not only for the mating of the sexes but also for the dissemination of the species, since the daughter queens, on descending to the ground, usually establish their nests at some distance from the parental colony.

It will be seen that the queen ant, like the queen wasp and bumble-bee, but unlike the queen honeybee, is the perfect female of her species, possessing not only great fecundity but in addition all the worker propensities, as shown by her ability to make a nest and bring up her young. But as soon as the first brood of workers appears, these propensities are no longer manifested. That they are not lost is shown by the simple experiment of removing the queen's first brood of workers. Then, provided she be fed or have a sufficient store of food in her body, she will at once proceed to bring up another brood in the same manner as the first, although she would have manifested no such behavior under normal conditions.

As already stated, this independent method of colony formation is the most universal and is followed alike by tropical and extratropical ants. It is undoubtedly the primitive method and, as we shall see, the one from which the dependent method has been derived. It differs from that of Vespa and Bombus, nevertheless, in leading to the formation of perennial colonies even in temperate and boreal regions. The queen ant may, in fact, live from 12 to 17 years and although, like other aculeates, she is fecundated only once, may produce offspring up to the time of her death. Unlike the queen honeybee she is never hostile to her own queen daughters, and in many species of ants some of these daughters may return after their marriage flight to the maternal colony and take a very active part in increasing its population. In this manner the colony may become polygynic or pleometrotic, and in some instances may contain a large number of fertile queens. When such a colony grows too large it may separate into several, the queens emigrating singly or in small companies, each accompanied by a detachment of workers, to form a new nest near the parental formicary. This behavior is exhibited by the well-known mound-building ant (Formica exsectoides) of our New England hills. You will notice that its mounds usually occur in loose groups or clusters and that the workers of the different nests are on friendly terms with one another and sometimes visit back and forth. We may, of course, call the whole cluster a single (polydomous) colony, but it really differs from a number of colonies only in the absence of hostility between the inhabitants of the different mounds. In certain tropical ants, like the Doryline (Figs. 59 and 60), however, I am inclined to believe that the only method of colony formation is by a splitting of the original colony into as many parts as it con

FIG. 59

Argentinian legionary (Doryline) ant Eciton (Acamatus) strobeli. Workers showing polymorphism and male, photographed to the same scale as the four smaller workers. (Photograph by Dr. Carlos Bruch.)

Dorsal and lateral view of the wingless queen (dichthadiigyne) of Eciton (Acamatus) strobeli. Same scale as Fig. 59.

Bruch.)

(Photograph by Dr. Carlos

tains young queens. These huge, clumsy creatures (Fig. 60) are always wingless and must therefore be fecundated in the nest, and since the colonies, which comprise hundreds of thousands of workers, are nomadic and keep wandering from place to place, they must become independent entities as soon as they are formed.

We possess no accurate data on the age that ant colonies may attain. Some of them certainly persist for 30 or 40 years and probably even longer. In such old colonies the original queen has, of course, been replaced by successive generations of queens, that is, by her fertile daughters, grand-daughters and great-granddaughters, and the worker personnel has been replaced at a more rapid rate, because the individual worker does not live more, and in most instances lives considerably less, than three or four years.

The feeding habits of ants are so varied and complicated that it will be advisable before considering them to describe the structure of the alimentary canal in both adult and larva. The mouthparts of the adult are of the generalized vespine type and consist

Sagittal sections through the heads of ants. A, of queen Lasius niger with the mouth open (After Janet). B, of queen Camponotus brutus with the mouth closed. t, tongue; o, oral orifice; ph, pharynx; h, infrabuccal pocket: pe, pellet in situ, made up of solid particles of food refuse and strigil sweepings. Note stratification in the substance of the pellet, indicating successive meals or toilet operations.

