Now such a picture of the original clan and of the original inter-clan organization is very pretty and easy to keep in one’s head.  And when one is simply guessing about the first beginnings of things, there is something to be said for starting from some highly abstract and simple concept, which is afterwards elaborated by additions and qualifications until the developed notion comes near to matching the complexity of the real facts.  Such speculations, then, are quite permissible and even necessary in their place.  To do justice, however, to the facts about totemic society, as known to us by actual observation, it remains to note that the clan is by no means the only form of social organization that it displays.

The clan, it is true, whether matrilineal or patrilineal, tends at the totemic level of society to eclipse the family.  The natural family, of course—­that is to say, the more or less permanent association of father, mother and children, is always there in some shape and to some extent.  But, so long as the one-sided method of counting descent prevails, and is reinforced by totemism, the family cannot attain to the dignity of a formally recognized institution.  On the other hand, the totemic clan, of all the formally recognized groupings of society to which an individual belongs in virtue of his birth and kinship, is, so to speak, the most specific.  As the Australian puts it, it makes him what he “is.”  His social essence is to be a Cockatoo or a Crow.  Consequently his first duty is towards his clan and its members, human and not-human.  Wherever there are clans, and so long as there is any totemism worthy of the name, this would seem to be the general law.

Besides the specific unity, however, provided by the clan, there are wider, and, as it were, more generic unities into which a man is born, in totemic society of the complex type that is found in the actual world of to-day.

First, he belongs to a phratry.  In Australia the tribe—­a term to be defined presently—­is nearly always split up into two exogamous divisions, which it is usual to call phratries.[5] Then, in some of the Australian tribes, the phratry is subdivided into two, and, in others, into four portions, between which exogamy takes place according to a curious criss-cross scheme.  These exogamous subdivisions, which are peculiar to Australia, are known as matrimonial classes.  Dr. Frazer thinks that they are the result of deliberate arrangement on the part of native statesmen; and certainly he is right in his contention that there is an artificial and man-made look about them.  The system of phratries, on the other hand, whether it carves up the tribe into two, or, as sometimes in North America and elsewhere, into more than two primary divisions, under which the clans tend to group themselves in a more or less orderly way, has all the appearance of a natural development out of the clan-system.  Thus, to revert to the imaginary case of the Cockatoos and Crows practising