governs all the facts of life, and its marks are found in all the principal rites. That food only may be eaten which is allotted, or free, to each particular division of the tribe. The Wakelbura of North Central Queensland are divided into two phratries, the Mallera and the Wutaru. A Mallera wizard can only make use for his art of objects which are reckoned Mallera. When a Mallera dies, the scaffold on which his body is exposed (a practice of the tribe in lieu of burial), and the branches which cover him, must be of some wood belonging to his phratry.
The simplest form of classification takes account only of the phratries. A more complete and characteristic system takes account also of the clans or totems. If totemism is on one side the grouping of men into clans according to natural objects, it is also inversely a grouping of natural objects according to the social groups. As Mr. Fison says, the South Australian savage considers the universe as a great tribe, to one of the divisions of which he belongs ; and all things animate or inanimate which are of his group are parts of the body corporate of which he him- self is part. They are absolutely parts of himself. But while this classification into phratries and clans is an advance on the previous one, it is still imperfect and bears witness to the state of initial confusion whence the human mind set out. The distinct groups are multiplied. The objects attributed to each phratry or each clan are clearly distinguished ; but inside the clan they are to a large extent indifferentiated. They are all of the same nature, and there are no strict lines of demarcation such as exist between ultimate varieties in our classification. The individual human members of the clan, the members of the totemic species, and those of the various species attached in the native view to the clan, are all only different aspects of one and the same reality. By applying the social divisions to the primitive mass of repre- sentations a certain number of pigeon-holes for these representa- tions have been made; but the contents of the pigeon-holes remain in a relatively amorphous state, witnessing to the slowness and difficulty with which the classificatory function is developed.
The conclusions drawn after a searching examination of Australian ideas are that if this manner of classifying objects is not necessarily implied in totemism, it is at any rate certain that it is frequently met with in societies organised on a totemic basis ; and that there is, therefore, a close bond and not an accidental