There are two species of these ants on Barro Colorado Island in the Panama Canal Zone. Each species has approximately 50 colonies and each colony consists of from a few hundred thousands to more than a million individuals. At the head of each colony is a single queen who lays all the eggs.
There is a new lot of larvae every 33 days—all workers or incompletely developed females. Development is restricted by the amount of food available. Since each brood consists of about 60,000 individuals, a colony theoretically might reach titanic proportions. However, it does little more than maintain its population. The death rate of soldier ants, in constant combat, is very heavy.
Once each year, at the start of the dry season in the tropics, a colony queen produces a sexual brood of about 3,000 males and six queens. The rest of the 60,000 eggs laid at this time are incapable of hatching and are fed to the new-born sexed individuals. They apparently have some of the nutritious properties of the royal jelly fed to queen bees.
This sexual brood is produced in what has been called a statory period in which the army maintains a fixed bivouac for about three weeks. During this time the new queens develop and around at least one of them a new group of workers, about half the whole, tends to congregate. A strange antagonism seems to develop between the old and new groups. Eventually the colony divides in two and each half starts moving in opposite directions. The other new queens are lost in the shuffle.
Most of the newly developed males are ‘excess baggage.’ During the winged, or mating, stage they fly into the forest where the great majority of them are eaten by birds. When the surviving ants light on a tree, on the ground or on some other object, the wings drop off. Then they apparently wander about aimlessly until they come to an army ant trail which they recognize by the odor and follow it until they come to the colony which has made it. If this happens to be a colony of their own relatives, they probably are killed by the workers. If it happens to be an entirely foreign colony, they may be accepted. This apparently is one of nature’s mechanisms for intruding new genes into a strain.
The raiding activities of a colony are carried out during the day from a central headquarters. During the daytime raiding individuals return to the colony from their forays and by dusk all have returned. At night the bivouac is changed, the whole colony moving forward along one of the trails blazed by the raiders. A new headquarters is thus established. A colony moves from six to seven hours before striking a new bivouac. Not infrequently, if no promising site is found, it moves from dusk to dawn.
This would seem like constant activity, too strenuous even for the constitution of an army ant. Actually the individual workers probably get plenty of rest. Each colony is divided into two units—the raiders and those that constitute the structural unit. The walls of the “headquarters” are made up of the bodies of the latter. These “living brick” do nothing throughout the day. They may be asleep. When the raiders return at dusk the structural unit breaks up and the members lead the migration to a new bivouac. The erstwhile raiders follow leisurely in the rear and in turn become the structural unit when a stopping place is selected.
When to rest? When to raid? There apparently is an irresistible war rhythm, like the rhythm of the tides, in the basic constitution of these ants. Some have postulated the same sort of thing, on a lesser scale, in man who goes to war every so often but camouflages the war tide with economic or political explanations.
These ants are remarkable not only as warriors but as architects. They build complex, air-conditioned, hanging houses out of thousands of their own suspended bodies. Within these structures the queen is sheltered, eggs laid, young hatched and reared. Much of the time the “houses” are constructed anew each night.
This home-building behavior is unique in nature, as Dr. T. C. Schneirla of the American Museum of Natural History has pointed out: