VIII. Ancient Systems of Kinship, and Their Influence on the Structure of Primitive Groups

The earliest human groups can have had no idea of kinship. We do not mean to say that there ever was a time when men were not bound together by a feeling of kindred. The filial and fraternal affections may be instinctive. They are obviously independent of any theory of kinship, its origin or consequences; they are distinct from the perception of the unity of blood upon which kinship depends; and they may have existed long before kinship became an object of thought. What we would say is, that ideas of kinship must be regarded as growths—must have grown like all other ideas related to matters primarily cognizable only by the senses; and that the fact of consanguinity must have long remained unperceived as other facts, quite as obvious, have done. In other words, at the root of kinship is a physical fact, which could be discerned only through observation and reflection—a fact, therefore, which must for a time have been overlooked. No advocate of innate ideas, we should imagine, will maintain their existence on a subject so concrete as relationship by blood.

A group of kindred in that stage of ignorance is the rudest that can be imagined. Though they were chiefly held together by the feeling of kindred, the apparent bond of fellowship between the members of such a group would be that they and theirs had always been companions in war or the chase—joint-tenants of the same cave or grove. To one another they would simply be as comrades. As distinguished from men of other groups, they would be of the group, and named after it.

Hence, most naturally, on the idea of blood-relationship arising, would be formed the conception of Stocks. Previously individuals had been affiliated not to persons, but to some group. The new idea of blood-relationship would more readily demonstrate the group to be composed of kindred than it would evolve a special system of blood-ties between certain of the individuals in the group. The members of a group would now have become brethren. As distinguished from men of other groups, they would be of the group-stock, and named after the group.

The development of the idea of blood-relationship into a system of kinship, must have been a work of time—at least the establishment over any great area of any such system as an institution of customary law must have been slowly effected. It is most improbable that that idea, when first formed, was anywhere at once embodied in a well-defined system of kinship.

We shall endeavour to show—

I. That the most ancient system in which the idea of blood-relationship was embodied, was a system of kinship through females only.

Once a man has perceived the fact of consanguinity in the simplest case—namely, that he has his mother’s blood in his veins, he may quickly see that he is of the same blood with her other children. A little more reflection will enable him to see that he is of one blood with the brothers and sisters of his mother. On further thought he will perceive that he is of the same blood with the children of his mother’s sister. And, in process of time, following the ties of blood through his mother, and females of the same blood, he must arrive at a system of kinship through females. The blood-ties through females being obvious and indisputable, the idea of blood-relationship, as soon as it was formed, must have begun to develop, however slowly, into a system embracing them. What further development this idea might have—whether it would simultaneously have a development in the direction of kinship through males—must have depended on the circumstances connected with paternity. If the paternity of a child were usually as indisputable as the maternity, we might expect to find kinship through males acknowledged soon after kinship through females. But however natural it might be that men should think of blood-ties as possible to be propagated through fathers, blood-ties through fathers could not find a place in a system of kinship, unless circumstances usually allowed of some degree of certainty as to who the father of a child was, or of certainty as to the father’s blood. A system of relationship through fathers could only be formed—as we have seen that a system of relationship through mothers would be formed—after a good deal of reflection upon the fact of paternity. And fathers must usually be known before men will think of relationship through fathers—indeed, before the idea of a father can be formed. There could be no system of kinship through males if paternity was usually, or in a great proportion of cases, uncertain. The requisite degree of certainty can be had only when the mother is appropriated to a particular man as his wife, or to men of one blood as wife, and when women thus appropriated are usually found faithful to their lords.

Considering that the history of all the races of men, so far as we know it, is the history of a progress from the savage state; considering the social condition of rude tribes still upon the earth,—remembering that the races which can be traced in history had all a previous history, which remains unwritten,—it cannot seem a very strange proposition that there has been a stage in the development of human races when there was no such appropriation of women to particular men—when, in short, marriage, as it exists among civilized nations, was not practised. We believe we shall show, to a sufficient degree of probability, that there have been times when marriage, in this sense, was yet undreamt of. Wherever this has been the case the paternity of children must have been uncertain; the conditions essential to a system of kinship through males being formed would therefore be wanting; no such system would be formed; there would be—there could be—kinship through females only.

Not to assume that the progress of the various races of men from savagery has been a uniform process, that all the stages which any of them has gone through have been passed in their order by all, we shall be justified in believing that more or less of promiscuity in the connection of the sexes, and a system of kinship through females only have subsisted among races of men among which no traces of them remain, when we have shown their existence in a considerable number of cases—if in these there appear nothing exceptional. After what has been said above, it must be plain that kinship through females only, if it exist at all, must be a more archaic system of relationship than kinship through males—the product of an earlier and ruder stage in human development than the latter—somewhat more than a step farther back in the direction of savagery. To prove its existence on such a scale as to entitle it to rank among the normal phenomena of human development, is, we may now say, to prove it the most ancient system of kinship. As customs tend to perpetuate themselves and die hard, it will not in any degree make against our explanation of the origin of kinship through females only, that it should be found in some cases along with marriage relations which allow of certainty as to fathers. It is inconceivable that any thing but the want of certainty on that point could have long prevented the acknowledgment of kinship through males; and in such cases we shall be able to conclude that such certainty has formerly been wanting—that more or less promiscuous intercourse between the sexes has formerly prevailed. The connection between these two things—uncertain paternity and kinship through females only, seems so necessary—that of cause and effect—that we may confidently infer the one where we find the other.

Let us see, then, what can be said for the proposition that there has been a stage in the progress of men in which a woman was not usually appropriated to a particular man as his wife.

All the evidence we have goes to show that men were from the beginning gregarious. The geological record distinctly exhibits them in groups—naked hunters or feeders upon shell fish leading a precarious life of squalid misery. This testimony is confirmed by all history. We hear nothing in the most ancient times of individuals except as being members of groups. The history of property is the history of the development of proprietary rights inside groups, which were at first the only owners, and of all other personal rights—even including the right in offspring—it may be said that their history is that of the gradual assertion of the claims of individuals against the traditional rights of groups.

We, of course, know nothing about the co-ordination of the sexes in the earliest groups. The reader knows already what must be our conjecture as to what it was. We can trace the line of human progress far back towards brutishness; finding as we go back the noble faculties peculiar to man weaker and weaker in their manifestations, producing less and less effect,—at last scarcely any effect at all—upon his position and habits. As we go back, we find more and more in men the traits of gregarious animals; slighter and slighter indications of operative intellect. As among other gregarious animals, the unions of the sexes were probably in the earliest times, loose, transitory, and in some degree promiscuous.

Before the invention of the arts, and the formation of provident habits, the struggle for existence must often have become very serious. The instincts of self-preservation, therefore, must have frequently predominated and shaped the features of society freely, as if the unselfish affections had no place in human nature. None of the races of mankind can have been spared the cruel experience of this initiatory stage; or can have escaped the effects of that experience on its character and customs. Even those most favourably situated must have had long periods of trial, and have suffered from the incessant hostility of neighbours. So, without supposing the course of human events to have been uniform, we must conceive of early human society as having been throughout affected by influences of the same general, unfriendly, character, and as having been determined, though perhaps by unequal pressures, towards one uniform type in all its parts.

Foremost among the results of this early struggle for food and security, must have been an effect upon the balance of the sexes. As braves and hunters were required and valued, it would be the interest of every horde to rear, when possible, its healthy male children. It would be less its interest to rear females, as they would be less capable of self-support, and of contributing, by their exertions, to the common good. In this lies the only explanation which can be accepted of the origin of those systems of female infanticide still existing, the discovery of which from time to time, in out of the way places, so shocks our humanity. It is of no consequence by what theories the races who practise infanticide now defend the practice. There can be no doubt that its origin is everywhere referable to that early time of struggle and necessity which we have been contemplating.

