[ii]

The Cambridge Archaeological and Ethnological Series is supervised by an Editorial Committee consisting of William Ridgeway, M.A., F.B.A., Disney Professor of Archaeology, A. C. Haddon, Sc.D., F.R.S., University Lecturer in Ethnology, M. R. James, Litt. D., F.B.A., Provost of King's College and C. Waldstein, Litt. D., Slade Professor of Fine Art.

[iii]

BY

NORTHCOTE W. THOMAS, M.A.
Diplomé de l'École des Hautes-Études,
Corresponding Member of the Société d'Anthropologie de Paris, etc.

CAMBRIDGE:
at the University Press
1906

[iv]

CAMBRIDGE UNIVERSITY PRESS WAREHOUSE,
C. F. CLAY, Manager,

London: FETTER LANE, E.C.
Glasgow: 50, WELLINGTON STREET.

Leipzig: F. A. BROCKHAUS.
New York: G. P. PUTNAM'S SONS.
Bombay and Calcutta: MACMILLAN AND CO., Ltd.

[All Rights reserved.]

[v]

DEDICATED
TO
MISS C. S. BURNE,
WHO FIRST GUIDED MY STEPS
INTO THE PATHS OF
ANTHROPOLOGY

[vi]

[vii]

It is becoming an axiom in anthropology that what is needed is not discursive treatment of large subjects but the minute discussion of special themes, not a ranging at large over the peoples of the earth past and present, but a detailed examination of limited areas. This work I am undertaking for Australia, and in the present volume I deal briefly with some of the aspects of Australian kinship organisations, in the hope that a survey of our present knowledge may stimulate further research on the spot and help to throw more light on many difficult problems of primitive sociology.

We have still much to learn of the relations of the central tribes and their organisations to the less elaborately studied Anula and Mara. I have therefore passed over the questions discussed by Dr Durkheim. We have still more to learn as to the descent of the totem, the relation of totem-kin, class and phratry, and the like; totemism is therefore treated only incidentally in the present work, and lack of knowledge compels me to pass over many other interesting questions.

The present volume owes much to Mr Andrew Lang. He has read twice over both my typescript MS, and my proofs; in the detection of ambiguities and the removal of obscurities he has rendered my readers a greater service than any bald statement will convey; for his aid in the matter of terminology, for his criticisms of ideas already put forward and for his many pregnant suggestions, but inadequately worked out in the present volume.[viii] I am under the deepest obligations to him; and no mere formal expression of thanks will meet the case. I have been more than fortunate in securing aid from Mr Lang in a subject which he has made his own.

I do not for a moment suppose that the information here collected is exhaustive. If any one should be in a position to supplement or correct my facts or to enlighten me in any way as to the ideas and customs of the blacks I shall be obliged if he will tell me all he knows about them and their ways. Letters may be addressed to me c/o the Anthropological Institute, 3 Hanover Sq., W.

NORTHCOTE W. THOMAS.

Buntingford,
Sept. 11th, 1906.

[ix]

[x]

[xi]

[xii]

[xiii]

[xiv]

[1]

INTRODUCTORY.

The passage from what is commonly termed savagery through barbarism to civilisation is marked by a change in the character of the associations which are almost everywhere a feature of human society. In the lower stages of culture, save among peoples whose organisation has perished under the pressure of foreign invasion or other external influences, man is found grouped into totem kins, intermarrying classes and similar organised bodies, and one of their most important characteristics is that membership of them depends on birth, not on the choice of the individual. In modern society, on the other hand, associations of this sort have entirely disappeared and man is grouped in voluntary societies, membership of which depends on his own choice.

It is true that the family, which exists in the lower stages of culture, though it is overshadowed by the other social phenomena, has persisted through all the manifold revolutions of society; especially in the stage of barbarism, its importance in some directions, such as the regulation of marriage, often forbidden within limits of consanguinity much wider than among ourselves, approaches the influence of the forms of natal association which it had supplanted. In the present day, however, if we set aside its economic and steadily diminishing ethical [2]sides, it cannot be compared in importance with the territorial groupings on which state and municipal activities depend.

If the family is a persistent type the tribe may also be compared to the modern state; it is, in most parts of the world, no less territorial in its nature; membership of it does not depend among the Australians on any supposed descent from a common ancestor; and though residence plus possession of a common speech is mentioned by Howitt as the test of tribe, it is possible in Australia, under certain conditions1, to pass from one tribe to another in such a way that we seem reduced to residence as the test of membership. This change of tribe takes place almost exclusively where tribes are friendly, so far as is known; and we may doubt whether it would be possible for a stranger to settle, without any rite of adoption, in the midst of a hostile or even of an unknown tribe; but this is clearly a matter of minor importance, if adoption is not, as in North America, an invariable element of the change of tribe. Although membership of a tribe is thus loosely determined, tribesmen feel themselves bound by ties of some kind to their fellow-tribesmen, as we shall see below, but in this they do not differ from the members of any modern state.

But in Australia the importance of the tribe, save from an economic point of view, as joint owner of the tribal land, is small compared with the part played in the lives of its members by the intratribal associations, whose influence is recognised without, as within the tribe. These associations are of two kinds in the lowest strata of human society; in each case membership is determined by birth and they may therefore be distinguished as natal associations. In the one case, the kinship groups such as totem kins, phratries, etc., an individual remains permanently in the association into which he is born, special cases apart, in which by adoption he passes out of it and joins another by means of a legal fiction2. The other kind of association, to which the name age-grades is applied, is composed of a series of grades, through which, concomitantly with the [3]performance of the rites of initiation obligatory on every male member of the community, each man passes in succession, until he attains the highest. In the rare cases where an individual fails to qualify for the grade into which his coevals pass, and remains in the grade of "youth" or even lower grades, he is by birth a member of one class and does not remain outside the age-grades altogether.

In the element of voluntary action lies the distinction between age-grades and secret societies, which are organised on identical or similar lines but depend for membership on ceremonies of initiation, alike in the lowest as in the highest grade. Such societies may be termed voluntary. The differentia between the natal and the voluntary association lies in the fact that in the former all are members of one or other grade, in the latter only such as have taken steps to gain admission, all others being simply non-members.

Although primâ facie all these forms of association are equally entitled to be classed as social organisations, the use of this term is limited in practice, at any rate as regards Australia, and is the accepted designation of the kinship form of natal associations only; for this limitation there is so far justification, that though they perhaps play a smaller part in the daily life of the people than the secret societies of some areas, with their club-houses and other features which determine the whole form of life, the kinship associations are normally regulative of marriage and thus exercise an influence in a field of their own.

Marriage prohibitions in the various races of mankind show an almost endless diversity of form; but all are based on considerations either of consanguinity or kinship or on a combination of the two. The distinction between consanguinity and kinship first demands attention; the former depends on birth, the latter on the law or custom of the community, and this distinction is all-important, especially in dealing with primitive peoples. With ourselves the two usually coincide, though even in civilised communities there are variations in this respect. Thus, according to the law of England, the father of an illegitimate child is not akin to it, though ex hypothesi there is a [4]tie of blood between them. In England nothing short of an Act of Parliament can make them akin; but in Scotland the subsequent marriage of the father with the mother of the child changes the legal status of the latter and makes it of kin with its father. These two examples make it abundantly evident that kinship is with us a matter of law.

Among primitive peoples kinship occupies a similar position but with important differences. As with us, it is a sociological fact; custom, which has among them far more power than law among us, determines whether a man is of kin to his mother and her relatives alone, or to his father and father's relatives, or whether both sets of relatives are alike of kin to him. In the latter case, where parental kinship prevails, the limits of the kin are often determined by the facts of consanguinity. In the two former cases, where kinship is reckoned through males alone or through females alone, consanguinity has little or nothing to do with kinship, as will be shown more in detail below.

Kinship is sociological, consanguinity physiological; in thus stating the case we are concerned only with broad principles. In practice the idea of consanguinity is modified in two ways and a sociological element is introduced, which has gone far to obscure the difference between these two systems of laying the foundations of human society. In the first place, custom determines the limits within which consanguinity is supposed to exist; or, in other words, at what point the descendants of a given ancestor cease to be blood relations. In the second place erroneous physiological ideas modify the ideas held as to actually existing consanguine relations, as we conceive them. The latter peculiarity does not affect the enquiry to any extent; it merely limits the sphere within which consanguinity plays a part, side by side with kinship, in moulding social institutions. If an Australian tribe, for example, distinguishes the actual mother of a child from the other women who go by the same kinship name, they may or may not develop on parallel lines their ideas as to the relation of the child and his real father. Some relation will almost certainly be found to exist between them; but it by no means follows that it arises from any idea of consanguinity. In other communities potestas and not consanguinity is held to [5]determine the relations of the husband of a woman to her offspring; and it is a matter for careful enquiry how far the same holds good in Australia, where the fact of fatherhood is in some cases asserted to be unrecognised by the natives. In speaking of consanguinity therefore, it must be made quite clear whether consanguinity according to native ideas or according to our own ideas is meant.

The customary limitations and extensions of consanguinity, on the other hand, cause more inconvenience. They are of course sometimes combined with the other kind, which we may term quasi-physiological, but with this combination we need not deal, as we are concerned to analyse only on broad lines the nature of these elements. Just as, with us, kinship and consanguinity largely coincide, so with primitive peoples are the kinship organisations immense, if one-sided, extensions of blood relationship, at all events in theory. In many parts of the world a totem kin traces its descent to a single male or female ancestor; and even where, as in Australia, this is not the case, blood brotherhood is expressly asserted of the totem kin3.

Entry into the totem kin may often be gained by adoption, though not apparently in Australia, and the blood relationship thus becomes an artificial one and partakes, even if the initial assumption be accepted as true, far more of the nature of kinship than of consanguinity. In Australia, and possibly in other parts of the world, there is a further extension of natal kinship. Although the tribe is not regarded as descended from a single pair, its members are certainly reckoned as of kin to each other in some way; the situation may be summarised by saying that under one of the systems of kinship organisation (the two-phratry), half of the members of the tribe in a given generation are related to a given man, A, and the other half to his wife. More than one observer assures us that there is a solidarity about the tribe, which regards some, if not all other [6]tribes as "wild blacks," though it may be on terms of friendship and alliance with certain neighbours, and feel itself united to them by a bond analogous to, though weaker than, that which holds its own members together.

If however a homonymous totem kin exists even in a hostile or absolutely unknown tribe, a member of it will be regarded, as we learn from Dr Howitt, as a brother. How this view is reconciled with the belief that the tribe in question is alien and in no way akin to that in which the other totem kin is found, is a question of some interest for which there appears to be no answer in the literature concerning the Australian aborigines.

Even if, therefore, we had reason to believe that all totem kins in a given tribe or group of tribes could make out a good case for their descent from single male or female ancestors, which is far from being the case, we should still have to recognise that kinship and not consanguinity is the proper term to apply to the relationship between members of the same group. For, as we have seen, it may be recruited from without in some cases, while in others, persons who are demonstrably not of the same blood, are regarded as totem-brethren by virtue of the common name.

Enough has now been said to make clear the difference between consanguinity and kinship and to exemplify the nature of some of the transitional forms. As we have seen, it is on considerations of either consanguinity or kinship that many marriage prohibitions are based.

Marriage prohibitions depend broadly on three kinds of considerations: (1) Kinship, intermarriage being forbidden to members of the same kinship group; a brief introductory sketch of the nature and distribution of kinship groups will be found below. (2) Locality. In New Guinea, parts of Australia, Melanesia, Africa, and possibly elsewhere, local exogamy is found. By this is meant that the resident in one place is bound to go outside his own group for a mate, and may perhaps be bound to seek a spouse in a specified locality. This kind of organisation is in Australia almost certainly an offshoot of kinship organisation (see p. 10), and is primâ facie due to the same cause in other areas. (3) (a) consanguinity, and (b) affinity. The first of these [7]considerations is regulative of marriage even in Australia, where the influence of kinship organisations is in the main supreme in these matters. We learn from Roth and other authorities that blood cousins, children of own brother and sister, may not marry in North-West Central Queensland, although the kinship regulations designate them as the proper spouses one for the other. (b) Considerations of affinity, the relations set up by marriage, do not affect the status of the parties, so far as the legality of marriage is concerned, till a somewhat higher stage is reached.

In the present work we are concerned with kinship groups and the marriage regulations based on them. A kinship group, whether it be a totem kin, phratry, class, or other form of association, is a fraction of a tribe; and before we proceed to deal with kinship organisations, it will be necessary to say a few words on the nature of the tribe and the family. In Australia the tribe is a local aggregate, composed of friendly groups speaking the same language and owning corporately or individually the land to which the tribe lays claim. A change of tribe is effected by marriage plus removal, and possibly by simple residence; children belong to the tribe among which their parents reside. In the ordinary tribe each member seems to apply to every other member one or other of the kinship terms; and this no doubt accounts for the feeling of tribal solidarity already mentioned. There are however certain tribes in which the marriage regulations, as with the Urabunna, so split the intermarrying fractions, that the tribe is, as it were, divided into water-tight compartments; how far kinship terms are applied under these circumstances our information does not say.

The tribe is defined by American anthropologists as a union of hordes or clans for common defence under a chief. The American tribe differs in two respects, at least, from the Australian tribe; in the first place, marriage outside the tribe is exceptional in America and common in Australia; in the second place, the stranger gains entrance to the American tribe only by adoption; and we may probably add, thirdly, that the American tribe does not invariably lay claim to landed property or hunting rights.

[8]

The tribe is subdivided in various ways. In addition to the various forms of natal and other associations, there is, at any rate in Australia, a local organisation; the local group is often the owner of a portion of the tribal area. This local group again falls into a number of families (in the European sense), and the land is parcelled out among them in some cases, in others it may be the property of individuals. But there is a great lack of clearness with regard to the bodies or persons in whom landed property is vested. The composition of the local group varies according to the customs of residence after marriage, and the rules by which membership of the kinship organisation is determined. These two forces acting together may produce two types of local group: (1) the mixed group, in which persons of various kinship organisations are scattered at random; (2) the kin group, in which either all the males or all the females together with the children are members of one kinship organisation.

Save in the rare instances of non-exogamous kinship groups, the family necessarily contains one member, at least, whose kin is not the same as that of the remainder; this is either the husband or the wife, according as descent is reckoned in the female or the male line; where polygyny is practised, this unity may go no further than the phratry or the class, each wife being of a different totem kin.

Although it frequently happens that the children belong to the kin which through one of the parents or otherwise exercises the supreme authority in the family, it is far from being the case that there is invariable agreement between the principles on which kinship and authority are determined. Three main types of family may be distinguished: (1) patripotestal, (2) matripotestal, (a) direct, and (b) indirect, in which the authority is wielded by the father, mother, and mother's relatives, in particular her brothers, respectively. Innumerable transitional forms are found, some of which will be mentioned in the next chapter, which deals with the rule of descent by which membership of natal groups is determined.

Turning now to kinship organisations, we find that the most widely distributed type is the totem kin, in fact, if we except the Hottentots and a few other peoples among whom no trace [9]of it is found, it is difficult to say where totemism has not at one time or another prevailed. It is found as a living cult to-day among the greater part of the aborigines of North and South America, in Australia, and among some of the Bantu populations of the southern half of Africa. In more or less recognisable forms it is found in other parts of Africa, New Guinea, India, and other parts of the world. In the ancient world its existence has been maintained for Rome (clan Valeria etc.), Greece, and Egypt, but the absence of information as to details of the social structure renders these theories uncertain.

