By far the greatest number of primitive peoples are monogamous. Only in relatively few cases is there polyandry. Polygyny often occurs among persons who are specially favoured, either economically or socially; but it is nowhere the form of marriage of the majority of the population. The polygyny reported among certain tribes generally refers only to chiefs, magic doctors, or some other special persons who have more than one wife. Sexual group communism at the side of monogamy or polyandry has been found in various places, but it is wrong to speak of it as "group marriage." This is evident from the previously quoted examples of the pirauru in Australia, the sex communities among the Chukchee, the Nandi, Masai, and others. It is possible, of course, that monogamy which now co-exists with certain cases of sex communism may have been a later addition, but this is not proven. It is more likely that the pairing instinct (not identical with the instinct of procreation) is characteristic of our sub-human ancestors. In fact, even in the animal world there are numerous examples of monogamy (P. Deegener).

It has been established that in Africa, Indonesia, Melanesia, and elsewhere, the small children remain with their parents, while the bigger children are lodged together in special boys' and girls' houses, and are, as it were, brought up communally. The relationship of the children to their own parents is not notably closer than that between them and other persons of the same age class. We must not look upon this child communism solely as a curiosity, but as the relic of a very ancient primitive institution. Most likely there is some connection between child communism and the interchange of children which is customary, for example, among the Dravidian races of India ("Ethnographical Survey of the Central India Agency") and on the Murray Islands, in the Torres Straits (Australia). According to W. H. R. Rivers (1907, p. 318), the interchange of children between families is very frequent here without the peoples being able to give any explanation of it. Nor do other social and religious institutions offer any indication as to the origin of this custom. Rivers surmises that it has been preserved from a social organisation in which "children were largely common to the women of the group so far as nurture was concerned." At any rate, this adoption en masse will help civilised man to understand that less civilised peoples have ideas about parenthood different from those that exist among us, and also that group motherhood is not absurd. The existence of group motherhood among primitive communities—whose members were much more dependent on each other in the struggle for existence than are the members of much more advanced societies—must often have been of considerable advantage to these communities. On the assumption of "group motherhood" it is easily explainable that children use the same mode of address for their own sisters and brothers as for all the other children of the group, and that all the women of equal ages are called "mother." Hence the classificatory system of relationship ceases to be puzzling. It becomes clear why under this system whole groups of persons designate each other as husbands and wives, and why the children of all the persons of these groups call each other brothers and sisters, etc. The assumption is justified that man in a low state of civilisation knew only group relationship; further distinctions were derived only later from these relationships, the present-day classificatory system arising ultimately from them. Among the peoples where Rivers could examine this system there were indications of a development in the direction of using it rather for the distinction of real blood and marriage relationship than for the distinction of social position, for which it was originally intended. A connection between marriage regulation and the classificatory system of relationships exists not only among the Dravidian races, but also among the North American Indians, and certainly among other branches of the human race. Rivers says: "The classificatory system in one form or another is spread so widely over the world as to make it probable that it had its origin in some universal stage of social development"; and further he says: "The kind of society which most readily accounts for its chief features is one characterised by a form of marriage in which definite groups of men are the husbands of definite groups of women." Rivers does not mean thereby institutions like the pirauru, but a permanent group marriage. It may be objected against this latter assumption that permanent (not occasional) sex communism does not necessarily need to be connected with communism of children. It is quite possible that monogamy and child communism may exist side by side, as, e.g., among the Murray Islanders.

