mr. m’lennan has remarked, in relation to the curious customs of capturing women for wives found among peoples in all parts of the world, that “in almost all cases the form of capture is the symbol of a group-act—of a siege, or a pitched battle, or an invasion of a house by an armed band, while in a few cases only, and these much disintegrated, it represents a capture by an individual. on the one side are the kindred of the husband; on the other the kindred of the wife.”250 whatever may be the true explanation of the origin of exogamy, with which the custom referred to is connected, there can be no doubt of the truth of the statement that the wife-capture is now usually, although it sometimes has relation solely to the individual, the symbol of a group-act. this may not be in the sense intended by mr. m’lennan, who looks upon exogamy and polyandry as referable to one and the same cause, and who regards “all the exogamous races as having originally been polyandrous.”251 the phenomena of wife-capture prove conclusively, however, that the family group to which the woman belonged possessed, or thought themselves entitled to, certain rights over her—rights of which they resisted the invasion, whether by an individual alone, or by193 a group of persons, or by an individual aided by the other members of a group. it is important to notice that the groups in question appear to consist, not of strangers to each other, or to the man or woman more immediately concerned, but of persons bound together by certain ties of blood. this is shown to be so by the fact that the capture is atoned for by the payment to the relations of the woman of the marriage-price, if this has not been agreed on beforehand.252 it is required, moreover, by the conclusion arrived at by mr. m’lennan, that the tribes among whom the system of wife-capture prevails are chiefly those whose marriages are governed by the law of exogamy.253 by exogamy is meant the practice of marrying out of the tribe or group of kindred,254 and it is founded on a prejudice against marriage with kinsfolk.255 there is some uncertainty as to the nature of m’lennan’s primitive group, but, judging from his statement that “promiscuity, producing uncertainty of fatherhood, led to the system of kinship through mothers only,”256 we may suppose that it consisted of a number of persons, all of whom, as the result of promiscuity, were related by blood. the first division into which he classes uncultured peoples, according to their marriage-rules, is that where tribes are separate, and all the members of the tribes are, or feign themselves to be, of the same blood.257 mr. morgan very properly criticises this definition, which, he says,194 “might answer for a description of a gens; but the gens is never found alone, separate from other gentes. there are several gentes intermingled by marriage in every tribe composed of gentes,”258 a fact which would seem to distinguish the primitive group of m’lennan, although consisting of consanguinei, from a gens or clan proper. moreover, as mr. morgan shows, exogamy has relation to a rule or law of a gens, considered as “the unit of organisation of a social system,” and therefore the gens (of which, as an institution, the rules are prohibition of intermarriage in the gens, and limitation of descent in the female line259), or rather the family from which it has sprung, may be regarded as the earliest social group of which we have any knowledge.

it is of the greatest importance to the discovery of the nature of the primitive human family to understand the origin of the gens or clan. as defined by morgan, it is “a body of consanguinei descended from the same common ancestor, distinguished by a gentile name, and bound together by affinities of blood.” mr. morgan affirms that the gens originated in three principal conceptions, “the bond of kin, a pure lineage through descent in the female line, and non-intermarriage in the gens.”260 the most essential feature is that of tracing kinship through females only, and the discovery of the origin of this custom will throw light on that of the clan-institution itself, and therefore on the nature of the primitive family.

