was that if the ants fell, as often happened, they were within the water boundary, and were able to return home. This plan answered fairly well, and saved space, but it did not quite fulfil my hopes, as the ants were so pugnacious, that I was obliged to be very careful which nests were placed on the same stand.

Of course it is impossible to force the ants into these glass nests. On the other hand, when once the right way is known, it is easy to induce them to go in. When I wished to start a new nest I dug one up, and brought home the ants, earth, &c., all together. I then put them over one of my artificial nests, on one of the platforms surrounded by a moat of water. Gradually the outer earth dried up, while that between the two plates of glass, being protected from evaporation, retained its moisture. Under these circumstances the ants found it more suitable to their requirements, and gradually deserted the drier mould outside, which I removed by degrees. In the earth between the plates of glass the ants tunnelled out passages, chambers, &c. (fig. 2, p. 43), varying in form according to the circumstances and species.

Even between the plates of glass the earth gradually dried up, and I had to supply artificial rain from time to time. Occasionally also I gave them an altogether new nest. They seem, however, to get attached to their old homes, and I have still (August, 1882) one community which has inhabited the same glass case ever since 1874.

It is hardly necessary to say that the individual

ants belonging to the communities placed on the stands just described, knew their own nests perfectly well.

These nests gave me special facilities for observing the internal economy of ant life. Another main difference between my observations and those of previous naturalists has consisted in the careful record of the actions of individual ants. The most convenient mode of marking them was, I found, either with a small dab of paint on the back, or, in the case of bees or wasps, by snipping off a fragment at the extremity of the wing. This, I need hardly say, from the structure of the wing, gives the insect no pain; in fact, as it is only necessary to remove a minute portion, not sufficient to make any difference in their flight, they seemed scarcely to notice it. I never found any difficulty in painting bees or wasps; if they are given a little honey they become so intent that they quietly allow the paint to be applied. Of course too much must not be put on, and care must be taken not to touch the wings or cover up the spiracles. Ants require somewhat more delicate treatment, but with a little practice they could also be marked without any real difficulty.

No two species of Ants are identical in habits; and, on various accounts, their mode of life is far from easy to unravel. In the first place, most of their time is passed underground: all the education of the young, for instance, is carried on in the dark. Again, ants are essentially gregarious; it is in some cases difficult to

keep a few alive by themselves in captivity, and at any rate their habits under such circumstances are entirely altered. If, on the other hand, a whole community is kept, then the greater number introduces a fresh element of difficulty and complexity. Moreover, even individuals of the same species seem to differ in character, and the same individual will behave very differently under different circumstances. Although, then, ants have attracted the attention of many of the older naturalists,-Gould, De Geer, Reaumur, Swammerdam, Latreille, Leuwenhoeck, Huber,--and have recently been the object of interesting observations by Frederick Smith, Belt, Moggridge, Bates, Mayr, Emery, Forel, McCook, and others, they still present one of the most promising fields for observation and experiment.

The life of an ant falls into four well-marked periods those of the egg, of the larva or grub, of the pupa or chrysalis, and of the perfect insect or imago. The eggs are white or yellowish, and somewhat elongated. They are hatched about fifteen days after being laid. Those observed by me have taken a month or six weeks.

The larvæ of ants (Pl. V. fig. 3), like those of bees and wasps, are small, white, legless grubs, somewhat conical in form, narrowing towards the head. They are carefully tended and fed, being carried about from chamber to chamber by the workers, probably in order to secure the most suitable amount of warmth

and moisture.

I have observed, also, that they are very often assorted according to age. It is sometimes very curious in my nests to see them arranged in groups according to size, so that they remind one of a school divided into five or six classes.

As regards the length of life of the larvae, Forel supposed1 that those of Tapinoma matured the quickest, and were full-grown in about six or seven weeks. Some of Myrmica ruginodis, however, observed by me, turned into pupæ in less than a month. In other cases the period is much longer. In certain species, Lasius flavus, for instance, some of the larvæ live through the winter.

When full grown the larvæ turn into pupa (Pl. V. fig. 4), sometimes naked, sometimes covered with a silken cocoon, constituting the so-called 'ant-eggs.' We do not yet understand why some larvæ spin cocoons, while others remain naked. As a general rule, the species which have not a sting, are enveloped in a cocoon, while those which have, are naked. Latrielle was the first to observe that in one species (F. fusca) the larvæ sometimes spin a cocoon, and sometimes remain naked. The reason for this difference is still quite unknown. After remaining some days in this state they emerge as perfect insects. In many cases, however, they would perish in the attempt, if they were not assisted; and it is very pretty to see the older ants helping them to extricate them

1 Les Fourmis de la Suisse, p. 420.

selves, carefully unfolding their legs and smoothing out the wings, with truly feminine tenderness and delicacy. Our countryman Gould long ago mentioned, and the fact has since been fully confirmed by Forel, that the pupæ are unable to emerge from the cocoons without the assistance of the workers. The ants generally remain from three to four weeks in this condition.

In the case of ants, as with other insects which pass through similar metamorphoses, such as bees, wasps, moths, butterflies, flies, and beetles, &c., the larval stage is the period of growth. During the chrysalis stage, though immense changes take place, and the organs of the perfect insect are more or less rapidly developed, no food is taken, and there is no addition to the size or weight.

The imago or perfect insect again takes food, but does not grow. The ant, like all the insects above named, is as large when it emerges from the pupa as it ever will be, though the abdomen of the females sometimes increases in size from the development of the eggs.

We have hitherto had very little information as to the length of life in ants in the imago, or perfect, state. So far, indeed, as the preparatory stages are concerned, there is little difficulty in approximately ascertaining the facts; namely, that while in summer they take only a few weeks, in some species, as our small yellow meadow ants (Lasius flavus), the autumn larvæ