that radiation of different parts of the body is not necessarily correlated; that is, that the adaptive divergence of the feet and limbs may take one direction, while that of the teeth and skull may take another direction. Thus great variety in combinations of characters may arise, bringing about the very antithesis of Cuvier's supposed "law of correlation"; for we find that while the end results of adaptation are such that all parts of an animal conspire to make the whole adaptive, there is no fixed correlation either in the form or rate of development of parts, and that it is, therefore, impossible for the paleontologist to predict the anatomy of an unknown animal from one of its parts only, unless the animal happen to belong to a type generally familiar. For example, among the land vertebrates the feet, which are associated with the structure of the limbs and trunk, may take one of many lines of adaptation to different media or habitats, either aquatic, terrestrial, arboreal or aerial; while the teeth, which are associated with the structure of the skull and jaws, also may take one of many lines of adaptation to different kinds of food or modes of feeding, whether herbivorous, insectivorous or carnivorous. Through this independent adaptation of different parts of animals to their specific ends there have arisen among vertebrates almost unlimited numbers of combinations of food and tooth structure.
Alternations of Habitat.—In the long vicissitudes of time and procession of continental changes animals have been subjected to alternations of habitat either through their own migrations or through the "migration of the environment itself," to employ Van den Broeck's epigrammatic description of the profound and sometimes sudden environmental changes which may take place in a single locality. The traces of alternations of anatomical adaptation corresponding with these alternations of habitat are recorded both in paleontology and anatomy. For example, Huxley in 1880 briefly suggested the arboreal origin of all the marsupials, a suggestion which has been confirmed abundantly by the detailed studies of Dollo and Bensley, according to which we may imagine that the marsupials have passed through a series of phases, as follows: (1) a very early "terrestrial or ambulatory" phase, (2) a "primary arboreal" phase as exemplified by the tree phalangers of the present day, (3) a "secondary terrestrial" phase as exemplified by the kangaroos and wallabies, (4) a "secondary arboreal" phase as exemplified by the tree kangaroos.
Each one of these phases has left its anatomical record in the structure of the feet and limbs, although this record is often obscured by adaptation.
Louis Dollo especially has contributed most brilliant discussions of this theory of "alternations of habitat" as applied not only to the interpretation of the anatomy of the marsupials but of many kinds of
