ceous animals, for in both cases the victimized species exists at the present time only because it has great reproductive powers or a margin of redundant vitality which can be exploited by its enemies and parasites; and the survival of these enemies and parasites themselves in turn depends on their refraining from overstepping this margin. In the case of sanguinea the enormous reproductive powers of the species must more than compensate for the destruction of colonies by the Lomechusa.

Ectoparasites. With the Lomechusini we may close our account of the true guests, although these include also several other interesting forms, like Lomechon among the Silphidæ, Amorphocephalus among the Brenthidæ and Pogonoxenus among the Tenebrionidæ. Turning to the parasites proper, we find it impossible to draw a hard and fast line. between symphiles and ectoparasites, owing to the existence of such intermediate forms as Thorictus, Antennophorus and Orasema. The group of ectoparasites as a whole is heterogeneous assemblage of mites, Hymenoptera, Diptera and Coleoptera. The Coleoptera, however, are represented only by certain species of Thorictus.

FIG. 248.

a

Some of the ectoparasites, of which the Phorid flies of the genus Metopina and the Gamasid mites of the genus Antennophorus may be taken as interesting examples, are hardly more than commensals. In two papers (1901e, 1907a) I have described the singular habits of Metopina pachycondyle, which lives in Pachycondyla harpar colonies in Texas. Its small larva clings to the necks of the ant-larva by means of a sucker-like posterior end and encircles its host like a collar (Fig. 243). Whenever the ant-larva is fed by the workers with pieces of insect placed on its trough-like ventral surface, within reach of its mouthparts, the larval Metopina uncoils its body and partakes of the feast; and when the ant-larva spins its cocoon it also encloses the Metopina larva within the silken web. The commensal, however, moves to the caudal end of its host and forms a small, flattened puparium which is applied to the wall of the cocoon. This is obviously an adaptation for preventing injury from the jaws of the worker ants when the cocoon is being opened and the callow extracted from its anterior end. The ant hatches before the Metopina and the empty cocoon with the puparium concealed in its posterior pole is carried to the refuse heap. There the fly emerges and escapes from the cocoon by the opening through

which its host emerged. The Metopina larva consumes so little food and is so considerate of its host, that it can hardly be said to produce any injurious effect on the colony; at any rate, the larvæ which have borne commensals develop into perfectly normal workers. The ants

B

D

FIG. 249.

Chalcidid ant parasites. (Original.) A, Isomeralia coronata, female; B, lateral view of same; C, Kapala floridana, male; D, female of same.

clean the commensals while they are cleaning their own progeny and show no signs of even being aware of their presence in the nest.

Another case of commensalism is that of the European Antennophorus species (A. uhlmanni, pubescens, foreli and grandis), studied by Janet (1897b), Wasmann (1902j) and Karawaiew (1905c, 1906a). These mites occur only in the nests of Lasius and cling to the workers

by means of their three posterior pairs of legs while the large fore pair is stretched out and moved about like antennæ. Janet found that these creatures, whether present in odd or even numbers, are always oriented in a symmetrical position with respect to their host (Fig. 244). When only one Antennophorus is present, it clings to the gula, or ventral surface of the ant's head, with its fore legs directed towards the ant's mouthparts. When two are present, there is one on each side of the

B

A

FIG. 250.

Orasema viridis. (Original.) A, Female; B, male.

head or one on each side of the gaster; in the former case the antenniform appendages are directed towards the anterior, in the latter towards the posterior end of the ant's body. When there are three mites, one attaches itself to the gula and the two others to the sides of the gaster. Four place themselves in pairs on the sides of the head and gaster. If

six are present, which rarely happens, four are arranged in pairs on the sides of the head and gaster, while of the two remaining individuals, one attaches itself to the gula, the other to the mid-dorsal surface of the gaster. Janet believes that these symmetrical arrangements are for the purpose of balancing the burden and thus making it easier for the ants to carry. When attached to the head the mite obtains its food by drinking from the regurgitated droplet as it is being passed to or from

FIG. 251. Development of Orasema viridis. (Original.) A. First larval stage of Orasema. B, pupal worker of Pheidole instabilis with Orasema larva (o) attached to side of neck; C, female Pheidole pupa with somewhat older Orasema larva (o) attached in sternal region; D, female Pheidole pupa with Orasema larva (o) in same stage as in preceding figure, attached behind head; E, female Pheidole pupa (phthisogyne) with older Orasema larva (o) in sternal region; F, Orasema larva (0) beginning to pupate, with vesiculate knobs on its surface; G, Orasema pupa fallen from its host and developing within the vesiculate skin; H, fully formed pupa; I, pupa pigmented and ready to hatch.

the mouthparts of the host, or it titillates the ant with its antenni form legs and induces her to regurgitate for its special benefit. The mites attached to the gaster obtain their food by stroking other ants in the vicinity or by reaching out and partaking of the droplets as they pass

from one ant to another. The ants try to rid themselves of the parasites when these first attach themselves, but after they have taken up their definitive, symmetrical positions, they seem to be tolerated with indifference. There is nothing to indicate that the ants, while cleaning one another, are even aware of the existence of the parasites. The relations of the allied American ectoparasitic mites to their host ants have not yet been studied. Echinomegistus wheeleri (Fig. 245, A and B), which occurs on Lasius aphidicola, is probably very similar to Antennophorus in its habits.1

The number of mites living in the nests or on the surfaces of ants seems to be very great. Berlese, in a recent work (1904) has described more than sixty species of myrmecophilous Gamasidæ alone. The habits of a few of these have been studied by Janet and Wasmann and may be very briefly described. Cillibano [Discopoma] comata is a peculiar tortoise-shaped mite which attaches itself to the workers of Lasius mixtus and its larvæ (Fig. 246). On the ants it always assumes a definite position. According to Janet, when only one is present, it places itself on the side of the second gastric segment. If there are two, they place themselves symmetrically, one on each side. If there is a third, it is attached in the mid-dorsal line of the same segment. Rarely as many as six may be present; in which case there are three also on the third gastric segment in positions corresponding to those on the second. The ants dislike the Cillibano and tear them to pieces whenever they can seize them. The mites, however, usually slip out of their mandibles or apply

038

FIG. 252. A, Normal pupa of Pheidole instabilis worker; B, and C, phthisergates, produced by parasitism of Orasema viridis larvæ on the larvæ of the same ant. (Original.)

the edges of their bodies so closely to the surfaces of the ants that they cannot be picked off. From scars (Fig. 246, B, C, n) left on the intersegmental membranes of the ant's gaster Janet infers that the Cillibano sucks the blood of its host. The types of another Cillibano (C. hirticoma), with long, flexuous dorsal hairs, were found by me in Texas on an Eciton schmitti queen (Figs. 147, c; 245, C, D). This mite attaches itself not only to the body, but also to the antennæ and legs of its host.

'I have recently found an undescribed species of Antennophorus on our North American Acanthomyops interjectus.