body is inclined to think that the process of copying models by entire societies is extinct, he should look at the way in which the British Constitution, which was once regarded by men more civilised than the English as an eccentric political oddity, has spread over nearly all Europe in less than seventy years. What Sir Alfred Lyall has shown from his own observations is the activity of this process of imitation in barbarous stages of society. Barbarous men will copy any successful or fashionable social typea Tribe, a Sept, a Gens, a Village Community, the rules of exogamy' or 'endogamy,' the practice of infanticide or suttee. The agency by which the imitation is carried out is Fiction, sometimes of the most audacious kind, and through it an old order is constantly giving place to a new, and even broken hordes, mere miscellanies of men, are transmuted into definite social forms, which afterwards might seem as if they had all sprung together from roots deep in the Past.

The important lesson is that in sociological investigation it is never possible to discover more than the way in which the Type has been formed. If an institution is once successful, it extends itself through the imitative faculty, which is stronger in barbarous than in civilised man. It follows from this that no universal theory, attempting to account for all social forms by supposing an evolution from within, can possibly be true. A person perfectly ignorant of European history might suppose that the British Constitution and the Belgian Constitution, which are extremely like one another, were produced by analogous courses of development; yet the Belgian Constitution is really the copy of a copy, and the true growth of constitutionalism can only be traced in the history of the English Constitution.

The eminent writers I have named, if I have rightly understood them, are of opinion that the Roman Gens and all similar bodies are derived, without exception, from older groups still to be observed among savages. Among the so-called aborigines of Australia, among the North and South American Indians, and elsewhere, but always among slightly advanced communities, there are found groups of men and women tracing kinship exclusively through the female parent, and not through the male. Wherever they have any tradition of human ancestry, they trace their parentage, according to Mr. Morgan, to a common ancestress, and not to an ancestor. Their most distinct characteristics are that they mark their bodies with some common mark or 'totem,' and that the members of the same group never intermarry; and thus, as I have said in the text, they resemble a Sex rather than any other combination of human beings now familiar to us. On the other hand, among several barbarous or semi-barbarous communities we can still observe, and in the ancient history of several civilised or semi-civilised societies we can still detect, another class of groups having a close resemblance to the Roman Gens. They attribute their origin to a single common male ancestor, and they trace kinship through the male parent, real or adoptive, alone. The members of such groups much more frequently intermarry than do the members of the savage group; but occasionally they will not intermarry, at all events as a matter of theory. Such is the Hindu theory with reference to kinsmen and kinswomen belonging to the same Gotra, and there is some faint evidence of a similar feeling once existing in the Roman Gens. Even among

savages, examples of groups tracing relationship through males are found intermingled with groups acknowledging female descents only; and Mr. Morgan

insists that the first groups are merely the last in a transmuted shape, and that the transformation occurred everywhere, in the societies now civilised as well as in those still savage or barbarous. He several times thus describes the process: 'Descent in the Gens was changed from the female to the male line,' giving the uame Gens indifferently to both groups.

Whatever the facts may have been, the language of Mr. Morgan seems to me to be open to much objection. One of these two groups did not really succeed the other, but the two co-existed from all time, and were always distinct from one another. We must be careful, in theorising on these subjects, not to confound mental operations with substantive realities. The 'Agnatic' Gentile groups, consisting of all the descendants, through males, of a common male ancestor, began to exist in every association of men and women which held together for more than a single generation. They existed because they existed in nature. Similarly the group consisting of the descendants, through women, of a single ancestress still survives, and its outline may still be marked out, if it be worth anybody's while to trace it. What was new at a certain stage of the history of all or of a portion of the human race must have occurred, not in connection with the Gens, but in connection with the Family. There was always one male parent of each child born, but prevalent habits prevented his being individualised in the mind. At some point of time, some change of surrounding facts enabled paternity, which had always existed, to be mentally contemplated; and further, as a consequence of its recognition, enabled the kinship flowing from common paternity to be mentally contemplated also. As to the new facts which led to this recognition, all that. in my opinion, can be said of them is that they

- sad by loking sly a male besean she untorSame Bere in tends to pa of the essenthefressen the rvs hard to we bow the sumage man be self-existing, or indeed anging but in aciematical purposes spread overal manity. The men born of the women belonging to in are themselves members of it, but the sons of these men leave it because they belong to the same group as their mothers. But the other group, formel by male descents from males, retains uninterruptely the fower of its masculine strength, and this strength is constantly reproducing itself. Hence it ten is to be a self-existent militant body. The famous exploit of the Fabian Gens at Rome, when they collectively attacked Veii, and were all but extirpated, is thus in itself perfectly credible. No doubt it is said that Australian savages will sometimes travel great distances to join in the quarrels of men having the same 'totem; but these contests between people who have hardly any interests in common can be scarcely more than faction-fights between men wearing different colours. At the same time, I admit that further information of the precise way in which these peculiar organisations affect the practical life of the communities subject to them is greatly needed; and I regret that Mr. Morgan's death prevented his communicating to me the result of some investigations

on the subject which he had promised to make. As to the South Slavonian communities, the actual origin of many of them has been recorded or is otherwise known. With the limitations mentioned in the text, they are composed of the descendants, through males, of a common male ancestor.

I have said above that workers in the new field of investigation opened by the life and usages of savage societies seem to me to be under some temptation to take mental operations for substantive realities. Mr. Morgan, it is well known, considered that the savage habit of grouping relatives in large classes, without reference to degree of grouping, for instance, a man's father and his uncles together and calling them all his fathers, or forming his brothers and male cousins in one class and calling them all his brothers-is a relic of a state of society in which the relations of the sexes were very unlike those to which we are accustomed. Earnest, and indeed bitter, controversies have already arisen on this theory of Classificatory Relationship, and ingenious efforts are from time to time made to identify and recover the lost forms of marriage. May I suggest that it is at least worthy of consideration whether all or part of the explanation may not lie in an imperfection of mental grasp on the part of savages? The reader of Dr. Macfarlane's remarkable 'Analysis of Relationships of Consanguinity and Affinity' ('Journal of Anthropological Institute,' xii. 1) will require no further proof that the comprehension of a large body of complex relationships demands a prodigious mental effort, even now requiring for its success the aid of a special notation. Some communities have surmounted a part of the difficulty by giving separate names to the nearer relationships, which is what Mr. Morgan calls the Descriptive System; but is there not ground

U