mr. m’lennan finds the origin of kinship through females only in the uncertainty of paternity, arising195 from the fact that, in primitive times, a woman was not appropriated to a particular man for his wife, or to men of one blood as wife.261 the children, although belonging to the horde, remain attached to their mothers, and the blood tie observed between them would, as promiscuity gave place to polyandry of the ruder kind in which the husbands are strangers in blood to each other, become developed into the system of kinship through females.262 an earlier writer, bachofen, was so much struck with certain social phenomena among the ancients, that he believed women to have, at an early period, been supreme, not only in the family but in the state. he supposed that woman revolted against the primitive condition of promiscuity, and established a system of marriage, in which the female occupied the first place as the head of the family, and as the person through whom kinship was to be traced. this movement, which had a religious origin, was followed by another resulting from the development of the idea that the mother occupied a subordinate position in relation to her children, of whom the father was the true parent. mr. m’lennan very justly objects to this theory that, if marriage was, from the beginning, monogamous, kinship would have been traced through fathers from the first.263 he adds that196 “those signs of supremacy on the woman’s part were the direct consequences (1) of marriage not being monogamous, or such as to permit of certainty of fatherhood; and (2) of wives not as yet living in their husband’s houses, but apart from them, in the homes of their own mothers.”264 the meaning of this is, that the phenomena referred to by bachofen were due to the former prevalence of a system of polyandry, such as still exists among the nairs of southern india. it is very improbable, however, that kinship through the female only could have had the origin supposed by mr. m’lennan. according to him one cause of the supremacy of woman referred to by bachofen was the fact of wives living apart from their husbands in the homes of their own mothers. this custom must, therefore, have preceded the supremacy of woman, assuming this to have existed, and the tracing of kinship through females which gave rise to it. we must believe that originally women lived alone with their daughters (and their sons also, until these set up a separate establishment for themselves, taking with them probably their favourite sisters, as with the nairs at the present day),265 there being no male head of the family. if, however, we trace our steps back in thought to the most primitive period of human existence, we shall see that such a domestic state as that here supposed cannot have been the original one. among savages there is never that subordination of the man to the woman which we should have to assume. we cannot suppose that the primeval group of mankind consisted of a woman and her children, and if the woman had a male companion we cannot doubt, judging from what we know of savage races, that he would be the head197 and chief of the group. the very notion, however, of the family group having a male as well as a female head is inconsistent with mr. m’lennan’s theory, and we must trace the origin of female kinship as a system to a different source from the polyandry to which he ascribed it.

the idea of a special relationship subsisting between a woman and her children might no doubt be originated during the period when the men of a group, “in the spirit of indifference, indulged in savage promiscuity,”266 if such a condition of things ever existed, but that alone would not be sufficient to establish kinship through females only. it may be questioned, indeed, whether there ever was a time when the uncertainty of paternity, which mr. m’lennan’s whole theory requires, was so pronounced as to prevent kinship through males being acknowledged. mr. morgan agrees with mr. m’lennan so far as to say that, “prior to the gentile organisation, kinship through females was undoubtedly superior to kinship through males, and was doubtless the principal basis upon which the lower tribal groups were organised.” he affirms truly, however, that “descent in the female line, which is all that ‘kinship through females only’ can possibly indicate,” is only the rule of a gens, and that relationship through the father is recognised as fully as that through the mother.267 i have elsewhere, however, given reasons for believing that this statement does not go far enough, and that the earliest forms of the classificatory system of relationships, on which mr. morgan198 bases his special theory, require actual kinship, and not relationship merely, through the male quite as fully as through the female.

it is surprising that mr. morgan says little as to the origin of descent in the female line. he says: “the gens, though a very ancient social organisation founded upon kin, does not include all the descendants of a common ancestor. it was for the reason that, when the gens came in, marriage between single pairs was unknown, and descent through males could not be traced with certainty. kindred were linked together chiefly through the bond of their maternity.”268 we have here apparently two reasons stated for the establishment of kinship through females, the absence of marriages between single pairs, and the uncertainty of paternity. both of these conditions are found by mr. morgan to exist in the consanguine family groups which he supposes to have been formed when promiscuity ceased. the polynesian peoples, among whom he finds traces of the consanguine family, have preserved the recollection of female kinship, although, according to mr. morgan, the gens is unknown to them.269 the classificatory system of relationships, the origin of which he traces to the consanguine family, can, however, receive a totally different interpretation, and the existence of that family itself is very doubtful. further, the difficulty of tracing descent through males, which mr. morgan supposes, is the result only of the polyandrous unions his theory requires, and if they ever really existed they could supply no further199 explanation of the origin of female kinship than the polyandry of the nairs. he would have done better to have sought to connect it, as mr. m’lennan does, with the special relation supposed to exist between a mother and her child.

mr. herbert spencer shows how this idea may have arisen. unlike the other writers i have referred to, he does not think that promiscuity in the relation of the sexes ever existed in an unqualified form.270 he thinks, indeed, that monogamy must have preceded polygamy, although, owing to the extension of promiscuity, and the birth of a larger number of children to unknown fathers than to known fathers, a habit would arise of thinking of maternal kinship rather than of paternal, and where paternity was manifest children would come to be spoken of in the same way.271 mr. spencer adds, that the habit having arisen, the resulting system of kinship in the female line would be strengthened by the practice of exogamy.272 the defect of this explanation lies in its requiring uncertain paternity, and i shall show that the system of female kinship has not arisen from the simple association in thought of a child with its mother in preference to its father. it is, moreover, inconsistent with the fact mentioned by mr. spencer himself, that where the system of female kinship now subsists “male parentage is habitually known.”273 it is true that he supposes male kinship to be disregarded, but this conclusion appears to me not to be supported by sufficient evidence.