What is now true in varying degrees of all the rudest races may be assumed to have been true of all the earliest groups. We may predicate of the primitive groups that they were all or nearly all marked by a want of balance between the sexes—the males being in the majority. The reader will have little difficulty in granting that we may do so when he reflects on the prevalence of exogamy, the origin of which must be referred to that want of balance. And we think he will be still more ready to make the concession when we shall have surveyed the facts connected with polyandry—the origin of which must be referred to the same cause.

What diminished the number of the female sex would increase the importance of women. The first result of the balance of the sexes being against the females, must have been to give every woman more than one, it might be several wooers. Apart from any disproportion of the sexes, we might expect the more engaging females of a horde to be surrounded by suitors. Savages are unrestrained by any sense of delicacy from a copartnery in sexual enjoyments; and, indeed, in the civilised state, the sin of great cities shows that there are no natural restraints sufficient to hold men back from grosser copartneries. But within a horde possessing few women, such copartneries would be a necessity. And as savages assert for themselves a high degree of independence, it is obvious that grave difficulties must have surrounded the constitution and regulation of such copartneries. And to the consideration of these difficulties we are led, the instant we conceive of the primitive groups as containing fewer women than men.

The men of a group must either have quarrelled about their women and separated, splitting the horde into hostile sections; or, in the spirit of indifference, indulged in savage promiscuity. That quarrels and divisions were of frequent occurrence cannot be doubted. These were the first wars for women, and they went to form the habits which established exogamy. And whether quarrels arose or not, we are led to contemplate groups—the horde or its sections—indulging in a promiscuity more or less general. The quarrels must have been between sections of the hordes rather than between individuals. No individual at that stage could well carry off a woman, isolate himself, and found a family. However brave and strong, he could scarcely maintain his independence for any time against numerous assailants. Unless these quarrels went the length of completely disintegrating the groups—a result which the gregarious nature of men tended to prevent—we must arrive at last at groups within which harmony was maintained through indifference and promiscuity.

These groups would hold their women, like their other goods, in common. And the children, while attached to mothers, would belong to the horde. We find traces of the former existence of groups of this description; and it is probable that before the rise of kinship, all the human groups were of that model. On the rise of kinship, the difficulty due to the scarcity of women would more easily be overcome. The first advance from a general promiscuity—assuming its existence—would naturally be to a promiscuity less general—to arrangements between small sets of men to attach themselves to a particular woman. Previous to the establishment of a system of kinship—when men were bound to each other only by the tribal tie—it is obvious that there would constantly be difficulties in the way of their forming such combinations. When, however, the system of kinship through females only, had been firmly established, every group stood resolved into a number of small brotherhoods, each composed of sons of the same mother. And within these, the feeling of close kinship would simplify the constitution of the polyandrous arrangement.

Now, here, at length, we are upon the firm ground of fact. We have examples of general promiscuity; and examples of modified promiscuity, in which, with a pretence of marriage, the woman may bestow her favours upon any one, under certain restrictions as to rank and family. We have numerous examples of polyandry, and they are such as to show that polyandry must be regarded as a modification of and advance from promiscuity. We have examples of polyandry in which the wife has several husbands, who are not necessarily relatives; and very many examples of polyandry in which the husbands are all brothers. We often find these two forms of polyandry in the same district, in different sections of the population: here, the husbands as a rule, are no relations; there, the husbands as a rule, are brothers. Farther, where the husbands are not brothers, we find the system of relationship through females only; and, so enduring is custom, we very often find that system where marriage has long been so regulated as to permit of kinship through males. In many cases we find traces of the system of kinship through females only, lingering about the laws of marriage and succession to estates and titles, even where male kinship has been long established. Moreover, in nearly all the cases in which traces are to be found of kinship through females only, traces of polyandry also remain. Thus, what we find is just what was to be expected if the account we have offered of the origin of polyandry were correct.

We repeat, that in showing the prevalence of polyandry, we shall be showing the prevalence of a modification of promiscuity. This is manifest as regards the ruder species of polyandry, in which the husbands are not relations. It is equally, though less obviously, true of the less rude polyandry in which the husbands are brothers. From the way in which polyandry is presented to us, we shall have a proof that the less rude polyandry was developed from the ruder by the help of the system of kinship through females only—was superinduced, that is, upon a promiscuity less qualified than itself. Promiscuity, producing uncertainty of fatherhood, led to the system of kinship through mothers only. This kinship paved the way for polyandry such as we commonly find it; and this form of polyandry introduced male kinship. That, along with the ruder polyandry, we always find the system of kinship through females only, and that where the less rude form prevails we can generally trace that system, is moreover a proof a posteriori of what we have shown must be the case, that the origin of kinship through females only is referable to uncertainty of male parentage.

We shall not concern ourselves with the direct evidence which might be adduced to show that there once prevailed among men a promiscuity less qualified than polyandry. We may however recall the fact, that tradition is found everywhere pointing to a time when marriage was unknown, and to some legislator to whom it owed its institution: among the Egyptians to Menes; the Chinese to Fohi; the Greeks to Cecrops; the Hindus to Svetaketu. And we shall proceed to show how much evidence remains to give verisimilitude to these traditions. Passing over communities in which, according to ancient historians, something like a general promiscuity prevailed—such as the Massagetæ, Agathyrsi, and the ancient Spartans; passing over also the numerous races now existing, which, according to modern travellers, have no conception of conjugal fidelity—we shall now go on to consider the regulated promiscuity known as polyandry, and see to what extent it exists, and what traces of its former existence still remain.

Let us first see what is the area over which polyandry now prevails. It prevails universally in Tibet, and is common in the Himalayan and Sub-Himalayan regions adjoining Tibet; in the valley of Kashmir; among the Spiti, in Ladak; in Kistewar and Sirmor. It occurs among the Telingese; in the Sivalik mountains, and in Kasia. There are unmistakeable traces of its existence till recently in Gurwhal, Sylhet and Cachar. Farther south in India we find polyandry among the Tudas of the Nilgherry Hills, the Coorgs of Mysore, and the Nairs, the Maleres, and Poleres of Malabar. We find it off the Indian coast in Ceylon; and going eastward strike on it as an ancient though now almost superseded custom in New Zealand, and in one or two of the Pacific Islands. Going northward we meet it again in the Aleutian Islands; and taking the continent to the west and north of the Aleutians we find it among the Koryaks to the north of the Okhotsk Sea. Crossing the Russian Empire to the west side we find polyandry among the Saporogian Cossacks; we thus have traced it at points half round the globe. This is not all, however. Polyandry is found in several parts of Africa and of America. We have the authority of Humboldt for its prevalence among the tribes on the Orinocco, and he also vouches for its former prevalence in Lancerota one of the Canary Islands.

From ancient history we learn that polyandry at one time existed over even a greater area. Traces of it remained in the time of Tacitus among the Germans. And while in certain cantons of Media, according to Strabo, polygunia was authorised by express law which ordained every inhabitant to maintain at least seven wives; in other cantons the opposite rule was in force: a woman was allowed to have many husbands, and they looked with contempt on those who had less than five. Cæsar informs us that in his time polyandry prevailed among the Britons. We find direct evidence of its existence among the Picts in the Irish Nennius, not to mention traces of it in the Pictish Laws of Succession. Further we find traditions of it among the Hindus—especially among the Rajputs. And we find it among the Getes of Transaxiana (the Yuti or Yuechi of the Chinese historians). To see where else it prevailed we must go back upon our authorities and examine the various phases of polyandry which they present, and obtain a test for detecting its presence where historical evidence of its existence is awanting.

