Bible Encyclopedias
Totemism

Religious Aspect of the Totem

As a rule, by no means invariable, the individual may not kill or eat the name-giving object of his kin, except under dire necessity; while less usually it is supposed to protect him and to send him monitory dreams. This is the " religious " or semi-religious aspect of the totem, or this aspect is, by some students, called " religious." We also hear of customs of burying and lamenting dead animals which are regarded with reverence by this or that " family," or " clan." This custom is reported among the Samoans, and one " clan " was said to offer first-fruits to its sacred animal, the eel; while the " clan " that revered the pigeon kept and fed a tame specimen.' But in Samoa, though the sacred animals of "clans " or " families " are, in all probability, survivals of totemism, they are now regarded by the people as the vehicles Long, Voyages and Travels of an Indian Interpreter (1791), p. 86.5 Turner, Samoa, p. 71.

of " clan " or " family " gods, and therefore receive honours not paid to the hereditary totems of Australia and North America, which have nothing godlike. It is to be presumed that " totem dances " in which some Australian tribes exhibit, in ballets d'action, the incidents of a myth concerning the totem, are, in a certain sense, " religious "; when they are not magical, and intended to foster and fertilize the species, animal or vegetable or other to which the totem belongs.

The magical performances for the behoof of the totem creatures may be studied in the chapters on " Intichiuma " in Messrs Spencer and Gillen's Native Tribes of Central Australia, and Native Tribes of Northern Australia. Among the many guesses at the original purpose of totemism, one has been that the primal intention of totem sets of human beings was to act as magical co-operative stores for supplying increased quantities of food to the tribe. But this opinion has gone the way of other conjectures. The " religious " status of the totem is lowest among peoples where its influence on social regulations is greatest, and vice versa, a topic to which we recur.

There are also various rites, in various tribes, connecting the dead man with his totem at his funeral; perhaps at his initiation, when a boy, into the esoteric knowledge and rules of his tribe. Men may identify themselves with their totems, or, mark themselves as of this or that totem by wearing the hide or the plumage of the bird or beast, or by putting on a mask resembling its face. The degree of " religious " regard for the revered object increases in proportion as it is taken to contain the spirit of an ancestor or to be the embodiment of a god: ideas not found among the most backward savages.

The supreme or superior being of low savage religion or mythology is never a totem. He may be able, like Zeus in Greek mythology, to assume any shape he pleases; and in the myths of some Australian tribes he ordained the institution of totemism. Byamee, among the Euahlayi tribe of north-west New South Wales, had all the totems in him, and when he went to his paradise, Bullimah, he distributed them, with the marriage rules, among his people.' In other legends, especially those of central and northern Australia; the original totem creatures, animal in form, with bestial aspect, were developed in a marine or lacustrine environment, and from them were evolved the human beings of each totem kin. The rule of non-intermarriage within the totem was, in some myths, of divine institution; in others, was invented by the primitive wandering totemic beings; or was laid down by the wisdom of mere men who saw some unknown evil in consanguine unions. The strict regard paid to the rule may be called " religious "; in so far as totemists are aware of no secular and social raison d'titre of the rule it has a mysterious character. But whereas to eat the totem is sometimes thought to be automatically punished by sickness or death, this danger does not attach to marriage within the totem save in a single known case. The secular penalty alone is dreaded; so there seems to be no religious fear of offending a superior being, or the totem himself: no tabu of a mystic sort.

Social Aspect of the Totem

The totem has almost always a strong influence on or is associated with marriage law, and except in the centre of Australia, and perhaps in the little-known West, men and women of the same totem may not intermarry, " however far apart their hunting grounds," and though there is no objection on the score of consanguinity.

This is the result, in Australia, of the custom, there almost universal, which causes each individual to belong, by birth, to one or other of the two main exogamous and intermarrying divisions of the tribe (usually called " phratries "). The phratries (often known by names of animals, as Eagle Hawk and Crow, Crow and White Cockatoo) contain each a number of totem kins, as Dog, Wild Cherry, Wombat, Frog, Owl, Emu, Kangaroo, and so on, and (except among the Arunta " nation " of five tribes in Central Australia) the same totem kin never occurs in both phratries. Thus as all persons except in the Arunta nation, marry out of their own phratry, none can marry into his or her totem kin.

