The length of time during which sleep is prolonged appears to vary according to circumstances and, perhaps, organism. The large head-soldiers of the Floridian harvesters appear to have a more sluggish nature than the smaller workers. Their sleep is longer and heavier. The former fact the watch readily determined. The latter appeared from the greater stolidity of the creatures under disturbance. While the ants of one group are taking sleep others may be busy at work, and these stalk among and over the sleepers, jostling them quite vigorously at times. Again, new members occasionally join the group, and, in their desire to get close up to the heat and light, crowd their drowsy comrades aside. I have seen ants who had been at work in the galleries drop their pellets, push thus into the cluster, and presently be apparently sound asleep. This rough treatment is invariably received with perfect good humour, as are like jostlings when the ants are awake. I have never seen the slightest display of anger or attempt to resent disturbance even under these circumstances, so peculiarly calculated to excite the utmost irritation in men. But of course some of the sleepers are aroused. They change position a little, or give themselves a brief combing, and then resume their nap, unless, indeed, they are satisfied. In watching these movements it was quite evident that the Florida soldiers were far less easily disturbed than their smaller fellows. They slept on stolidly while all the others were in agitation around them. Moreover, their very appearance, particularly when awaking out of sleep, indicated the greater sluggishness of their temperament in this respect. The ordinary duration of sleep MacCook takes to be about three hours. Ants, like many other insects, are in the habit of cleaning themselves, being, like them, provided by nature with combs and brushes, &c., for the purpose. But, unlike other insects, several species of ants are also in the habit of assisting each other in the performance of their toilet. The author last quoted gives the following account of this process in the genus _Atta_:--