The ruder form of polyandry, as we have said, is that in which the husbands are not brothers; the less rude is that in which they are brothers. The polyandry of the Kasias, the Nairs, and the Saporogian Cossacks, appears to be purely of the ruder sort, and is attended by the system of kinship through females only. It is left doubtful what is the form of the institution in some instances, as in the Aleutian Islands, and among the Koryaks. But in all the other cases in which polyandry occurs, the authorities show that the ruder form occurs among the lower classes wherever the less rude occurs, except in Tibet, where polyandry is universal and the husbands are always brothers; except in Malabar, where polyandry is universally practised by all classes, saving the Brahmans only, but is of the ruder species among the high caste Nairs, and of the less rude among the lower castes, the Teers, Maleres, and Poleres. It is in the nature of the case that all the possible forms of polyandry must lie in between, or be embraced in, the Nair and Tibetan forms.

Let us attend then to the accounts we have of these two forms. Of the Nair polyandry we have three accounts. The account in the “Asiatic Researches” is that among the Nairs it is the custom for one woman “to have attached to her two males, or four, or perhaps more, and they cohabit according to rules.” With this account that of Hamilton agrees, excepting that he states that a Nair woman could have no more than twelve husbands, and had to select these under certain restrictions as to rank and caste. On the other hand, Buchanan states that the women after marriage are free to cohabit with any number of men, under certain restrictions as to tribe and caste. It is consistent with the three accounts, and is directly stated by Hamilton, that a Nair may be one in several combinations of husbands; that is, he may have any number of wives. The accounts, however, differ in regard to one important particular. Buchanan represents the wife as living in family with her mother or brother, while Hamilton represents her as having “an house built for her own conveniency” on being married to the first of her husbands. In the “Asiatic Researches” the wife is represented as living with her mother or brother. The probability is that both arrangements are occasionally adopted, the more usual course being for the wife to remain in the family of her mother and brothers. In Ceylon, where the higher and lower polyandry co-exist, marriage is of two sorts—Deega or Beena—according as the wife goes to live in the house and village of her husbands, or as the husband or husbands come to live with her in or near the house of her birth. And among the Kandyans the rights of inheritance of a woman and her children are found to depend on whether the woman is a beena or a deega wife.

The three accounts which we have of the Nair polyandry are agreed that the Nair husbands are usually not brothers—usually not relatives—and that the institution leaves male parentage and the father’s blood quite uncertain. “In consequence of this strange manner of propagating the species,” says Buchanan, “no Nair knows his father, and every man looks upon his sister’s children as his heirs. He indeed looks upon them with the same fondness that fathers in other parts of the world have for their own children, and he would be considered as an unnatural monster were he to show such signs of grief at the death of a child, which from long cohabitation and love with its mother, he might suppose to be his own, as he did at the death of a child of his sister. A man’s mother manages his family; and after her death his eldest sister assumes the direction. Brothers almost always live under the same roof; but if one of the family separates from the rest he is always accompanied by his favourite sister. A man’s movable property, after his death, is divided among the sons and daughters of all his sisters; and if there are lands, their management falls to the eldest male of the family.”

Now here, derived from the ruder polyandry, is an exceedingly rude, the rudest, form of family system with which we are acquainted. And it is a sort of family system which is found, more or less modified in some of its features, in several cases, where marriage is now either monogamous or polygamous. Its chief features are the absence of a paternal head, and the system of female succession. Among the Kocch, with whom marriage is now monogamous, we find the same system, excepting that the family circle includes the daughter’s husband, as a subordinate member of the family. A Kocch man goes, on his marriage, like the beena husband of Ceylon, to live in family with his wife and her mother; on his marriage all his property is made over to his wife; and on her death her heirs are her daughters. Here we conclude that the advance from the ruder polyandry to monogamy took place in some way consistent with the preservation of the main features of the family system peculiar to the ruder polyandry consistent with the mother’s maintaining her position as the head of the family, and with an increase of the influence of women as connecting links in the social and proprietary systems. We shall presently see that the advance in this direction must be counted exceptional; at the same time it cannot well be doubted that such a family system as we find among the Kocch had its origin in the ruder species of polyandry.

What, then, was the normal line of progress? We think that we shall be able to show what it was—that it lay between the lower and higher polyandry. In the accounts we have, we can detect stages of preparation for the change from the former species of polyandry to the latter. We must regard as the rudest cases those in which the wife lives not with her husbands, but with her mother or brothers. In these cases a woman’s children are born in and belong to her mother’s house. In the cases next in order of rudeness, the wife passes into cohabitation, according to fixed rules, with the husbands, in a house of her own—becoming thus detached from her family, though still connected with it through the right of her children to become heirs to the family estate. Her children would still belong to her mother’s family—the want of a community of blood and interests among the husbands preventing the appropriation of the children to them. Such cases, however—detaching the woman from her family—would prepare the way for a species of marriage still less rude, in which the woman passed from her family, not into a house of her own, but into the family of her husbands, in which her children would be born, and to which they would belong. This could only happen when the husbands were all of one blood, and had common rights of property—in short, when they were brothers.

This last was a most important step in advance. The girl of a house no longer remained at home with her mother and brothers—aiding in and succeeding to the management; she passed into another family, associating with the sons thereof as wife; while her place at home was assumed by a stranger—as wife to her brothers. There being now a community of blood and interests in the husbands, there was nothing to prevent the appropriation to them of her children—an appropriation which would disqualify the children for being heirs to the property of her mother and brothers. To give effect now to the old law of succession, would be, not to keep property in families, but to introduce a system of exchanges of family estates. Moreover, when this form of marriage became general, and when conjugal fidelity was secured by penalties, we should expect to find that the system of kinship through males would appear—this species of marriage allowing of certainty as to the father’s blood, though not of certainty as to fathers. A woman’s children would become the heirs of the husband’s family in which they would be born, and to which they would belong.

Now it is this highest development of polyandry, and of the family system which polyandry admitted of, which we find in Tibet.

“Here,” says Turner, speaking of Tibet, “we find a practice—that of polyandry—universally prevailing; and see one female associating her fate and fortune with all the brothers of a family, without any restriction of age or of numbers. The choice of a wife is the privilege of the elder brother. The number of husbands is not, as far as I could learn, defined or restricted within any limits; it sometimes happens that in a small family there is but one male; and the number, perhaps, may seldom exceed that which a native of rank, during my residence at Teshoo Loomboo, pointed out to me in a family resident in the neighbourhood, in which five brothers were then living very happily with one female, under the same connubial compact. Nor is this sort of league confined to the lower ranks of people alone: it is found also frequently in the most opulent families.”

Let us now see to what extent polyandry of the Tibetan type can be traced elsewhere than in Tibet; and what evidence there is of its being an advance from the Nair species of polyandry. The authorities already cited exhibit the Tibetan as the prevailing species of polyandry in nearly the whole of the Himalayan and sub-Himalayan regions: Kashmir, Ladak, Kinawer, Kistewar, and Sirmor. It is the general form of polyandry in Ceylon. It is the form which Humboldt found among the red-men. “Among the Avaroes and the Maypures,” he says, “brothers have often but one wife.” It is the form which Cæsar found among the Britons. “Uxores habent deni duodenique inter se communes, et maxime fratres cum fratribus, et parentes cum liberis; sed si qui sunt ex his nati, eorum habentur liberi a quibus primum virgines quæque ductæ sunt.” And to show that the two forms of polyandry are stages in a progress, we repeat that almost everywhere, outside Tibet, we find the lower form accompanying the higher. In some quarters the lower only is known—as in Kasia and among the Nairs. In others—Kooloo, for example—the lower form is prevalent; the higher also is known, but is exceptional. Again, in numerous quarters the higher is the general form, and the lower the exceptional—as in Ceylon; and lastly, in some quarters, as in Tibet, we lose sight of the lower form altogether. The higher polyandry has become a national institution.

