circumstances they remained in perfect health, while, but for the slaves, they would have perished in two or three days.”

In this matter, we have in different species successive stages in the development of the instinctive behaviour which is thus carried so far in Formica rufescens. Our English ants, of the species Formica sanguinea, have fewer slaves and are less dependent on them; they can feed and forage for themselves, and during migration carry their slaves-which are of the same species as in the other case-instead of being carried by them. In the nests of the common wood ant or horse ant (Formica rufa) there are occasionally a few slaves. Lord Avcbury thinks it likely that they are developed from larvæ or pupa, originally taken for food, which have by chance come to maturity in the nest of their captors.

But one more incident in the social life of ants can here be noticed-though many others could be given did spaco permit. The leaf-cutting ants of America form paths from their nests to suitable trees, from which to obtain the small coin-like leaf fragments, which they carry in the mandibles, and hence have gained the name of umbrella or parasol ants. These paths are sometimes underground; and Mr. McCook measured one which ran at a depth of some 18 inches beneath the surface for 448 feet, and was then continued for another 185 feet to the tree which the ants were stripping. The whole path was in an almost perfect straight line from nest to tree. The leaf fragments are stored in large quantities in the nest, and it was long a matter of uncertainty for what purpose they were collected. The problem was solved by Alfred Möller, who found that the leaves, which are subdivided and masticated by a special set of workers within the nest, form the appropriate material in which the threads of a fungus ramify and flourish. This fungus is tended by the ants with great care, and is made to produce a specially modified form of growth, not found under other circumstances, in the form of white aggregations, termed by Möller "Kohlrabi clumps." These form the principal food of the ants; and the spongy mass of earth and * Lord Avebury (Sir John Lubbock), "Scientific Lectures," pp. 78, 79.

leaves is called the fungus garden. "If a nest be broken into and the fungus garden scattered, the ants collect it as quickly as possible, especially the younger parts, taking as much trouble over it as over the larva. They also cover it up again as soon as possible to protect it from the light."*

Again, it may be asked with regard to the social life of ants as with respect to that of bees-How far is their complex behaviour instinctive? How far is it due to imitation? What part does intelligence play, and under what conditions of acquisition? Is reason, in the restricted sense of the word, a factor in the development of the behaviour? I cannot answer these questions, and am of opinion that much detailed observation is yet needed before we can do much more than speculate in the matter. Much indeed has been done, but yet more remains for future investigation.

The conditions under which much of the behaviour is carried out seem to indicate strong instinctive tendencies which give an hereditary trend to the direction which the social behaviour takes. Dr. Bethe,† indeed, goes so far as to regard the behaviour as almost entirely instinctive, affording little evidence of that modifiability of reactions which indicates intelligent guidance. He shows as the result of careful experiment that the behaviour of ants to friends and enemies. are direct reactions to smell. Enemies washed with the excretions of members of the nest are treated as friends, notwithstanding their different colour, size, and general appearance. By scent, too, they follow the lead of others and retrace their way to the nest; this, he says, is not the result of a mental process, but is the reaction of a complicated reflex mechanism. As the outcome of careful observation, Dr. Bethe's conclusions are of great value and interest. But he seems to go too far in denying to ants any power of intelligent accommodation to circumstances. If we admit intelligence, then the fact that

Nature, vol. xlvii., p. 393 (Aug., 1893), where A. Möller's investigations are described by J. C. Willis.

†A. Bethe, "Dürfen wir den Ameisen und Bienen psychische Qualitäten Zuschreiben," Pflüger's Archiv., lxx., 1898.

the insects come forth in the midst of a community in full social activity would tend to the imitative or intelligent acquisition of like modes of procedure. It is difficult to distinguish the share taken by these two factors which may well cooperate. And if natural selection is exercising its influence through the elimination of those which do not fall into line in social behaviour, there would be ample opportunity for the survival of coincident variations. If one may be allowed to speculate, it seems probable that the interaction of instinct and intelligence will be found with fuller knowledge to suffice. for the explanation of the facts, without calling in the known but here improbable factor of rationality or any factors unknown elsewhere in psychology.