But even if group marriage did really exist in some places, and if the existence of child communism would prove this, it still cannot be asserted that it is a phase of development through which all human races have passed. For the assumption of a parallel development of all races is untenable. It is true the basic psychic organisation is the same for all human beings, being due to the common descent of mankind. But owing to the continual adaptation to changing environmental conditions, it was not preserved, but underwent different changes. There is no ground for the assumption that, while environmental changes brought about bodily modifications, mental changes did not take place also, therewith leading at the same time to differences in social culture. On the contrary, we must rather assume that together with anthropological variations among the races there also arose variations in social development, the different civilisations resulting from differentiated mental dispositions and deviating more and more from each other. Certain elements of the original primitive civilisation have been preserved in the various later developments, but not everywhere the same elements, nor were the differentiations that did take place all of the same degree. Certain fundamental conceptions may remain unchanged for long periods, and may produce analogous phenomena in different civilisations. Since deviations from monogamy are extremely rare among primitive peoples, the assumption is justified that monogamy is one of the fundamental factors of human civilisation. How could its practically universal occurrence be explained otherwise? There can be no question of convergence, nor has a world-wide transmission of a cultural element that has arisen later been proved up to the present.

The opinion, first expressed by L. H. Morgan, that the classificatory relationship system is evidence of the existence of group marriage (not merely in the form of pirauru existing at the side of monogamy), is contradicted by the etymological meaning of the terms used by primitive people, which are generally translated by "father," "mother," "grandfather," "brother," "sister," "child," etc. These collective names show nowhere an allusion to procreation, but only to age differences: father and mother are the "elder," the "big ones," the "grown-ups"; the children are the "little ones," the "young ones"; brothers and sisters are the "comrades." We often find that among the Australian negroes and the South Sea islanders no distinction is made between father and mother. All persons of an older generation of a horde or a totem (or of a phratry respectively) are simply the "elder," the "big ones." If a native wishes to indicate more clearly the sex of a person of an older class, he must add the word "man" or "woman" (or the adjective "male" or "female"). It often happens that grandparents and grandchildren use the same form of address, which in no way refers to descent (Cunow). Other facts point to the same conclusion. Where the pirauru exists in Australia, the same form of address is used for persons standing in pirauru relationship to the speaker as for members of the same age class who have no such relationship. This could not be so if the appellation had originated from common sexual relationship. Cunow rightly concludes: "Sexual communities can be proved to exist here and there among primitive peoples, but the nomenclature of the classificatory relationships has not grown out of such group relationships. These so-called group marriages are rather adventitious growths, playing only a secondary rÔle in the history of the family."

Buschan (1912, p. 254) looks upon the pre-marital sexual freedom of girls among many primitive peoples (most probably among the majority of them) as a relic of communal marriage from earlier times. He assumes that the girls had promiscuous relationships with the other sex. This, however, is not the case. As a rule, couples meet together for a time, and only rarely does a person have relationship with several persons at the same time. The conditions are essentially the same as in Europe, except that amongst "savages" a love affair going as far as intercourse is not considered immoral. The assumption of many authors that man is polygynous is far from being proved, at least not in the sense that the majority of men are inclined to have relationship with several women at the same time. It cannot, however, be disputed that after some time the relationship between two people tends to lose its attraction, often causing a breaking of the marriage vow.

There is a custom among many peoples that a man's widow falls to his younger brother (or cousin)—the levirate. According to another custom, a man has the right to marry the sisters of his wife. Both these customs have been explained as being relics of a form of marriage in which brothers married several sisters or sisters married brothers at the same time (Frazer, II., p. 144). But it seems much more likely that we have here before us merely a case of property rights.

Even if constancy in marriage is not the rule, especially among primitive people, yet we must still regard the permanent living together of one man and one woman as a state that has always prevailed amongst human beings (Westermarck). Many of the speculations, at first sight so learned, about the apparently intricate paths in the development of marriage, remain merely speculations which cannot stand the test of modern ethnological research. Heinrich Schurtz (p. 175) makes the pertinent remark that nothing excited the hostile camps of the sociological idealists and naturalists more than the dispute about promiscuity in primitive times. While the one party painted with zest the indiscriminate and irregular sex relationship of primitive races, claiming it as an established original stage in human development, the adherents of idealism rose in indignation against a theory that places primitive man far below the level of the higher animals, and that leaves the riddle unsolved how such a chaos could lead to the idea of sexual purity and a spiritualisation of the sexual impulse. In this battle for and against promiscuity even facts were unfortunately too often not respected, attempts being made to disregard them at any cost. This cannot be good for the ultimate victory of truth. Facts should not be passed over, but should be taken into full consideration. In this conflict of opinions the institution of pirauru especially has fared particularly badly. Some anthropologists wanted to do away with it altogether at any price (for instance, Josef MÜller); others drew conclusions from it that are utterly unjustified. But even if this were not so, even if the pirauru could be used as a proof of previous sexual promiscuity, it still does not follow that it was a general custom in man, for the majority of the peoples show no trace of it.

