only be solved forty years later by Darwin's theory of selection. A genuine inductive process alone could reveal these remarkable relations between the morphological and physiological characters of organs. But it is at the same time true that De Candolle could not have made this discovery, if his predecessors had not already established a large number of affinities. It was while he was engaged in an exact comparison of forms already recognised as undoubtedly related to one another, that that which he called the plan of symmetry, and which was afterwards named a type, revealed itself to him; and as he examined it more closely, and compared it with peculiarities of habit in different plants formed on the same plan, he discovered certain causes, by means of which the deviations were to be explained; these were abortion, degeneration, and adherence. By attending to these he succeeded in discovering affinities that had been hitherto doubtful or unknown; this was at all events the true inductive way of advancing the system, and whatever the earlier systematists had effected that was really valuable had been effected virtually in the same way, only they never arrived at a clear understanding of their own mode of proceeding; they had followed unconsciously the method which De Candolle clearly understood and consciously pursued.
The majority of De Candolle's successors were far from fully appreciating the entire significance of his theory, its importance as a matter of method and principle; on the contrary in the search for affinities they continued to surrender themselves to a blind feeling rather than to a clearly recognised method, and the same must be said unhappily of De Candolle himself, when he was dealing with the establishment of the large divisions of the vegetable kingdom. With equal surprise we find him in the book before us, in which he has developed the true method in systematic botany, expressing the opinion that the most important physiological characters must be employed for the primary divisions of the system, and this
