A more important point to notice is that, if we believe the Hindoo system to date from pro-ethnic times, we must also assume that the Hindoo system of naming is pro-ethnic, _i.e._, we must assume that each Aryan had two names, one distinguishing him personally from other people, the other indicating what kin he belonged to; and in this event, the Near and the Remote Kin must, in pro-ethnic times, have had a common name. There is, however, very little evidence to show that this was the case: gentile names are found among the Hindoos and the Romans alone of Aryan peoples. It is, of course, possible that, before the dispersion, the Aryans had gentile names, and that, after the dispersion, all the Aryans, with the exception of the Romans and the Hindoos, lost them entirely. On the other hand, if there was a time when gentile names had not yet been invented, if they have had a history and growth, we must consider it as at least possible that gentile names had not been evolved at the time of the dispersion, and were only developed subsequently by the Romans and Hindoos. Whether the undispersed Aryans had gentile names, and at the same time an aversion to marriages between persons of the same name, is a question on which it were vain to pronounce confidently. We may more safely consider both these equally possible alternatives, together with the consequences which flow from each. Let us assume that marriage was, amongst the Aryans as amongst the Hindoos, prohibited between persons of the same gentile name: is there anything in the social organisation presupposed by this prohibition incompatible with the patriarchal system? According to Mr. D. M‘Lennan there is: not only are there "numerous societies of which the patriarchal theory does not even attempt to give any account," but "in the societies upon contemplation of which it was formed, a most serious difficulty for it is presented by the tribes, which consist of several clans, each clan considered separate in blood from all the others. The patriarchal theory, of course, involves that the clans are all of the same blood."[128] Mr. M‘Lennan's difficulty seems to be this: where inheritance (of family name, property, sacra, &c.) is confined to the male line, the descendants of a common ancestor must all have the same family or gentile name; persons having different names cannot be descended from the same ancestor—that is to say, different _gentes_ or clans cannot have a common origin. A tribe, therefore, which consists of several clans cannot consist of descendants of a common ancestor. Yet, these clans believe they have an ancestor, however remote, in common. If their belief is incorrect (if the _gentes_ have not a common origin), how did the error arise? If, on the other hand, the different _gentes_ of the same tribe have a common origin, how came they to have different names?