larity that women should use wae in speaking of their brother's children. But the twofold expressions applied to these degrees show that it is due to considerations of the different sex of the brothers and sisters in this first generation. The position of individuals within the patriarchal family is the same or different, in accordance with the fact that they are of the same or of another sex; on the other hand, their relation to their grandfather's generation does not depend upon the sex of the individuals, but upon the fact that they are connected with that generation through the father or mother. This shows that all the expressions in question are influenced by the same idea, namely, that the children of the father's house are aliens to their mother, and belong to their father's family.

In this way we have a simple explanation of the use of the particle, of which the meaning has been given above as stranger woman." The degrees of kinship which are thus designated are all connected with the mother's sister, either as her children, grandchildren, or great-grandchildren. No bond unites the families to which the children of two married sisters belong, while both are united with the family of these sisters' brother, who represents their original family. Only the children of the speaker's mother's sister, not those of his mother's brother, are of such remote kinship as to be designated by the particle e. There is only a single case which throws doubt upon this usage. The woman calls her sister's children e-sung, but the man calls them wae-sung. Morgan, as we saw above, holds that e and wae are used in the same sense, but he does not explain how it is that the sister's children in this case obtain a designation which is in other instances reserved for the family of the mother's sister. It is only needful to refer to the conditions of kinship to see that a male speaker is mother's brother to his sister's children, while the female speaker is the mother's sister. There is therefore a strongly

marked difference between e and wae, which is due to the legal connection of the two families, and it is quite logical that the woman should place her sister's children in the category of e.

The word tang is used when the generations are connected through brothers, of whom the speaker is not one. The word peaon is used when the brothers or sisters are of a different sex from that of the speaker.

The principle dominating the Chinese nomenclature is therefore the same as that which prevails in the Turanian, and in some cases in the Ganowanian systems; namely, that the position of the individual is decided by his parents' generation and not by his own.

In the whole nomenclature the word chih presents the only difficulty. It does not occur in the designations for sisters' children, and for the daughter's children of the father's brother. This may lead us to surmise that its meaning became superfluous owing to the use of wae-sung or e-sung; but the two words are used in connection for the daughter's children of the father's sister. This is perhaps because peaon-chih, which always occurred together, had been almost fused into one word, so that the expression peaon-chih-wae-sung for the daughter's child of the father's sister merely includes a superfluous definition.

If we are to define the character of the Chinese nomenclature, we should say that it resembles our own; that it is a classifying, descriptive nomenclature, which has been polished by frequent use in daily intercourse. The great power exerted by groups of kinship and the life of a community partly founded on economical considerations, decide the selection of the classifying categories which are employed. As Morgan observes, the posterity of a married pair does not in theory go beyond the class of brothers, and hence there arises a recognized connection of kinsfolk which never becomes extinct, although practically it is not considered beyond the fifth

degree. If the father dies intestate, the property is usually not divided during the widow's lifetime, but it remains under the control of the eldest brother. After the widow's death, the eldest son divides the property between himself and his brothers, and the portion of the younger brothers is altogether dependent on the will of the elder.1

A consideration of the whole series of our researches into nomenclature will show that it affords no warrant for the far-reaching conclusions of Morgan, McLennan, Lubbock, and others. The nomenclature was in every respect the faithful reflection of the juridical relations which arose between the nearest kinsfolk of each tribe. Individuals who were, according to the legal point of view, on the same level with the speaker, received the same designation. The other categories of kinship were formally developed out of this standpoint. The writers in question hold that the correct understanding of nomenclatures must be found in considerations of the circumstances of marriage and descent which are concealed in the categories of nomenclature, but the assertion is altogether unproved. We must, however, admit that the correct interpretation which we now flatter ourselves to have given, diminishes the significance of nomenclatures as a contribution to the means of historical research to such an extent that it ceases to possess the interest which would entitle us to dwell further on the subject. Lubbock, in his remarks on Morgan's book, states that while he does not accept his most important conclusions, yet he cannot avoid declaring that Morgan's work is one of the most important contributions to ethnological science which has appeared for many years.2 With all respect for Morgan's diligence as a collector of facts, I am more disposed to agree with McLennan that his work is altogether unscientific, and that his hypotheses are a wild dream, if not the 1 Morgan, Systems, p. 424; Appendix ix., p. 425; Appendix xvi. 2 Lubbock, Orig. of Civ., p. 157.

7

delirium of fever.1 His statements throughout are based on such vague analysis and such irrational psychology, that they can only confuse the question, unless they are altogether ignored.

McLennan, Studies p. 360.

CHAPTER VI.

EXOGAMY AND ENDOGAMY.

Conception of exogamy and endogamy-Their relation to incest-Modern ideas of incest-Immorality and incest-Various explanations of exogamy-Symbol of rape-Its bearing-Causes of rape of women -Symbol of rape and modesty-Criminal ties-Desire of trophiesClanless tribe endogamous-Exogamy and incest-Australian legend -Crime and punishment-Marriage and sexual intercourse-Marriage of royal brother and sister-Breach of privilege and marriage -King and his sister-Legal character of marriage-EndogamyArabs-Castes and classes-Karens-Kookas-Ceremonial intercourse of parents and children-in-law-Ceremonial and marriage by violence - Modesty of Bechuanas and Beni-Amirs-Reverence of Caribs-Symbol of rape.

In the preceding pages we have repeatedly had to consider exogamy, and also some of the facts of endogamy. We need only remind our readers that exogamy signifies the prohibition to contract marriages within the group, while endogamy forbids marriage outside the group. It is not too daring to hope that a closer examination of these two customs may enlighten us with respect to the ideas which underlie primitive man's view of marriage, whether these ideas agree with our own, or display the same difference between barbarism and civilization as we find to exist in the relations between parents and children.

The strong moral force which characterizes these prohibitions, the inexorable strictness with which they are obeyed, and the deep abhorrence with which a

Р