First of all, it must be noticed that even the pirauru possesses various restrictions upon marriage with persons outside certain groups, which alone exclude unrestrained promiscuity. Furthermore, individual marriage, the binding force of which is undoubtedly even stronger and closer, is well known to exist beside it. There is a good deal of probability for the assumption of Schurtz that marriage regulations establishing the right of several men to one wife may first have arisen from mere friendly acts, or the original sexual licentiousness may have developed occasionally under specially favourable circumstances into the institution of pirauru, while at other places such a systematic development did not take place. It is easily to be understood that lower civilisations will show a looser standard of the marriage bond than those where many interests of a rich cultural development require the strengthening of this bond. Sexual needs may also have brought about the origin of the pirauru institutions. Thus there exist in Australia tribes among which the loan of wives was customary owing to the scarcity of women. There is only one step from this state of affairs to the pirauru. Among many tribes complicated marriage restrictions make a "legitimate" marriage very difficult, and this may easily lead to other sex relationships taking the place of marriage.

It is a mistake to assume hastily that customs among primitive people that appear strange to us must therefore be ancient and be relics of a primitive state. Every primitive race has a long history behind it, and it is not likely that it has remained static all the time. Primitive people are not stationary in development; there is much change among them in the course of generations. This applies also to customs and habits which seem absolutely stable. External conditions may produce new developments, or result in foreign influences. Not everything, therefore, that is peculiar to uncivilised races of the present day must be looked upon as primitive.

Polyandry deserves our special consideration. As a recognised social institution it has so far been definitely established only among the Indian peoples and castes, as well as in Tibet, on the borders of Northern India. In exceptional cases polyandry occurs among the Eskimos and the Asiatic Polar races. The older accounts of polyandry occurring in Australia are not confirmed by the new ethnographical literature. The reports about polyandry among the American Indians are also incorrect. John Roscoe (1907, pp. 99 et seq.) has proved its existence among the Bahima and Baziba tribes of Central Africa, though here polyandry is not the rule, but is only practised occasionally. If a man is poor, if he cannot get together the number of cows required for the bride price, or if he is unable to support a wife, he can combine with one or several of his brothers and take a wife in common with them. It is easy to get the women for this purpose. Furthermore, among these tribes the housewife may be claimed by a guest, while exchange of wives also occurs.

In India polyandry is prevalent among the peoples of the Himalayan mountains and among some Southern Indian tribes. Some cases of this curious form of marriage are already mentioned in the ancient Indian literature. It may be assumed, therefore, that it was more prevalent formerly than at present. This institution was certainly never very general nor of great importance in the life of the people of India. At the present time it is restricted to a number of comparatively small tribes and castes. Two forms of polyandry can be distinguished among them, namely, the fraternal form, where several brothers or cousins have one wife in common, and the matriarchal form, where a woman has several husbands, not necessarily related to each other.

In Northern India polyandry is general among the Tibetans and Bhotias of the Himalayan border districts. Here, when the oldest of several brothers takes a wife, she has the right—but not the duty—to have sexual relationship with the other brothers living in the same household. If a younger brother also marries, the other still younger brothers have the choice in which household they wish to live. The surplus women become nuns. This system is said to be due to the poverty of the country. The Himalayan peoples, being intent on preventing the increase of the population and a further reduction of the means of existence, consign many women to celibacy and childlessness. Yet at the same time they make it possible, by this system, for the socially privileged man to satisfy his sexual needs. The children of polyandrous marriages belong, as a rule, legally to the oldest brother. But it also occurs that each brother in turn, according to his age, has a child assigned to him regardless of whether the brother concerned was on the spot at the time of the child's conception. Sometimes the mother has the right to name the father of each of her children.

