Association is found in the animal world at all degrees of evolution; and, according to the grand idea of Herbert Spencer, so brilliantly developed in Perrier's Colonies Animales, colonies are at the very origin of evolution in the animal kingdom. But, in proportion as we ascend the scale of evolution, we see association growing more and more conscious. It loses its purely physical character, it ceases to be simply instinctive, it becomes reasoned. With the higher vertebrates it is periodical, or is resorted to for the satisfaction of a given want--propagation of the species, migration, hunting, or mutual defence. It even becomes occasional, when birds associate against a robber, or mammals combine, under the pressure of exceptional circumstances, to emigrate. In this last case, it becomes a voluntary deviation from habitual moods of life. The combination sometimes appears in two or more degrees--the family first, then the group, and finally the association of groups, habitually scattered, but uniting in case of need, as we saw it with the bisons and other ruminants. It also takes higher forms, guaranteeing more independence to the individual without depriving it of the benefits of social life. With most rodents the individual has its own dwelling, which it can retire to when it prefers being left alone; but the dwellings are laid out in villages and cities, so as to guarantee to all inhabitants the benefits and joys of social life. And finally, in several species, such as rats, marmots, hares, etc., sociable life is maintained notwithstanding the quarrelsome or otherwise egotistic inclinations of the isolated individual. Thus it is not imposed, as is the case with ants and bees, by the very physiological structure of the individuals; it is cultivated for the benefits of mutual aid, or for the sake of its pleasures. And this, of course, appears with all possible gradations and with the greatest variety of individual and specific characters--the very variety of aspects taken by social life being a consequence, and for us a further proof, of its generality.(24)