of a small, flap-like upper lip, or labrum, a pair of strong, usually toothed mandibles, a pair of small maxillæ and a broad lower lip, or labium. The maxillæ and labium are each provided with a pair of jointed, sensory appendages, the palpi. The mandibles, which are really the ant's hands, vary greatly in shape in different genera and are used not only in securing the food but also in many other activities, such as digging in the earth or wood, transporting other ants or the young, fighting, leaping, etc. Liquids are, of course, merely imbibed and swallowed, but solid food is seized and crushed with the mandibles and the juices or smaller particles licked up with the tongue, which is a roughened pad at the tip of the lower lip (Fig. 61t) just anterior to the opening of the duct of the salivary glands. The small particles thus collected are carried back into a small chamber or sac, the infrabuccal pocket (Fig. 61h), which lies immediately below and anteriorly to the mouth-opening (0). This pocket is an important structure since it serves as a receptacle not only for the more solid particles of food but also. for the dirt, fungus-spores, etc., which the ant collects during her toilet operations, for the ant is an exquisitely cleanly insect and devotes much of her leisure to licking and burnishing her own smooth or finely chiseled armor and that of her nest-mates. Moreover, the tip of the fore tibia is furnished with a beautiful comb or strigil which can be opposed to another comb on the concave inner surface of the fore metatarsus. The ant cleans her legs and antennæ by drawing them between these combs, which are then drawn across the mouth, with the result that any adhering dirt is carried off into the infrabuccal pocket. In this manner the dirt and the solid or semisolid food particles are combined and the whole mass moulded in the infrabuccal pocket into the form of a roundish oblong pellet (Fig. 61B pe). After any liquid which it may contain has been dissolved out and sucked back into the mouth, the pellet is cast out, so that no solid food actually enters the alimentary canal. All adult ants therefore subsist entirely on liquids.

The alimentary canal proper is a long tube extending through the body and divided into sections, each with its special function. The more anterior sections are the mouth cavity, the pharynx (Fig. 61 ph), which receives the ducts of certain glands, and the very long, slender gullet, which traverses the posterior part of the head, the whole thorax and the narrow waist, or pedicel of the abdomen as far as the base of its large, swollen portion, the gaster. Here the gullet expands into a thin-walled, distensible sac, the crop, which is used for the storage of the imbibed liquids. At its posterior end the crop is separated from the ellipsoidal stomach by a peculiar valvular constriction, the proventriculus. The hindermost sections of the alimentary tract are the intestine and

the large, pear-shaped rectum. The crop, proventriculus and stomach are the most interesting of these various organs. Forel calls the crop the "social stomach," because its liquid contents are in great part distributed by regurgitation to the other members of the colony and because only a small portion, which is permitted to pass back through the proventricular valve and enter the stomach, is absorbed and utilized by the individual ant. That the crop functions in the manner described can be readily demonstrated by permitting some pale yellow worker ant to gorge herself with syrup stained blue or red with an aniline dye. The ant's gaster will gradually become vividly colored as the crop expands. Now if the insect be allowed to return to the nest, other workers will come up to it, beg for food with rapidly vibrating antennæ and protrude their tongues, and very soon their crops, too, will become visible through the translucent gastric integument as they fill with the stained syrup. Then these workers in turn will distribute the food by regurgitation in the same manner till every member of the colony has at least a minute share of the blue or red cropful of the first worker.

The alimentary tract of the helpless, legless, soft-bodied ant grub or larva is much simpler than that of the adult. The mouthparts are similar but more rudimentary. As a rule, the mandibles are less developed but in some larvæ they are strong, dentate and very sharp. The lower lip is fleshy and protrusible and provided with sensory papillæ instead of palpi, and the unpaired duct of the long, tubular and more or less branched salivary glands opens near its tip. The mouth-opening is broad and its lining in many species is provided with numerous transverse ridges beset with very minutes spinules (Fig. 62C). Larger, pointed projections

or imbrications may also cover the basal portions of the mandibles. All these spinules and projections are probably used in triturating the food but perhaps when rubbed on one another they may also produce shrill sounds for the purpose of apprising the worker nurses of the hunger or discomfort of their charges. The gullet is long and very slender and opens directly into the large stomach, which throughout larval life is closed behind, that is, does not open into the intestine. A communication with the more posterior portion of the alimentary tract is not established till the larva is about. to pupate. Then all the undigested food which has accumulated in the stomach since the very beginning of larval life is voided as a large black pellet, the meconium.

In the larvæ of the Pseudomyrminæ (Figs. 62, 63 and 64) there are certain very peculiar additional structures which may be briefly described. The head is not at the anterior end of the body as in other ant larvæ but pushed far back on the ventral surface so that it is surrounded by a great hood formed from the three thoracic