1 Mrs Langloh Parker, The Euahlayi Tribe. In some parts of North America the same rule prevails, with this peculiarity that the phratries, or main exogamous divisions, are not always two, as in Australia, but, for example, among the Mohegans three - Wolf, Turtle, and Turkey . 2 In Wolf all the totems are quadrupeds; under Turtle they are various species of turtles and the yellow eel; and under Turkey all the totems are birds.

Clearly this ranking of the totems in the phratries is the result of purposeful design, not of accident. Design may also be observed in such phratries of Australian tribes as are named after animals of contrasted colours, such as White Cockatoo and Crow, Light Eagle Hawk and Crow. It has been supposed by Mr J. Mathew, Pere Schmidt and others that these Australian phratries arose in an alliance with connubium between a darker and a lighter race. 3 But another hypothesis is not less probable; and as we can translate only about a third of Australian phratry names, conjecture on this subject is premature.

Both in Australia and America the animals, as Eagle Hawk and Crow, which give their names to the phratries, are almost always totem kins within their own phratries.4 The Moquis of Arizona are said to have ten phratries, by Captain Ulick Bourke in his Snake Dance of the Moquis, but possibly he did not use the term "phratry " in the sense which we attach to it.

Among the Urabunna of Southern Central Australia, and among the tribes towards the Darling River, a very peculiar rule is said to prevail. There are two phratries, and in each are many totem kins, but each totem kin may intermarry with only one totem kin which must be in the opposite phratry. 5 Thus there are as many exogamous divisions as there are totems in the tribes, which reckon descent in the female line; children inheriting the mother's totem only. Corroboration of these statements is desirable, as the tribes implicated are peculiarly " primitive," and theirs may be the oldest extant set of marriage rules.

The existence of two or more main exogamous divisions, named or unnamed, is found among peoples where there are either no totem kins,. or where they have fallen into the background, as in parts of Melanesia, among the Todas and Meitchis of India and the Wanika in East Africa.6 An extraordinary case is reported from South Australia where people must marry in their own phratry, while their children belong to the opposite phratry. ? This awaits corroboration.

We now see some of the numerous varieties which prevail in the marriage rules connected with the totems. Even among a tribe whose members, it is reported, may marry into their own phratries, it appears that they must not marry within their own totem kins. This is, indeed, the rule wherever totemic societies are found in anything approaching to what we deem their most archaic constitution as in south-east Australia and some tribes of North America.

Exogamy: The Arunta Abnormality

Meanwhile, in Central Australia, in the Arunta " nation," the rule forbidding marriage within the totem kin does not exist. Totems here are not, as everywhere else, inherited from either parent, but a child is of what we may call " the local totem " of the place where its mother first became conscious of its life within her. The idea is that the spirits of a primal race, in groups each of one totem only (" Alcheringa folk "), haunt various localities; or spirits (ratapa) emanating from these primal beings do so; they enter into passing married women, and are incarnated and born again.8 2 Morgan, Ancient Society, p. 174.

3 Mathew, Eagle Hawk and Crow; Schmidt, Anthropos (2909).

4 See Lang, The Secret of the Totem, pp. 154, 170; and N. W. Thomas, Kinship and Marriage in Australia, pp. 9, 31.

Howitt, Native Tribes of South-East Australia, pp. 93, 181, 188; Spencer and Gillen, Native Tribes of Central Australia, pp. 60, 61, Northern Tribes, p. 71; Lang, Anthropological Essays; Tylor's Festschrift, pp. 203-210.

s Thomas, ut supra, p. 10. See, for numerous examples, T. G. Frazer, Totemism (1910).

MS. of Mrs Bates.

is necessary to state here the sources of our information about the central, north, north-western and south-eastern forms of Thus if a woman, whatever her own totem, and whatever her husband's may be, becomes conscious of her child's life in a known centre of Wild Cat spirits, her child's totem is Wild Cat, and so with all the rest.