Some interesting observations of Lord Avebury's are sometimes quoted as evidence that ants are lacking in intelligence, but (if we accept the distinction already drawn †) they seem rather to show the lack of reason. "I placed food," he says, ‡ "in a porcelain cup, on a slip of glass surrounded by water, but accessible to the ants by a bridge, consisting of a strip of paper two-thirds of an inch long and one-third wide. Having then put an ant (Formica nigra) from one of my nests to this food, she began carrying it off, and by degrees a number of friends came to help her. When about twenty-five ants were so engaged, I moved the little paper bridge slightly, so as to leave a chasm just so wide that the ants could not reach across. They came to the edge and tried hard to get over, but it did not occur to them to push the paper bridge, though the distance was only about one-third of an inch, and they might easily have done so. After trying for about a quarter of an hour they gave up the attempt and returned home. - This I repeated several times. Then, thinking that paper was a substance to which they were not accustomed, I tried the same with a bit of straw one inch long and one-eighth of an inch wide. The result was the same. Again, I placed particles of food close to and directly over the nest, but * See p. 37. † Supra, p. 138. "Scientific Lectures," pp. 80-82.

connected with it only by a passage several feet in length. Under these circumstances it would be obviously a saving of time and labour to drop the food on to the nest, or at any rate to spring down with it, so as to save one journey. But though I have frequently tried the experiment, my ants never adopted either of these courses. I arranged matters so that the glass on which the food was placed was only raised one-third of an inch above the nest. The ants tried to reach down, and the distance was so small that occasionally, if another ant passed underneath just as one was reaching down, the upper one could step on to its back, and so descend; but this only happened accidentally, and they did not think of throwing the particles down, nor, which surprised me very much, would they jump down themselves. I then placed a heap of mould close to the glass, but just so far that they could not reach It would have been quite easy for any ant, by moving a particle of earth for a quarter of an inch, to have made a bridge by which the food might have been reached, but this simple expedient did not occur to them."

across.

Now, when we remember that the method of intelligence is to profit by chance experience, while the method of reason is, with foresight and intention, to adapt means to ends, we shall see that to move a straw even a quarter of an inch, or to make a bridge with particles of mould, would require rational and not merely intelligent powers. Chance experience would not supply the necessary data to be utilized by intelligence when repetition had established an association in the conscious situation. Granting that the ants were intelligent but not rational, they could not be expected to overcome the difficulties, simple as they seem to us, which Lord Avebury placed in their path. Had they been overcome the fact would be more difficult to explain than the use of a stone tool by the sand wasp, since this could more readily be hit upon by chance experience. And what these valuable experiments, of which kind more are needed, seem to show is, that the ant, probably the most intelligent of all insects, has no claim to be regarded as a rational being.

IV. ANIMAL TRADITION

In that interaction between instinct and intelligence which, when further detailed work has sifted and purified our knowledge of the psychology of animal communities, may prove sufficient to account for the well-established facts, animal tradition will probably have to be recognized as of no little importance. When a newly emerging ant or bee, or a young bird or mammal is born into a community where certain modes of behaviour are already in full swing, an imitative tendency of the follow-my-leader type may lead it to fall in line with the traditional habits. It is said that young ants follow the older workers about the nest, and are "trained to a knowledge of domestic duties, especially in the case of larva." On the other hand, we have seen that, in certain observed cases, the queen ant is the solitary starting-point of a new community, and that the division of labour follows with the increasing numbers of the newly formed social group; so that, in such cases, whatever part tradition may play in the later phases of + social life, it cannot afford a sufficient account of the division of labour in the earlier history of the community. We need, however, fuller information concerning the continued lifehistory of such communities under natural conditions, and as to how far they remain self-contained without any incorporation of older members from adjoining nests. In the case of bees, where the old queen departs with a swarm, there may be greater continuity of tradition. But how far this is a necessary factor in social development is at present a matter of conjecture. In the herd of mammals and the flock of birds, and in all the family and social life in these classes of animals, the example of elders, without any imitation of the higher reflective type, can scarcely be without its influence on the behaviour of the young which, one would suppose, would tend to fall in with the ways which had become traditional in the species. Professor Wesley Mills tells us that a mongrel pup, whose psychical development he carefully watched, showed "extraordinarily