Ordinarily the tubes expand when the ants are absent from the last segments, and are certainly withdrawn when they come near. These tubes, from all appearances, serve as signals to the ants. When the latter discover them expanded, they know that a refection is ready, and rush to the opening in the eleventh segment where it is to be found. The tubes certainly serve no other purpose. No visible duct appears in the dome of the tube when largely magnified, and the ants seek nothing from it or the twelfth segment. They cannot be used to intimidate, or to frighten away enemies, for in the younger stages, when the larvæ have the most to dread, neither the tube nor the organ in the eleventh segment is available. The outward openings, and the orifice in the eleventh segment, exist in the youngest larval stages, but are functionless until the larva has nearly attained maturity. Ants seldom attempt to caress or solicit young larvæ, but pass them by with indifference, seemingly knowing that they cannot emit the secretion. When an ant approaches one of these immature larvæ, the larva manifests considerable annoyance, throwing up the hinder segments, as though the ant was an enemy which it was desirous to get rid of. If the tubes could now be thrust out, the ant would be attracted, rather than repelled. But when the period arrives that the tubes are free, and the secretion is ready to be ejected, which is perhaps just after the third skin-moulting, and it cannot be earlier, the larva grows now quiet and submissive, inviting the attentions of the ants, and rewarding their antennal caresses. Four species of parasites affect these larvæ. Two are dipterous. These, which are of the size of the common house-fly, deposit their eggs, during the second larval stage, on the back, and near the junction of the second and third segments. In process of time the grubs hatch and eat their way into the larva, to emerge when the latter has become fully grown, thus destroying its life. Another of these enemies is a minute hymenopterous insect, whose egg is placed in the very young larva, probably in the first stage of its life. The grub, in this case, eats its way out of the half-grown larva, spins a silken cocoon, from which in a few days issues the newly-matured parasite. The destruction of larvæ by these, and very likely by other similar parasites, is doubtless immense. But no parasite attacks, it does seem, the mature larva, for, if it did, the grub of the former would live within and destroy the chrysalis, and instead of a butterfly emerging therefrom would come forth the parasite. Multitudes of chrysalids of other species of butterflies are thus destroyed, but Pseudargiolus, at this stage, appears to enjoy a singular immunity from enemies.