Aberrant cases apart, totemism is understood to involve (1) the existence of a body of persons claiming kinship, who (2) stand in a certain relation to some object, usually an animal, and (3) do not marry within the kin.

Passing over the classes, which are peculiar to Australia and will be fully dealt with below, we come to a more comprehensive form of kinship organisation in the phratries. These are a grouping of the community in two or more exogamous divisions, between which the totem kins, where they exist, are distributed. The essential feature of a phratry is that it is exogamous; its members cannot ordinarily marry within it, and, where there are more than two phratries, there may exist rules limiting their choice to certain phratries.4

This dual or other grouping of the kins is widely found in North America, the number of phratries ranging from two among the Tlinkits, Cayugas, Choctaws, and others, to ten among the Moquis of Arizona. As in Australia, the totem kins bearing the same eponymous animal as the phratry are usually, e.g. among the Tlinkits, found in the phratry in question. Exceptions to this rule are found among the Haida, where both eagle and raven are in the eagle phratry.

The Mohegan and Kutchin phratries call for special notice. The kins of the former are arranged in three groups: wolf, turtle, and turkey; and the first phratry includes quadrupeds, the second turtles of various kinds and the yellow eel, and the third birds. We find a parallel to these phratries in the groups [10]of the Kutchin, but in the latter case our lack of knowledge of the tribe precludes us from saying whether totem kins exist among them, and, if so, how far the grouping is systematic; the Kutchin groups, according to one authority, are known by the generic names of birds, beasts, and fish. As a rule, however, no classification of kins is found, nor are the phratry names specially significant.

Dual grouping of the kins is also found in New Guinea, the Torres Straits Islands, and possibly among the ancient Arabs5; but evidence in the latter case has not been systematically dealt with.

Other peoples have a similar dichotomous organisation; but it is either not based on the totem kins or they have fallen into the background.

In various parts of Melanesia we find the people divided into two groups, each associated with a single totem or mythological personage, and sexual intercourse, whether marital or otherwise, is strictly forbidden between those of the same phratry6. In India the Todas have a similar organisation7, and the Wanika in East Africa8.

Customs of residence and descent affect the distribution of the phratries within the tribe, no less than the composition of the local group. With patrilineal descent they tend to occupy the tribal territory in such a way that each phratry becomes a local group. With the disappearance of phratry names this would be transformed into a local exogamous group, which is, however, indistinguishable from the local group of the same nature which is the result of the development of a totem kin under similar conditions.

As a rule kinship organisations descend in a given tribe either in the male line or in the female. Among the Ova-Herero, however, and other Bantu tribes, there are two kinds of organisation, one—the eanda—descending in female line and regulative of marriage, is clearly the totem kin; property remains [11]in the eanda, and consequently descends to the sister's son. The other—the oruzo—descends in the male line; it is concerned with chieftainship and priesthood, which remain in the same oruzo, and the heir is the brother's son.9

This dual rule of descent brings us face to face with the question of how membership of kinship groups is determined.

[12]

DESCENT.

In discussions of the origin and evolution of kinship organisations, we are necessarily concerned not only with their forms but also with the rules of descent which regulate membership of them. Until recently the main questions at issue were twofold: (1) the priority or otherwise of female descent; (2) the causes of the transition from one form of descent to another. Of late the question has been raised whether in the beginning hereditary kinship groups existed at all, or whether membership was not rather determined by considerations of an entirely different order. Dr Frazer, who has enunciated this view, maintains that totemism rests on a primitive theory of conception, due to savage ignorance of the facts of procreation.10 But his theory is based exclusively on the foundation of the beliefs of the Central Australians and seems to neglect more than one important point which goes to show that the Arunta have evolved their totemic system from the more ordinary hereditary form. Whether this be so or not, it is difficult to see how any idea of kinship could arise from such a condition of nescience. If we take the analogous case of the nagual or "individual totem" there seems to be no trace of any belief in the kinship of those who have the same animal as their nagual, but are otherwise bound by no tie of relationship. Yet if Dr Frazer's theory were correct, this is precisely what we ought to find.

[13]

This is, however, no reason for rejecting the general proposition that kinship, at its origin, was not hereditary; or, more exactly, that the beginnings of the kinship groups found at the present day may be traced back to a point at which the hereditary principle virtually disappears, although the bond of union and perhaps the totem name already existed. If, as suggested by Mr Lang, man was originally distributed in small communities, known by names which ultimately came to be those of the totem kins, we may suppose that daily association would not fail to bring about that sense of solidarity in its members which it is found to produce in more advanced communities. In the case of the tribe an even feebler bond, the possession of a common language, seems to give the tribesmen a sense of the unity of the tribe, though perhaps other explanations may be suggested, such as the possession in common of the tribal land, or the origin of the tribe from a single blood-related group. However this may be, it seems reasonable to look for one factor of the first bond of union in the influence of the daily and hourly association of group-mates. On the other hand, if, as Mr Lang supposes, the original group was a consanguine one, the claims of the factor of consanguinity and perhaps of foster brotherhood and motherhood cannot be neglected. It may be true, as Dr Frazer argues, that man was originally and still is in some cases ignorant of physiological facts. But all races of man and a great part of the rest of the animal kingdom show us the phenomena of parental affection, of care for offspring and sometimes of union for their defence. This does not, it may be noted, imply any predominance of the mother.11

[14]

We may suppose that the idea of kinship or the recognition of consanguinity, whichever be the more correct term to apply to these far-off developments of the factors of human society, extended only by degrees beyond the limits of the group. First, perhaps, came the naming of the group, already, it may be, exogamous; then came the recognition of the fact that those members of it, viz. the women, who passed to community B after being born and having resided for years in community A, were in reality, in spite of their change of residence, still in fact the kin of community A; finally came the step of assigning to their children the group names which were retained by their mothers from the original natal groups. This brings us face to face with the first of the fundamental questions of descent, to which allusion has been made.

It is commonly assumed by students of primitive social organisation that matrilineal descent is the earlier and that it has everywhere preceded patrilineal descent; but the questions involved are highly complicated and it can hardly be said that the subject has been fully discussed.

Much of what has been said on the point has been vitiated by the introduction of foreign factors. Thus, the child belongs to the tribe of the father where the wife removes to the husband's local group or tribe. But though it may be taken as a mark of matrilineal institutions, often associated with matria potestas or its analogue the rule of the mother's brother, that the husband should remove and live with the wife, we are by no means entitled to say that the removal of the wife to the husband implies a different state of things. Customs of residence are no guide to the principles on which descent is regulated. Consequently it is entirely erroneous to import into the discussion with which we are concerned, viz. the rules by which kinship is determined, any considerations based on the rules by which membership of a tribe is settled.

Similarly, no proof of the existence of paternal authority in the family throws any light on the question of whether the children belong to the kin of the father rather than of the mother. Where the mother or mother's brother is the guardian, we are usually safe in assuming that descent is or has been until [15]recently matrilineal. But from the undisputed existence of patria potestas no similar inference can be drawn.

Again, as will be shown below, not even the tie of blood between parent and child, confined though it may be in the opinion of the people whose institutions are in question, to a single parent, is an index to the way in which is determined the kinship organisation to which the child belongs.

It is therefore clear that the utmost discrimination is necessary in dealing with these questions; rules of descent must be kept apart from matters which indeed influence the evolution of the rules but are in no way decisive as to their form at any given moment.

Returning now to the alleged priority of matrilineal descent in determining the kinship organisation into which a child passes, it may be said that whereas evidences of the passage from female to male reckoning may be observed,12 there is virtually none of a change in the opposite direction. In other words, where kinship is reckoned in the female line, there is no ground for supposing that it was ever hereditary in any other way. On the other hand, where kinship is reckoned in the male line, it is frequently not only legitimate but necessary to conclude that it has succeeded a system of female kinship. But this clearly does not mean that female descent has in all cases preceded the reckoning of kinship through males. Patrilineal descent may have been directly evolved without the intermediate stage of reckoning through females.

The problem is probably insoluble. No decisive data are available, for the mere absence of traces of matrilineal descent does not necessarily prove more than that it had long been superseded by reckoning of kinship through males. All that can be said is that in the kinship organisations known to us female descent seems to have prevailed in the vast majority of cases and probably existed in the residual class of indeterminable examples.

With patria potestas it is, of course, different. There can be little doubt that it might and probably did develop in the absence [16]of kinship organisations and in a state of society where consanguinity is no real bond after the children have reached puberty. If therefore under such circumstances a kinship organisation were to come into existence, either independently or by transmission, it might well be that patrilineal principles prevailed from the first. But of such a case we have no knowledge. It may perhaps be questioned whether the actually existing peoples who appear to have no kinship organisations, such as the Hottentots, the Bushmen, the Veddahs and perhaps the Fuegians, are not in this state rather as a result of the break-up of their former organisation than because they have never evolved kinship groups. But our knowledge in these matters is lamentably small and the problem is not one which calls for discussion here.

The second fundamental problem relating to rules of descent is that of the cause of the transition from matrilineal to patrilineal descent. The subject needs to be discussed in detail for each particular area before general conclusions can be formulated; it is quite possible that the causes will be found to differ widely; for no general rule can be laid down as to the relations between matrilineal descent and other cultural conditions.

All that can be attempted here is an examination of the various elements in the problem so far as it affects Australia. To this may be prefixed a further discussion of the origin of matrilineal descent with especial reference to Australian conditions.

It is commonly assumed that in a pure matrilineal community, the husband removes to the wife's local group (matrilocal marriage), or if not that, that at any rate the authority in the family rests in the hands of the mother's brothers, who are also the heirs to the exclusion of the children. But of any such custom of removal there is but the very slenderest evidence in Australia. According to Howitt it occurs occasionally in Victoria and among the Dieri; among the Wakelbura it is done only if a man elopes with a betrothed woman and the man to whom she was betrothed dies; among the Kuinmurbura it seems to have been a recognised thing for a man who married a woman of another tribe to remove, but in this case he took no part in [17]intertribal warfare13. In all these cases, the Kurnai excepted, descent is reckoned in the female line.

If however Dr Howitt's informant, who does not seem to have been particularly accurate in many cases, is to be relied on, the removal of the husband to the wife's group is also found among the patrilineal Maryborough tribes, though only if the woman belonged to a distant tribelet, whatever that may be14. To this information is added the statement that in such cases the husband joined his wife's tribe for purposes of hostilities also and that it has happened that a son has come into conflict with his father under these circumstances and endangered his life with full knowledge of what he was doing. There is, it is true, no definite statement to the effect that children in these tribes take their totems from the father, but we may assume that it is the case. If therefore the statement in question is accurate, it is a pretty clear proof of the break-up of the social system; for under no circumstances does the totem-kinsman, as a rule, violate the sacro-sanctity of his own flesh. It cannot therefore be argued that the fact of removal in the Maryborough tribes is any very strong evidence of the primitive nature of the custom. In the other tribes, on the other hand, it is distinctly stated that the practice prevails only when marriage takes place between members of two different tribes, and among the Wakelbura only exceptionally even when the wife is of an alien folk. Whatever else the custom proves in these cases, it certainly evidences the existence of friendly relations between the tribes in question; for if it were otherwise the man would hardly be disposed to give up the security of his own people for the perils of a strange community; on the other hand it is hardly likely that the man's tribe would allow him to pass over to the ranks of the strangers, nor would they view with equanimity the loss of effective fighting strength which would result from the fact that his children too would be numbered against them, not for them, if it came to hostilities. The custom is therefore clear evidence of [18]fairly permanent friendly relations in the district in question; and it is plain that we cannot assume these to have existed in more primitive times. It is therefore difficult to see in what way the present day practices lend support to the theory that the original usage was for the husband to remove to his wife's group. For, be it noted, there is not a single case, unless we include the anomalous Kurnai, in which the husband removes to his wife's group within his own tribe; but clearly this is the custom to which the removal theory applies. So far, therefore, as Australia is concerned, the removal theory falls to the ground; it cannot of course be disproved, but we are not justified in assuming that matrilineal descent and matria potestas are due to a custom of removal.

Inasmuch as patrilocal15 marriage involves descent of group and tribal property rights in the male line, it might appear that in rejecting the hypothesis of a prior stage of matrilocal marriage, we are involving ourselves in difficulties; for it is clearly not easy to see how descent could come to be reckoned through the mother, while property descended through the father. But it is obviously unnecessary in the first place to regard the individual rights of property as originating simultaneously or under the same conditions as the rules as to kinship or even communal property; there is nothing to show how long the present system of land tenure in Australia has held good, and it is clearly one which points to a certain growth of population; for if the local group were remote from their neighbours, there would be little need to encroach; moreover, the exact delimitation of territory now in practice is a thing of long growth.

Further consideration however shows that it is only by a confusion of thought that we can speak of land descending in the male line (that is, of course, in respect of group rights, not private property, to which we return later); strictly speaking the descent of landed property is neither in the male nor the female line but local. A man who removes to his wife's tribe is, so far as we can see, as truly part owner of the tribal land as if he were himself a member of the tribe by birth within its limits. The suggested difficulty, therefore, does not exist, and the conclusion as to removal customs holds good.

[19]

We may now examine the relation of matriliny to the seat of authority in the family. Questions of potestas naturally range themselves under more than one head. We have (1) the relation of the husband (a) to the wife and (b) to the children; (2) the relation of the mother to the children, and closely connected with this the influence of the mother's brother; finally (3) we have the position of the widow, a matter indeed more intimately connected with inheritance from a legal point of view but in Australia more closely connected with potestas than in countries where slavery is a recognised institution.

Small as is our information on Australian jurisprudence, it is certain that the husband enjoys practically unrestricted rights over the person of his wife, pirrauru and similar customs apart. He may at will lend her or hire her out to strangers; he may punish her infidelity, disobedience or awkwardness by chastisement, not stopping short of the infliction of spear or club wounds; he may even, according to Roth16, go so far as to kill her and yet get off scot free, his only duty in such a case being to provide a sister for the brothers of his dead wife to kill in retaliation.

This custom suggests that the kin to which the woman belongs claim a certain property in her even after she is married, and this partial proprietorship naturally implies a slight protecting influence; for it would clearly not be in every case easy for the homicidal male to find a sister ready to go out and be killed as a set-off to his murdered wife. We should not, it is true, overlook the fact that the customs of the Pitta-Pitta differ from those of many of the Australian tribes, in that exchange of sisters is not practised. Otherwise it would be tempting to argue that this proprietorship in the women of their kin may go back to the time of Mr Lang's connubial groups and help to explain the reckoning of descent through females. For clearly, if a woman still belongs in a sense to the group she has left, so may her children belong to the same group, inasmuch as their relationship to her is, to us at any rate, unmistakeable. If any evidence could be produced for the widespread existence of the custom (found in various parts of the globe, though not, up to [20]the present, in Australia), according to which the widow and her children remove to her own district, some probability would be imparted to this hypothesis.

The ordinary rule as regards punishment inflicted by the husband on the wife seems to be that he may go any length short of doing her a mortal injury, without being liable to be called to account. The punishment of death however may only be inflicted for adultery and certain specified offences without incurring a blood-feud with the woman's relatives.

It is by no means improbable that under the influence of the custom of exchanging sisters there may be a tendency for the control of the kin in this respect to diminish; in fact the Boulia example is only explicable on this hypothesis. At the same time we cannot overlook the fact that elopement, or real marriage by capture, as distinguished from formal abduction, would, so far as we can see, have a similar effect, and the rise of the custom of exchange of sisters would in that case tend to re-establish rather than weaken the power of the woman's kin, at any rate in the first instance.