And finding the higher polyandry a national institution, we observe that we are in a position to show that most probably polyandry formerly prevailed over a still vaster area than that within which we have hitherto found it. We have seen that with polyandry, of the Tibetan type, wherever it was long and generally established, kinship through males must have been introduced; the father’s blood, though not the father, being certain, where the wife was faithful. We have also seen, in the case of the Britons, that the children of the woman were accounted to belong to the husband who first espoused her; and that in Tibet, the right of choosing the wife belongs to the eldest brother, to whom, also, the children of the marriage are held to belong. We must now, to obtain what we have been in search of—a test of the former presence of polyandry—look at the Tibetan form of polyandry in a state of decadence. We find it in such a state in Ladak. “In Ladak,” says Moorcroft, “when an eldest son marries, the property of his father (more properly the family estate) descends to him, and he is charged with the maintenance of his parents. The parents may continue to live with him, if he and his wife please; if not, a separate dwelling is provided for them. A younger son is usually made a Lama. Should there be more brothers, and they agree to the arrangement, the juniors become inferior husbands to the wife; all the children, however, are considered as belonging to the head of the family. The younger brothers have no authority; they wait upon the elder as his servants, and can be turned out of doors at his pleasure, without it being incumbent upon him to provide for them. On the death of the eldest brother, his property, authority, and widow, devolve upon his next brother.” And that whether the younger brother has agreed to the polyandrous arrangement or not. He has a customary right of succession to his brother’s property, and to his widow, and he cannot take the one without taking the other.

Here we are brought to consider the meaning and origin of the legal obligation which we find laid on younger brothers, among certain peoples, to marry in their turn the widow of their deceased elder brother. There can be no doubt that that obligation was in its origin the counterpart of a legal right of succession. It is so with the Kirghiz, Aenezes, and Mongols—the next brother being heir even where the elder leaves issue.

When history begins, the Hebrew law preferred the issue to the next brother; but when he or the next of kin succeeded, it was on the old footing. This is clear from the book of Ruth. The hereditatis emptor of the deceased took to wife at the same time his widow, “to raise up the name of the dead upon his inheritance.” The obligation to marry the widow was the counterpart of the right of succession. And we can see the connection between the obligation and heirship dropping slowly out of view. In Deuteronomy it is provided that the husband’s brother shall “perform the duty of an husband’s brother” to the widow, only when the brethren dwell together, and one of them dies childless. The obligation is here presented pure—as a duty falling on the brother, which it was disgraceful to neglect.

In India, by the time when the Institutes of Menu were compiled, the obligation was laid on the brother only in case the deceased left no son. Grave doubts had arisen as to the extent and propriety of the obligation, the number of sons to be begotten on the widow, and the terms on which the brother should live with her. “The first object of the appointment being obtained, according to law, both the brother and the widow must live together like a father and daughter by affinity.” Again, it is doubted whether the obligation extends to the twice-born classes. “Such a commission to a brother or other near kinsman is nowhere mentioned in the nuptial texts of the Veda.... This practice, fit only for cattle, is reprehended by learned Brahmans; yet it is declared to have been the practice even of men while Vena had sovereign power.” Yet elsewhere in the code the obligation is contemplated as legal, and provision is made for the rights of succession of the issue of the Levirate union. “Should a younger brother have begotten a son on the wife of his deceased elder brother, the division (of the estate) must then be made equally between that son, who represents the deceased, and his natural father: thus is the law settled.” We repeat, that in Menu’s time the obligation had not only been, to some extent, dissociated from the corresponding right of inheritance, but was falling into disrepute. We see it also falling into desuetude among the Hebrews. In the earliest age the Levir had no alternative but to take the widow; indeed she was his wife without any form of marriage! By the Mosaic Law, however, he might get quit of her if he chose by submitting to the ceremony of “loosing the shoe.”

It is impossible not to believe that we have here presented to us successive stages of decay of one and the same original institution; impossible not to connect the obligation, in its several phases, with what we have seen prevailing in Ladak; impossible not to regard it as having originally been a right of succession, or the counterpart of such a right, derived from the practice of polyandry. Regarded as in its origin a right of succession, it exhibits the next younger brother as succeeding to the universitas of the elder—taking up all his rights and obligations—inter alia, his widow. But how came the right of succession to open, as in the ruder cases, to the brother in preference to the son of the deceased? We repeat, that the only explanation that can be given of this is, that the law of succession was derived from polyandry. The succession of brothers to one another, in order of age, is a feature of the law of succession under both forms of polyandry. Under the ruder, brothers succeed one another; and failing brothers, the sister’s children come in: under the less rude, brothers succeed one another; and failing brothers, comes in the eldest son of the brotherhood. And nowhere, excepting where there is or has been polyandry, have we such a system of succession—brothers succeeding in preference to sons.

The same conclusion is forced upon us from another point of view. In the lowest cases of polyandry the children belong to the mother; in the more advanced to the eldest brother (an approach towards agnation). Now the peculiarity of the obligation is, in all cases, that it was an obligation “to raise up seed” to the elder brother. The children begotten by the younger brother were accounted the children of the elder deceased. It is obvious that it could more easily be feigned that the children belonged to the brother deceased, if already, at a prior stage, the children of the brotherhood had been accounted the children of the eldest brother, i.e., if we suppose the obligation to be a relic of polyandry.

Curiously enough, Dr. Latham would invert the order of development by producing the ruder fact—polyandry—from the less rude obligation. But, clearly, this is an inversion of the order of nature—which is progressive—in which the ruder gives birth to the less rude, not the less rude to the ruder.

Assuming the correctness of this view of the origin of that obligation, we must hold that polyandry in the Tibetan form prevailed at one time throughout India. Among the race from which the ancient Hebrews were descended; among the Moabites and the ancient Persians. Among the Druses and all the Arab tribes in Syria; the Mongols, Ostiaks, Kirghiz Turks, and tribes of the Caucasus. Among the Makololo, and, we may believe, many other peoples in Africa. It is needless to repeat that we must also conclude that among the peoples just enumerated the Tibetan form of polyandry was preceded by the Nair, and, at a still earlier date, by utter promiscuity.

We have found polyandry in so many lands, among so many races, and in such phases of progressive development, that we are surely justified in classing it among the phenomena most distinctive of—the most likely to occur at—the earlier stages of the progress of any race of men. Its origin can only be ascribed to a scarcity of women as compared to men. And the vast area over which it anciently prevailed, can leave no doubt in the mind that in former times the balance of the sexes must have been seriously disturbed (artificially), and that we were right in predicating of the primitive groups that they usually contained fewer women than men. When the phenomena of exogamy—also due to a scarcity of women—are contemplated along with the phenomena of polyandry, the impression of this fact produced on the mind is almost as strong as the feeling produced by demonstration. To whatever extent a want of balance between the sexes prevailed, to that extent certainty as to male parentage was in the earlier stages of progress excluded. With polyandry itself there is uncertainty upon this point. In the lower cases the uncertainty is absolute. And regarding, as we must do, the higher as an advance from the lower, we are forced to conclude that wherever we have found polyandry, or traces of it, there must anciently have prevailed the system of kinship through females only.