200

that there may have been a short period of barbarism in which the intercourse between the sexes was unrestrained by any law of marriage is possible. probably, as female chastity before marriage is even now but slightly regarded among most uncultured peoples, all sexual alliances were allowable, so long as the rule as to consanguinity was not infringed, and so long as no offspring resulted from the alliance,274 where this was entered into without the consent of parents. this consent would be necessary in all cases where such alliances were formed by females for marital purposes, and the sanction required would be that of the family head at the early period we are treating of. judging from what we observe among modern savages we cannot doubt that self-interest chiefly would govern the father in connection with his daughter’s marriage. he would make certain requisitions as the price of his consent. whether the marriage was to be a permanent or a terminable engagement, the father would stipulate that his daughter should continue to live with or near him, and that her children should belong to the family group of which he is the head. in this case not only would the children form part of the family to which their mother belonged, but the husband himself would become united to it, and would be required to labour for the benefit of his father-in-law.

a custom still prevalent among the new zealanders may be cited in illustration. the reverend richard taylor says:201 “sometimes the father simply told his intended son-in-law he might come and live with his daughter; she was thenceforth considered his wife, he lived with his father-in-law, and became one of his tribe or hapu to which his wife belonged, and in case of war was often obliged to fight against his own relatives.” mr. taylor adds, that so common is the custom of the bridegroom going to live with his wife’s family, that it frequently occurs; when he refuses to do so, she will leave him, and go back to her relatives.275 when the wife left her father’s house to reside with her husband he had to purchase the privilege by giving her father and other relations handsome presents.276 as among the new zealanders, children belonged to their father’s family, the fact of the wife going to reside among her husband’s relations meant the loss by her father’s family of the children. the presents may, therefore, be supposed to represent the price given by a man for his wife’s offspring to her relations. this opinion is confirmed by reference to the marriage customs of a west african people. mr. john kizell, in his correspondence with governor columbine, respecting his negotiations with the chiefs in the river sherbro, says:202 “the young women are not allowed to have whom they like for a husband; the choice rests with the parents. if a man wishes to marry the daughter, he must bring to the value of twenty or thirty bars to the father and mother; if they like the man, and the brother likes him, then they will call all their family together, and tell them, ‘we have a man in the house who wishes to have our daughter; it is that which makes us call the family together, that they may know it.’ then the friends inquire what he has brought with him? the man tells them. they then tell him to go and bring a quantity of palm wine. when he returns, they again call the family together; they all place themselves on the ground, and drink the wine, and then give him his wife. in this case, all the children he has by her are his, but if he gives nothing for his wife, then the children will all be taken from him, and will belong to the woman’s family; he will have nothing to do with them.”277

mr. taylor says that the ancient and most general way of obtaining a wife among the new zealanders was “for the gentleman to summon his friends, and make a regular taua, or fight, to carry off the lady by force, and ofttimes with great violence.”278 a fight also took place if, when a girl was given in marriage, the friends of another man thought he had a greater right to her, or if she eloped with some one contrary to her father’s or brother’s wish. even if all were agreeable, “it was still customary for the bridegroom to go with a party, and appear to take her away by force, her friends yielding her up after a feigned struggle; a few days afterwards, the parents of the lady, with all her relatives, came upon the bridegroom for his pretended abduction; after much speaking and apparent anger, it ended with his making a handsome present of fine mats, &c., and giving an abundant feast.”279 in this case the affair ended in the same manner as the203 african marriage already referred to, and the idea was no doubt the same in both—the giving of compensation to the parents and relations of the woman for the loss sustained by them through her offspring being removed from the family group; probably the widespread custom of pretended forcible marriage was originally connected with the rights of the woman’s relations, although sometimes the capture is due to the desire to obtain for nothing what could otherwise be acquired only by a purchase fee.