However this may be, the woman's kin exercises, primâ facie, some kind of protectorship. At the present day the kinship may be matrilineal or patrilineal without affecting their right. But if, before kinship was reckoned at all, this protectorship were exercised for the benefit of the children, we clearly have a possible cause of matriliny.

For a discussion of the question of the inheritance of the deceased's wife by his brother we have more facts at our disposal. As a matter of fact it is a not infrequent custom in Australia for the widow to pass to the deceased husband's brother17; or if she does not become his wife, he decides to whom she shall be allotted18. In no case do the woman's kin seem to have a voice in the selection of her new husband. On the whole therefore the proprietary rights found in the Boulia district seem to be the product of exceptional local conditions. If this is so, it is clear that in the matter of potestas the rights of the woman's kin are now absolutely restricted to protecting her from a death [21]which she has not according to native law deserved and to avenging such a death when it is inflicted by the husband.

The so-called levirate, or right of succession to the widow, is clearly of much importance, so far as questions of dominion are concerned; but as regards the problems of descent the evidence is less easily interpreted. It has sometimes been assumed that the succession of the brother and not the son is a mark of matriliny; but it is clear that where the right of appropriating the widow is concerned, this is very far from being the case, for the simple reason that the real matria potestas would put her at the disposal of the kin from whom she originally came; on the other hand, inasmuch as the son is naturally debarred from marrying his own mother or his tribal mother, who commonly belongs to a class into which he does not marry, there might easily arise in a purely patripotestal and patrilineal tribe a custom of handing over the widow to the father's brother.

On the whole however it seems simplest to regard the matter as one in which the rights are determined by no considerations of inheritance or descent but simply by the rule that the property in the woman remains vested in the body of purchasers. For it must be remembered that not only an own but also a tribal sister may be given in exchange for a wife. From this it follows that, theoretically at any rate, the contracting parties are corporations rather than individuals, and in this case the death of the individual on whose behalf the transaction has been effected does not extinguish the proprietary rights acquired by handing over a woman, standing in the relation of sister to the one corporation, in exchange for another woman standing in the relation of sister to the other corporation.

If this solution is correct, it is unnecessary to go into the complicated question of the relation of brother-inheritance to matriliny and patriliny. For it is by no means clear that it is an exemplification of the former rather than the latter principle. It may, of course, be argued that brothers succeed as children of the same mother; but against this must be set the fact that they are also children of the same father; for uncertain paternity can only be a vera causa where pirrauru and similar customs are found; and even here the pre-eminence of the primary husband [22]might well be held to determine the legal paternity of the children, which is, of course, especially in Africa, a matter of potestas rather than procreation. However this may be, the position of the widow does not appear to invalidate the guardianship origin of matriliny.

We now turn to the question of why male tends to take the place of female descent. The possible factors are (1) authority in the family, (2) the rise of chieftainship and inheritance generally, and (3) the organisation of the family group. Of the authority of father or mother over the children, there is not much trace in Australia except in the most youthful period of the pre-adult life. It is for example exceptional for a parent to correct a child. As to who decides in cases of infanticide we have unfortunately too little information to be able to generalise. Only in one important step—that of betrothal—have we anything like adequate information, and the interrelations between rule of descent and potestas are found to be in this case sufficiently clear, though it is not clear on what principle it is decided who shall exercise the right.

Taking first tribes with matrilineal descent, we find that the Barkinji, the Wakelbura, the Dieri, and in some cases the Wollaroi, assign the right of betrothal to the mother or mother's brother19. In other cases, transitional forms, the father, his elder brother, or the girl's brothers decide, or else the parents or two of these persons jointly20. Among the Mukjarawaint the betrothal rested in part with the paternal grandparents21; it may be noted that the grandfather had to decide also whether a child should be brought up or killed. Among the Kuinmurbura it falls to the mother's brother's son or the father's sister's son, who is, apparently, entitled to marry the girl himself22.

Turning now to tribes with male descent, we find that the father, his brother, or the parents, almost invariably make the decision23. Among the eight-class tribes, Spencer and Gillen assert in one place24 that the mother's brother betroths a girl; [23]but this is contradicted in two other passages25, and cannot be regarded as reliable.

On the whole therefore it appears that while there are some survivals of matria potestas into patrilineal descent, and in the matrilineal stage transitional forms are found, the right of betrothal tends to pass from the mother's to the father's side, when the rule of descent changes; but there is little to show how far a change in the right of betrothal tends to cause a change in the rule of descent.

A curious fact may be noted here, which goes far to demonstrate the absolutely heterogeneous nature of kinship and consanguinity, and suggests that descent is not reckoned in the female line on account of any supposed specially close connection between the mother and her offspring. Of the four tribes among which, according to Howitt, the child is regarded as the offspring of the father alone26, the mother being only its nurse, two, the Yuin and Kulin, have male descent; two, however, the Wolgal and Tatathi, have female descent, and among the latter, in addition, the right of betrothal lies with the mother or mother's brother.

On the whole, therefore, it may be said that no questions of potestas seem to have exercised any influence in bringing about the transition from matrilineal to patrilineal descent. It does not appear necessary, therefore, to do more than allude in passing to a fact which may well have had something to do with the decay of matria potestas, at any rate, so far as the mother's brother is concerned, even if it did not actively hasten the coming of patria potestas. This fact is the considerable size of the area over which, with the rise of the so-called nations, it is possible to select a wife. The more remote geographically the mother's relatives, the less their influence. Allowance must of course be made for the opportunities of discussion afforded by the great gatherings of the tribes; but the wider area of bride-choice must have shaken the authority of the brother.

It has been remarked above that there is no well-established case of the right of betrothal being assigned on patrilineal prin[24]ciples in a matrilineal tribe. The influence of the father's brother is not necessarily a mark of patrilineal tendencies, except in so far as all patria potestas is such. That the elder brother has authority in this case is no more decisive than that the elder brother has authority in cases of betrothal; it is no more an exemplification of the simple patria potestas, which has already been shown to be universal and under but slight limitations so far as the wife is concerned. From the point of view of potestas, it is a great advance that the father should be able to dispose of his own daughter in marriage; but if we may judge by the survival of matria potestas into patriliny, the cases of patria potestas under matriliny cannot have exercised an important influence in bringing about a change in the rule of descent.

The case of the power of the girl's own brother is somewhat different. Primâ facie it appears to owe its origin to the fact that it is the brothers who are mainly interested in the transaction, inasmuch as it is to them that wives come in exchange for the sisters given in marriage. Consequently we cannot, as has already been the case with the so-called levirate, assign the practice definitely either to matripotestal or patripotestal customs, for father's and mother's authority are alike overruled.

It has already been stated that we have but few data for estimating the influence of the right of betrothal on the rule of descent. Clearly the father has little to gain from the fact that his daughter follows him rather than the mother, when the inevitable effect of the marriage regulations is to make her children of the phratry and totem of her husband, and consequently to make them of a different phratry and totem from her father. Under matriliny on the other hand there is nothing to prevent the grandchildren from being of the same totem as the grandfather, and they are necessarily of the same class in a four-class tribe. If considerations with regard to the phratry and totem of the grandchildren played any part in bringing about a change in the rule of descent, this must have been based on a review of the changes that would be brought about in the position of the son's and not the daughter's offspring. But this is unlikely.

But on the other hand the father's disposal of the daughter's [25]hand is indirectly a means of increasing his influence both with his son and in general. If the son gains his wife by an exchange of sisters, the father's authority is obviously increased. But we do not know how far this factor of the right of betrothal has operated.

Turning now to questions of inheritance, we find that properly speaking the hereditary chief is unknown in Australia. There is a tendency for the son of the tribal headman to succeed his father, but it is subject to exceptions. Moreover, it is by no means a universal rule for the tribe to have an over-headman; it may be ruled by the council of district headmen. In any case the influence of the quasi-hereditary character of the over-headmanship upon the rule of descent cannot but have been comparatively slight.

It is, on the other hand, usual for the local group and the totem kin to have headmen. In the case of the latter, age is often the qualification, as among the Dieri27; in such cases there is no possible effect on the rule of succession. But among some of the Victorian tribes with matrilineal descent the rule is for the son to follow the father in the headmanship28; and the same is the case, as we should expect, among the patrilineal eight-class tribes29. The most important tribe in which hereditary headmanship is combined with female descent is the Wiradjeri30; their neighbours, the Kamilaroi, showed marked respect to the son of a headman, if he possessed ability, though they did not, apparently, make him his father's successor31.

On the whole, then, we cannot assign much weight to this element in the list of possible causes of the transition.

Of inheritance of chattels or land and fixtures we know little. From Spencer and Gillen we learn that among the Warramunga the mother's brother, or daughter's husband, succeeds to the boomerangs, and other moveable property32. Among the Kulin and the Kurnai inheritance in the male line seems to have been the rule. In the Adelaide district, as we learn from Gerstaecker33, individual property in land was known; it descended in the [26]male line. Among the Turribul there was individual property in bunya-bunya trees; these too devolved from father to son34.

On the other hand on the Bloomfield property in zamia nut grounds has vested in women and descends from mother to daughter35; but in this remarkable variant we see, of course, not the influence of the mother's kin, but female influence or rather the right of females to the produce of their labour. In respect of other property, inheritance in North Queensland is in the male line, for it descends to blood brothers and remains in the same exogamous group from generation to generation.

This brings us to the question of the part played by the local group in causing the change from female to male descent. Under ordinary circumstances, with female descent, the local group is made up of persons of different phratries and totems; in any case, just as the phratry and totem of the members of the individual family change from generation to generation, the complexion of the local group is liable to be completely changed; though in practice the changes in one direction are no doubt counterbalanced by changes in the other, so that the net result may be nil, when the original differences were small. But we cannot suppose that the group was often evenly balanced; and a change in the rule of descent would in that case have important results for the local group and in any case for the individual family.

The importance of the difference in the constitution of the local group under descent in the male line is seen when we reflect that in the normal tribe the totem kin is practically the unit for many purposes. If, for example, an emu man has killed, let us say, an iguana man, it is the duty of the iguana men to avenge the death of their kinsman. Their vengeance need not, however, fall on the original perpetrator of the deed; according to the rules of savage justice all the emu men are equally responsible with the culprit; consequently it suffices to kill the first emu person whom they can find. Conversely, those to whom an emu man looks for defence, when he is attacked, or assistance, when he wishes to abduct a wife or anything of that sort, are his fellow emu [27]men. It is therefore clear that the rule of male descent gives far greater security to the members of a local group; for they are surrounded by kinsmen. Under the rule of female descent, on the other hand, they probably have some kinsmen in the same group but equally a considerable number of members of other totem kins.

Self-interest therefore, no less than the natural sympathy between fathers and children, as well as between members of the same group (quite apart from forays and fighting), must have tended to bring about a change in the laws of descent.

The late Major J. W. Powell has already described the transition from matria potestas to patria potestas among the Pueblo peoples. He put it down to economic conditions, which lead the groups to scatter, each under the headship of a male, who is also the husband; this naturally resulted in a weakening of the influence of the mother's brother. It is, however, less clear that it would bring about the decay of the power of the mother herself, which in Australian tribes, at any rate, seems to be independent of the support she obtains from her male relatives.

In Australia, as we have seen, the change from matria to patria potestas had but little influence in bringing about a change in the rule of descent. Here, too, the change in the rule of descent may be put down in the main to economic causes also in a broad sense. Dumping was not in those days a question of practical politics; the problem was to prevent the neighbours from pursuing the policy of the free and open port. The necessity of protecting tribal and group property in land and game would naturally tend to bind men closer and closer, in proportion as the pressure from without became greater. It is perhaps hardly accidental that the main area of male descent is that which has also developed the Intichiuma ceremonies.

If Prof. Gregory's view36 that the occupation of Victoria by the natives dates back no more than 300 years is correct, we may perhaps see in the migration one cause of the rise of patriliny. Anything which tended to shake the influence of the mother's kin would increase the father's power; and the need of [28]protecting newly established groups from the incursions of their neighbours would be more urgent than in older districts. As we have seen, the first mentioned cause has elsewhere had little direct effect; but it may well have played a larger part under the novel conditions of migration and occupation of fresh territory.

In South Queensland the fractionation of tribes seems to have gone further than elsewhere, unless we suppose that we have here an area, where, as in California, pressure from without has crowded together the remnants of many tribes. Although it is not obvious how the multiplication of distinct tribes has favoured patrilineal descent, we may, at any rate, say that the conditions in the area are exceptional; possibly it was more fruitful than the greater part of the continent; if so personal property in the shape of trees, etc., which we have already seen in existence in this area, would play a more important rôle here, and may well have determined the transition to patrilineal descent.

[29]

DEFINITIONS AND HISTORY.

Before proceeding to deal with the Australian facts it will be well to define the terminology to be employed, and give a brief survey of a typical organisation. Looking at the population from the territorial point of view in the first place, we find aggregates of tribes; these may be termed nations. The component tribes are friendly, one with another; they may and often do hold initiation ceremonies and other ceremonials in common; although the language is usually syntactically the same, and though they contain many words in common, the vocabularies differ to such an extent that members of different tribes are not mutually intelligible. How far the occurrence of identical kinship organisation and nomenclature should be taken as indicating a still larger unity than the nation is a difficult question. Primâ facie the nation is a relatively late phenomenon; but the distribution of the names of kinship organisations, as will be shown later, indicates that communication, if not alliance, existed over a wide area at some periods, which it is difficult to suppose were anything but remote.

The idea of the tribe has already been defined. It is a community which occupies a definite area, recognises its solidarity and possesses a common speech or dialects of the same.

Between the tribe and the family occur various subdivisions, known as sub-tribes, hordes, local groups, etc., but without any very clear definition of their nature. It appears, however, that the tribal area is sometimes so parcelled out that property in it [30]is vested, not in the tribe as a whole, but in the local group, which welcomes fellow-tribesmen in times of plenty, but has the right of punishing intruders of the same tribe who seek for food without permission; for a non-tribesman the penalty is death. In some cases the local group is little more than an undivided family including three generations; it may then occupy and own an area of some ten miles radius. In other cases the term is applied to a larger aggregate, the nature and rights of which are not strictly defined; it may number some hundreds of persons and form one-third of the whole tribe; it seems best to denominate such an aggregate by the name of sub-tribe.

The term family may be retained in its ordinary sense.

Superposed on the tribal organisation are the kinship organisations, which, in the case of most Australian tribes, are independent of locality. Leaving out of account certain anomalous tribes, it may be said broadly that an Australian tribe is divided into two sets, called phratries, primary classes, moieties, etc. by various authors; the term used in the present work for these divisions is phratry. Membership of a phratry depends on birth and is taken directly from the mother (matrilineal descent) or father (patrilineal descent).

In Queensland and part of N. S. Wales the phratry is again subdivided, and four intermarrying classes (sometimes called sub-phratries) are formed, two of which make up each phratry. In North Australia and Queensland a further subdivision of each of these classes is found, making eight in all. Descent in the classes is indirect matrilineal or indirect37 patrilineal, the child belonging to the mother's or father's phratry as before, but being assigned to the class of that phratry to which the mother or father does not belong. The classes of father and son together are called a couple. The parent from whom the phratry and class name are thus derived is said to be the determinant spouse.

These phratries and classes regulate marriage. It is forbidden to marry within one's own phratry. This custom is termed exogamy. When the husband removes and lives in his wife's group the marriage is matrilocal; if the wife removes it is patrilocal.