Fraternal polyandry also exists in Cashmir and among certain Sudra castes of the Punjab mountains. In the Punjab, however, the Rajputs and other castes of that neighbourhood are also influenced by polyandry. The ceremonies which take place at marriage in the Punjab bear traces of "marriage by capture." The dwellings of the polyandrous castes of this district consist of two rooms, one for the woman and one for the group of brothers. In Tibet, as also among the polyandrous Southern Indians, they have, however, mostly one room. The surplus women in the Punjab become objects of commerce. In the native State of Bashar, for instance, an active export trade is carried on with the surplus women, for whom sums up to 500 rupees are given.

Among the Dyats in the Punjab, the Gudyars in the United Provinces, as among all the Hindu castes in the mountain districts of Ambala, polyandry existed until lately; but it is said not to do so there any longer. In Ambala not only brothers, but also first cousins, were considered to be husbands of the oldest brother's wife.

Further, in East India the Santal caste (2,138,000 persons in Bengal, Bihar and Orissa) is the only community among which a similar custom exists. Among the Santals not only have the younger brothers access to the wife of the older brother, but the husband also may have relations with the younger sisters of his wife. This state of affairs may perhaps be looked upon as sexual communism among a small group. In Ladakh, too, and in other places of Cashmir, the wife common to several brothers may bring with her her sister into the marriage as co-partner. In the Punjab the fraternal husbands may also marry a second and third wife.

Among Indian migratory labourers it seems to have been formerly the rule that the brother remaining at home served as a conjugal substitute for the husband temporarily absent. Nowadays this custom has almost disappeared.

In Southern India polyandry is a recognised institution among the Toda and Kurumba of the Nilgiri mountains, as also among a number of the lower castes, especially on the coast of Malabar. Here polyandry and polygyny occasionally co-exist side by side.

The polyandry among the Toda has been described in detail by W. H. R. Rivers. The whole tribe is divided into two endogamous groups, which, again, are split up into a number of exogamous sub-groups. The husbands shared in common by a woman are in most cases brothers; they are rarely other members of the same exogamous group and of the same age class. When the husbands are brothers, there never ensue any quarrels about access to the wife. All the brothers are reckoned as fathers of a child. Yet it often occurs that a Toda only calls one man his father. It is exclusively external circumstances that are here decisive; often one of the fathers is more influential and more respected than his brothers, and naturally the sons prefer to speak of him as their father. If only one of the fathers is alive, the offspring always describe him as their father. If the husbands are not real brothers, they live, like these, in one household, but the children are allotted to single definite fathers. That man is considered the father of a child who in the seventh month of the mother's pregnancy has gone with her through the ceremony of the presentation of bow and arrow (which is also customary in fraternal polyandry). The husbands may take turns in the practice of this ceremony at every pregnancy; it results, therefore, frequently that the first two or three children belong to one and the same man, the other husbands acquiring formal father-right only at the later births. If the husbands separate and give up the common household, each one takes with him the children belonging to him by right of the bow-and-arrow ceremony. As everywhere else in India, polyandry has fallen into decay among the Toda. It may happen that several men have in common several wives, or that of a group of brothers each has his own wife. But polyandry has remained up to the present time the prevalent form of marriage among these hill-folk. The surplus girls used formerly to be killed without exception; and it is certain, says Rivers, that girl infanticide is still practised to some extent, although the Toda themselves deny this. It must be noted that child marriage exists among the Toda.

