and each of these seasons corresponds to a division of the primi- tive four and of the cardinal points. Nor is this all. For the divisions of time two cycles, one of twelve and the other of ten, are employed concurrently for years, months, days, and hours. Five revolutions of the former cycle and six of the latter are equivalent, thus forming in combination a sexagenary cycle, and serving to mark time with exactitude. The two cycles with their divisions are closely connected with the compass, and by that means with the five elements ; and thus the Chinese have arrived at the extraordinary notion of time not homogeneous, symbolised by the elements, the cardinal points, the colours, and so forth ; and in its different parts the most various influences are held to predominate. The twelve years of the cycle are related also to twelve animals, the rat, cow, tiger, hare, dragon, serpent, horse, goat, ape, cock, dog, and pig, which are assigned three by three to the cardinal points ; and thus again this division of time is connected with the general system. The years are thus sub- sumed under the elements and under regions represented by the animals. Evidently we have here a multitude of interlaced classi- fications. The system dominates the entire life of China, and is an instance, a specially typical instance, in which the collective thought has in a fashion reflechie et savante worked over themes obviously primitive. We have not the means of tracing any his- torical connection between it and the systems previously referred to, though it is clear enough that it reposes on the same principles as they — a conclusion confirmed by the remains of totem-clans in China.
The authors proceed to refer to traces in other civilised countries of systems of classification which recall those previously discussed, instancing ancient Greece and India. They finally argue that primitive classifications are not exceptional and without analogy with those of the most cultivated peoples, but on the contrary there is no solution of continuity between the latter and the former. Primitive classifications are the first efforts to unify knowledge ; they are truly a philosophy of nature. It is not for the purpose of regulating his conduct or justifying his practice that the Austra- lian divides the world among the totems of his tribe. It is because the notion of totem is cardinal for him that he is compelled to group all his other knowledge by relation with it. The conditions on which these ancient classifications depend have in fact played