In a preceding chapter we found the system of kinship through females only, universally prevailing among the Australian Blacks; prevailing among the majority of the nations of the American Red-men; and among the South Sea Islanders. That is, we found it among peoples now practising polygunia, and which have advanced far towards the patriarchal system. We infer that with these peoples the unions of the sexes were originally promiscuous or polyandrous. With regard to the Red-men, indeed, there is little room for doubting that they formerly all practised polyandry. It is now occasionally to be found among them, and their system of relationship—their names for kinsmen and kinswomen—point to its having been their universal custom. Mr. Morgan of Rochester, New York, whose account of the Indian nations we have already had occasion to refer to, gives the following as radical features of the system of relationship prevailing among them: 1. “All the brothers of a father are equally fathers to his children (this where there is now no polyandry). 2. All the children of several brothers are brothers and sisters to each other; all the grandsons of a man’s brothers are his grandsons.” These features of the system bear the stamp of a polyandrous origin; they are features of the system of relationship which might be expected to accompany the higher polyandry. The schedules returned to Mr. Morgan show that among the Tamul and Telugu, peoples of Southern India, numbering about twenty-four millions, “all the brothers of a father are usually called fathers, but in strictness, those who are older than the father are called great fathers, and those who are younger, little fathers.” And both the Tamul and Telugu are still, as we have seen, to some extent polyandrous. The same system of relationship is found among the Puharies, a people on the skirts of the Tibetan region, and that manifestly practised polyandry till a late date. With the Puharies, all the brothers of a father are equally fathers to his children.

We have seen that the Kasias, the Nairs, and the Saporogian Cossacks, have the system of kinship through females only. We find that system in Tulava, in the neighbourhood of the Nairs. “Among the Buntar”—the highest rank of Sudras in Tulava—“a man’s children,” says Buchanan, “are not his heirs. During his lifetime he may give them money; but all of which he dies possessed goes to his sisters and to their children.” The cause must be the same in either case, though marriage in Tulava has shifted from polyandry to polygunia. Among the Rajputs we have traces of the system of female kinship. The Kocch have kinship and succession through females only; and so have the But (Bodo). Farther, we find that system among the Banyai, in Ashanti, Aquapim, and Congo, and are assured that traces of it are to be found all over Africa. We have already had occasion to notice its occurrence among the Chinese.

Let us now see what evidence there is of the former existence of the system of kinship through females only. We recal the fact that, in an earlier chapter, we saw reason to believe that it anciently prevailed among the Celts. We infer that among the Celts there was anciently no certainty of male parentage. We now notice that we find traces of such a system in India in the Sutras of Gautama. In these, marriage with the daughter of a maternal uncle—a cousin on the mother’s side—is emphatically prohibited as being clearly against the principles of the sacred writings. Such a prohibition, found with an exogamous race—and almost all the Indian races were and are, as we have seen, exogamous—can be referred only to the system of kinship through females only. And it is impossible to avoid connecting with this the tradition that the five Pandava princes—brothers so called—were husbands of one wife. “How is it,” asks Max Muller, in discussing the character of the Mahabharata, “that the five Pandava princes who are at first represented as receiving so strictly Brahmanic an education, could afterwards have been married to one wife? This is in plain opposition to Brahmanic law, where it is said ‘they are many wives of one man; not many husbands of one wife.’ Such a contradiction can only be accounted for by the admission that in this case epic tradition in the mouth of the people was too strong to allow this essential and curious feature in the life of its heroes to be changed.” In other words, we have here the tradition “that the races among whom the five principal heroes of the Mahabharata were born and fostered,” practised polyandry. This is confirmed by all that is related of the Pandava princes. They were the reputed sons of Pandu,—but, in fact, three of them were sons of one of his wives by three different gods, and the other two were sons of another wife by the Aswini-Kumaras. Pandu himself was the son of a marriage with a brother’s widow. When the five princes married one wife, the eldest was first married to her by the family priest, and then the other four in their order, according to priority of birth. The princes are represented as living in family with Kauli, their mother—the head of their house. In the poem, Bishma, their granduncle—grandfather’s brother—is often styled their grandfather; and though Bishma was really the uncle—father’s brother—of Pandu, he is sometimes styled his father. All these circumstances point to a system of polyandry of the Tibetan type. The very terminology is that of polyandry, and which polyandry has left behind it among the Tamul, the Telugu, the Puharies, and the red-men of America. In short, though the original tradition has obviously been tampered with, enough of it remains to oblige us to acknowledge it as a genuine tradition of a stage of Aryan civilization, when the marriage system was polyandrous as it is now in Tibet. It is almost needless to point out that we have, in this tradition, a confirmation of our view of the origin of the obligation which, in the code of Menu, is recognised as imposed on brothers in turn to marry the widow of a brother deceased. We shall find a further confirmation of that view in the case of the Hebrews.

We are not without evidence of the existence in early times of the system of female kinship among the Semitic races. It would appear that while Abraham still lived, his tribesmen as yet recognised only that primitive kinship in some important relations in life—e.g., as affecting the right of intermarriage. Between the times of Abraham and the promulgation of the Levitical law, a complete revolution took place in Jewish custom. The patriarch himself married his sister-german, or by the same father; and his brother Nahor married his niece, the daughter of a brother. So Amram, the father of Moses and Aaron, married his father’s sister (Exod. vi. 20). These women were not relatives, in a full legal sense, of their husbands. They were connected with them through males only, and through males in those times there was not, as yet, a perfect kinship. We have similar evidence of the existence of the system of kinship, through females only, among the Phoenicians.

Among the Greeks, traces of this early law remained in historic times. To pass over the tradition that in Greece before Cecrops children always bore the names of their mothers, we have the fact that at Athens a brother might marry a sister-german, but not a sister-uterine or consanguineous. Here again we have a relic of the doctrine that a child had no paternal relatives. A sister-uterine was a near kinswoman, but a sister-german was no kinswoman at all. Montesquieu ascribes this Athenian rule to a device of the legislature for regulating successions; but he belongs to the class of philosophers who make more of enactments than of popular usages. As Bunsen has pointed out, there can be no doubt that the true meaning and origin of the rule were what we have indicated.

There is one case which might be cited to throw doubt upon some of the conclusions at which we have arrived in this chapter. This is the report of Philo, that the Spartans allowed a man to marry his sister-uterine, but not his sister-german, or by the same father. This may have been circulated for the sake of the contrast which it presented to Athenian custom; at any rate, we hold it to be incredible—as discordant with old law as with the habits of the Lacedæmonians. It is beyond belief that there was this superior regard for the father’s blood in ancient Sparta, where the marriage tie was so loose that men lent their wives to one another, and cared little by whom children were begotten, provided they turned out strong and healthy. It is incredible, that in a community where any sort of importance was attached to blood, the unquestionable blood tie between children of the same mother should be so disregarded. If we are to credit the report at all, it must be on the supposition that the Spartans were exceptional in their development like the ancient Persians (from whom the Druses derived their customs). And we do not regard the case of the Persians as of weight against our reasoning, but the contrary. The Persian customs were just those of hordes who consecrated an incestuous promiscuity into a system. If they allowed the marriages of brothers and sisters consanguineous, they also sanctioned the unions of sons and mothers, and of fathers and daughters, and in some cases required them for the purposes of religion.