Matriarchal polyandry, which, in contradistinction to fraternal polyandry, goes with descent through the mother, still occurs among the Munduvars of the Travancore plateaus, the Nayars in some parts of Travancore and Cochin, the Western Kallan, and also among some other Southern Indian communities. Among numerous other races having mother descent, but not among all, relics of the former existence of matriarchal polyandry have been established. The secular authorities, and no less the European missions, are trying hard to exterminate this form of marriage.

It is difficult to trace any connection between the polyandry in the north and that in the south of India. It is most probable that this custom was carried into Southern India by the Tibetan conquerors in ancient times. Many Southern Indian polyandrous races, like the Toda and the Nayar, are distinguished from their real Dravidian neighbours by their more powerful build, lighter colouring, higher noses, etc. Furthermore, the architecture of the Malabar temples bears traces of Tibetan influence. The demon masks carved thereon show almost the same faces as the Tibetan masks. Among the Kallan the tradition of northern descent has been preserved up to the present time, and they bury their dead with their faces turned towards the north.

Exogamy is the custom which forbids the choice of partners for marriage within a certain group, and which has the effect of preventing near relations from sexual intercourse. It is found very frequently among primitive people, and is very prevalent, as Sir J. G. Frazer shows in his book "Totemism and Exogamy." This, however, does in no way justify the assumption that it was a general stage of civilisation of all mankind, and that it once existed even in those places where it is not found to-day.

Although European travellers, colonists and scientists had long been in contact with coloured races, it was the Scotsman J. F. McLennan who first discovered the existence of exogamy. He was led to this discovery by the study of that peculiar marriage custom which consists in the pretence of forcible bride capture, though the marriage of the couple concerned has been agreed to by both families beforehand. McLennan tried to find an explanation for this custom, and came to the conclusion that capture of women, which only took place in pretence, must once have been practised in reality to a large extent. In searching for facts confirmatory of this assumption, he was struck by the fact that among savage and barbarous people the men married women not of their own, but of another, tribal group. He described this as "exogamy," in contradistinction to "endogamy," by which marriage partners are restricted in their choice to their own group. In a tribe or other social group both sexual arrangements may exist side by side, in such a manner that the tribe is closely endogamous and is divided into several exogamous groups.

The theory put forward by McLennan as an explanation of the origin of exogamy is very simple and on superficial examination very convincing. He assumed that exogamy arose from a scarcity of women, which forced men to obtain wives by capture from other groups and thus gradually led to a general preference for strange women. The cause of this assumed scarcity of women was considered to be the infanticide of new-born females, which was carried on systematically, for savage people foresaw that in the struggle for existence it would be a hindrance to have a great number of women, who could take no share in the battle with enemies, and who presumably would contribute less to the food supply than the men.

H. Cunow also traces back the origin of exogamy to the scarcity of women and wife capture. He starts from the assumption that among the Australian and other uncivilised races the number of persons in a horde is very limited. "If one assumes that the number of members of a horde is sixty, the youngest class would contain, according to present-day reckoning, about twenty-five persons, the middle class twenty, and the oldest class about fifteen persons. In the middle class there would, therefore, be only about ten women. Among these a young man entering the middle class would often not find a single woman that he could take for his wife, for, after pairing marriage had become general, the few existing women had already found a spouse; they had already been disposed of. There was nothing left for the young man but to capture a woman from a strange horde as soon as possible, or to try to persuade a comrade of the same age class to let him share in his marriage relationship on the understanding that his hunting bag would contribute towards the 'household of the three.' This multiple conjugal partnership is customary among most of the Australian tribes even to-day." To this it must be added that the man needs to show much less consideration for a captured strange woman than for one of his own tribe, who would run away if badly treated. Nor can the young man remain single, for he himself would then have to drag his property about, which would hinder him in the hunt and expose him to the ridicule of his companions. (In reality there are many unmarried men even in Australia.) The search for wives led ultimately, according to Cunow, to wife capture and exogamy.