what those rights are may be ascertained from the information given us by mr. morgan as to the privileges and obligations associated with the membership of a gens. among them is an obligation not to marry in the gens, mutual rights of inheritance of the property of deceased members, and reciprocal obligations of help, defence, and redress of injuries. “the functions and attributes of the gens,” says morgan, “gave vitality as well as individuality to the organisation, and protected the personal rights of its members,”280 who, as being connected by the ties of blood relationship, may be regarded as forming an enlarged family group, or rather a fraternal association based on kinship.

the gens would, however, form too large a group for ordinary social purposes, and a smaller group would be composed of those more immediately allied by blood. thus, although theoretically the effects of a deceased person were distributed among his gentile relations, yet morgan admits that204 “practically they were appropriated by the nearest of kin.”281 among the iroquois, if a man died leaving a wife and children, his property was distributed among his gentiles in such a manner that his sisters and their children, and his maternal uncles, would receive the most of it. his brothers might receive a small portion. an analogous rule prevailed when a woman died. the property remained in the gens in either case,282 although its division was restricted to a small number of gentiles. it could not have been otherwise where the members of the gens are numerous or widely distributed. the same principle would apply in relation to rights over children, who in a low social stage are looked upon in the light of property. among the aborigines of america each gens had personal names that were used by it alone, and, says morgan, a gentile name conferred of itself gentile rights. now, although a child was not fully christened until its birth and name had been announced to the council of the tribe, its name was selected by its mother with the concurrence of her nearest relatives. morgan says nothing of any right of the gens over the marriage of its members, and it would seem not to have any voice in the matter. the formation of the alliance is usually left to the two individuals more immediately concerned or to their near relations,283 and the marriage price belongs to the parents and near kin of the wife. this, in the absence of the marriage price, would be the case also with the children born of her marriage, on the principle that205 “children are the wealth of savages.” reference to the custom of blood revenge confirms the view that, for certain purposes, a smaller family group than the gens is recognised by the peoples having that organisation. mr. morgan thinks the practice of blood revenge had “its birthplace in the gens,” which was bound to avenge the murder of one of its members. he says further that it was “the duty of the gens of the slayer, and of the slain, to attempt an adjustment of the crime before proceeding to extremities.” it rested, however, with the gentile kindred of the slain person to decide whether a composition for the crime should be accepted, showing that they were considered the persons more immediately concerned. the crime of murder is, as mr. morgan says, “as old as human society, and its punishment by the revenge of kinsmen is as old as the crime itself.”284 this is hardly consistent with the preceding statement that the practice of blood revenge had its birthplace in the gens. it preceded the development of the gens, and originated with the smaller family group which, as we have seen, is more immediately connected with property and children and the marriage of its female members. those who are liable to the obligations of the law of blood revenge in any particular case must be identified, and, as they can hardly comprise all the members of the gens, we must suppose them to be restricted to the smaller group consisting of near blood relations. judging from what we know of the habits of the australian aborigines in relation to the lex talionis, we cannot doubt that the persons subject206 to retaliation in any particular case are well defined.

the example of the polynesian islanders, who are said not to have risen to the conception of the gens, shows that before this was developed, not only was the lex talionis recognised, but the law of marriage and the rights of parents over their children were fully established. these are, therefore, not dependent on the gens, but are incidental to a simpler group of blood relations—that on which the gens itself is based. the idea of “brotherhood” is at the foundation of all these early social organisations. mr. morgan says, in relation to the iroquois phratry, that “the phratry is a brotherhood, as the term imports, and a natural growth from the organisation into gentes. it is an organic union or association of two or more gentes of the same tribe for certain common objects. these gentes were usually such as had been formed by the segmentation of an original gens.”285 so also, a gens forms a fraternal association, as it consists of “a body of consanguinei descended from the same common ancestor, distinguished by a gentile name, and bound together by affinities of blood.”286 if we trace the ascent until we come to the common ancestor, we shall have a group of kinsmen who compose the simplest form of “brotherhood,” that of a parent and his or her children. originally this would be a mother and her daughters, as when the sons formed marriage associations the daughters only and their children would be left under the parental roof. it is evident, therefore, that the primitive family cannot have207 originated within the gens or clan. on the contrary, the clan was based on the family or group of kinsmen, without which it could not have existed. moreover, it by no means follows that, because the common ancestor of the members of the gens or clan was a female, the primitive group of kinsmen had not a male as well as a female head. considered as a “fraternal association,” the father may have been excluded, but for the purposes of the brotherhood it was of no importance whether paternity was certain or uncertain. the result would have been the same in either case. for other than brotherhood purposes kinship to the father may have been fully recognised. the obligations of the lex talionis, the right to property, and the control of children in marriage, may have concerned only the kinsmen by the mother’s side, but those on the father’s side may have been equally affected by the law of marriage. that such was the case i have sought to establish elsewhere, as evidenced by the classificatory system of relationships, and that view is confirmed by various facts showing that kinship by the male side is fully recognised among savages.