[31]

In addition to the division into classes each phratry is further divided into a number of totem kins. A totem is usually a species of animals or plants; a body of human beings stands in a certain peculiar relation to the totem species and is termed the totem kin; each member of a totem kin is termed a kinsman. Membership of the totem kin usually descends directly from parent to child.

The existence of these kinship organisations is universally recognised. Mr R. H. Mathews has recently asserted the existence of yet another form and at the same time controverted the accepted views as to the operation and meaning of those described above. He distinguishes in certain tribes of New South Wales kinship organisations running across the phratries; these are of two kinds, according to the author, but they do not seem to differ in function. They are termed by Mr Mathews "blood" and "shade" divisions, and are held by him to be the names of the really exogamous groups. The subject is discussed in detail below.

In order to make the working of these regulations plain, let us take as an example the Kamilaroi tribe of N. S. Wales, with two phratries, four classes and various totem kins. The phratries are named Dilbi and Kupathin; Dilbi is divided into two classes, Muri and Kubi; Kupathin into Kumbo and Ipai. The Dilbi totems, which may belong to either of the classes, are kangaroo, opossum and iguana; those of Kupathin are emu, bandicoot and black snake. Every member of the tribe has his own phratry, class and totem; these all come to him by descent.

We have little or no information as to the local grouping of the Kamilaroi tribes, but it was possibly not unlike that of some of the tribes to the north-west. In the case of the latter the tribal area was some 3000 sq. miles in extent, it was split up into smaller areas, thirty or more in number, which were the property of the local groups; a local group consisted frequently of three generations of relatives. When we come to deal below with marriage regulations it will be shown that husband, wife and child under the four-class system all belong to different classes; there were therefore in each group at least three classes, if not four, and consequently members of two [32]phratries. If we assume that the same conditions prevailed among the Kamilaroi, the local groups would then be made up of members of both the Dilbi and Kupathin phratries; and probably all four classes, Muri, Kubi, Ipai and Kumbo, would be found in each group, which in Australia varied in size according to local conditions from 20 or 30 to 200; under special conditions, such as prevailed in the neighbourhood of Lake Alexandrina, the number might run up to 600 or more, but this was exceptional.

From the fact that the totems are divided between the phratries it is clear that the local group may also have members of all the six totem kins mentioned above, among its members.

The rules by which marriage and descent are regulated are apparently very complicated but practically very simple. Taking the Kamilaroi tribe again, the rule is that Muri marries Butha (a female Kumbo) and their children are Ipai and Ipatha: Kubi marries Ipatha and their children are Kumbo and Butha; in each case the children belong to the same phratry as the mother but to the other class in that phratry. This is termed indirect matrilineal descent.

The rule of descent for the totem among the Kamilaroi was simpler; membership of a totem kin descends directly from a mother to her child. The combined effect of these rules is that if, for example, a male Dilbi of the Muri class and iguana totem wants to marry, he must choose a wife of the Kupathin phratry, the Kumbo class, and either the emu, bandicoot, or black snake totems; suppose he marries an emu woman; then his children are of the Kupathin phratry, the Ipai (or Ipatha) class, and the emu totem. These regulations are naturally more complicated among the eight-class tribes; on the other hand, where only phratries and totems are found, but no classes, descent is much simpler; for in each case the child takes the phratry and totem of its mother, where matrilineal descent prevails, or of its father, where patrilineal descent is found.

The general rule in Australia is that the wife goes to live with her husband; in other words, she leaves the local group in which she was born and becomes a member of her husband's local group. The effect of this is very different according as [33]descent is reckoned through the mother or through the father. Taking the Kamilaroi again, the Muri-iguana man brings into his group a Butha-emu woman; their children are Ipatha-emu. If, therefore, a local group is made up of the descendants of a single family, the phratry, class, and totem names vary from generation to generation; for the girls go to other groups, and the men bring in wives of a phratry, class, and totem different, as a rule, from their own; the children of the next generation take their kinship names directly or indirectly from the mother.

If, on the other hand, descent is reckoned through the father, the phratry and totem names are always the same from generation to generation; from this it follows that the phratry of the wife, who comes from without, is also the same from generation to generation, though her totem name does not of necessity remain the same. The class name alternates both in the case of the family and of the wives in successive generations. It has already been pointed out that reckoning of descent in the male line tends to bring about local grouping of the kinship organisations. In the eight-class tribes, and in parts of Victoria, the phratries, elsewhere the totem kins, tend to be or are actually limited to certain portions of the tribal area.

Our knowledge of these matters has not, of course, been gained at a bound. Before indicating the present extent of our information, it may be well to give an historical sketch of early discoveries in this field.

Some seventy years ago the attention of students of primitive social institutions was drawn to the marriage regulations of the Indian tribes of North America by an article in Archaeologia Americana38; in which the author, drawing his conclusions partly from earlier writers, partly from his own investigations, showed that the totem kin was an exogamous group, while in some cases the kin bearing the name of a given totem were not only exogamous, but not even permitted to choose their wives from any of the other kins at will, being restricted in their choice to certain groups or, in many cases, to a single group of totem kins, according as the tribe was arranged in two or more phratries.

[34]

At least two observers had detected the existence of Australian organisations of the same nature as the American phratries, so far as our scanty information from West Australia goes, even before the publication of Archaeologia Americana. The honour of being the first to publish information on the subject belongs to Nind, who had spent some time in the neighbourhood of King George's Sound in 1829, and published his observations on native customs in the Journal of the Royal Geographical Society39 for 1832. Close on his heels came the authors of Journals of Explorations in West Australia, which appeared in 1833, and described journeys undertaken between 1829 and 1832.

The phratries were discovered in South Australia by the Rev. C. W. Schürmann, whose Vocabulary40, published in 1844, contains a mention of the Parnkalla phratries, without, however, any indication of their connection with marriage customs and exogamy. Five years earlier, however, Lieutenant, afterwards Sir George Grey, had observed institutions of the nature of totem kins, phratries, or intermarrying classes in West Australia, and had detected their connection with the marriage laws of the natives41.

In 1841 and 1842, G. F. Moore42 called attention to the grouping of the native divisions or kins, and anticipated Schürmann, as will be shown later. Grey, before the publication of his Journal, had read the Archaeologia; but though he mentions the naming of "families" after animals, he makes no mention of any grouping, but merely distinguishes between "families" and "local names." Some of the names which he gives seem to be those of phratries, and if he had been led by his study of Archaeologia Americana to the discovery of exogamic regulations dealing with the relations of individual totem kins to one another, it seems on the whole probable that he would not have overlooked the grouping of the kins which is, with certain exceptions, of a more or less local character, common to the whole of Australia, so far as our information goes.[35] Singularly enough this information, very full, relatively, for the eastern and central tribes, has, so far as South-West Australia is concerned, only just been completed, although more than sixty years have elapsed since Grey wrote, the last twenty of which have seen much additional light thrown on the organisation of the tribes of the remainder of the continent.

The American tribes, where simple totemic exogamy is not the rule, are organised in two and sometimes three or more, up to ten, phratries. It is possible that Grey, in spite of his attention having been drawn to the bi- or trichotomous organisation of American totem kins, failed to understand the Australian system owing to the presence of an element, discovered a few years later at a point remote from the scene of Grey's researches, to which no American analogue exists. In addition to the grouping of the kins into phratries, the Australian tribes over a large part of the continent subdivide each phratry into two or four classes or "castes," as they were frequently termed by the early investigators. The effect of the class system is to further limit the choice of a given individual, restricted to one-half of the women of the tribe under the simple phratry system, to one-fourth of them or one-eighth, as the case may be. Probably the first person to publish the fact of the existence of these classes, which he regarded as differing in rank, was C. P. Hodgson43, who found them in 1846 among the blacks of Wide Bay. From a letter of Leichardt's however it appears that the discovery must have been made nearly simultaneously by several observers. Writing in 184744, he says that the castes are the most interesting and most obscure feature among the tribes to the northward, and mentions F. N. Isaacs as having noticed the existence of the classes among the natives of Darling Downs, adding that Capt. Macarthur had also found them among the Monobar tribes of the Coburg Peninsula. "These castes," he adds, "are probably intimately connected with the laws of intermarriage."

If Leichardt's words mean, as apparently they do, that the Monobar classes are regulative of marriage, and if his information was correct, the first mention of classes in Australia is [36]found, not in Hodgson's work, but in Wilson's account45. Neither he, however, nor Stokes46, who mentions them as existing among the Limba Karadjee, makes any mention of their connection with marriage regulations. And Earl, at a later period, omits in like manner to say what constituted membership of a caste, though he states that they differed in rank. The names—Manjarojally (fire people), Manjarwuli (land people), and Mambulgit (makers of nets, perhaps, therefore, water people), as well as the anomalous number of the classes, seem to indicate that they are of a somewhat different nature to the real intermarrying classes found elsewhere47. It is of course well known that the initiation ceremonies and totemic system of the northern tribes on both sides of the Gulf of Carpentaria differ somewhat widely from the normal Australian form.

None of the observers hitherto mentioned can be said however to have applied himself to the scientific study of the questions raised by the facts which they recorded. Anthropology was in those days in its infancy. The first to make a really serious effort to clear up the many difficult questions, some of them still matters of controversy, which a closer study of the native marriage customs brought to the surface, was a missionary anthropologist, a class of which England has produced all too few. In 1853 the Rev. William Ridley published the first of many studies of the Kamilaroi speaking tribes, and, thanks to the impetus given to the investigation of systems of relationship and allied questions by Lewis Morgan, was the pioneer of a series of efforts which have rescued for us at the nick of time a record of the social organisation of many tribes which under European influence are now rapidly losing or have already lost all traces of their primitive customs, if indeed they have not, like the tribes formerly resident at Adelaide and other centres of population, been absolutely exterminated by contact with the white man with his vices and his civilisation, or by the less gentle method euphemis[37]tically termed "dispersion," which, if other nations were the offenders, we should term massacre.

After Mr Ridley, Messrs Fison and Howitt turned their attention to the Kamilaroi group of tribes. The progress of these investigations is traced, historically and controversially, in the second series of Maclennan's Studies in Ancient History, and it is unnecessary to deal with it in detail. More and more light was thrown on totemism, marriage regulations, and intermarrying classes by the persistent efforts of Mr Howitt, by Dr Frazer's little work on Totemism, and by other students, until it seemed that the main features of Australian social organisation had been clearly established, when in 1898 the researches of Messrs Spencer and Gillen seemed to do much to overthrow all recognised principles, so far as the totemic regulation of marriage was concerned. How far this is actually the case it is unnecessary to consider here. It may be said however that the work of these two investigators and the enquiries of Dr Roth in North Queensland make it more than ever a matter for regret that the British Empire, the greatest colonial power that the world has ever seen, will not afford the few thousand pounds needed to put such researches on a firm basis.

Having defined the various terms, and shown the actual working of the system by the aid of the best known example, we may now pass, after this brief historical sketch of the development of our knowledge, to the task of giving the broad outlines of the phratry and class organisations.

If our knowledge of Australian phratries and classes is far from exhaustive, we have at any rate a fair knowledge of the distribution of the various types whose existence is generally recognised; that is to say, we can delimit the greater part of the continent according to whether the tribes show two phratries only, or two phratries, which may be anonymous, with the further subdivision into four classes, or into eight classes. We also know approximately the limits of the matrilineal and patrilineal systems. New South Wales, Victoria, the southern portion of Queensland and Northern Territory, the eastern part of South Australia, and the coastal regions of West Australia, [38]are now known with more or less accuracy from the point of view of kinship organisations. On the other hand, from the Cape York Peninsula, and the part of Northern Territory north of Lat. 15°, we have little if any information. The south coast and its hinterland from 135° westwards, as far as King George's Sound, is virtually a terra incognita; in fact beyond the south-western corner and the fringe which lies along the coast we know little of the West Australian blacks, and the frontiers between the various systems must in these areas be regarded as purely provisional.

Broadly speaking, the tribes of the whole of the known area of Australia, certain coast regions of comparatively small extent excepted, have a dichotomous kinship organisation. The accompanying map (Map II) shows how the various forms are distributed. Along most of the south coast, and up a belt broken perhaps in the northern portion, running through the centre of the continent in Lat. 137°, are found two phratries without intermarrying classes; for the area west of Lat. 130° we have, it is true, only one datum, which gives no information as to the area to which it applies; this portion of the field therefore is assigned only provisionally to the two-phratry system. On the Bloomfield River, which runs into Weary Bay, associated with the name of Captain Cook, is an isolated two-phratry organisation, unless indeed we may assume that the class names have either been overlooked or have passed out of use.

The four-class system extends over the greater part of New South Wales, and Queensland; a narrow belt runs through the north of South Australia and broadens till it embraces the whole coastline of West Australia, the north-eastern area excluded. An isolated four-class system, which does not regulate marriage, is found in the Yorke Peninsula of South Australia.

The eight-class system forms a compact mass, between the Gulf of Carpentaria and Roebuck Bay, extending south as far as Lat. 25° in the centre of Australia.

In reality the rule of the eight-class system extends considerably further south, but the classes are nameless or altogether non-existent. Thus, the southern Arunta have nominally four classes, but each of these has two sections, so [39]that the final result is as though they were an eight-class tribe. In the same way the marriage regulations of the two-phratry Dieri are such that choice is limited among them precisely as it would be if they had eight classes. The same may be true of the remainder of the western branch of the four-class system, which is closely allied in name to the Arunta type; the boundary between the related sets of names is unknown.

Among the Narrinyeri and the Yuin the kinship organisation, which is confined to totemic groups, takes a local form; here the regulation of marriage depends on considerations of the residence of the pair. Local exogamy also prevails among the unorganised Kurnai. The Chepara appear to have had no organisation, and among the Narrangga ties of consanguinity constituted the sole bar to marriage. We are not however concerned with the problems presented by these aberrant types of organisation, to which no further reference is made in the present work.

The area covered by the dichotomous organisations is divided almost equally between matrilineal and patrilineal tribes. The latter occupy the region north of Lat. 30° and west of an irregular line running from Long. 137° to 140° or thereabouts. In addition a portion of Victoria and the region west of Brisbane form isolated patrilineal groups. The problem presented by these anomalous areas has already been discussed in the chapter on the Rule of Descent. Where local exogamy is the rule, kinship is also virtually patrilineal.

In the remainder of Australia, non-organised tribes of course excepted, the rule of descent is matrilineal, save that in North Queensland a small tribe on the Annan River prefers paternal descent. The accompanying map shows the distribution of the two forms.

[40]

[41]

TABLES OF CLASSES, PHRATRIES, ETC.

In order to facilitate reference and to diminish the necessity for footnotes a survey of classes and phratries is here given. It will be well to explain how they are arranged.

In the two-phratry system the rule of intermarriage is clear; a man of phratry A marries a woman of phratry B and vice versa. The direct descent of the kinship name is obviously the rule.

The four classes are arranged according to the phratries; the normal rule is that a man A1 marries B1, A2 marries B2; their children are in matrilineal tribes A2 and B2, in patrilineal B2 and A2. In the patrilineal Mara and Anula, by exception, the rule of descent is direct; it will be remembered that a dichotomy of the classes prevails, so that they really belong to the eight-class system.

In the eight-class system and among the nominally four-class southern Arunta the intermarriage and descent is as follows, according to Spencer and Gillen;

In each case the male is the numerator, the woman the denominator, and the = shows the child.