Infanticide, which McLennan assumes, is at present a rare exception among primitive people. Almost all explorers praise their great love for children, and even malformed children are not always killed. Even where infanticide does occur, the sex of the child is certainly not the factor that decides whether it is to be killed or not. The assumption that scarcity of women is brought about by girl infanticide is not correct. The female sex is, indeed, in the minority among uncivilised natives where they have been counted; but the excess of men is only small. Mutual capture of women could not alter this disparity, for it is unlikely that some tribes permitted the capture of their women without retaliation. Besides, even among primitive people men are careful in risking their lives. Capture of women is, therefore, nowhere the rule, but is everywhere the exception. Had it been the rule anywhere, the continuous fighting would have led to the extermination of the tribes in question. Frazer is right when he says: "If women are scarce in a group, many men will prefer to remain single rather than expose themselves to the danger of death by trying to capture women from their neighbours." This is what really happened among many tribes of the Australian natives who lived on a friendly footing with each other. It even happens that the old men who claim the women expressly forbid the young men to steal women from other tribes, because that will lead to bloodshed. Further, scarcity of women is most likely overcome, as previously mentioned, by several men's sharing one wife, which arrangement, unlike the capture of women, avoids arousing the hostility of neighbours. Among peaceable tribes, therefore, a numerical preponderance of men results not in exogamy, but in polyandry. But admitting that a warlike tribe has not sufficient women and therefore captures them from their neighbours, it is still unexplainable why the men should altogether avoid sexual relationship with their own women, few as they are, and have no desire for them whatsoever. This will certainly not be the result; on the contrary, the few women obtainable without force will be all the more in demand.

Frazer thinks that the origin of exogamy has been rightly explained by the American ethnologist L. H. Morgan, who for many years lived among the exogamic Indians as one of them, and thus came into direct contact with exogamy. Morgan assumed that sexual promiscuity was general at a very early period in the history of mankind, and that exogamy was instituted for the deliberate purpose of preventing cohabitation between blood relations, particularly between brothers and sisters, as was previously customary. This struck promiscuity at the root; it removed its worst peculiarity, and resulted at the same time in a powerful movement towards the establishment of sexual monogamy.

Frazer, in supporting Morgan's theory, relies exclusively on the Australian natives, who, according to him, though extremely primitive savages, "carry out the principle of exogamy with a practical astuteness, logical thoroughness, and precision such as no other race shows in its marriage system."

Frazer finds that the effects of the Australian marriage class system are in complete harmony with the deeply rooted convictions and feelings of the natives as regards sexual intercourse, and concludes that the successive tribal subdivisions have been brought about deliberately in order to avoid marriage of blood relations. According to him, it is not going too far to assert that "no other human institution bears the stamp of deliberate purpose more clearly than the exogamous classes of the Australians. To assume that they serve only accidentally the purpose that they actually fulfil, and which is approved by them unreservedly, would be to test our credulity nearly as much as if we were told that the complicated mechanism of a watch has originated without human design."

Nearly all Australian tribes have the system of division into marriage classes. Every tribe consists of two main groups (called in ethnographical literature phratries or moieties), and each of these groups is again divided into two, four, or eight classes. Sometimes the phratries and classes have special names, but not always. In the latter case it may be assumed that the names have been lost, while the division of the tribes into marriage groups remains. These groups are strictly exogamous. In no case are the members of the main group of the tribe (phratry) or of the same class allowed to marry each other. Only members of two given classes may marry, and their children are again assigned to given classes. Among some of the tribes there exists paternal descent, among others maternal descent. Which of the two modes of descent prevails in Australia can hardly be determined. Among some tribes property is inherited in the female line. Other rights of the female sex connected with mother descent are unknown. An example of the Australian marriage classes is given here, namely, that of the tribe Warrai, who live on the railway line running from Port Darwin to the south. Among this tribe indirect paternal descent is the custom; i.e., the children belong to the main group (phratry) of the father, but to other marriage classes.

The female marriage classes are marked with an asterisk.