i have already had occasion to refer to mr. m’lennan’s admission that, if “marriage was, from its beginning, monogamous, kinship would certainly (human nature being as it now is) have been traced through fathers, if not indeed through fathers only, from the first.”287 mr. herbert spencer, although apparently thinking that promiscuity in the relations of the sexes was originally extensive, yet supposes that it was accompanied by monogamic connections of208 a limited duration. he says that “always the state of having two wives must be preceded by the state of having one,” and he looks upon the preference for the maternal kinship rather than paternal kinship as a habit, arising from the fact that the former is observed in all cases, whilst the latter is inferable only in some cases.288 mr. spencer’s admission that where the system of female kinship now subsists, “male parentage is habitually known, though disregarded,” greatly weakens his position, the more so as we are not told why or when it is disregarded.289 mr. morgan goes far towards supplying an explanation of the fact, although his theory is defective. he affirms that gentile kin were superior to other kin only because it conferred the rights and privileges of a gens, and not because no other kin was recognised. “whether in or out of the gens, a brother was recognised as a brother, a father as a father, a son as a son, and the same term was applied in either case without discrimination between them.”290 mr. morgan does not, however, admit of certainty of paternity, although he states that “they did not reject kinship through males because of uncertainty, but gave the benefit of the doubt to a number of persons—probable fathers being placed in the category of real fathers, probable brothers in that of real brothers, and probable sons in that of real sons.”291 this explanation is plausible but insufficient, if, as mr. morgan says, descent in the female line is only a rule of a gens.292 in this case, female descent209 cannot have existed before the gens, and recognition of kinship through the father may have subsisted prior to the formation of the gens, together with that of the relationship between mother and child on which such descent is founded. this would seem to be required by the facts mentioned by mr. morgan in relation to the social institutions of the american aborigines. he says “an indian tribe is composed of several gentes developed from two or more, all the members of which are intermingled by marriage, and all of them speak the same dialect. to a stranger the tribe is visible and not the gens.”293 originally, therefore, the tribe consisted of two gentes, that is of the descendants from two female common ancestors, and, as the gentes are not visible to a stranger, we must suppose that the tribe originally represented the male head of the primitive family group to which the female common ancestors belonged. on this supposition the primitive group consisted of a male and two females, the former being the recognised representative of the group, although the descent of its members is traced through the latter. this view is quite consistent with the explanation i have elsewhere given of the classificatory system of relationship, which undoubtedly requires the full recognition for certain purposes of blood relationship through both the father and the mother.

the conclusion thus arrived at is confirmed by what we know of the opinions entertained by peoples among whom the gentile organisation is fully developed. carver, as quoted by sir john lubbock, states that among the hudson’s bay indians, children always take210 the name of their mother. the reason they give for this is, “that as their offspring are indebted to the father for their souls, the invisible part of their essence, and to the mother for their corporal and apparent part, it is more rational that they should be distinguished by the name of the latter, from whom they indubitably derive their being, than by that of the father, to which a doubt might arise whether they are justly entitled.”294 the reason given by the hudson’s bay indians why children are called after their mothers shows that the system of female kinship is quite consistent with the recognition of kinship through the male. no doubt the mother is regarded by savages as having a closer physical relationship to her child than their father, but it is incredible to suppose that the latter could ever be looked upon as having no closer relationship to it than a stranger in blood. if the mother had several husbands the actual paternity may not be certain, but, as the father must be one of several well-ascertained individuals, the paternity is only rendered less certain, and the child may be regarded as having several fathers, and claim kinship through them all. if they are sons of the same father, that kinship will be with the same persons as though its mother had but one husband. under the conditions i have supposed, however, where a woman takes, as her husband, a man who lives with her among her own relations, there would not be any uncertainty as to paternity, and therefore the stronger relationship supposed between mother and child must have originated in the close physical connection211 observed to subsist between them. this does not, however, explain the origin of clan relationship based on kinship through females only, which is connected with the fact of the members of a woman’s clan possessing certain rights over her and her children. these rights would not be affected, even if the primitive custom of the woman continuing to live among her relations after marriage were departed from. before this took place, the system of female kinship would have become firmly established, and it would be confirmed, although it could not be originated, by the idea that, as the wife may not be faithful to her husband, there is more certainty about maternity than paternity.

