose are provisioned from day to day. The queen mother is obese but capable of flight; she does not allow the young queens near her, and after they issue, swarms with a party of workers. All the species are not, however, in the same evolutional stage. Apis dor sat a makes only one comb, in which all the cells are alike ; the little Apis florea also constructs a single comb, but the cells are of two kinds, small ones for the workers, larger ones for the males, and according to Buttel-Reepen, it seems, indeed, that the royal cells are constructed at the expense of the latter. In the Apis indica and in our honey-bee (Apis mellifica) the combs are multiplied, more or less parallel, with cells of three kinds, of which the royal cells are constructed of several of the others, or at the expense of the worker cells. This is the highest social type in the group of Mellifera.

1 H. von Ihering, Biologie der stachellosen Honingbienen Brasiliens, 1903. (Zool. Jahrh. Syst., v. 19. pp. 179-287).

One would think that the higher bees are related to the solitary bees through the bumblebees, the Meliponas, and the Trigonas, but this is not true, as we see it, and it seems rather that these types have diverged on leaving their solitary sources.

In fact, we conclude with Emery (1894), "we know no precursor stage" to the kind of social form of our honey-bee, "but only stages which resemble these precursor stages, and which we may utilize to understand the biological phylogeny of the societies" of the Mellifera. The same is true, moreover, of wasps.