Each member of a certain male marriage class may only marry a member of a marriage class of the other phratry, placed opposite in the table. Thus, for instance, an Adshumbitch man marries an Alpungerti woman, an Apungerti man an Aldshambitch woman, etc. The children always belong to the phratry of the men, but to another marriage group of theirs. Thus, for instance, the boys born from the union of an Adshumbitch man with an Apungerti woman belong to the Apularan class, and the girls born of this marriage belong to the Alpularan class. Further complications arise in consequence of the totem system, which exists among most of the Australian tribes. As the local groups of a tribe are numerically weak and consist of members of all marriage classes, the choice of mates is restricted to quite a small number of persons, being further limited to a great extent by the marriage of girls in childhood. But even when adults marry, they can rarely decide according to their own will, but are dependent on the circumstances of relationship. On the northern coast of Australia the marriage class system does not exist, but exogamy exists there, the members of certain local groups not being allowed to marry each other. The now extinct tribes in the south-east of the continent also had no marriage class system.

But it still remains a mystery how it was found out that marriages of blood relations were harmful. One objection is, that some of the Australians are ignorant of the process of generation; they do not even know that pregnancy is the result of cohabitation. It is also doubtful whether the Australian natives can in any case be considered as typical representatives of primitive man. If this were so, all mankind would still be in a very low state of civilisation, for the Australians appear incapable of progressive development. And further, if exogamous classes were purposely instituted in order to prevent cohabitation between blood relations, how is it that other people also are excluded from sexual intercourse who are not blood relations? Frazer's comparison with a watch is also badly chosen. We must take into consideration the intellectual stage of development of mankind at the time when exogamy arose, and when the watch was invented. Even if we do not admit that exogamy was instituted with a conscious purpose, this does not by any means, as Frazer says, do away altogether with will and purpose from the history of human institutions. There is no need to doubt that the Australian system of exogamy became more and more complicated through the deliberate action of man.

Frazer himself assumes that the Australians had an aversion to cohabitation between brothers and sisters even before it was definitely fixed by binding rules. Sexual aversion between parents and children, according to him, is universal among them, whether there be in vogue the two-, four- or eight-classes system, i.e., whether incest between parents and children is expressly forbidden or not. "In democratic societies like those of the Australian natives, the law sanctions only thoughts that have already been long the mental possession of the majority of people." Hence the agreement of the marriage class system with the feelings of the people becomes explainable.

Since the aversion to sexual intercourse within certain classes was already in existence before the formation of marriage classes, the classificatory system being merely the formal expression of it, we have to find some explanation for it. For the appearance of this aversion marks the real beginning of exogamy, which cannot be explained by the complicated system of the Australians. It is possible that the sexual aversion towards blood relations is already a characteristic trait of the human race before its truly human development, and that it may have to be looked upon as an instinct. This is the opinion of F. Hellwald, which has also been upheld of late by A. E. Crawley. It is assumed that among brothers and sisters, as among boys and girls who have lived together from childhood, the pairing instinct generally remains in abeyance, because the conditions are wanting that are likely to awaken this instinct. Courting the favour of a person of the other sex is the process that gradually brings about the sexual excitement necessary for union. The possibility of sexual excitation between people who have lived together from childhood is decidedly lessened through habituation, if not completely inhibited. In this respect brothers and sisters reach already at puberty that state towards each other to which people married for a long time approach gradually, through the constant living together and the exhaustion of youthful passion. If brother and sister sometimes show passion for each other, it is generally the result of the same circumstances that are necessary to arouse it under normal conditions, e.g., a long separation. As the absence of sexual attraction between brother and sister who have grown up together is a natural thing, it is strange that cohabitation between them should have to be specially prohibited and enforced by strict measures among primitive peoples. The explanation, according to Crawley, is simple. "In many departments of primitive life we find a naÏve desire to, as it were, assist Nature, to affirm what is normal and later to confirm it by the categorical imperative of custom and law. This tendency still flourishes in our civilised communities, and, as the worship of the normal, is often a deadly foe to the abnormal and eccentric, and too often paralyses originality. Laws thus made, and with this object, have some justification, and their existence may be due, in some small measure, to the fact that abnormality increases pari passu with culture. But it is a grave error to ascribe a prevalence of incest to the period preceding the law against it." All the facts tend to show that the most primitive people procured their wives by friendly arrangements. From this standpoint it would be most practical if each tribe were divided into two groups, the men of each group marrying wives from the other group. This state of affairs is actually to be found among many uncivilised peoples that are divided into two exogamous groups or phratries. It has still to be discovered how this bipartition arose. It is unthinkable that a division into two groups was intentionally brought about by the members of the groups for the purpose of preventing marriages between blood relations of a certain grade. No tribe has ever been divided in such a manner; the division must therefore be explainable in another way. The phratries are large families (in the broad sense of the word); they descend from families (in the narrower sense of the word), reciprocally supplying each other with wives. The names of the phratries are generally unintelligible, in contradistinction to the names of the totem groups, and therefore most probably older. The totem groups, of which a phratry consists, are to be considered as younger branches of the original double family, which have arisen through wives being taken from other groups whose children again received the name of their mothers. If it should be asked why the members of two phratries should constantly intermarry, it should be pointed out that among communities in the lowest stage of civilisation women are not easily procurable, and the force of external circumstances would favour the unions just mentioned (Crawley, pp. 54 et seq.).

