In the _De Partibus Animalium_ there occurs the following passage:--"Groups that only differ in degree, and in the more or less of an identical element that they possess, are aggregated under a single class; groups whose attributes are not identical but analogous are separated. For instance, bird differs from bird by gradation, or by excess and defect; some birds have long feathers, others short ones, but all are feathered. Bird and Fish are more remote and only agree in having analogous organs; for what in the bird is feather, in the fish is scale. Such analogies can scarcely, however, serve universally as indications for the formation of groups, for almost all animals present analogies in their corresponding parts."[8] It is thus similarity in form and structure which determines the formation of the main groups. Within each group the parts differ only in degree, in largeness or smallness, softness and hardness, smoothness or roughness, and the like (_loc. cit._, i., 4, 644^b). These passages show that Aristotle had some conception of homology as distinct from analogy. He did not, however, develop the idea. What Aristotle sought in the variety of animal structure, and what he found, were not homologies, but rather communities of function, parts with the same attributes. His interest was all in _organs_, in functioning parts, not in the mere spatial relationship of parts. This comes out clearly in his treatise _On the Parts of Animals_, which is subsequent to, and the complement of, his _History of Animals_. The latter is a description of the variety of animal form, the former is a treatise on the functions of the parts. He describes the plan of the _De Partibus Animalium_ as follows:--"We have, then, first to describe the common functions, common, that is, to the whole animal kingdom, or to certain large groups, or to members of a species. In other words, we have to describe the attributes common to all animals, or to assemblages, like the class of Birds, of closely allied groups differentiated by gradation, or to groups like Man not differentiated into subordinate groups. In the first case the common attributes may be called analogous, in the second generic, in the third specific" (i, 5, 645^b, trans. Ogle). The alimentary canal is a good example of a part which is "analogous" throughout the animal kingdom, for "all animals possess in common those parts by which they take in food, and into which they receive it" (Cresswell, _loc. cit._, p. 6).