Tribes with conterminous territories usually know what phratries and classes are equivalent in their systems. In the [42]tables which follow the phratries and the classes of matrilineal tribes are arranged to show this correspondence so far as it is known. A * shows that no information on the point is to hand. A rearrangement of patrilineal classes is necessary to make them equivalent to the organisations of matrilineal tribes; this cannot be shown in the tables; but full details will be found in the works of Spencer and Gillen. A † indicates patrilineal descent.

Where the names of phratries and classes are translated, the meanings are shown in the tables; where the authorities do not give the translation but a word of the same form is in use in the tribe or group of tribes the meanings are given in round brackets; words in use in neighbouring tribes are put in square brackets.

The Class Names.

These class names are found in the following tribes:

Kamilaroi (Howitt, p. 107); Wiradjeri (ib. 107); Wonghi (ib. 108); Euahlayi (Mrs L. Parker, Euahlayi Tribe, p. 13); Ngeumba (Mathews in Eth. Notes, p. 5); Murawari (id. in Proc. R. G. S. Qu., 1906, 55); Moree (R. G. S. Qu. X, 20); Turribul (R. S. Vict. I, 102); Wollaroi (Howitt, 109); on Narran R. (Curr, I, 117); Pikumbul (ib.); Unghi (Howitt, 217); Peechera (Curr, III, 271); Wailwun (ib. I, 116); Wonnaruah (Sci. Man, I, 180); Geawegal (Howitt, 266).

Associated with these class names are the following phratry names:

[43]

The proper arrangement of these names is unknown.

Tribe: Kombinegherry (J. A. I. XIII, 304; Howitt, 105).

Science of Man (IV, 8) gives:

For the Anaywan, Thangatty, etc., R. H. Mathews gives (J. R. S. N. S. W. XXXI, 169):

Tribes: Maryborough tribes (Howitt, 117); Kabi (Curr, III, 163): Kiabara (J. A. I. XIII, 305); ? (Hodgson, 212; Mathew, Eaglehawk, 100); Wide Bay (Curr, I, 117).

For the Emon, Howitt (p. 109) gives:

With these classes are associated the phratries:

are the forms given by Mathews (Proc. Am. Phil. Soc. XXXVIII, 329).

[44]

Tribes: Kuinmurbura (J. A. I. XIII, 341; Howitt, 111). The Taroombul have the form Koodala (Proc. R. S. Qu. XIII, 41).

For the Kangulu, Mathews (J. R. S. N. S. W. XXXIII, 111) gives:

With these may be compared Howitt's (p. 111):

The phratries associated with these are:

Tribes: Ungorri (Howitt, 109); Kogai (Curr, I, 117; J. A. I. XIII, 337); Yuipera etc. (Curr, III, 45, 64; J. A. I. XIII, 302); Akulbura, Bathalibura (Howitt, 113, 141); Wakelbura (Howitt, 112); on Belyando (Curr, III, 26); Dalebura (Howitt, 113), Buntamurra (Howitt, 113, 226); Purgoma (Roth, 66); Jouon (ib. 67); Pitta-Pitta, Goa, Miorli (Roth, 56-7); Ringa-Ringa (J. A. I. XIII, 337); Mittakoodi (Roth, 56-7); Woonamurra (ib.); Yerunthully (Mathews in R. G. S. Qu. X, 30); Badieri (id. ib. 1905, 55).

[45]

With these class names are associated the phratries

Aberrant forms, probably inaccurate, are given by Curr (II, 424) for Halifax Bay: Korkoro, Korkeen, Wongo, Wotero; by Lumholtz (p. 199) for the Herbert R.: Gorilla, Gorgero, Gorgorilla, Otero, by Curr (II, 468) for the Yukkaburra: Utheroo, Multheroo, Yungaroo, Goorgilla.

On the Tully R. Roth (Ethn. Bull. V, 20) found the following:

With these may be compared the names given by Mathews for the Warkeman (J. R. S. N. S. W. XXXII, 109, 251):

On the Annan R. we find (Howitt, 118) with male descent:

With these are associated the phratries:

Tribes: Wollongurma (Roth, 68); Goothanto (Mathews in J. R. S. N. S. W. XXXIII, 109).

[46]

Connected with these forms are:

Tribes: Koogobathy (J. A. I. XIII, 303); Koonjan etc. (Mathews in J. R. S. N. S. W. XXXIII, 110, XXXIV, 135). Probably Perrynung and Ahjereenya should be transposed.

Tribes: Miappe (Roth, 56-7); Mycoolon (J. A. I. XIII, 302); Workoboongo (Roth, ib.).

For the Kalkadoon, Roth (ib.) gives:

With these are associated the phratries:

Tribe: Mara (Northern Tribes, 119).

With these the phratry names:

In this tribe is male descent, and, as in the S. Arunta, the classes are themselves divided; for equivalence the numbers of the eight-class system are arranged (Nor. Tr. 123), 1, 4; 3, 2; 5, 7; 6, 8.

[47]

Leichardt (Journal, 447) reports from the Roper R., Gnangball, Odall, Nurumball, which from their form seem to be class names and identifiable with some of the Mara names.

Tribe: Anula (Nor. Tr. 119).

XII. For the eight-class system see Table Ia; in which it is assumed that patrilineal descent prevails in all the tribes.

With these are associated the following phratries:

Spencer and Gillen, Nor. Tr. pp. 100-102, 119. On p. 102 is a statement about the Bingongina inconsistent with that on the following page; according to the former the phratry names are Illitchi, Liaritchi, as among the Umbaia.

Tribe: S. Arunta (Nat. Tr. 90).

Tribe: Yoolanlanya etc. (R. G. S. Qu. XVI, 75).

The arrangement suggests that matrilineal descent prevails, but there is probably some error.

[48]

Tribes: Gnamo, Gnalluma (Int. Arch. XVI, 12); Nickol Bay and Kimberley have the alternative forms of 1, 2, and 4 (Curr, I, 296; Kamilaroi, 36, Mathews in J. R. S. N. S. W. XXXV, 220), Weedokarry (id. in Proc. Am. Phil. Soc. XXXIX, 89) have third form of 2; at Murchison R. Boorgarloo comes into use (West Australian, Ap. 7, 1906).

Tribes: S. W. Australia, Tarderick etc. (West. Aust., loc. cit.; Moore, Desc. Voc., Col. Mag. V, 422.

The phratries are

The equivalence is unknown.

Tribe: Joongoongie of N. Queensland (Mathews in Proc. Am. Phil. Soc. XXXIX, 93).

Associated with them the phratries:

The equivalence is unknown.

Tribe: Narrangga of Yorke Peninsula (Howitt, p. 130).

Phratry Names.

The feminine terminations are -egor, -gurk or -jarr.

For South-West Victoria Dawson (Aborigines, p. 26) gives two groups and an odd totem kin (?):

The equivalence is not known.

[50]

*The equivalence is unknown.

Allusion has been made in Chapter III to kinship organisations denominated "bloods" and "shades" by Mr R. H. Mathews. Whether it is that some observers have mistaken these for phratries or vice versâ, it seems that the names of the two classes of organisation are at present inextricably intermingled, as the following table shows:

[51]

The areas covered by the different class and phratry names are not co-extensive, that is to say a class is associated with more than one phratry and vice versâ. The Undekerebina98 and Yelyuyendi99 have phratries (No. 29) which are usually associated with classes but in their case none have been noted. On the other hand it is not uncommon to find classes without the corresponding phratry names; this is the case in the eight class area, among the tribes of N. S. Wales, S. Queensland, etc.; but no special significance attaches to it unless we are certain that it is not the negligence of the observer nor the disuse of the names which has produced this state of things. On the other hand the relation of phratry and class areas is of the highest importance, as is shown in Chapter V. The following table shows the anomalies:

[52]

PHRATRY NAMES.

It has been shown in Chapter III that from the point of view of kinship organisations Australia falls into three main areas—occupied by the classless two-phratry, the four-class and the eight-class organisations. The total number of phratry names, thirty-three pairs in all, does not of course fall solely to the count of the two-phratry tribes, but is divided between the three kinds of organisation, the two-phratry having twelve pairs with one anomalous area, the four-class sixteen, and the eight-class five such sets. As regards the relative size of the areas thus organised, the largest seems to be that occupied by the Matteri-Kiraru system, though the Muquara-Kilpara (5) probably runs it close, especially if we take into account the names of like meaning (1-4) in the East Victorian area. The remainder of the two-phratry systems do not range over a wide extent of country, so far as is known; but 10, 11, and 33 are of unknown extent.

In the four-class area are two extensive systems, ranking next after those of South Australia and N. S. Wales; these are Mallera-Wuthera (27) and Pakoota-Wootaro (29); they have a single phratry name in common, which is also found in two other systems; if we add these together, as we may perhaps do on this evidence of a common basis, we have by far the largest phratric system in Australia as the result. Almost equal in extent to either of the two areas occupied by 27 and 29 [53]is that claimed by the better known Kamilaroi system—Dilbi-Kupathin, which spreads over a long, comparatively narrow region, but had possibly at one time a wider field from which at the present time only the corresponding class names can be recovered. Of the remaining thirteen in the two-class region, only 28, one of the Wuthera systems already mentioned, has more than a restricted field of influence. Of moderate size are the four areas in the eight-class system proper, that of the Mara being small in comparison.

Taking now the native names, we find that, in addition to the Wuthera (Ootaroo) sets already mentioned, the Dieri and Kurnandaburi have Matteri (Mattera) in common, while the latter have in the Baddieri tribe a neighbour which shares the Yungo phratry name with them. The fact, if correct, that with the Badieri Yungo is associated with Wutheru, and takes the place of the more usual Yungaru, suggests that we may equate the latter with Yungo. In the eight-class area Uluuru is common to two systems, while a third has Wiliuku, and the fourth Illitchi, all of which seem to be allied, if we may take it that uru, uku, and tchi are suffixes; that they are is borne out by the corresponding names Liaritchi and Liaraku. Other possible equations are Mukula—Mukumurra, and Cheepa—Koocheebinga, but in the latter case, even if koo is a prefix, the distance of the two systems makes any such correspondence improbable. In Victoria the Malian-Multa equation is indisputable; it is interesting to note that the former is found in N. S. Wales as the name of the bird, while Multa belongs to Yorke Peninsula.

As regards the meaning of these names, we find that of the fifty-eight names which remain after deducting those which occur in more than one system, nineteen can be translated with certainty, and we can guess at the meaning of some half dozen more. Of translateable names the most widely spread are various titles of Eaglehawk and Crow, which appear in five different systems in Victoria and New South Wales100. Crow reappears in West Australia under the name of Wartung, with white cockatoo, also a Victorian phratry name, as its fellow. In North Queensland, as a parallel to the black and white cockatoo of the south, [54]we find on the Annan River two species of bee giving their names to phratries; and the Black Duck phratry of the Waradjeri suggests that here too might be found another contrasting pair, if we could translate the other name. For the Euahlayi phratry names, on which more will be said in discussing the "blood" organisations, Mrs Parker gives the translation "Light-blooded" and "Dark-blooded," which comes near that suggested by Mr Mathews—slow and quick blooded. In the Ulu, Illi, and Wili of Northern Territory we seem to recognise Welu (curlew). Koolpuru (emu), Yungaru and Yungo (kangaroo), and Wutheroo (emu) are also possible meanings.

The problems raised by the phratriac nomenclature are complex and probably insoluble. They are in part bound up with the problem of the origin of the organisation itself; of this nature, for example, is the question whether the names correspond to anything existing in the pre-phratriac stage, or whether the organisation was borrowed and the names taken over translated or untranslated into the idiom of the borrowers. If the latter be the solution, we have a simple explanation of the wide-spread Eaglehawk-Crow system as well as of other facts, to which reference is made below.

If on the other hand the names have not been much spread by borrowing,—and the increasing number of small phratry areas known to us tells in favour of this, though it also suggests that the widely-found systems have gained ground at the expense of their neighbours,—then we obviously need some theory as to the origin of the organisation, before we can frame any hypothesis as to the origin of the names.

The prominent part, however, played by the Eaglehawk among phratry names raises some questions which can be discussed on their merits. One of these is the age of phratry names. Some of the earliest records of initiation ceremonies in New South Wales mention that the eaglehawk figured in them101. In West Australia this bird is the demiurge, and the progenitors of the phratries, of which crow is one, are his nephews. This is not the only case in which these birds figure in mythology.

[55]

As the Rev. John Mathew has pointed out in his work, Eaglehawk and Crow, there are found in Australia, especially in the south-eastern portion, a number of myths relating to the conflicts of these birds. These myths he interprets as echoes of a long-past conflict between the aboriginal Negrito race and the invading Papuans, and traces the origin of the phratries to the same racial strife. As an explanation of exogamy the hypothesis is clearly insufficient, but it is evident that no theory of the origin of the phratries can leave exogamy out of the question. The point, however, with which we are immediately concerned is the myth on which in the main Mr Mathew based his theory. Unfortunately, he did not think it necessary to attempt to define either the area covered by the different phratry names—an omission which is remedied by the present work—nor yet the limits within which the myth in question or its analogues are part of the native mythology. These analogues to the story of the battle of Eaglehawk and Crow, ended in the Darling area according to tradition by a treaty between the contending birds, are myths in which birds are said to have destroyed the human race, or a large portion of it, to have contended with Baiame, or one of the other gods, or to have figured in some other conflict102. The bird of this myth—the bird conflict myth, as it may be termed—is the Eaglehawk. Possibly, as I have pointed out in the note in Man, both bird conflict myths and Eaglehawk-Crow myths—they may be termed collectively bird myths—may go back to a common origin. So far as Mr Mathew's evidence goes, bird myths do not seem to be told outside the colony of Victoria and the Darling area of New South Wales.

A little research, however, shows that this idea is altogether erroneous. There are unfortunately large areas in Australia, as to the mythology of which we know absolutely nothing. Therefore it must not be supposed that the bird conflict myth is confined to the districts in which we have evidence of its existence. We may rather infer that a myth so widely distributed—it ranges from the head of the Bight, 129° E., to the coast north of Sydney, and probably as far as Moreton Bay; [56]to the north it is found among the Urabunna, and probably elsewhere—is common property of the Australian Tribes.

A glance at the map will show that the eaglehawk and crow myth covers but a small portion of the area in which the bird conflict myth is found. On the other hand we find within the eaglehawk-crow myth district the phratry names Cockatoo, three names of unknown meaning, and the doubtful Kiraru—Kirarawa. Now if a racial conflict is indicated by the names eaglehawk and crow, this must be either because the contending races were already known by these names, or because the two birds in question are proverbially hostile to each other. In either case we are left without any explanation of the two cockatoo phratries. It may indeed be argued that the locality in which the eaglehawk-crow phratry names are found tells strongly in favour of the racial conflict hypothesis; for it is precisely in this area that the last stand of the aborigines against the invaders may, on the theory put forward by Mr Mathew and accepted by some anthropologists103, be supposed to have taken place. But against this must be set the fact that in this area also we find two cockatoos, and on the Annan River two bees, arrayed against one another; unless it can be shown that these two birds are also proverbial foes, or that the Australian native had reached a point in his biological investigations at which he recognised that the presence of two closely allied species in a district involves a particularly keen struggle for existence (which they would, however, regard in such an advanced stage of knowledge as appropriate to the designation of intra-racial rather than inter-racial feuds), the two sets of facts balance one another, and leave us still engaged in a vain quest for a conclusion.

