hordes, forces them to prey upon one another for wives.  Usage, induced by necessity, would in time establish a prejudice among the tribes observing it, a prejudice strong as a principle of religion—­as every prejudice relating to marriage is apt to be—­against marrying women of their own stock.’

Mr. M’Lennan describes his own hypothesis as ’a suggestion thrown out at what it was worth.’ {258} In his later years, as we have said, he developed a very subtle and ingenious theory of the origin of exogamy, still connecting it with scarcity of women, but making use of various supposed stages and processes in the development of the law.  That speculation remains unpublished.  To myself, the suggestion given in ‘Studies in Ancient History’ seems inadequate.  I find it difficult to conceive that the frequent habit of stealing women should indispose men to marry the native women they had at hand.  That this indisposition should grow into a positive law, and the infringement of the law be regarded as a capital offence, seems still more inconceivable.  My own impression is, that exogamy may be connected with some early superstition or idea of which we have lost the touch, and which we can no longer explain.

Thus far, the consideration of exogamy has thrown no clear light on the main question—­the question whether the customs of civilised races contain relics of female kinship.  On Sir Henry Maine’s theory of exogamy, that Aryan custom is unconnected with female kinship, polyandry, and scarcity of women.  On Mr. M’Lennan’s theory, exogamy is the result of scarcity of women, and implies polyandry and female kinship.  But neither theory has seemed satisfactory.  Yet we need not despair of extracting some evidence from exogamy, and that evidence, on the whole, is in favour of Mr. M’Lennan’s general hypothesis. (1.) The exogamous prohibition must have first come into force when kinship was only reckoned on one side of the family.  This is obvious, whether we suppose it to have arisen in a society which reckoned by male or by female kinship.  In the former case, the law only prohibits marriage with persons of the father’s, in the second case with persons of the mother’s, family name, and these only it recognises as kindred. (2.) Our second point is much more important.  The exogamous prohibition must first have come into force when kinship was so little understood that it could best be denoted by the family name.  This would be self-evident, if we could suppose the prohibition to be intended to prevent marriages of relations.  Had the authors of the prohibition been acquainted with the nature of near kinships, they would simply (as we do) have forbidden marriage between persons in those degrees.  The very nature of the prohibition, on the other hand, shows that kinship was understood in a manner all unlike our modern system.  The limit of kindred was everywhere the family name:  a limit which excludes many real kinsfolk and includes