At the outset of our argument we saw that if the system of kinship through females only could be shown to exist, or to have existed, it must be accounted a more archaic system of kinship than the system of relationship through males,—the product of an earlier and ruder stage in human development; and that to prove its existence on such a scale as to entitle it to rank among the normal phenomena of human development would be to prove it the most ancient system of kinship. We now submit that we have amply established our proposition. We have collected abundant evidence of the non-existence in many places of the conditions necessary for the rise of kinship through males; in many of these cases—some of them cases of great populations—we have been able to adduce evidence of the existence of the system of kinship through females only. We have seen that polyandry must be accepted as a stage in the progress towards marriage proper and the patriarchal system. The lower forms of polyandry we have found to be accompanied by the system of kinship through females only. We have seen polyandry change its form till it allowed of kinship through males, and then die away into an obligation on younger brothers in turn to espouse the widow of the eldest brother; and in some cases, Indo-European as well as Semitic, in which we found that relic of polyandry, we have found, or found traces of, the system of kinship through females only. Had the facts bearing on our inquiry ever been systematically observed, noted, and collected, it is probable our case might be made to appear stronger than it does. But as it is, we submit that we have done quite enough to establish the truth of our proposition.

Before leaving this subject we would observe that, whether the system of kinship through females only prevailed universally at the first or not, it must have prevailed wherever exogamy prevailed—exogamy and the consequent practice of capturing wives. Certainty as to fathers is impossible where mothers are stolen from their first lords, and liable to be re-stolen before the birth of children. And as exogamy and polyandry are referable to one and the same cause—a want of balance between the sexes—we are forced to regard all the exogamous races as having originally been polyandrous. While polyandry supplied a method whereby the want of balance might be the less felt, and may thus have retarded, and in some cases prevented, the establishment of exogamy, wherever exogamy took root polyandry must have been practised. Therefore we must hold it to be beyond dispute that among exogamous races the first system of kinship was that which recognised blood-ties through mothers only.

We may be pardoned for here adverting to the views of ancient kinship advanced by Mr. Maine. We have already pointed out that Mr. Maine seems not to have been able to conceive of any social order more primitive than the patriarchal. And as he found agnation—or kinship exclusively through males—to be a common concomitant of the patriarchal system, he has committed himself to the opinion that that was the only kinship known to primitive times. He argues, indeed, against the possibility of kinship through females in early times as being inconsistent with social order and stability. The learned and ingenious writer must be held to have taken up the threads of legal history where they began to unwind themselves, of new, after the completion of a social revolution. It is quite undoubted, as he says, that few indigenous bodies of law belonging to communities of the Indo-European stock do not exhibit peculiarities which are referable to agnation. With the advance of society—the growth of marriage laws—the superiority of the male sex must have everywhere tended to establish that system. But, before that result could be reached, many stages of progress had to be traversed. And while traces of agnation are to be found in the early customs of most of the Indo-European races, we have seen that the indigenous customs of most early communities—whether of the Indo-European, Turanian, or Semitic race—exhibit peculiarities intelligible only on the supposition that kinship and succession through females were the rule before the rise of agnation. Farther, we have seen that wherever non-advancing communities are to be found—isolated in islands or maintaining their savage liberties in mountain fastnesses—there to this day exists the system of kinship through females only. The state of old, says Mr. Maine, recognised as its units not individuals, but families. True. But at a yet older date we must conclude that neither the state, nor the family, properly speaking, existed. And at that earlier time the unnamed species of kinship—the counterpart and complement of agnation—was the chief determinant of social phenomena.

We now go on to show—

II.—That the primitive groups were, or were assumed to be, homogeneous.

It appeared at the outset that individuals must have been primarily affiliated not to persons but to groups, and that the first effect of the rise of the idea of kinship must have been to give birth to the conception of stocks; farther, that the establishment over any great area of the system of kinship through females only must have occupied a considerable period of time. Until that system was firmly established, there could be no such interference with the homogeneity of the groups as to be worth consideration. An amount of heterogeneity short of that which would introduce at least the germ of a system of betrothals may fairly be overlooked. While as yet there was no system of kinship, the presence of captive women in a horde, in whatever numbers, could not introduce a system of betrothals. Heterogeneity as a statical force can only have come into play when a system of kinship led the hordes to look on the children of their foreign women as belonging to the stocks of their mothers; that is, when the sentiments which grew up with the system of kinship became so strong as to overmaster the old filiation to the group (and its stock) of the children born within it. We may depend upon it that this was a stage of progress which it took long to reach, and thus that it was long before the original homogeneity of the groups was substantially impaired. That in the stage of progress we are contemplating, adoption was practised (the adoption of one group by another to which some writers ascribe such great effects), is altogether unlikely. If it was, it would most probably—as in later times—proceed on the fiction that the uniting groups were of the same original stock. But looking to the state of hostility between groups at the stage we are considering, and the degree of advancement implied in the conception of adoption, we cannot believe that the groups then tended to amalgamate, however they may have tended to divide. We conclude that we must regard the primitive groups as having been, or having been assumed to be, homogeneous up to that stage when, through the joint operation of exogamy and the system of kinship through females only, foreigners recognised as such began to be systematically born within them.

III.—The system of kinship through females only tended to render the exogamous groups heterogeneous, and thus to supersede the system of capturing wives.

We may here be very brief. We have already seen the effects of the joint operation of exogamy and this system of kinship among the Australians. Indeed, what their effects must have been is exceedingly obvious. Owing to exogamy, the mothers in each horde were foreigners, and, owing to the system of kinship, the children born to them were esteemed foreigners also. Thus, so far as the system of infanticide allowed, the hordes contained young men and women accounted of different stocks, who might intermarry consistently with exogamy. Hence grew up a system of betrothals, and of marriage by sale and purchase. In Australia and America we saw that in spite of the law of blood-feud, the heterogeneity is now such that the system of betrothals is well established, and that of the original stock-groups the names alone appear to remain.

IV.—As civilization advanced, the system of kinship through females only was succeeded by a system which acknowledged kinship through males also; and which in most cases passed into a system which acknowledged kinship through males only.

It is obviously needless to say anything in support of the first branch of this proposition. The difficulty was to show, as we have done in our first proposition in this chapter, that there was a more archaic system which did not acknowledge kinship through males. With the fact of kinship through males in advanced communities, every reader is familiar. Farther, as to the second branch of the proposition, it is unnecessary to adduce evidence, or to do more than give some explanations. Those who are acquainted with Mr. Maine’s (in many respects admirable) chapter on primitive society and ancient law, will see from the terms of our proposition, that we have not altogether adopted his view that agnation at one time or other prevailed everywhere in the advancing communities; but it is beyond dispute that its prevalence was most general. As it will be convenient hereafter to speak of agnation as a familiar system, we must here say something of its nature. This system, as it long prevailed in Rome, may be best explained by using the terminology of Roman Jurisprudence—to which indeed its name belongs. Its general description is that it embraced only the ties of blood through males. But it will be well to see who were thus included in the kindred. Those united by ties of blood through descent from the same married pair being called cognates; the agnates were those cognates who traced their connection exclusively through males. By a fiction, adopted persons and their descendants through males were within the agnatic bond. All the children of a married pair were agnates, as well as all the grandchildren through sons, but the grandchildren through daughters were not in the number of agnates. The children of the same father by different mothers were kindred, but the children of the same mother by different fathers were not relations to any legal effect. The sons of brothers were kinsmen, but the sons of sisters or of brother and sister were no relations; for a woman’s children were held to be not of the kin of their mother but of their father. And in no case was there a tie of kinship between a woman’s children and her natural relatives unless there was an affinity between her and her husband. It is needless to say that the cognates who remain over, after deducting the agnates, are those who would be relatives under the system of kinship through females only. If the one system involved anomalies so did also the other. Where female kinship prevails, a Rajah’s son may become a hodman—taking the state of his mother—while the son of the Rajah’s sister mounts the throne. The nephew—a sister’s child—is a relation of the Rajah, but his son is none at all. No more is his brother’s son; for through a male under that system there is no blood-tie.

