right to the habit of harvesting, for it is clear that a given tract of country can only afford supplies of grain to a limited number of colonies; so that, if these ants have taken up the ground and are strong enough to maintain possession, no others would have a chance. However this may be, I find that the more insight I gain into the distinctive habits and relations of animals, the more the belief impresses itself upon me that wherever we find many closely-allied species inhabiting restricted areas, there we may safely look for important differences among these species in respect of their modes of life, and in the development of their instinct and intelligence. And indeed this may be considered as a corollary of the great law of natural selection, which uniformly tends to secure the greatest possible amount of divergence in this respect, and to prevent the co-existence in close proximity to each other of distinct species having the same requirements and manner of life.
Thus, for example, even Atta barbara and structor, though most closely related as species, differ in habit; the former leading a much more active life during the winter months at Mentone than the latter, and seeking its home rather in wild than cultivated ground. Then what differences different ants present in respect of strength, speed, powers of offence and defence, numerical strength of colonies, timidity, date and frequency of departure of winged ants from the nest, odour emitted, combativeness, architecture and selection of localities, nature of food, nocturnal and diurnal habits, and in many other properties and conditions! It is doubtless owing to dissimilarity in these and other respects that it becomes possible for
