docile than the typical European fusca and rufibarbis. This latter is, indeed, very far from being a gentle and tractable ant.

The slave-making tactics of our sanguinary ants are in the main very similar to those of the European form. They usually start on their raids in the morning and may return laden with booty before noon, or their expeditions may drag along for the remainder of the day or even over the following day if the colony to be pillaged is at some distance, of large size and belligerent, or contains a great number of larvæ and pupæ. Sometimes, however, the sortie is postponed till the afternoon. This was the case in the following instance which I take from a number of similar expeditions of which I have kept notes:

Rockford, Ill., July 14. At 4 P. M. I located a large colony of F. fuscata which was nesting under a piece of wood in a loose hazel thicket. On removing the wood I found a large, flat chamber, from the bottom of which a single opening 2 cm. in diameter led down into the subterranean galleries of the nest. The chamber was full of fuscata workers, winged females, larvæ and naked pupa and the whole assemblage hastily poured down the opening out of sight. Looking up I saw a scattered army of rubicunda rapidly approaching the nest. When they reached the circle of grass immediately surrounding the earth just exposed by the removal of the wood, they stopped and completely surrounded the spot. They waited or kept advancing and retreating, but never entered the hole until the rear detachment had arrived. Even after the whole army, numbering at least 400 rubicunda, had assembled, they kept up this advancing and retreating movement for fully fifteen minutes, as if fearing the fuscata, which in the meantime were hiding in their nest. Now and then a rubicunda, bolder than her sisters, would enter the hole, but dart out again immediately. After twenty minutes more of this manoeuvring, however, the slave-makers grew bolder and began to pour into the opening. For some time longer and at intervals of three to five minutes a rubicunda would emerge from the nest with a larva or pupa and start for home. As soon as one of these lucky individuals appeared, four or five of the workers on the outside of the nest would try to wrest away her booty. Sometimes one of them was successful and at once started off for her nest. Finally, at 4.35 P. M., thirty-five minutes after the nest had been surrounded, a winged fuscata female shot out of the opening, immediately followed by fully fifty others and a flood of fuscata workers carrying larvæ and pupæ in their jaws. They scattered at once in all directions, breaking through the rubicunda cordon and making for the grass beyond. The rubicunda instantly fell upon both females and workers. and tore the larvæ and pupæ from the jaws of the latter. The long

legged fuscata, however, all managed to escape unscathed and with a few of their young. The wildest excitement prevailed till all the fuscata were out of the nest, but not one of them remained on the premises ten minutes after the first winged female had emerged from the opening. The rubicunda then proceeded to pillage the nest at their leisure, bringing out the deserted larvæ and pupæ and making for home. I followed them as they hurried off over a very tortuous path under the hazel bushes to their formicary, which was covered by a pile of twigs and dead leaves, some 40 meters from the nest they had pillaged. Loitering about the rubicunda nest were a number of slaves, large subsericea and an occasional small fuscata. These seemed to show great interest in the larvæ and pupa with which the rubicunda were constantly arriving. I returned to the fuscata nest. It was now 5.25 P. M. and the last straggling rubicunda were just starting home with the last of the pupæ. In the meantime the fuscata had established themselves under a bunch of dead leaves around the roots of a hazel bush about two meters from their old quarters. They had transported thither the rescued larvæ and pupa and were very busy carrying in the workers and females that had strayed about in the grass. This was done with marvellous dispatch and precision. The whole raid had been accomplished in an hour and a half, without the death or injury of a single ant, showing that the rubicunda, like her European congener, accomplishes her purpose by surprising and terrorizing rather than by killing the colonies on which she preys. The unharmed fuscata could at once set to work to raise another large brood to be pillaged in turn at a later date, and this is as it should be-from the rubicunda point of view.

F. fuscata, like the other members of the pallide-fulva group, is even more cowardly than subsericea, so that the raid above described is not typical in all respects. Large subsericea or neoclara colonies offer a much more hostile resistance to the invading slave-makers, and the battle may continue for hours or even days before the latter succeed in pillaging the nest. At such times the sanguinea will not hesitate to use her mandibles and the ground may be strewn with the corpses of both species. Colonies that have been attacked and plundered repeatedly season after season seem to submit to the affliction more passively than those attacked for the first time. Owing to the great differences in the size and condition of the colonies of both the slaves and the slave-makers, the forays of the latter present an enormous range of variability, and it would be desirable to record many more observations on them, both in Europe and North America.

Like the typical sanguinea, our American forms may also pillage the nests of ants belonging to strange genera. I once witnessed a

ridiculous foray of a large rubicunda colony on a woodland variety of Myrmica scabrinodis near Rockford, Ill. The foray was carried out exactly as if it had been directed against one of the normal auxiliary species. After killing or putting to flight the scabrinodis, the rubicunda returned to their nest with the small larvæ and pupæ of an ant belonging to an entirely different subfamily. In another rubicunda nest in the same wood, I found two of the flat chambers full of uninjured pupæ of scabrinodis. These had evidently been set apart from the rubicunda young and from those of the normal auxiliaries (in this case F. subanescens). Forel (1874) made a similar observation on a sanguinea nest in which Lasius niger and L. flavus cocoons had been stacked up in a chamber by themselves. Near Rockford, Ill., a large number of subintegra workers were seen one morning to make a normal assault on a Lasius americanus colony and to return with a number of cocoons in their jaws and many Lasius workers hanging to their legs and antennæ. These forays, which are probably not at all infrequent and are, moreover, undoubtedly undertaken by colonies of considerable size and of some experience in capturing the normal auxiliaries, point to hunger as one of the impulses which compels them to undertake their expeditions. We can hardly suppose that sanguinea workers, even after some practice in making slaves, have any definite ideal association between the kidnapped pupa and the slaves that hatch from them or they would not make forays on such unsuitable species.

