CLASSIFICATION OF ANTS.—Ants are considered by many entomologists to form a series called Heterogyna. They can, however, be scarcely considered as more than a single family, Formicidae, so that the serial name is superfluous. Their nearest approach to other Aculeates is apparently made, by _Amblyopone_, to certain Mutillides (_e.g._ _Apterogyna_) and to the Thynnides, two divisions of Scoliidae. Emery considers Dorylides rather than Amblyoponides to be the most primitive form of ants, but we are disposed to consider Forel's view to the effect we have above mentioned as more probably correct. The point is, however, very doubtful. The condition of the peduncle is in both the sub-families we have mentioned very imperfect compared with that of other ants. Both these sub-families are of very small extent and very imperfectly known. We shall also follow Forel in adopting six sub-families, Camponotides, Dolichoderides, Myrmicides, Ponerides, Dorylides, and Amblyoponides. Emery rejects the Amblyoponides as being merely a division of the Ponerides. This latter group displays the widest relations of all the sub-families, and may be looked on as a sort of central form. The Camponotides and Dolichoderides are closely allied, and represent the highest differentiation of the families in one direction. The Myrmicides are also highly differentiated, but are not allied to the Camponotides and Dolichoderides.[62] SUB-FAM. 1. CAMPONOTIDES.—_Hind body furnished with but one constriction, so that only a single scale or node exists on the pedicel. Poison-sac forming a cushion of convolutions, on which is situate the modified sting, which forms merely an ejaculatory orifice for the poison._ The members of this very extensive division of ants can be readily distinguished from all others, except the Dolichoderides, by the absence of a true sting, and by the peculiar form of the hind body; this possesses only a single scale at the base, and has no {145}constriction at all on the oval, convex and compact mass of the abdomen behind this. The cloacal orifice is circular, not, as in other ants, transverse. These characters are accompanied by a difference in habits. The Camponotides, though they do not sting, produce poison in large quantity, and eject it to some distance. Hence, if two specimens are confined in a tube they are apt to kill one another by the random discharges they make. Janet suggests that in order to neutralise the effect of this very acid poison, they may have some means of using, when they are in their natural abodes, the alkaline contents of a second gland with which they are provided. We shall mention the characters by which the Camponotides are distinguished from the small sub-family Dolichoderides when we deal with the latter.