the fact that a man’s heirs are usually his sister’s children, shows that consanguinity is of great importance in the eyes of uncultured peoples, and what has been advanced is quite sufficient to account for that fact without assuming the existence of a state of promiscuity in the relations between the sexes. such a state is not consistent with the abhorrence which even savages show to the marriage of persons of near blood relationship, and it has no support at all in the observed phenomena of savage life. the punalua custom of the polynesian islanders, which has its counterpart among the todas of the neilgherries, and traces of which may perhaps be found, on the one hand, in the fraternal polyandry of the tibetans, and, on the other hand, in the sororal polygamy of the north american aborigines, is neither promiscuous nor incestuous in the proper sense of these words. the possession by several brothers of wives in com212mon, who may themselves be sisters, or by several sisters of husbands in common, who may be brothers, may, as i have elsewhere suggested, have originally been due to the feeling that marriage has a spiritual as well as a physical significance. punalua was really an application of the idea of brotherhood to marriage, and it is not surprising that, among uncultured peoples, the having wives or husbands in common should be considered a high mark of friendship.

it is important to notice that among the peoples who have developed or perfected the gentile institution, a rule of which is descent in the female line, the husband is the head of the household, and the wife little more than a servant, so long as they continue to live together. it is true, as lahontan states,295 that the wife has the same power of divorce as the husband, but so long as she remains in his cabin she is treated by him as a drudge and a mere child-bearer. as women they have some influence in the tribe, but this is only when they have children to give them dignity. the polynesian islanders not having risen to the conception of a gens, it is, perhaps, not surprising that woman is usually regarded by them as an inferior creature. her position as a woman is, however, better than that of a wife, in which capacity she is cared for as little as among the american aborigines. her condition is mitigated only under the influence of the areoi institution, and where she enters into the punaluan engagement. if it is true, as mr. morgan states, that213 “the australians rank below the polynesians, and far below the american aborigines,” we cannot wonder that the position of woman among the australian aborigines is one of great inferiority. in fact, among them wives are considered as articles of property, and not only do they suffer great privations, but they are most barbarously treated. the last-named people practice the simplest form of obtaining wives, that of capture by cunning and personal violence, but in most of their tribes descent is in the female line, and the gens or clan is developed more or less perfectly. and yet the australian aborigines possess marriage regulations which seem formed for the express purpose of preventing the intermarriage of blood relations, and which fully recognise kinship by the male line.