As a consequence, a totem sometimes here appears in what the people call the " wrong " (i.e. not the original) exogamous division; and persons may marry within their own totem name, if that totem be in the " right " exogamous division, which is not theirs. Each totem spirit is among the Arunta associated with an amulet or churinga of stone; these are of various shapes, and are decorated with concentric circles, spirals, cupules, and other archaic patterns. These amulets are only used in this sense by the Arunta nation and their neighbours the Kaitish, " and it is this idea of spirit individuals associated with churinga and resident in certain definite spots that lies at the root of the present totemism. About the central Arunta tribe with its neighbours, the Urabunna, we have the evidence very carefully collected by Mr Gillen, a protector of the aborigines, and Professor Baldwin Spencer (Native Tribes of Central Australia). Concerning the peoples north from the centre to the Gulf of Carpentaria, the same scholars furnish a copious account in their Northern Tribes. These two explorers had the confidence of the blacks; witnessed their most secret ceremonies, magical and initiatory; and collected their legends. Their books, however, contain no philological information as to the structure and interrelation of the dialects, information which is rarely to be found in the works of English observers in Australia. As far as appears, the observers conversed with the tribes only in " pidgin English." If this be the case that lingua franca is current among some eighteen central-northern tribes speaking various native dialects. We are told nothing about the languages used in each case; perhaps the Arunta men who accompanied the expedition arranged a system of interpreters.

For the Dieri tribe, neighbours of the Urabunna, we have copious evidence in Native Tribes of South-East Australia by the late Mr A. W. Howitt, who studied the peoples for forty years; was made free of their initiatory ceremonies; and obtained intelligence from settlers in regions which he did not visit. We have also legends with Dieri texts and translations from the Rev. Mr Siebert, a missionary among the Dieri. That tribe appears now to exist in a very dwindled condition under missionary supervision. The accounts of tribes from the centre to the south-east by Mr R. E. Mathew, are scattered in many English, Australian and American learned periodicals. Mr Mathew has given a good deal of information about some of the dialects. His statements as to the line of descent and on other points among certain tribes are at variance with those of Messrs Spencer and Gillen (see an article by Mr A. R. Brown in Man, March 1910). Mr Mathew, however, does not enable us to test the accuracy of his informants among the northern tribes, which is unfortunate. For the Aranda (or Arunta) of a region apparently not explored by Messrs Spencer and Gillen, and for the neighbouring Loritja tribe,'we have Die Aranda and Loritja Stdmme, two volumes by the Rev. C. Strehlow (Baer, Frankfurt am Main, 1907, 1908). Mr Strehlow is a German missionary who, after working among the Dieri and acquiring their language, served for many years among a branch of the Arunta (the Aranda), differing considerably in dialect, myths and usages from the Arunta of Messrs Spencer and Gillen. In some points, for example as to the primal ancestors and the spirits diffused by them for incarnation in human bodies, the Aranda and Loritja are more akin to the northern tribes than to Mr Spencer's Arunta. In other myths they resemble some south-eastern tribes reported on by Mr Howitt. Unlike the Arunta of Messrs Spencer and Gillen, but like the Arunta described by Mr Gillen earlier in The Horn Expedition, they believe in " a magnified non-natural man," Altjira, with a goose-foot, dwelling in the heavens. Unlike the self-created Atnatu of the Kaitish of Messrs Spencer and Gillen, he is not said to have created things, or to take any concern about human beings, as Atnatu does in matters of ceremonial. Mr Strehlow gives Aranda and Lortija texts in the original, with translations and philological remarks.

Mr Frazer, in his Totemism, makes no use of Mr Strehlow's information (save in a single instance). To us it seems worthy of study. His reason for this abstention is that, in a letter to him (Melbourne, March 10, 1908), Mr Spencer says that for at least twenty years the Lutheran Missions have taught the natives " that altjira means ' god '; have taught that their sacred ceremonies and secular dances are ` wicked '; have prohibited them, and have never seen them. Flour and tobacco, &c., are only given to natives who attend church and school. Natives have been married who, according to native customary law, belong to groups to which marriage is forbidden. For these reasons Mr Frazer cannot attempt " to filter the native liquor clear of its alien sediment," (Totemism, i. 186, note 2).

Against this we may urge that, as regards the goose-footed skydweller, Mr Strehlow reports less of his active interest in human affairs than Mr Gillen does concerning his " Great Ulthaana of the totemic system of the Arunta," says Messrs Spencer and Gillen.' Every Arunta born incarnates a pre-existent primal spirit attached to one of the stone churinga dropped by primal totemic beings, all of one totem in each case, at a place called an oknanikilla. Each child belongs to the totem of the primal beings of the place, where the mother became aware of the child's life.

Thus the peculiar causes which have produced the unique Arunta licence of marrying within the totem are conspicuously obvious.