Under each system, while it prevailed, the effects of kinship were confined to those who according to it were relations. And often it is in the laws of succession that we find the best evidence of the former existence among a race of either system. The rule preserved in the customs of Normandy, which prohibited uterine brothers from succeeding to one another’s lands, attests the former prevalence there of agnation; and, in some quarters, as in Congo, the descent of the crown from the uncle to the sister’s son is nearly all that remains to witness to the former prevalence of the system of kinship through females. It will be a curious chapter in history which successfully narrates the progress of the revolution by which the passage from the earlier to the later of these systems was effected; exhibiting the stages in the development of the family system, as based upon the patria-potestas, and of agnatic kinship as deduced therefrom.

Let us see whether we cannot in a few sentences suggest some of the steps in that progress. The reader must suppose the progress to commence as soon as, through the joint operation of exogamy and the system of female kinship, the groups have been rendered so far heterogeneous as to permit of marriage within the group.

Children having been affiliated to mothers instead of to groups, the first approach to a family system would be through a separation of residences—all of a group having no longer a common haunt or dwelling, but at first all of the same stock within a group associating as a gens or house; and next, mothers and their children occupying separate homes. With this separation of residences would come a closer knitting together of the kindred first in the gens, and next of mothers and children in the family—rude proprietary rights distinct from the tribal, distinct from the gentile—in the common home, weapons, and garnered food. There would be introduced such a species of family system as we find among the Nairs—the rudest that can be imagined. And from this to the family system, peculiar to polyandry of the Tibetan type, we have seen the stages of development.

In the Nair stage, kinship would be of importance chiefly in two respects—(1), as determining the right of intermarriage; (2), as determining the right of succession. It might be expected that the system of kinship through females only would first lose importance in regard to successions.

While the Nair family system lasted, we may assume that the common home of the brothers and mother would not often be such as to permit (conveniently) of a general succession—failing the brothers—of all the sister’s children, where there was more than one sister; and that the first advance to a restricted system of succession would be through the limitation of the right of succession, primo loco, to the children of the eldest sister. We have seen the practice growing of fathers making gifts, inter vivos, to women’s children whom they had reason to think their own; and as this practice grew with the number of cases in which there was a degree of certainty of male parentage, there would be a farther practical restriction of the right of succession through females. And with the practice of gifts, inter vivos, to putative children, would grow a feeling against allowing estates to pass from the house of the brothers to that of, or to the putative children of, the polyandrous husbands of their sister, and a corresponding disposition towards a system of marriage which would allow of the property passing to the brother’s own children. The system which would suggest itself would be the Tibetan system—to have a wife in common in the house of their mother. This system would produce certainty of the children being of their own blood; they would be born in the house, and would become its heirs.

The next step in advance is obvious. The succession of the younger brother to the elder was a feature of the earliest law—sister’s sons only succeeded failing their uncles. Now, everything conspired to invest the eldest brother, when he came into the succession, with some of the attributes of a paterfamilias. This he did only on the death of all the polyandrous fathers. But—in his relation to his younger brothers—in respect of his being the first to marry, reaching puberty first, and choosing the future wife of himself and brothers—in respect of the first-born, and frequently more than one of the children of his marriage being unquestionably his offspring—it was natural that the fiction should be formed that his were all the children. Women had already, and in the recourse to Tibetan polyandry, been deposed from the sovereignty and management of families; and now, with additional guarantees for the wife’s fidelity, parentage was either become certain, or feigned to be so; the elder brother was a sort of paterfamilias, the right to succeed him being in his younger brothers in their order; after them, in their eldest son. Thus, the idea of fatherhood—formed under the system of Nair polyandry—attained something like maturity under the Tibetan, and took its place in customary law. And so far as it was a step towards, or accompanied by kinship through males, it was a step away from kinship through females, and especially as regards rights of succession.

Apart from such certainty of fatherhood as was incidental to the marriage of eldest sons, the earliest examples of such married life as would give certainty of male parentage would probably be furnished by the chiefs of tribes, who might have the power to secure to themselves one, or perhaps several wives. In Kandya, Ceylon, where polyandry is universal among the lower and middle classes, the chiefs are strictly monogamists, apparently regarding polyandry as a low practice, unworthy of men in their position. As settled habits arose, as property accumulated, and the sexes became more evenly balanced, the example of the chiefs would find more and more imitators, and their cases would furnish a model for an improved system of succession. Thus would arise a practice of monogamy or of polygamy. Brothers would not now always be co-husbands; the Tibetan form of polyandry would die out, and the marriage of a brother to his elder brother’s widow would become first an act of succession necessary for the assumption of the brother’s place as head of the family; next, as the succession of sons was introduced as in right prior to the brother of a father—chiefly through the brothers leaving the house and contracting separate marriages—it would become an obligation founded on usage, and which, being unproductive of material advantages, would not unfrequently, from men’s other marriages, be found irksome and inconvenient. Finally, the obligation itself would die out under the influence of ideas of propriety, which grew up with the improved marriage system. In the Institutes of Menu the obligation is seen in a state of decadence under the influence of such ideas.

Paternity having become certain, a system of kinship through males would arise with the growth of property, and a practice of sons succeeding, as heirs direct, to the estates of fathers; and as the system of kinship through males arose, that through females would—and chiefly under the influence of property—die away. The cases of Abraham and Nahor, however, show that it would be long before that system ceased to be influential as regards intermarriages; that it might, as regards them, linger to some effect even after men had reached the patriarchal state. From the patriarchal state, with such a customary law as prevailed in Abraham’s time, to the system of agnation as it prevailed in Rome, is still a long progress. Every step in it, we may be sure, was affected by considerations derived from property. While wives were captured, if there was any sense of property at all, wives would be regarded as property. When at a later stage they came to pass from the houses of their birth into alien houses—by purchase—they would still be property. And with the wives considered as property, it is easy to conceive how there would have arisen a sense of property in children. Hence, additional features of the patria potestas. And when a woman had been sold to her husband by her father, and had thus come to be considered the husband’s property, it is easy to see how neither her original family, nor that into which she had married, should be able to inherit any property through her. But the right of inheritance, as property became abundant, tended to become, and did become, the test of kinship. And in course of time, the notion of kinship derived through females would disappear among a people who cared nothing for a kinship barren as regarded patrimonial advantages. The result would be the system of agnation.

It is not necessary for us here to do more than repeat that we have, in numerous cases, found agnation, or at least kinship through males, preceded by the system of kinship through females only. It was so in the cases of the Hebrews, Hindus, Celts, and Greeks; it was presumably so in all those cases in which we find kinship through males accompanied by that relic of polyandry—the obligation laid on younger brothers in turn to marry the widow of the elder brother deceased. And since we have shown special cause for believing that all the exogamous races had originally the system of kinship through females only, we are entitled to assume that it was so among those exogamous peoples with which, as with the Khonds, the Circassians, and the Kalmucks, we find relationship to be agnatic.

V.—The system of kinship through males tended to rear up homogeneous groups, and thus to restore the original condition of affairs among exogamous races, as regards both the practice of capturing wives and the evolution of the form of capture.