3. The Founding of the Sanguinea Colony.-How do the mixed colonies of the facultative slave-makers arise? As no one had been able to observe the behavior of the sanguinea queen just after descending from her nuptial flight and while establishing her colony, Forel and Wasmann supposed that she must either be adopted by some colony of the slave species or bring up unaided a brood of her own which could then by dulosis make the mixed colony. During the summer of 1905 I performed a number of experiments on young, artificially deälated queens, introducing them into nests containing several subsericea workers with their brood. I here transcribe the account of one of these experiments from my paper "On the Founding of Colonies. by Queen Ants" (1906c):

July 8, 9 A. M. A rubicunda female was placed in a nest containing 33 subsericea workers, small and large, 150 cocoons, and a few larvæ. The workers at once seized their cocoons and fled into the light chamber. One or two of them attacked the female, but she shook them off and killed one of them. In the meantime some of the workers kept stealing into the dark chamber for the purpose of securing cocoons and carried them to the remotest corner of the light chamber. As the

morning wore away the female gradually became more and more excited. By I P. M. she had killed five more workers and was busy carrying the cocoons back from the illuminated into the dark chamber, where she had already stored most of them in a corner. In a few minutes she had secured all the cocoons in the light chamber, 36 in number. She interrupted this task twice, each time for the purpose of killing a worker that came within her reach. Finally she retired to the dark chamber and began to collect the cocoons into a more compact pile. Two of the workers persisted in stealing in and hurrying back with a cocoon taken from the edge of the pile. The female soon perceived this, however, and dispatched both of them. The whole performance resembled a dulotic expedition in miniature, carried out by a single virgin queen instead of by an army of rubicunda workers. In killing the subsericea workers she was quite as ruthless as the workers of her own species, but more sure on account of her larger size and greater strength. She exhibited very beautifully what may be called the "prancing" movement, so characteristic of the females in this stage of their activities. She moved in a jerky fashion, taking a few steps in one direction, then turning her body and taking a few steps more. July 9, 8 A. M., only two of the workers survived. They had regained possession of 30 of their cocoons, however, and were guarding them in a remote corner of the light chamber, while the female was watching over the great bulk of the brood in a corner of the dark chamber. By 10.30 she had entered the light chamber, recaptured all but 6 of the cocoons, carried them into the dark chamber and placed them on her pile. The two workers were wandering about in a state of "abulic dejection." At 11.30 one of them was seen to enter the dark chamber and approach the female, but the latter opened her mandibles and the worker fled. The female had stacked her cocoons in a compact heap and was bent on defending them. Apparently she had not forgotten the 6 cocoons still remaining in the light chamber. At any rate, she secured 4 of them by 12 M. She took up her position on the pile of cocoons, and whenever light was admitted into the dark chamber, opened her mandibles and went to prancing about as if looking for an enemy. By 1.15 P. M. she had secured one of the remaining cocoons in the light chamber. July 10, 6 A. M. In the night the female had killed the two remaining workers and had taken their last cocoon. Throughout the day she kept closely to the brood, prancing whenever the light was admitted into the chamber and fiercely seizing a straw or my finger whenever either was held near her. She seemed to display a much greater interest in the pupæ than in the larvæ. July 11 to 15 she remained in statu quo. Whenever the nest was uncovered she

July 16, 7 A. M., five
One larva had been

hastily took up a cocoon and tried to conceal it. callow workers had hatched during the night. partially eaten by the female. At 1.40 she was surprised in the act of opening a cocoon. She used her fore and middle feet to hold the cocoon while she tore a large, elliptical hole with her mandibles in the portion of its wall overlying the concave ventral surface of the pupa. Through this hole the worker was later drawn after it had thrust out its antennæ and legs. Whenever the nest was uncovered throughout this and the following of the first days, the female could nearly always be detected in the act of either opening a cocoon or removing the pupal envelope from a callow just released. By the afternoon of July 16 some of the callows began to assist the female in releasing their sister workers so that the number of callows now began to increase rapidly. On the morning of July 17 there were 19 altogether, by 5 P. M. 24, by 7.30 A. M., July 18, 30, and by 7.30 A. M., July 19, 50. On the following days the numbers ran thus: July 20, about 60; July 21, about 75; July 22, about 100; July 23 and 24, about 130. This completed the callow brood, as some of the cocoons failed to hatch. The female took the greatest interest in her black family, and they bestowed on her every attention. Soon after they had begun to feed and clean her another marked change supervened in her instincts. Instead of defending herself and brood when the nest was uncovered she slunk away, or at any rate attempted to conceal herself among the mass of workers. She had become highly photophobic and behaved exactly like the old queens, that invariably make for the galleries whenever the nest is disturbed or illuminated. This experiment was concluded and the ants were liberated in the garden on July 26.

The above experiment shows very clearly that the female rubicunda, when placed with a small number of subsericea workers and their pupæ, displays a chain of instincts that result in her gaining possession of the latter. To all appearances she is quite ready to be amicably adopted by the subsericea, but when received with marked hostility, as is probably almost invariably the case, her animosity is very quickly kindled, and she slays the subsericea with all possible dispatch, thus manifesting instincts very similar to those of her own workers when engaged in a dulotic raid. Owing to her powerful mandibles and closely knit frame she is always a match for several workers and may kill as many as twenty-one of these in a very short time. Before she has killed them all, however, she becomes much interested in their brood, eagerly collects and secretes it in some favorable corner and guards it with open mandibles till the callows are ready to hatch. These she skilfully divests of their cocoons and pupal envelopes. Their advent in consid