A biological explanation of the origin of exogamy is given by Herbert Risley. Without basing it on the assumption that primitive people have a knowledge of the harmfulness of incest, he gives the following exposition: "Exogamy can be brought under the law of natural selection without extending it too far. We know that among individuals or groups of individuals there exists a tendency to vary in their instincts, and that useful variations (such as are suitable to the conditions of life) tend to be preserved and transmitted by inheritance. Let us assume now that in a primitive community the men varied in the direction towards choosing wives from another community, and that this infusion of fresh blood was advantageous. The original instinct would then be strengthened by inheritance, and sexual selection would be added in the course of time. For an exogamous group would have a greater choice of women than an endogamous one, ... and in the competition for women the best would fall to the strongest and most warlike men. In this way the strengthened exogamous groups would in time exterminate the endogamous neighbours, or at least take away their best marriageable maidens. Exogamy would spread partly through imitation, partly through the extermination of endogamous groups. The fact that we cannot explain how it came about that the people varied in the aforesaid direction is not fatal to this hypothesis. We do not doubt natural selection in the case of animals because we cannot give the exact cause of a favourable variation."

E. Westermarck holds a similar theory about the cessation of incest. He thinks that "among the ancestors of man, as among other animals, there was, no doubt, a time when blood relationship was no bar to sexual intercourse. But variations here, as elsewhere, would naturally present themselves; and those of our ancestors who avoided in-and-in breeding would survive, while the others would gradually decay and ultimately perish. Thus an instinct would be developed which would be powerful enough, as a rule, to prevent injurious unions. Of course it would display itself simply as an aversion on the part of individuals to union with others with whom they lived; but these, as a matter of fact, would be blood relations, so that the result would be the survival of the fittest. Whether man inherited the feeling from the predecessors from whom he sprang, or whether it was developed after the evolution of distinctly human qualities, we do not know. It must necessarily have arisen at a stage when family ties became comparatively strong, and children remained with their parents until the age of puberty or even longer."

It may be surmised that the impulse towards the appearance of the exogamous tendency arose through economic progress, which led to an increase of the means of existence, and this in its turn produced a more friendly relationship between neighbouring groups that previously had quarrelled about food. The men thus came into contact with strange women, and this awakened a heightened sexual feeling, in other words the instinct which is said to have led to the avoidance of incest. Thus among the peoples on a very low economic level (e.g., the Pigmies) no laws for the prevention of incest are to be found, a fact that may be held to confirm this idea. Primitive people could in any case not understand the harmfulness of incest, while it is certain that strange members of the opposite sex could exert a stronger attraction, and thus render the sexual impulse permanent, which previously was periodical, as among the animals.