'Malgré l'exactitude avec laquelle il décrit ce fait, j'avais peine à y croire avant de l'avoir vu moi-même, mais une fourmilière pratensis m'en donna l'exemple à plusieurs reprises lorsque je l'approchai avec précaution. Des (ie. workers) se saisissaient par les pattes ou par les mandibules, se roulaient par terre, puis se retachaient, s'entraînaient les unes les autres dans les trous de leur dôme pour en ressortir aussitôt après, etc. Tout cela sans aucun acharnement, sans venin; il était évident que c'était purement amical. Le moindre souffle de ma part mettait aussitôt fin à ces jeux. J'avoue que ce fait peut paraître imaginaire à qui ne l'a pas vu, quand on pense que l'attrait des sexes ne peut en être cause.'

Bates, also, in the case of Eciton legionis, observed behaviour which looked to him like simple indulgence in idle amusement, the conclusion,' he says, 'that the ants were engaged merely in play was irresistible.'1

Lastly, I may observe that ants are very cleanly animals, and assist one another in this respect. I have often seen them licking one another. Those, moreover, which I painted for facility of recognition were gradually cleaned by their friends.

1 Loc. cit., vol. ii. p. 362.

30

CHAPTER II,

ON THE FORMATION AND MAINTENANCE OF NESTS, AND ON THE DIVISION OF LABOUR.

Iris remarkable that notwithstanding the researches of so many excellent observers, and though ants' nests swarm in every field and every wood, we did not know how their nests commence.

Three principal modes have been suggested. After the marriage-flight the young queen may either— 1. Join her own or some other old nest;

2. Associate herself with a certain number of workers, and with their assistance commence a new nest; or

3. Found a new nest by herself.

The question can of course only be settled by observation, and the experiments made to determine it had hitherto been indecisive.

Blanchard, indeed, in his work on the Metamorphoses of Insects' (I quote from Dr. Duncan's translation, p. 205), says :- Huber observed a solitary female go down into a small under-ground hole, take off her own wings, and become, as it were, a worker; then she constructed a small nest, laid a few eggs, and brought

up

the larvæ by acting as mother and nurse at the same time.'

This, however, is not a correct version of what Huber says. His words are:-"I enclosed several females in a vessel full of light humid earth, with which they constructed lodges, where they resided, some singly, others in common. They laid their eggs and took great care of them; and notwithstanding the inconvenience of not being able to vary the temperature of their habitation, they reared some, which became larvæ of a tolerable size, but which soon perished from the effect of my own negligence.' 1

It will be observed that it was the eggs, not the larvæ, which, according to Huber, these isolated females reared. It is true that he attributes the early and uniform death of the larvæ to his own negligence, but the fact remains that in none of his observations did an isolated female bring her offspring to maturity.

Other entomologists, especially Forel and Ebrard, have repeated the same observations with similar results; and as yet in no single case had an isolated female been known to bring her young to maturity. Forel even thought himself justified in concluding, from his observations and from those of Ebrard, that such a fact could not occur.

Lepeletier de St. Fargeau2 was of opinion that ants' nests originate in the second mode indicated above, and

1 Natural History of Ants, Huber, p. 121.

2 Hist. Nat. des Ins. Hyménoptères, vol. i. p. 143.

it is, indeed, far from improbable that this may occur. No clear case has, however, yet been observed. M. de St. Fargeau himself observes that les particularités qui accompagnent la formation première d'une fourmilière sont encore incertaines et mériteraient d'être observées avec soin.'

Under these circumstances I made the following experiments :

la. I took an old, fertile, queen from a nest of Lasius flavus, and put her to another nest of the same species. The workers became very excited and attacked her.

b. I repeated the experiment, with the same result. c. Do. do. In this case the nest to which the queen was transferred was without a queen; still they would not receive her.

d and e. Do. do. do.

I conclude, then, that, at any rate in the case of L. flavus, the workers will not adopt an old queen from another nest.

The following observation shows that, at any rate in some cases, isolated queen ants are capable of giving origin to a new community.

On August 14, 1876, I isolated two pairs of Myrmica ruginodis which I found flying in my garden. I placed them with damp earth, food, and water, and they continued perfectly healthy through the winter.

1 Hist. Nat. des Ins. Hyménoptères, vol. i.

P. 144.

In April one of the males died, and the second in the middle of May. The first eggs were laid between April 12 and 23. They began to hatch the first week in June, and the first larva turned into a chrysalis on the 27th; a second on the 30th; a third on July 1, when there were also seven larvæ and two eggs. On the 8th there was another egg. On July 8 a fourth larva had turned into a pupa. On July 11 I found there were six eggs, and on the 14th about ten. On the 15th one of the pupa began to turn brown, and the eggs were about 15 in number. On the 16th a second pupa began to turn brown. On the 21st a fifth larva had turned into a pupa, and there were about 20 eggs.. On July 22 the first worker emerged, and a sixth larva had changed. On the 25th I observed the young worker carrying the larvæ about when I looked into the nest; a second worker was coming out. On July 28 a third worker emerged, and a fourth on Aug 5. The eggs appeared to be less numerous, and some had probably been devoured.

This experiment shows that the queens of Myrmica ruginodis have the instinct of bringing up larvæ and the power of founding communities. The workers remained about six weeks in the egg, a month in the state of larvæ, and twenty-five to twenty-seven days as pupæ.

Since, however, cases are on record in which communities are known to have existed for many years, it seems clear that fresh queens must be sometimes adopted. I have indeed recorded several experiments

D