a modern french writer of great authority, fustel de coulanges, affirms that the ancient family was constituted chiefly by religion, the first institution of which was marriage. the family gives rise to the gens, and “with its elder and younger branches, its servants and dependents, formed possibly a very numerous group of persons.” such a family, says de coulanges, “thanks to the religion which maintained its unity, thanks to its special privileges which rendered it indivisible, thanks to the laws of protection which retained its dependents, formed in time a widespread society under an hereditary chief.”296 this view of the primitive family possesses much truth, although it leaves out of sight one of the most essential features of the family among uncultured peoples. the same may214 be said in relation to the patriarchal family of sir henry maine. this writer says that “the earliest tie which knitted men together in communities was consanguinity or kinship,” and that “there was no brotherhood recognised by our savage forefathers, except actual consanguinity regarded as a fact.”297 he adds, that “kinship, as the tie, binding communities together, tends to be regarded as the same thing with subjection to a common authority.” the notions of power and consanguinity are blended, a mixture of ideas which is seen “in the subjection of the smallest group, the family, to its patriarchal head.”298 “this group,” says sir henry maine, “consists of animate and inanimate property, of wife, children, slaves, land and goods, all held together by subjection to the despotic authority of the eldest male of the eldest ascending line, the father, grandfather, or even more remote ancestor. the force which binds the group together is power. a child adopted into the patriarchal family belongs to it as perfectly as the child naturally born into it, and a child who severs his connection with it is lost to it altogether.” the patriarchal family of maine thus differs from the ancient family of de coulanges in its binding force, which in the one case is power, and in the other religion, forces which are, nevertheless, reconciled by the fact that the chief element in this religion is the ancestral idea which is at the base of the patriarchal family. this view of the nature of the ancient family would be complete if215 it provided for the fact, revealed by the study of primitive institutions as now exhibited among uncultured peoples, that descent was originally traced by the female line in preference to the male line. the defect thus revealed will, however, be removed if it can be shown, as i have endeavoured to do, that descent through the male is, for certain purposes, recognised equally with that through the female. mr. herbert spencer, in his “principles of sociology,” refers,299 as follows, to a suggestion made by mr. fiske, which contains an important truth bearing on the subject of this paper: “postulating the general law that, in proportion as organisms are complex, they evolve slowly, he infers that the prolongation of infancy which accompanied development of the less intelligent primates into the more intelligent ones, implied greater duration of parental care. children, not so soon capable of providing for themselves, had to be longer nurtured by female parents, to some extent indeed by male parents, individually or jointly; and hence resulted a bond holding together parents and offspring for longer periods, and tending to initiate the family. that this has been a co-operating factor in social evolution is very probable.” the bond thus formed shows its influence even among the lowest savages, in the natural affection which subsists between a mother and her children, when these escape the not unusual fate of infanticide. natural affection is less operative with male parents, but there are other feelings which have relation chiefly to male children which tend to form an equally binding tie. mr. spencer216 remarks that “to the yearnings of natural affection are added, in early stages of progress, certain motives, partly personal, partly social, which help to secure the lives of children, but which, at the same time, initiate differences of status between children of different sexes. there is the desire to strengthen the tribe in war; there is the wish to have a future avenger on individual enemies; there is the anxiety to leave behind one who shall perform the funeral rites and continue oblations at the grave.”300 these motives must have been influential from the earliest period at which mankind consisted of more than a few small and isolated groups, and, therefore, we must assume that in these groups the male element was equally as strong as the female element, if, indeed, they had not a male head. mr. spencer remarks further that those motives, “strengthening as societies passed through the earlier stages, gradually gave a certain authority to the claims of male children, though not to those of females.”301 these ideas are quite inconsistent with the notion that the family group ever consisted only of a female ancestor and her children, or that the woman was originally the head of, and supreme in, the family. the custom of tracing descent by the female line shows, however, that for certain purposes the woman occupied an important position, although it may, when the practice of wives going to reside among their husband’s relations become established, have tended to confirm that of female infanticide, as the children would be lost to the217 mother’s family group. one of the motives referred to by mr. spencer would, after the idea of special kinship through females had become established, affect more especially the persons bound together by a maternal tie. where the gentile organisation is established the duty of revenging private injuries is confined to the other members of the common gens. the duty of defence against the external enemy belongs, however, to the tribe, which here undoubtedly stands in the place of the original family group, in which both male and female kinship, with their special duties, was recognised, represented by its male head. this group we must suppose, therefore, had much in common with sir henry maine’s patriarchal family. under the head of the oldest living male ancestor, it embraced wife or wives, children and dependents. the repugnance to marriages between blood relations, which seems almost instinctive to man, would prevent such alliances between the members of the group. the male children, when they reached the age of manhood, would leave the paternal roof, and obtain wives from other groups, with which they would become associated on the principle of adoption, while, on the other hand, young men from other groups would take their places as the husbands of the female children. it would be during this primitive period that the idea of a special relationship subsisting between a mother and her children, on which the custom of tracing descent through the female is founded, would become formed, as already mentioned. the importance attached to female kinship would be increased by the development of a fraternal feeling among the children of the same218 mother, a feeling which would be strengthened if, as would probably not seldom be the case, men, after some years of cohabitation with their wives, left their children solely to the mother’s care. under the influence of these various ideas and circumstances the custom of tracing kinship for certain purposes in the female line would be developed by the time that the habit had been formed of wives leaving their parents to reside among their husband’s family. as when this took place, the custom would be firmly established under the influence of polygamy, the development of the gentile organisation would almost necessarily follow. the primitive idea of kinship through the father would, however, still remain in full force with the attributes which originally appertained to it—namely, the headship in the family group of the eldest male ancestor, whose authority is practically represented by the tribe, and the non-intermarriage of those thus connected.