Putting theories as to racial conflicts aside, and dealing with the facts as we find them, we seem to have a choice of two hypotheses. Either the eaglehawk-crow myths were told before the phratry names came into existence, or they were invented to explain the existence of the phratry names. Let us assume that none of the unknown names mean eaglehawk or crow, and that the eaglehawk-crow area has remained approximately the [57]same size, or has, at any rate, not diminished (excluding, of course, those cases where it seems to have lost ground owing to the disappearance of phratry names altogether, as among the Kurnai); we must then, on the second theory, assume that the story of the combat spread to tribes with completely different phratry names like the Urabunna, and got mixed up with their ceremonies of initiation (the most sacred part of the mythology of the Australian natives, and one not likely to be much influenced by chance intruders); and that it came even in some cases to be told of Baiame, the creator and institutor of the rites of initiation, who is represented as himself taking part in the conflict and gaining a victory over the foes of mankind104. On the whole, therefore, this view of the case appears improbable.

To the theory that the Eaglehawk-Crow story was originally independent of the phratry names no such objections apply. We are indefinitely remote from the period at which the anthropologist will be able to do for Australia what Franz Boas has done for the North-West of America—draw up a table showing the resemblances and differences between the stock of folktales of the different tribes, or, which is more important for our present purpose, of the main divisions, eastern, central, and western, which the analysis of initiation ceremonies gives us—a tripartite division which Curr also makes on the linguistic side, though Mathew's map shows considerable intermixture in this respect. Until we know to what extent the Urabunna or the Ikula have folktales in common with the Victorian area, or,—which is perhaps more important, though we do not seem to hear of any communication on this line,—how far there is a stock of folktales common to the Darling district and the central area, it is obviously idle to speculate as to how it comes that an Eaglehawk myth is told in both areas. The physical anthropology of the Australian natives is at present a little-worked field, in which, singularly enough, the French have done more than the English, to our shame be it said. Possibly a somatological survey might disclose to what extent the central tribes are distinct from the eastern group, and how far we may assume movements of [58]population, subsequent to the original peopling of the country by the stocks in question, in either or both directions. In the absence of such data, and until an Australian Grimm has arisen to bring order into the present linguistic chaos, the evidence from folktales seems to promise most light on the question of migrations.

We are, of course, confronted by the difficulty that this evidence may simply disclose the lines along which tribal intercommunication has been most easy, whether in the way of simple interchange of commodities, evidence of which we have over considerable areas in Australia, or in the way of intermarriage, which, as we see by the example of the Urabunna and the Arunta, is found in spite of fundamental differences of tribal organisation. A common stock of folktales due to this cause would leave unexplained the prominence of the bird myth in the sacred rites, and leave the present hypothesis, in this regard, on a par with that of post-phratriac dissemination, in respect of probability. On the other hand we have the Scylla of tribal property in land, an idea so firmly rooted in our own day in the minds of the Australians as to make wars of conquest unthinkable to them, and to transform the practical part of their intertribal feuds into mere raids. If, therefore, investigation showed that the central and eastern tribes are in possession of a stock of folktales with many items in common, we should always have to take into consideration the possibility that these tales antedate the complete occupation of Australia, and go back to a period when the eastern and central divisions were in close relation. The probability of this view would, of course, depend on the extent of the resemblance between the two stocks of tales, or, perhaps, rather on the extent of the resemblance between those tales which they have in common; for it is clear that a close resemblance between comparatively few items would be more effective proof of intercommunication than a less marked general resemblance between the tale-stocks as a whole.

In spite of the deficiencies of our evidence we may perhaps incline to the view that the bird myth dates back to a very early period. Until it has been shown that intrusive elements are not only taken up into the tribal stock of tales, but also [59]incorporated in the more sacred portion of those tales, which are told at the tribal mysteries, it will always remain more probable that the myth belongs to the two divisions as a result of lineal and not lateral transmission. If this is so the differences between the initiation ceremonies, no less than the anthropomorphic form of the myth in the eastern division, as compared with the purely theriomorphic story of the central division and the mixed form of the Ikula, will enable us to say that the period when the separation of the divisions took place must be very remote.

There is, therefore, no inherent improbability in supposing that the bird myth was told before the phratry names were invented or adopted, and that the latter were in some cases taken from the principal characters in the myth. This conclusion is supported by the fact that the phratry names seem to be subsequent to the present grouping, if we may take as our guide the fact that the frontiers of the phratry names correspond with the boundaries between the central and eastern divisions. The fact that there is a cross division, if we base our reasoning on the class organisation, need not of course be taken into account, for we have every reason to believe that the classes are subsequent to the phratries.

In favour of the derivation of the phratry names from the myth tells also the five-fold division of the eaglehawk-crow groups into Muquara and Kilpara, Bunjil and Waa, Merung and Yuckembruk, Multa or Malian and Umbe. For it is clearly more probable that the names should have been taken from a common object than that they should have been in their origin identical in form and subsequently differentiated, as the languages changed; we have in fact direct evidence of a tendency to preserve the old names, which we may perhaps regard as the sacred names, after the bird has been rebaptised in the terminology of daily life. Over and above this we have of course the fact that the sacred language has, generally speaking, both in Australia and elsewhere, this unchanging character. But this simple name-borrowing theory, it is clear, is equally valid as an explanation of the facts.

Although we cannot determine the meaning of the names [60]the quadripartite division of the Mallera-Wuthera105 and allied phratries in the north is evidence of a similar tendency. It is by no means impossible that Mallera, Yungaroo, and Pakoota all mean the same thing. (This ignorance of the meaning of the phratry and class names is primâ facie evidence of their high antiquity.) In the newly-discovered phratry names of the eight-class tribes we have yet another instance of tripartite division. If we may assume that Illitchi, Uluuru, and Wiliuku are from the same root (which, as we have seen, is probably welu, the terminations -uku, -itchi, and -uru (= -aree) being formative suffixes), we have here too a single phratry name on the one side and three sister names on the other. While it is clear that the names cannot be in any sense of the term recent, from the fact that linguistic differentiation had already gone some distance in what we may call, for want of a better term, groups speaking a stock language (in proof of which we have only to look at the formative suffixes), it seems equally clear that the present phratry names must be considerably later than the final settlement of the country. At the same time it must not be forgotten that the existence of numerous small phratries, the number of which may yet be largely increased by more exact research, is primâ facie a proof that the groups which adopted them had not reached the stage at which anything like that tribal (still less national) organisation was known, which is at the present day characteristic of the Arunta, and, perhaps, we may say, of all groups organised on a class system with class names known and used over an area far beyond that over which the (in a restricted sense) tribal language extends.

The recurrence of crow in the phratry name of the far west lends further support to the view that the phratry names were selected in some way, and were not due to some accident of savage wit. The view has been taken that the phratry animals were originally totems, or animals that became totems at a later stage. In view of the large number of totems found in many tribes, or even restricting their number to six or eight in each phratry, it is not difficult to estimate the probability that cockatoo and crow would recur in different areas, and that [61]an opposition of characters should be found in other cases. The hypothesis needs at any rate to be combined with a theory, firstly, of borrowing of phratry names, a process which must indeed have played a large part in the development of the present system, but which does not necessarily involve the supposition that the borrowed names replaced previously existing home-made names; and, secondly, of selection of such names as were not borrowed.

It has been mentioned that the principle of tribal property in land or, to be strictly accurate, in hunting grounds, is, at the present day, a fundamental one in native Australian jurisprudence. But, as is shown by the map, in some cases the phratries are split into two or more segments106, more or less remote from one another, geographically speaking. Now this apparent segmentation must be due to migration; it can hardly arise from the chance adoption of identical names; for the groups in which the names occur are, though separated by a considerable distance, not so remote as, on the theory of chance selection, we should expect them to be, in other words the probability is in favour of the segmentation of an original group or its cleavage by an intrusive element. Of the causes of this drift of population, which on a large scale, and under pressure of any kind, might well overrule even the rights of property, we have naturally no idea. In a homogeneous mass like the population of Australia, and especially in a mass whose level of culture is so low as to leave no remains behind which we could use for the purposes of chronology, it is hopeless to expect any solution of any of the problems connected with drift of population. One thing only seems clear, and on this point we may hope for some light from the data of philology, namely that the migration was long subsequent to the original Volkerwanderung; for this must have preceded the rise of phratry names, which again must have preceded the migration of which the segmentation of groups, evidenced by the names themselves, is at present, and in default of the aid of philology, our only proof.

The migrations of which we are speaking must, if the possession of one phratry name in common be worth anything [62]as evidence of a closer connection between the groups, have been internal to a group or, if the term be preferred, to a nation occupying the south of Queensland. For in the absence of evidence that phratry names are to be found outside their own linguistic groups, we cannot but infer from the quadripartite division of the Wuthera phratries both the linguistic unity (and language must be in Australia the ultimate test of racial relationship on a large scale) and the internal movements of the group in which they occur.

In favour of the primitive unity of the Wuthera groups, is the fact that with small exceptions, and those on the outskirts of the district, the area occupied by the assumed homogeneous pre-phratry group has the same class names throughout—which is at the same time a proof that the class names are posterior to the phratry names; for the later the date, the more extensive the group, may be taken to be the rule in savage communities; if the phratry names came later than the class names we should expect them to be identical, and the class names different instead of the reverse. But to the relative age of classes and phratries we return at another point of our argument.

The available data being few, it could hardly be expected that a discussion of them would be very fruitful. In the present chapter we have, however, shown that the phratry names and organisation are probably of very early date, that considerable movements of population took place within the linguistic groups subsequent to the adoption of the phratry names, and that these names have been selected for some explicit reason and not adopted at haphazard.

[63]

ORIGIN OF PHRATRIES.

If a pre-phratry organisation developed into the system as we find it, it is a little difficult to see how selection can have operated, unless, indeed, as Mr Lang suggests, the phratries are transformed connubial groups, in which case they may have received new names. It is perhaps simpler to suppose that the cases of selection of phratry names cited above are those in which the organisation has been borrowed with full knowledge of its meaning. If this view is correct, no criticism of theories of the origin of phratries is possible from the point of view of the names actually existing, for we cannot say which, if any, are those which were evolved in the organisation which served as a model to the remainder.

Broadly speaking the theories of origin at present in the field may be reduced to two: in the first place, the conscious reformation theory, which supposes that man discovered the evils of in-and-in breeding, a point on which some discussion will be found in a later portion of this work. In the second place, there is the unconscious evolution theory put forward by Mr Lang, whose criticism of the opposing view makes it unnecessary to deal with the objections here107.

Mr Lang's original theory took for its basis the hypothesis, put forward by the late Mr J. J. Atkinson, in Primal Law, of [64]the origin of exogamy. His starting-point was mankind in the brute stage. At the point in the evolution of the human race at which Mr Atkinson takes up his tale, man, or rather Eoanthropos, was, according to his conjecture, organised, if that term can be applied to the grouping of the lower animals, in bodies consisting of one adult male, an attendant horde of adult females, including, probably, at any rate after a certain lapse of time, his own progeny, together with the immature offspring of both sexes. As the young males came to maturity, they would be expelled from the herd, as is actually the case with cattle and other mammals, by their sire, now become their foe. They probably wandered about, as do the young males of some existing species, in droves of a dozen or more, and at certain seasons of the year, one or more of them would, as they felt their powers mature, engage the lord of their own or of another herd in single combat, until with the lapse of time the latter either succumbed or was driven from the herd to end his days in solitary ferocity, his hand against everyone, just as we see the rogue elephant wage war indiscriminately on all who approach him.

In process of time, so Mr Atkinson suggests, with the lengthening childhood conditioned by the progress of the race, maternal love of a more enduring kind developed, than is found among the non-human species of the present day. This led eventually to the presence of a young male, perhaps the youngest born of a given mother, being permitted to remain, on conditions, in the herd after he had attained maturity. The original lord and master of the herd retained, Mr Atkinson supposes, his full sovereignty over the females born in the herd as well as over those whom his prowess had perhaps added to it from time to time. The young male on the other hand was not condemned to a life of celibacy as a condition of his non-enforcement of the traditional decree of banishment. He was permitted to find a mate, but she must be a mate not born in the herd, nor one of the harem of his sire; he had, if he wished to wed, to capture a spouse for himself from another herd. For the detailed working out of this ingenious theory we must refer our readers to Mr Atkinson's work, Primal Law. Here it [65]suffices to state the primal law which resulted from the process sketched above. This primal law was "thou shalt not marry within the group." This law, at first enforced by the superior strength of the sire, came in the process of time to be a traditional rule of conduct, almost an instinct. And with this we reach the theory put forward in Social Origins by Mr Andrew Lang, according to which local groups received animal names, perhaps from their neighbours. These local groups being exogamous for the reason just given, and the group name being eventually108 given, not only to the actual members of the group, but also to the women, captured or otherwise, who became the mates of the men of the adjoining groups, it necessarily resulted that the men of a group, so long as the mother's group name did not descend to her children, were of one name, while their wives were of another, or more probably of many other names. The group became definitely heterogeneous when the maternal group name descended to the children born in the alien group, and in process of time these maternal group names became totem names.

Meanwhile the original group names had been retained and applied, along with the totem or quasi-totem names, to the members of the group; the name being probably, in the first place, that of the group in which they were born, but, with the rise of the matrilineal descent, which has been discussed above, eventually taken from the group to which the mother belonged.

During these processes the custom had sprung up to select a wife, not at random from any of the probably more or less hostile surrounding groups, but from one particular group with which the group of the candidate for matrimony had in the course of time come to be on friendly terms.

The names of these two groups, which drew in other smaller groups, became the phratry names of the newly-formed aggregate, the largest unit known to primitive society at that stage of its evolution, and corresponding roughly to what we have defined as a tribe; for it was united by bonds of friendship, and in the course of time the language, originally very different no doubt, how different we can, indeed, hardly say, [66]must have so far coalesced, owing to the interchange of wives (in so far as a distinct woman's language, traces of which are found among some savage tribes, was not developed), as to produce a single tongue.

This theory Mr Lang has now fortified and elaborated in The Secret of the Totem, the most important new point being the demonstration of the fact that totem kins which bear names of the same significance as the phratry names are almost invariably in the eponymous phratries—a clear proof that law and not chance has determined their position.

As an explanation of the distribution of phratry names Mr Lang adopts a theory which combines the hypotheses of evolution and borrowing, and thus explains both the wide area covered by some systems, and the increasing multitude of organisations confined to small districts, which more minute research reveals. This does not, it is true, explain the geographical remoteness of different parts of the same system or of allied systems, shown to be so by the identity of phratry animal or name. Not only is Wuthera-Mallera split into two sections; but a portion of Wuthera-Yungaru seems to be in the same position; if we may take the Badieri Yungo as equivalent to Yungaru, dispersion alone suffices to explain the case; but if Yungo is derived from the Kurnandaburi, who have Mattera as the sister phratry, then we have the Badieri phratry names borrowed each from a different tribe, at any rate in appearance.

In reality this state of things affords the strongest possible support to Mr Lang's hypothesis, if only we can suppose that the formation of tribes is subsequent to the elaboration of the phratriac system. For it might well happen that an original Yungo local group divided, from economic causes, but that each half retained its original name. Under these circumstances the two portions formed connubial alliances with other groups; and in the tribes as we see the names of these split groups are found as phratry names, combined in each case with a different sister phratry name. We find for example Wuthera-Yungo, Yungo-Mattera, Matteri-Kiraru in the central area. The same theory will explain the appearance of Wuthera beside three other sister names, though here we must call in the borrowing and migration [67]theories as well, to explain the wide area over which the names are found. We have seen that in the northern tribes one of the phratry names appears to be in each case from the same root; if this is so, we can apply to them too the split-group hypothesis.