Contradictory Theories about the Arunta Abnormal Totemism.- At this point theories concerning the origin of totemism begin to differ irreconcilably. Mr Frazer, Mr Spencer, and, apparently Dr Rivers, hold that, in Australia at least, totemism was originally " conceptional." It began in the belief by the women that pregnancy was caused by the entrance into them of some spirit associated with a visible object, usually animal or vegetable; while the child born, in each case, was that object. Hence that class of objects was tabued to the child; was its totem, but such totems were not hereditary.

Next, for some unknown reason, the tribes were divided into two bodies or segments. The members of segment A may not intermarry; they must marry persons of segment B, and vice versa. Thus were evolved the primal forms of totemism and exogamy now represented in the law of the Arunta nation alone. Here, and here alone, marriage within the totem is permitted. The theory is, apparently, that, in all other exogamous and totemic peoples, totems had been, for various reasons, made hereditary, before exogamy was enforced by the legislator in his wisdom. Thus, all over the totemic world, except in the Arunta nation, the method of the legislator was simply to place one set of totem kins in tribal segment A, and the other in segment B, and make the segments exogamous and intermarrying. Thus it was impossible for any person to marry another of the same totem. This is the theory of Mr Frazer.

Upholders of the contradictory system maintain that the Arunta nation has passed through and out of the universal and normal system of hereditary and exogamous totemism into its present condition, by reason of the belief that children are incarnations of pre-existing animal or vegetable spirits, plus the unique Arunta idea of the connexion of such spirits with their stone churinga. Where this combination of the two beliefs does not occur, there the Arunta non-hereditary and non-exogamous totemism does riot occur. It would necessarily arise in any normal tribe which adopted the two Arunta beliefs, which are not " primitive." Arguments against Mr Frazer's Theory. - There was obviously a time, it is urged, when all totems were, as everywhere else, heavens " among the Arunta. Mr Strehlow's being, Altjira, has a name apparently meaning " mystic " or sacred, which is applied to other things, for example to the inherited maternal totem of each native. His names for Altjira (god) and for the totemic ancestors (totem gods), are inappropriate, but may be discounted. Many other tribes who are discussed by Mr Frazer have been long under missionary influence as well as the Aranda. According to Mr Frazer the Dieri tribe had enjoyed a German Lutheran mission station (since 1866) for forty-four years up to 1910. About 150 Dieri were alive in 1909 (Totemism, iii. 344). Nevertheless the Dieri myths published by Mr Siebert in the decadence of the tribe, and when the remnant was under missionaries, show no " alien sediment." Nor do the traditions of Mr Strehlow's Aranda. Their traditions are closely akin, now to those of the Arunta, now to those of the northern tribes, now to those of the Euahlayi of Mrs Langloh Parker (The Euahlayi Tribe) in New South Wales, and once more to those of Mr Howitt's south-eastern tribes. There is no trace of Christian influence in the Aranda and Loritja matter, no vestige of " alien " (that is, of European) " sediment," but the account of Atnatu among the Kaitish reported on by Messrs Spencer and Gillen reads like a savage version of Milton's " Fall of the Angels " in Paradise Lost. For these reasons we do not reject the information of Mr Strehlow, who is master of several tribal languages, and, of course, does not encourage wicked native rites by providing supplies of flour, tobacco, &c., during the performances, as Mr Howitt and others say that they found it necessary to do. Sceptical colonists have been heard to aver that natives will go on performing rites as long as white men will provide supplies.

1 Native Tribes of Central Australia, p. 123.

in what the Arunta call "the right" divisions; Arunta, that is, were so arrayed thatno totem existed in more than one division. Obliged, as row, to marry out of their own exogamous division (one of four sub-classes among the Arunta) into one of the four sub-classes of the opposite side, no man could then find in it a woman of his own totem to marry. But when Arunta ceased to be hereditary, and came to be acquired, as now, by the local accident of the totem spirits - all, in each case, of one totem name, which haunt the supposed place of a child's conception - some totems inevitably would often get out of their original sub-class into another, and thus the same totems are in several divisions. But granting that a man of division A may legally marry a woman of division B, he is not now prevented from doing so because his totem (say Wild Cat) is also hers. His, or hers has strayed, by accident of supposed place of conception, out of its " right " into its " wrong " division. The words " right " and " wrong " as here used by the Arunta make it certain that they still perceive the distinction, and that, before the Arunta evolved the spiritual view of conception, they had, like other people, their totems in each case confined to a single main exogamous division of their tribe, and therefore no persons could then marry into their own totems.