The first effect of kinship through males must have been to arrest the progress of heterogeneity. The introduction of foreign women into a tribe no longer brought into it children accounted foreigners; for either the children were no longer of the mother’s stock, but of the father’s, or, if of the mother’s, they were yet of the father’s also. Where, then, in a tribe, a balance of persons of different stocks had not been reached, it was henceforth unattainable. Farther, with kinship through males would arise the habit of feigning a common descent from some distinguished man—a fiction which would lead in many cases to the denial or neglect of such heterogeneity as existed. Of the new groups that were formed, the homogeneity was perfectly secured. The family now tended to grow into the tribe of kinsfolk. The children born to a polygamist husband were all kinsfolk, of whatever stocks their mothers were. The children of brothers, though they married women of different stocks, were kinsfolk. And however the family increased by the addition of generations, its members were all within the kindred. And that, as well where the exogamous prejudice survived as where it perished. Where it survived the women of a family could find no mates within its bounds, and marrying into other groups would follow, and their children with them, the kindred of their husbands. They would be out of the group. Thus, within such exogamous groups as were remodelled, and within such new exogamous groups as were formed, under this species of kinship there could be no marriage within the group. There would be no place for a system of betrothals; and except where friendly relations subsisted between the groups to allow of marriage by purchase, their members would once more be able to get wives only by capturing them. Thus, even if, in the first stage, the system of capture had—as we see in Australia that it partially has—been superseded, the exogamous races, in entering on a new phase of advancement, had reserved for them a farther experience of that system, to confirm or re-establish the old association between marriage and the act of rapine. And we cannot doubt that many exogamous peoples have had this twofold experience. We know of several exogamous races which, after having had kinship through females, had kinship through males; and we cannot doubt that the same was the case with the other exogamous races that have had the form of capture and agnatic relationship. And indeed, as in the later stage, the experience must have been more uniform and continuous, there being nothing, in the absence of friendliness between the groups, to interfere with the system of capture, so it is observable that the form of capture is now most distinctly marked and impressive just among those races which have male kinship. It might be doubted, but for the case of the Fuegians and traces of the symbol, as if of a thing decayed, occurring in America, whether the experience of the earlier stage could generate the form. There is no doubt it can, and has frequently done so; but the question whether it could have done so, might, on mere general reasoning, have been decided in the negative.

VI.—Under the combined influence of exogamy and the system of female kinship, a local tribe might attain a balance of persons regarded as being of different descent, and its members might thus be able to intermarry with one another, and wholly within the tribe, in consistency with the principle of exogamy.

This sufficiently appears from what has preceded; and it farther appears from what has preceded, that such a balance of the sexes as would render a group independent of other groups in the matter of marriage would more speedily be reached in respect of the practice of polyandry.

VII.—A local tribe having reached the stage contemplated in the last proposition, and having grown proud through successes in war, might become a caste.

It is obvious that the feeling of superiority to other tribes, concurring with independence of them as regards marriage, might lead a tribe first to avoid, and then to decline and prohibit, intermarriage with the tribes which it esteemed inferior. And a tribe with a marriage-law restricted by such a prohibition is a caste, or has made an approach towards being a caste. That castes have, in fact, been produced in this way, is rendered certain by the fact already referred to—that nearly all the Indian castes, from the highest to the lowest, are divided into gotrams or families, and that marriage is prohibited between persons of the same gotram, who, according to the rule of Menu, are shown by their common name to be of the same original stock. We hold that this at once shows the caste to have been composed of members of different original stocks, and the stocks themselves to have been originally exogamous. There can, we think, be little doubt that all castes of this description were formed by the processes which we have been explaining. The Kamilaroi among the Australians appear to be such a caste. And were the natives of Australia to be left to themselves, their system of kinship remaining what it is, we might expect hereafter to find among them numerous caste tribes of this description.

It is a rider on what has preceded that caste may appear at that stage of a people’s progress while they are yet polyandrous. And of caste among people at that stage we have several instances.

VIII.—On kinship becoming agnatic, the members of a caste, formed as above explained, might—yielding to a common tendency of rude races—feign themselves to be all descended from a common ancestor, and thus become endogamous.

On the appearance of kinship through males, much confusion must for a time have prevailed in the application of the principle of exogamy. Some marriages that were formerly allowable would become illegal: as, for instance, the marriage of brother and sister-german, or of the children of brothers; on the other hand, some marriages that were before illegal would become allowable, as for instance the marriages of sisters’ children. At least the marriages last mentioned would be consistent with exogamy where relationship became agnatic. And as the old rules thus became inapplicable new rules would be formed, which in some cases might ignore the principle upon which the old proceeded. At any rate it is manifest that, while the application of the principle in the new circumstances was in dubio, the fiction of a common descent from an illustrious ancestor, should it be put forward, would come in aid of the confusion to destroy or render obsolete the principle of exogamy. The members of the caste already restricted to marriages among themselves, and now feigning themselves all to be kindred, would become endogamous.

Indeed, all that is necessary for the production of an endogamous tribe is, that a caste tribe composed of members of different stocks should, anyhow, at any time, yield to the tendency to eponomy. We see, however, that in the stage of transition from the system of female kinship to agnation, or to a system of male kinship, there must have been a time highly favourable for the introduction of the fiction of a common descent, and of the destruction thereby of exogamy. Of the tendency of rude races to employ that fiction, it is unnecessary that we should say anything; it is familiar to all students of early history. Nothing is more common than to find the belief in a common descent among peoples obviously heterogeneous; sometimes the belief is found even in communities which are not only heterogeneous, but the composition of which is known to be entirely artificial.

And if we anywhere find the form of capture among an endogamous race, as we do among the Bedouin Arabs, and appear to do among the Hebrews, it is not, we think, too much to say that the presence of the form is confirmatory of the supposition that the race became endogamous through employing this fiction, and by the processes which we have been explaining. At least this is the only explanation which we can offer of the appearance among an endogamous people of the form of capture.

We have now gone over the ground laid out at the close of the last chapter. It will be seen that some of the propositions mutually support one another. For example, the observed heterogeneity of certain castes which are subject to the rule of exogamy, goes to show that their ancestors must have had the system of kinship through females only; for exogamy, by itself, will not explain the welding together, in a group, of persons of different original stocks. And to those who, from the earlier chapters, have formed the opinion that we are right in our theory of the origin of the form of capture, the appearance of the form among an endogamous race will strengthen the supposition that many endogamous races, which have not—or which we do not know to have—the form, may originally have been exogamous.

We may fitly close this chapter with some surmises—thrown out for what they are worth—as to some of the effects on early society of the law of blood-feud which exists everywhere, so far as we know, among rude races, and which of course grew up with, and out of kinship. So far as the law of blood-feud retarded the production of heterogeneity within the groups in any district, it was unfriendly to progress. From another point of view it appears that it must have favoured progress, especially among exogamous races, at that stage when kinship was through females only. It bound all the kindred to avenge the death of any one, and the obligation was a point of religion. At first, probably, the protection to the person which this law afforded may have extended only to adults, but in time it came to be extended even to infants. This extension indeed was a logical necessity. And when infants came within the benefit of the law, their lives must often have been spared to avoid the blood-feud with their mother’s kindred—a body of protectors, as we have seen, usually living outside and foreign to the gens or house of birth. Thus the law of blood-feud must be credited with a mitigation, perhaps in some cases with the suppression, of infanticide, male as well as female, in exogamous societies, at that stage when kinship is through mothers only. And by checking this practice it tended to restore the balance of the sexes, to allow of the rise of polygunia and the decay of polyandry. It is a curious fact that nowhere now, that we are aware of, is infanticide a system where exogamy and the earliest form of kinship co-exist.

When, however, with agnation, groups became homogeneous, containing none but kindred, and containing in fact all the kindred, this beneficial action of the blood-feud must have ceased. Where necessity or convenience prompted to infanticide among agnatic groups, the law of blood-feud opposed no impediment to the practice. If the children perished it was at the hands of their kindred. Accordingly the most impressive systems of infanticide—chiefly systems of female infanticide—now existing, occur among exogamous races which have male kinship.

On the one hand, then, the law of blood-feud would seem to have played an important part in introducing monogamy, polygunia, and the patriarchal system; on the other hand, it would seem at a later stage to have favoured the perpetuation of exogamy, and of systems of infanticide.