So strongly has the family impressed itself upon the thought of the age that students of man in his earliest ages are found stating that "the family is the most ancient and the most sacred of human institutions."[327] This proposition, however, is not only denied by other authorities, as, for instance, Mr. Jevons, who affirms that "the family is a comparatively late institution in the history of society,"[328] but it rests upon the merely analytical basis of research, separated entirely from those facts of man's history which are discoverable by the means just now suggested. One is, of course, quite prepared to find the family among civilisations older than the Indo-European, and yet to find that it is a comparatively late institution among Indo-European peoples. As a matter of fact, this is the case; for the two kinds of family, the family as seen in savage society and the family as it appears among the antiquities of the Indo-European people, are totally distinct in origin, in compass, and in force; while welded between the two kinds of family is the whole institution of the tribe. It is no use introducing the theory adopted by Grote, Niebuhr, Mommsen, Thirlwall, Maine, and other authorities who have studied the legal antiquities of classical times, that the tribe is the aggregate of original family units. Later on I shall show that this cannot be the case. The larger kinship of the tribe is a primary unit of ancient society, which thrusts itself between the savage family and the civilised family, showing that the two types are separated by a long period of history during which the family did not exist. It has taken me some time to explain these points in anthropological science, which appear to me not to have received proper consideration at the hands of the masters of the science, but which are essential factors in the history of man and are necessary to a due consideration of the position occupied by folklore. The chief results obtained are:-- (1) Migratory man would deposit his most rudimentary social type not at the point of starting his migration, but at the furthest point therefrom. (2) Custom due to the migratory period would continue after real migratory movement had ceased, and from this body of custom would be derived all later forms of social custom. (3) Non-kinship groups are more rudimentary than kinship groups, and are still observable in savage anthropology. (4) Anthropological evidence must be based upon the whole of the characteristics of human groups, not upon special characteristics singled out for the purpose of research. It is with these results we have to work. They will help us to see how far the facts of anthropology, which begin far behind the historical world, have to do with the problems presented by folklore as a science having to deal with the historical world. II