The case of Eaglehawk-Crow is less simple. Separated from the Darling area by a considerable space lie four systems of the same name in the east of Victoria. Here it is hardly possible to assume that the latter systems have migrated; on the other hand the area covered by the Darling group suggests that it is unlikely to have been forced from its original home by pressure from outside. Perhaps it is simplest to suppose that the Wiradjeri have gradually forced their way in, wedge fashion, between the different sections, and either swallowed up the intervening members or driven them before them; this would account for the existence of the anomalous groups to the south-west.

In this area, too, we seem to have a case of the split group; but the identity of meaning of the other phratry names (Malian and Multa both mean Eaglehawk) makes it clear that it is simply a case of translation—a possibility which must be kept in mind in the other cases also. It is a common phenomenon for two tribes to have the name of one animal in common, while for that of another entirely different words are in use. The four Victorian groups appear to have borrowed the phratry names, but the centre from which they took them must remain uncertain.

It may be noted in passing that the view of Prof. Gregory, who holds that the occupation of Victoria by the blacks dates back no more than 300 years, is hardly borne out by the distribution of the phratriac systems. It is clearly improbable that they were developed in situ, for this would make the organisation of very much more recent date than we have any warrant for supposing. On the other hand it is improbable that four tribes, all with the same phratriac names, should have taken their course in the same direction, and settled in proximity to one another, at any rate, unless the natural features of the country made this course the only possible one.

[68]

To return to Mr Lang's theory, it obviously suggests, if it does not demand, that such phratries as are spread over wide areas should in the main follow the lines of linguistic or cultural areas. Our knowledge of these is hardly sufficient to enable us to say at present how far the presumption of coincidence is fulfilled; but it is certain that in more than one large area the facts are as Mr Lang's theory requires them to be.

On the other hand in New South Wales we find an area in which we fail to discern the lines on which the phratriac systems are distributed. Here, however, we are at a disadvantage in consequence of the uncertainty introduced by the unsettled question of "blood" organisations109. Further research may show that the supposed phratriac areas, which are apparently only portions of the Wiradjeri territory, are in reality to be assigned to the "blood" organisations, which we may probably assign to a later date than the phratries and classes.

Perhaps Mr Lang's theory hardly accounts for the fact that eaglehawk and crow figure not only as phratry names but also in the myths and rites. It is not apparent why eaglehawk and crow groups should take the lead and give their names to the phratries unless it was as contrasted colours; on the other hand, if they were selected as the names from among a number of others this difficulty vanishes, but then we do not see why these names are not more widely found, unless indeed the untranslated names mean eaglehawk and crow; but possibly all express a contrast of some sort.

On the whole, however, it may be said that Mr Lang's theory holds the field. Not only is it internally consistent, which cannot be affirmed of the reformation theory, but it colligates the facts far better. This may be illustrated by a single point.

On the reformation theory, unaccompanied, as it is, by any hypothesis of borrowing of phratry names, we should primâ facie find the latter, where they are translateable, to be those of the animals which are most frequently found as totems. Now in the area covered by Dr Howitt's recent work, omitting those tribes for which our lists of totems are admittedly not complete, [69]we find that emu, kangaroo, snake, eaglehawk, and iguana are found as totems in about two-thirds of the cases; then, after a long interval, come wallaby and crow, less than half as often, with opossum rather more frequently, in half the total number. But it is clearly outside the bounds of probability that four of the commonest totems should not give their names, so far as is known, to phratries, while eaglehawk recurs five, crow six, and cockatoo three times, the two latter in one case in a remote area. Not only so, but the opposition between the phratry names—black and white or the like—is unintelligible, if, as on Dr Durkheim's theory, the phratries are simply the elementary totem groups which intermarried and threw off secondary totem kins. But criticism of other theories opens a wide field, into which it is best not to diverge.

On the development theory the phratries came into existence perhaps as the result of the persistence of an old custom of exogamy, non-moral in its inception, or, it may be, as a result of the rise of totemic tabus. The reformation theory, on the other hand, makes the conscious attainment of a better state of society the object of the institution of a dichotomous organisation. It will therefore be well to see what results in practice from the phratriac organisation.

In the two-phratry area (other rules, which usually exist, apart) it is impossible for children of the same mother or father, or of sisters or of brothers, to marry, nor can one of the parents, either mother or father, according to the rule of descent, take her or his own child in marriage. Now if the object of the reformation was to prevent parents from marrying children, it was clearly not attained. If, on the other hand, it was intended to prevent children of the same mother or father from intermarrying, the result could have been attained far more simply, either by direct prohibition, such as is found in other cases, or by the institution of totemic exogamy, which, in the view of some authorities, already existed, and consequently made the phratry superfluous.

According to Dr Frazer's 1905 theory, phratries were introduced to prevent brother and sister marriage and exogamous bars began in the female line110. Against this hypothesis may be [70]urged not only the objections first stated but also the fact that for Dr Frazer the Arunta are primitive and yet reckon descent (of the class) in the male line. If, as he conceives, conceptional totemism was transformed in the central tribes into patrilineal totemism, I fail to see why the phratries or classes should descend in the female line.

If in the third place, it was proposed to prevent children of sisters or of brothers from intermarrying, it is completely mysterious why children of brothers and sisters should not only not have been prevented in the same way, but absolutely be regarded as the proper mates for each other. Even if a single community reformed itself on these lines, it is hardly conceivable that many should have done so, even if we suppose that the advantages of prohibition were preached from tribe to tribe by missionaries of the new order of things. Ex hypothesi, cousin marriage was not regarded as harmful; and it is highly improbable that any people in the lower stages of culture should have discovered that in-and-in breeding is harmful, for the results, especially in a people which contained no degenerates, would not appear at once, even if they appeared at all.

On this point therefore the probabilities are wholly on the side of development as against reformation.

An additional reason against the reformation theory is found in the fact that phratries, on this theory, would never exceed two in number, but in practice there are, as shown in Chapter II, wide variations.

[71]

CLASS NAMES.

The priority of phratries over classes is commonly admitted and it is unnecessary to argue the question at length. The main grounds for the assumption are: (1) that it is a priori probable that the fourfold division succeeded the twofold division, exactly as the eightfold division has succeeded, and apparently is still gaining ground, at the expense of the four-class system. (2) Over a considerable and compact area phratries alone are found without a trace of named classes, if we except the anomalous organisation recorded by Dawson in S. W. Victoria. On the other hand, while we find certain tribes among whom no phratry names have yet been discovered, it is inherently probable that this is due to their having been forgotten and not to their never having existed. It is possible that the encroachments of an alien class system have in some cases helped on the extinction of the phratry names. (3) We find classes without phratry names, not in a compact group, but scattered up and down more or less at random, suggesting that chance and not law has been at work to produce this result. (4) Where class names are found without corresponding phratry names, they are invariably arranged in what may be termed anonymous phratries; that is to say, in pairs or fours, so that the member of one class is under normal circumstances not at liberty to select a wife at will from the [72]other three, but is usually limited to one of the other classes. This state of things clearly points to a time when the phratries were recognised by the tribes in question.

(5) While the classes are arranged in pairs or fours, according to whether the system is four- or eight-class, the totems, on the other hand, are distributed phratry fashion; in other words, one group of totems belongs to each pair or quadruplet of classes. This divergent organisation of the classes (four or eight for the whole tribe) and totems (two groups for the whole tribe) can only be explained on the supposition that the phratry everywhere preceded the class organisation.

The spatial relations of the phratries and classes are sufficiently clear from the map; and a table shows how far cross divisions are found.

The main area of disturbance of the normal relations is, as shown in Table IV (p. 51), the district occupied by the Koorgilla class-system and its immediate neighbourhood. The Yungaroo-Witteru group has three representatives in the Koorgilla class and one in the Kurpal class. The Pakoota-Wootaroo phratry has likewise three in the Koorgilla class, a fourth being in the Yowingo organisation. A large area is occupied by the Mallera-Witteru phratry in the Koorgilla class, and one tribe is again found in the Yowingo group. No class names are recorded for the Undekerebina in the Pakoota group, and no phratry names for the Mycoolon and Workobongo in the Yowingo group, nor for the Yerunthully in the Koorgilla group, which in addition to tribes belonging to the three Wuthera phratries also embraces within its limits the small Purgoma and Jouon tribes.

The only other anomaly recorded in addition to those mentioned is among the tribes on the south and south-east of the area just dealt with, which have the Barang class names with the Kamilaroi phratry names, or the Kamilaroi class names with tribal phratry names. In four cases therefore the phratry is found outside the limits of the class usually associated with it, or, in other words, it is associated with a strange class system. In one case, that of the Kalkadoon, this is sufficiently explained by the fact that the tribe is itself now remote geographically [73]speaking from its fellows, owing to the interposition of Pitta-Pitta and allied tribes. In the other three cases the facts seem to point to a change in the intertribal relationships in the period intervening between the adoption of phratry names and the introduction of the class system. If the lines of intercourse and intermarriage had suffered a revolution in the interval, the names, the origin of which we have yet to consider, would naturally show a different grouping of the tribes; for it is on the grouping of the tribes that the spread of the names, whether of phratries or classes, must have depended.

The main mass of the tribes organised on the four-class system lies in Queensland and New South Wales, and whereas only two sets of names are found in the latter colony, no less than fifteen (some of which are, however, of more than doubtful authenticity) are reported from various parts of Queensland. From Northern Territory two (Anula and Mara) of small extent are reported111; a considerable area of this colony, as well as of South and West Australia, is occupied by the Arunta system, and the closely allied classes to the north-west of them. The only other four-class system in West Australia of which we have definite information is that west and north of King George's Sound and eastwards for an unknown distance.

Covering nearly the whole of New South Wales outside the area occupied by the two-phratry tribes of the Darling country, and extending far up into Queensland, we find the well-known Muri-Kubbi, Ippai-Kumbo classes (1) of the Kamilaroi nation112. The Kamilaroi system appears to have touched the sea in the neighbourhood of Sydney. According to Mr Mathews, the Darkinung, who inhabited this part of New South Wales, substituted Bya for Muri. (1a) In like manner the Wiradjeri are stated by Gribble to have replaced Kumbo by Wombee; this may however be no more than a dialectical variant.

Lying along the sea coast north-east of the Darkinung and east of the main mass of Kamilaroi tribe were the Kombinegherry and other tribes, whom Mr Mathews denominates the Anaywan. Their classes are given by him as Irrpoong, Marroong, Imboong, and Irrong; but an earlier authority [74]gives the forms Kurbo, Marro, Wombo, and Wirro (2); at Wide Bay we find Baran, Balkun, Derwen, and Bundar (3) with an alternative form Banjoor.

North of them, still on the coast, we find the Kuinmurbura with Kurpal, Kuialla, Karilbura, and Munal (4); for the Taroombul, which I am unable to locate, Mr Mathews gives Koodala in place of Kuialla and Karalbara for Karilbura. For the Kangoollo, lying inland from this group, Mr Mathews gives Kearra, Banjoor, Banniar, and Koorpal. This suggests that there is some confusion, for the names include two from 4, and one or two from 3.

A very large area is occupied by tribes with the classes (5) Koorgila, Bunburi, Wunggo, and Obur (and variants). They include the Yuipera and allied tribes, the Kogai, the Wakelbura and allied tribes, the Yambeena, the Yerunthully, the Woonamurra, the Mittakoodi, the Pitta-Pitta, etc., together with the Purgoma of the Palm Islands and the neighbouring Jouon, whose headquarters are at Cooktown. In the southern portion of this group a correspondent of Curr's has reported the classes Nullum, Yoolgo, Bungumbura, and Teilling. We have class names analogous in form to the third of these names, it is true, but it resembles tribal names so closely as to suggest that the observer in question was really referring to a tribe and not to a class. If this is so we may perhaps identify Teilling with the Toolginbura. There seems to be no reason for admitting these four names to a place among the other groups of class names. In like manner we may dismiss the class names assigned to the Yukkaburra by an inaccurate correspondent of Curr's, who gives Utheroo, Multheroo, Yungaroo, and Goorgilla. It seems clear that the first and third of these are really phratry names; possibly the second is a dialectical form for Utheroo.

From Halifax Bay and Hinchinbrook Island are reported the names Korkoro, Korkeen, Wongo, and Wotero (with variants). Among the Joongoongie of North Queensland we find Langenam, Namegoor, Packewicky, and Pamarung (15); and among the Karandee Curr gives an anomalous and probably defective set, Moorob, Heyanbo, Lenai, Roanga, and Yelet.

[75]

The Goothanto and Wollungurma have Ranya, Rara, Loora, and Awunga (8); allied to these perhaps are the Jury, Ararey, Barry, and Mungilly of the Koogobathy; the Ahjeerena, Arrenynung, Perrynung, and Mahngal of the Koonjan are clearly variants of the latter set. East of the Koogobathy lie the Warkeman with Koopungie, Kellungie, Chukungie, and Karpungie (6), with an allied tribe on the Tully River with classes, Kurongon, Kurkulla, Chikun, Karavangie, the two latter obviously corresponding to Warkeman classes, the second to Koorgilla.

The Miappe, Mycoolon, Kalkadoon, and Workoboongo have Youingo, Maringo, Badingo, and Jimmilingo (9), with alternatives Kapoodingo, Kungilingo, and Toonbeungo.

The Yoolanlanya and others have Deringara, Gubilla, Koomara, and Belthara, possibly a defective list, for Mr Mathews adds to these for the Ullayilinya Lookwara and Ungella (probably a defective set) in another communication. Two of these are obviously identical with the Arunta Koomara and Bulthara, with which are associated Purula and Panungka (13), while Ungilla and Gubilla are taken from the eight-class system to which we may probably assign the tribe. North-west of the Arunta, outside the eight-class area, the class names are almost identical with, though they differ widely in form from the Arunta names. They are Burong, Ballieri, Baniker, and Caiemurra (13). The form Boorgarloo is given as a variant. Mrs Bates has found a system (14) in S. W. Australia.

On the western shores of the Gulf of Carpentaria we find the Mara with Purdal, Murungun, Mumbali, and Kuial (10); and the Anula with Awukaria, Roumburia, Urtalia, and Wialia (11).

The only two remaining four-class systems of which the names are known are on the Annan River with Wandi, Walar, Jorro, and Kutchal (7)—the Ngarranga of Yorke Peninsula, with Kari, Wani, Wilthi, and Wilthuthu.

Attention has been called in the course of the above exposition to various cases in which the class names found among one group of tribes are in part if not entirely identical with those found among their neighbours. A close examination [76]discloses other possible though hardly probable points of contact besides those already enumerated. The variant form Banjoora in 3 seems to be the same as the Banjoor of the Kangulu, which again has Koorpal in common with 4, and also Kearra, if we may equate the latter with Kuialla. This again is perhaps the Kuial of the Mara tribe (9).

The Marroong of 2 seems to be the Maringo of 9, and we may perhaps also equate the Kurbo of this group with the Kurpal of 4. Irroong resembles the roanga of the Karandee which is probably the Arawongo of the Goothanto.

In 5 Wongo suggests the Youingo of 9; it reappears in the Halifax Bay list, as also does Koorgilla in one of the variants. Again Kubi (1) corresponds to Koobaroo (5), and Kumbo (Wombee) to Bunburi (Unburi), but we can hardly regard them as the same words. Koodalla and Koorpal (4) may be the same as Kellungie and Koopungie (6); the other pair shows no resemblance.

