recognizes a symbolical marriage in the ceremony we have described. We need not now consider the surprising assertion that the Nairs' marriage contract serves as the preliminary condition of his claim to sexual liberty; this follows from the mythological theory, which we shall discuss later on.

Among the northern Nairs, the women hold intercourse with the men as their lovers, and are very faithful to them, but they have nothing to do with the household management as long as the man's mother is alive. He may, if he chooses, send the woman away at any time, and on his death she goes back with her children to her brother, takes charge of his house, and her children become his heirs.1 The same mode of life prevails among the Buntar, the highest class of the Sudras Tulavas; except that the eldest daughter of a Tulavan Rajah never marries, but takes now one, now another Brahman as her lover; her sons become Rajahs, and her eldest daughter carries on the family.2 Similar conditions exist among the Mogayen (Tulavan fishermen) and the Biluares, who extract juice from palm-trees. These are certainly transition forms either of a waning or of a waxing Nair-type; and since it is difficult to obtain any certainty about them, they are of no use for our present purpose.

Lubbock's account of the Limboos will serve as an example of the inaccurate way in which these kinds of customs have sometimes been described. Limboo sons

2 Ibid., vol. iii. pp. 16-18.

66

1 Buchanan, vol. ii. p. 513. Ibid., vol. ii. p. 492; vol. iii. pp. 22, 53. The wife works for her husband's master, who must maintain her and her children, until they are able to work; the eldest son then belongs to him, but all the other children belong to their mother's master, and return to the hut of her parents." The children of Catal or Curumbal slaves all belong to the mother's master (Ibid., vol. ii. p. 498). We are told of the Corar slaves to the Tulavas, who were formerly in possession of the country: "The master pays the expense of the marriage feast. When a man dies, his wives, with all their children, return to the huts of their respective mothers and brothers, and belong to their masters" (lbid., vol. iii. p. 101).

belong to their fathers if a small sum of money has been paid to the mother; the child then receives a name and is admitted into his father's tribe, while the daughters abide with their mother.1 In this custom Lubbock traces a survival of an extinct female line of descent. It is hardly necessary to say that we should rather trace in such a custom the dawning of a female line; but the custom itself has no existence. Campbell, to whom Lubbock refers,2 only states that the Limboo bride is purchased, and, if such a stipulation has been previously made, is taken to her husband's home. Labour is often substituted for the purchase money. His account goes on to say that children born out of wedlock, and those born of a connection between Limboo men and Lepcha women, are called Koosaba. The father may obtain possession of the boys by purchase and by naming them, but the girls belong to the mother. Here we are, in the first place, only concerned with children born out of wedlock, and, secondly, with the children of persons of distinct tribes, which are by the mode of naming them placed on a level with the illegitimate. It should be noted that the Limboos are strictly endogamous, the only exception being made in favour of the Lepchas.*

The existence of the female line among the peoples of Western Asia has been ascertained with respect to some ancient tribes; in other cases the inference of its existence is not wholly justified.5 It does not now prevail in any part of that country. It would be lost labour

1 Lubbock, Origin of Civ., p. 149. Appendix XXIV.

2 Lubbock quotes Campbell, Trans. Ethno. Soc., New Series, vol. vii., which I have had no opportunity of consulting; but I think the same reference may be found in Journal Asiat. Soc. of Bengal.

3 Journal Asiat. Soc. of Bengal, 1840, vol. ix. p. 603; Campbell. Appendix XXV.

1 Ibid., p. 596.

Lycians (Herodotus, i. 173). In Northern Africa we may also cite the Ethiopians (Bachofen, Mutterrecht, p. 21), the Bega ancestors of the Bischari, and the Messophites (Ibid., p. 108), Barea and Bazer (Munzinger, pp. 481, 484), Touaregs (Duveyrier, pp. 337-340).

to say more of these various accounts, since they afford no material of any value in solving the question as to the motives for preferring the one or the other line. They are capable of being interpreted in whichever sense has previously been adopted as the right one.1

1 I may here direct attention to Wilken's attempt to trace the female line among the Arabs. His conclusions are based upon the following facts: 1. That polyandry and unchaste customs throw doubt upon the paternity, so that it would be most natural to adopt the female line. 2. That the Arabs call the clan Batu, or belly, which points to the mother. 3. That some clans are called after the mother's tribe. 4. That the character of the mother's brother is supposed to devolve upon the sister's children. Wilken himself is of opinion that each fact, taken singly, proves nothing, but that in the aggregate they are of weight. We shall subsequently show that Wilken's opinions are uutenable.

CHAPTER V.

POLYNESIA.

Polynesian classes-Marriage and class-Fijian Vasu.

THE Polynesian definition of kinship must be added to the conditions we have described in the last chapter. The dividing line which is usually drawn between the Polynesian and Melanesian is in this case without significance.

In Polynesia, the distinct classes constitute a similar state of things to the family group in the peoples we have just considered, since they form an exclusive organization, holding property in common. It is not very clear how these classes arose, but we may assume that they are connected with an earlier distribution into clans, so that the chief represents the eldest line of the posterity of their common ancestor. In some cases this ancestor is supposed to be of Divine origin; but we lay no stress on such a supposition, since it probably arose after the chief's position was established. The people are usually in possession of small plots of ground, either as comparatively independent proprietors, or as serfs; the nobles are owners or rulers of small districts, and the king is ruler of the whole. The conditions are in many respects confused and indefinite, yet the type is undoubtedly that of the joint family, or village community.

The classes differ from clans in a natural way. The

nobles of different clans belong to one class, and while the clan is usually exogamous, the class always tends to become endogamous. In Polynesia, the definition of the class depends upon the line of kinship, and the classes are not isolated with the exclusiveness of castes in India: marriages between the different classes are not absolutely forbidden.

The position of a child born from a marriage between persons of unequal rank may be decided in several ways. The child may either be always assigned to the superior or inferior class, or always either to the father or mother. Polynesia offers us examples of all kinds.

If the father or mother alone belongs to the ruling class, the child is, in the Caroline Isles, assigned to that class.1 In the Tonga Isles, the highest class—the Egi, or nobles-inherits rank and property through the mother; the children of the common people (Mataboulas and Tuas) inherit from the father, but belong to the mother's class.2 In Otaheite, the children of a marriage between a noble (Hui-Arii) and a woman of a lower class are set aside, unless numerous ceremonies are performed in the temple at the time of the wedding, so as to raise the rank of the inferior person. Both among the nobles and in the intermediate class of landowners the father abdicates in favour of his new-born son, because the son has an additional ancestor, and is therefore of higher rank than his father.4

3

Marriages are dissolved in the Sandwich Isles at the wish of either party; only in the case of the chiefs there is no divorce, but they form a connection with other women, and their wives take other lovers. These are usually of inferior rank, and the children begotten of 1 Chamisso, vol. ii. p. 241.

2 Martin, vol. ii. p. 101. Rienzi, vol. iii. p. 45. Morgan, Systems,

p. 559.

3 Ellis, vol. iii. p. 98.

4 Ibid., vol. iii. p. 100.

Cook, vol. i. Hawkesworth, vol. ii. p.