Possibly the Wiradjeri Wombee is the Kombinegherry Wombo; it is at any rate significant that the name is found in the portion of the tribe nearest the Kombinegherry.

We have seen that the Arunta and their north-western neighbours have a four-class system, the component names of which are found with little variation over a range of nearly 25° of longitude. In the forms Kiemarra, Palyeri, Burong, and Baniker, the class names in vogue among the southern Arunta meet us again near the North-West Cape, thus covering a larger area than even the widespread Koorgila-Bunburi class names of Queensland, and forming a striking contrast to the narrow limits of the majority of the four-class system. This peculiarity is reproduced in the compact area of the central eight-class tribes, north and north-east of the Koomara four-class area, though with much greater variations in the names. Bulthara however in the form Palyeri is found in more or less disguised shapes in the whole of the eighteen tribes, whose class names are shown in Table Ia; Koomara is found in shapes which are on the whole harder to recognise, and Panunga and Purula in two or three cases, either replaced by another word or so changed as to be unrecognisable. Of the supplementary names [77]belonging to the eight-class Arunta, Uknaria, Ungalla, Appungerta, Umbitchana, Ungalla is found in the whole of the tribes under consideration, and Appungerta undergoes on the whole but little change; Uknaria is practically not found outside the Arunta area, and Umbitchana is in six cases replaced by Yacomary, which seems to be a form of Koomara (to this point we recur later).

Although this suggests that the names were in the first case taken from the Arunta a comparison of them shows that it is not among this tribe that the greatest number of forms common to the whole group and the greatest general resemblance of the names is to be found, as is shown by the comparative tables below. Judged by the standard of resemblance the Oolawunga of the north-west, on the Victoria River, have preserved the names nearest their original forms. Judged by the standard of least deviation from the common stock of names and basing the comparison, not on resemblances but on differences, the Koorangie of the upper waters of the same river take the first place, with the Oolawunga not far behind. In each case the Inchalachee, the most easterly of the group, take the last place, followed in the table of resemblances by the Walpari and the Worgaia; and in the table of differences by the Worgaia and, though at a considerable distance, the Mayoo and the Walpari.

Figure of Resemblance113.

[78]

Figure of Difference114.

Attention has already been drawn to the resemblance between the Arunta four-class names and the names of the eight-class group. It is clearly of high importance to determine whether the resemblance is on the whole between the names of the western group and the eight-class names, or whether the latter can more readily be derived from those of the Arunta. In the latter case it is obvious that the position of the Oolawunga and Koorangie in the comparative tables is due, not to their having been the tribes from which all the others derived their names, but rather to movements of population subsequent to the adoption of the class names. If on the other hand it appears that the names came in the first instance from the more western portion of the Koomara group, we have some grounds for supposing that the names and the system reached the eight-class area from the west and not from the south.

We have already seen that in the case of Palyeri-Bulthara all the evidence points to the name having come from the west. In the case of Panunga the evidence is weaker, certain of the forms being derivable from either Baniker or Panunga, but with the exception of the Warramunga, and possibly the[79] Tjingili, there are no tribes of whom we can definitely say that they took the name from the Arunta, whereas there are at least four cases where the resemblance is distinctly with the western class names, and several more in which it can more readily be derived from them. The resemblance between Koomarra and Kiemarra or Kiamba is already considerable, and makes it difficult to estimate the probabilities in most cases; the problem is complicated by the question of prefixes, which will come up for discussion later, and on the whole there appears to be no certain solution of the problem, though the Mayoo seem to have taken over and varied the western form. In the case of Purula-Burong there appear to be indeterminate cases; six seem to tell in favour of a southern origin; three suggest a western origin; and one word Chupil (f. Namilpa) seems to be from a different root.

The problem is further complicated by the anomalous class name Yakomari, to which allusion has already been made. As will be seen later, cha or ja seem to be prefixes, and if that is so we can hardly avoid the conclusion that Yakomari is Koomara or Kiemara. But in the table it takes the place of Umbitchana, with which it is not even remotely connected philologically; Jamara and its various forms take the place in the table occupied by Koomara among the Arunta when Yakomari holds the eighth place as well as in other cases. If therefore ku, ja, and ya are simply prefixes, as seems to be the case, we have this class name duplicated among five of the tribe—the Umbaia, Yookala, Binbinga, Worgaia, Yangarella, and Inchalachee, of which one comes near the top, and two fairly high in the comparative table. It is however worthy of notice that these six tribes form the eastern group, and are consequently precisely those among which we should, on the hypothesis that the class names originated in the western portion of the area, expect to find the greatest amount of variation and the most numerous anomalies. Dividing the six tribes into two groups, western and eastern, each of three tribes, we find that the cumulative resemblance of the western group to the Arunta is 132, to the Oolawunga 186; the same figures for the eastern group, more remote from the Oolawunga, but practically equidistant with the western group from [80]the Arunta, are 91 and 112. This again seems to lend support to the hypothesis of a western origin. It is perhaps simplest to suppose that the majority of the names came from the west; but that Yakomari, travelling upwards from the south-west, displaced the more usual eighth class name, or perhaps we should say, replaced it, when the eight-class system was adopted, for a name is not likely to have gone out of use when it had once been applied as a designation.

Attention has been called in connection with the phratries to the suffixes such as um, itch, aku115, etc. Their precise meaning is usually uncertain. An attentive consideration of the class names seems to show that similar suffixes have been used in forming them. If we compare Panunga and Baniker, it seems a fair conclusion that the ban or pan is compounded with iker (aku) or unga, for among the Yookala, the nearest neighbours of the Bingongina, who have it as a phratriac suffix, the -agoo of the class names is unmistakeably independent of the root word, whatever that may be. In addition to unga we find inginja, angie, inja, itch (recalling the itji of the phratries), itchana, and the form anjegoo which seems to have a double suffix. Ara, yeri, aree, um, ana, ula (as we see by comparing Purula with Burong), ta, and the possibly double form tjuka, seem to be further examples.

The feminine forms Nalyirri for Thalirri (= Palyeri), Nala for Chula, Ninum for Tjinum, Nana for Tjana or Thama, etc. suggest that prefixes are also to be distinguished. They seem to be choo, joo, ja, ya, n-, yun, u-, ku, pu, bu, nu, etc. We are however on very uncertain ground here, for the feminine forms may be deliberate creations. Allowance has to be made too for the personal equation of the observer, which is by no means inconsiderable. Possibly this factor, together with ordinary laws of phonetic change, the most elementary principles of which have yet to be established for the Australian languages, will suffice to account for the variations in the names as recorded. Otherwise the words are in most cases reduced to monosyllabic roots from which it seems hopeless to attempt to extract a meaning.

[81]

These questions of suffixes and prefixes are intimately connected with the very difficult problem of the origin of the classes. The languages of these tribes are at present, if not distinct linguistic stocks, at any rate very far from being mere dialectical variations of a common tongue, for the members of two tribes appear to be mutually unintelligible, unless, contrary to the custom of the American Indians, they are bilingual. But if each tribe added a suffix, and thus adopted into their own language words which, from the general agreement among the class names of this group, seem to have come to them from outside, it is a reasonable hypothesis that the word which they adopted had some meaning for them. Of course we may suppose that the class names were all adopted in the far off time when all spoke a common language. But apart from the difficulty that this presupposes the existence of an eight-class system at that early period, it is clear from the Queensland evidence that class names have been handed on from tribe to tribe, and it is reasonable to suppose this to have been the case with the northern tribes. This conclusion is borne out by the forms of the suffixes, which do not appear to have been developed from one root determinative, as must have been the case if we suppose that the names originated when the language spoken by these tribes was undifferentiated; and by the facts as to the apparent duplication of Koomara, to which allusion has already been made.

The important point about the class, as distinguished from the phratry systems, is the great extent covered by the former. The north-west area of male descent is virtually one from the point of view of class names; two other areas are very large, six are of medium size, three are small, and the remaining one is probably medium.

Although the question of the meaning of the class names is closely bound up with that of their origin, the problem is closely bound up with some of the points discussed in this chapter. The meaning of the eight-class names is connected with the area of origin of the system, and linguistic questions, such as those relating to suffixes, come in. We may therefore briefly discuss at this point the meaning of the class names.

[82]

On the whole it may be said that we know the meaning of the class names only in exceptional cases. The Kiabara, Kamilaroi, Annan River, Kuinmurbura, Narrang-ga, and two of the West Australian names can be translated (see Table I). But with these exceptions we have no certain knowledge of the meaning of the single class names.

Conjectures are of comparatively little value. For in the first place the number of words recorded from any given tribe is as a rule very small, and little or no indication of the pronunciation is given even in the latest works on Australian ethnography. The variations, evidently purely arbitrary and due to the want of training in phonetics, are frequently very considerable. And finally the area over which the names prevail is sufficiently great to give us our choice from half a dozen or more different tribal languages, which combined with the variation in the form of the words, adds very considerably to the probability that there will be found somewhere within the area a word or words bearing a deceptively close resemblance to the class names. How far this is the case may be made clear by one or two instances of chance resemblances between animal names (it seems on the whole probable that if the names are translateable they will turn out to be animal names) in the same or neighbouring tribes. The meaning of Arunta seems to be white cockatoo116, but we also find a word almost indistinguishable from it in sound—eranta—with the meaning of pelican117. Kulbara means emu and koolbirra kangaroo118. Malu (= kangaroo), mala (= mouse), and male (= swan) are found in tribes of West Australia, though not of tribes living in immediate proximity one to another119. But perhaps the best example is that of Derroein, which, as we have seen, means kangaroo. In addition to durween (young male kangaroo) we find at no great distance the words dirrawong (= iguana) and deerooyn (= whip snake), either of which bears a sufficiently close resemblance to the class name to be accepted as a translation for it in the absence of other competitors120.

[83]

With these facts in mind such suggestions as an attentive study of vocabularies has disclosed are naturally put forward with a full sense of their uncertainty, they are of a purely tentative nature.

For the Koobaroo (var. Obur) of the Goorgilla set I find in the same group the homophone obur (gidea tree), which is also a totem of the group of tribes in question121. The Wotero of Halifax Bay suggests Wutheru, for which I am unable to find a meaning, unless it be emu, as given by one observer, who however on another occasion gave a different translation. Korkoro in the same set may be the same as korkoren (opossum) of a tribe some 150 miles away122. The muri123 and kubbi of the Kamilaroi and Turribul (?) mean kangaroo and opossum in the latter language, and ibbai means Eaglehawk in Wiraidhuri124. The Kamilaroi bundar (= kangaroo) may give us a clue to the meaning of the Dippil Bundar125; the Kiabara Bulcoin has a homophone in the Peechera tribe, where it means kangaroo; on the Hastings River it means red wallaby. Balcun however means native bear according to Mathew126.

If we turn to the eight-class tribes the results are hardly more striking. The Dieri Pultara, Palyara and Upala127, are homophones of the class names which we have seen as alternative forms; but this very fact makes it certain, or nearly so, that one of the homophones is due to chance coincidence. Bearing in mind that the Arunta alone have the form Bulthara, we may perhaps see in the change undergone by the word in their language the result of attraction, though it must be confessed that the hypothesis is far-fetched in the case of a non-written language. On the other hand it tells against the Palyeri = Palyara equation that the Arunta, who are by far the nearest to the Dieri, use the form Bulthara. The equation Kanunka = Panunga is not backed by any evidence that the p-k change is [84]admissible. Finally three of the four words mentioned seem to be compounded with a suffix; and if this is so it is clearly useless to equate them with words in which this suffix is a component part.

One class name only, Ungilla, is found in the Arunta area itself (and far beyond it, as far as the Gulf of Carpentaria) with the meaning crow128. If we may regard the j and k of the forms jungalla, kungalla, as a prefix, the equation seems justified; otherwise it seems an insuperable difficulty that not the original form of the class name, but the derivative and shortened form is the one to which the equation applies. Our very defective knowledge of the languages of the eight-class tribes makes it possible that when we know more of them other root words may be discovered. At present it can only be said that in very few instances have we either in the four-class or the eight-class areas any warrant for saying that we know the meaning of the class names, much less that we know them to be derived from the names of animals.

One piece of evidence on the subject we need mention only to reject. The Rev. H. Kempe, of the Lutheran Mission among the southern Arunta, has on two occasions stated that the classes in signalling to each other use as their signs the gestures employed to designate animals129. On one occasion however he assigns to the Bunanka class the eaglehawk gesture, on another the lizard gesture; the remaining three, which he added only on the second occasion, were ant, wallaby and eaglehawk. It may be noted that the eaglehawk sign is attributed by him to the two classes which would form the main part of the population of a local group; in the second place all four animals are among the totems of the tribe; it seems therefore probable that Mr Kempe has merely confused the sign made to a man of the given kin with a sign which he supposed to be made to a man of a certain class. If he paid little attention to the subject, and especially if on the second occasion he gained his information at a large tribal meeting, the large number of totems would [85]render it improbable that conflicting evidence would lead him to discover his mistake. If he pursued his enquiries far enough he might, it is true, get more than one sign for a given class; but if he contented himself with asking four men, one of each class, the probability would be that he would get four separate gestures. In any case we have no warrant for arguing that the gesture in any way translates the class name.

[86]

THEORIES OF THE ORIGIN OF CLASSES.

In dealing with the origin of the classes it is important to bear in mind that they are undoubtedly later than the phratries. This is clear, not only from the considerations urged on p. 71, but also from the fact that the areas covered by the same classes are in the three most important cases immensely larger than any covered by a phratriac system. We may therefore dismiss at the outset Herr Cunow's theory, which makes the classes the original form of organisation.

To explain the origin of the classes, as of the phratries, two kinds of theories have been put forward, which are in this case also classifiable as reformatory and developmental respectively. The former labour under the same disadvantages, so far as they assume that particular marriages were regarded as immoral or objectionable, as do the similar hypotheses of the origin of phratries.

What is the effect of dividing a phratry into two classes? Firstly and most obviously, to reduce by one half the number of women from whom a man may take his spouse. Secondarily, to put in the forbidden class both his mother's generation and his daughters' generation. It must however not be overlooked that it is the whole class of individuals that are thus put beyond his reach and not those only who stand to him in the relation of daughters in the European sense. Now it is certain that the savage of the [87]present day distinguishes blood relationship from tribal relationship; of this there are plenty of examples in Australia itself130. In fact the hypothesis that the introduction of class regulations was due to a desire to prevent the intermarriage of parents and children, more especially of fathers and daughters, the mothers being of course of the same phratries as their sons in the normal tribe, depends for its existence on the assumption that consanguinity was recognised. But it is clearly a clumsy expedient to limit a man's right of choice to the extent we have indicated solely in order to prevent him from marrying his daughter, when the simple prohibition to marry her would, so far as we can see, have been equally effective.

Dr Durkheim has suggested that phratries and classes originated together.

If we start with two exogamous local groups in which the determinant spouse removes, the result is two groups in which both phratries are found, as is evident from the following graphic representation. The two sides represent the local grouping, the letters A and B the phratry names, and m or f male or female; the = denotes marriage, the vertical lines show the children, the brackets show that the person whose symbol is bracketed removes, and the italics that the symbol in question is that of a spouse introduced from without.

We see from this that the alternate generations are in each group A and B, whose spouses are in the same alternation B and A, the male remaining in the group, the female removing in each case, if we assume that the matrilineal kinship is the rule. The permanent members of each group therefore, and in like manner the imported members, are by alternate generations[88] A and B, though of course there is no difference of age actually corresponding to the difference of generation.