But when the theory of spiritual conception arose, and was combined, in the Arunta set of tribes alone (it is common enough elsewhere in northern and western Australia), with the churinga doctrine, which gave totems by accident, these two factors, as Messrs Spencer and Gillen say, became the causes - " lie at the root " - of the present Arunta system by which persons may marry others of " the right " division, but of " the wrong " totem. That system is strictly confined to the group of tribes (Ilpirra, Loritja, Unmaterja, Kaitish, Arunta) which constitute " the Arunta nation." Elsewhere the belief in spiritual conception widely prevails, but not the belief in the connexion of spirits of individuals with the stone churinga of individuals. Consequently the Arunta system of marriage within the totem exists nowhere, and the non-exogamous non-hereditary totem exists nowhere, except in the Arunta region. Everywhere else hereditary totems are exogamous.' Thus the practice of acquiring the totem by local accident is absolutely confined to five tribes where the churinga doctrine coexists with it. That the churinga belief, coexistent with the spiritual theory of conception, is of relatively recent origin is a demonstrable fact. Had it always been present among the Arunta the inevitable result, in the course of ages, would be the scattering of the totems almost equally, as chance would scatter them among the eight exogamous divisions.

This can be tested by experiment. Take eight men, to represent the eight exogamous divisions, and set them apart in two groups of four. Take four packs of cards, 208 cards, to represent the Arunta totems, which are over 200 in number. Deal the cards round in the usual way to each of the eight men; each will receive 26 cards. It will not be found that group A has " the great majority " of spades and clubs, while group B has " the great majority " of diamonds and hearts, and neither group will have " the great majority " of court cards. Accident does not work in that way. But while accident alone now determines the totem to which an Arunta shall belong, nevertheless " in the Arunta, as a general rule, the great majority of the members of any one totemic group belong to one moiety of the tribe; but this is by no means universal. .. " - that is, of the totems the great majority in each case, as a rule, belongs to one or the other set of four exogamous sub-classes.2 The inference is obvious. While chance has now placed only the small minority of each totem in all or several of the eight exogamous divisions, the great majority of totems is in one or another of the divisions. This great majority cannot come by chance, as Arunta totems now come; consequently it is but lately that chance has determined the totem of each individual. Had chance from the first been the determining cause, each totem 1 N.T.C.A. p. 257; cf. Frazer, Totemism, i. 200-201.

2 Northern Tribes, pp. 151 sqq.

would not be fairly equally present in each of the two sets of four exogamous divisions. But determination by accident has only existed long enough to affect " as a general rule " a small minority of cases. " The great majority " of totems remain in what is recognized as " the right," the original divisions, as elsewhere universally. Arunta myth sometimes supports, sometimes contradicts, the belief that the totems were originally limited, in each case, to one or other division only, and, being selfcontradictory, has no historic value.

A further proof of our point is that the northern neighbours of the Arunta, the Kaitish, have only partially accepted Arunta ideas, religious and social. Unlike the Arunta they have a creative being, Atnatu, from whom half of the population descend; the other half were evolved out of totemic forms. 3 In the same way the Kaitish totems " are more strictly divided between the two moieties " (main exogamous divisions) " of the .tribe." 4 Consequently a man may marry a woman of his own totem if she be in the right exogamous division. " She is not actually forbidden to him, as a wife becomes of this identity and totem, as she would be in the Warramunga neighbouring tribe. " " It is a very rare thing for a man to marry a woman of the same totem as himself," 5 naturally, for the old rule holds, in sentiment, and a totem is still very rarely in the wrong division. The Arunta system of accidental determination of the totem has as yet scarcely produced among the Kaitish any of its natural and important effects.

This view of the case seems logical: Arunta non-exogamous non-hereditary totemism is the result, as Messrs Spencer and Gillen show, of the theory of spiritual conception and the theory of the relation of the spirit part of each individual to his churinga. These two beliefs have already caused a minority of Arunta totems to get out of the original and into the wrong exogamous Arunta divisions. The process is not of old standing; if it were, all totems would not be fairly distributed among the divisions by the laws of chance. In the Kaitish tribe, on the other hand, the processes must be of very recent operation, for they have only begun to produce their necessary effects. The totemism of the Arunta is thus the reverse of " primitive," and has but slightly affected the Kaitish.

Precisely the opposite view of the facts is taken by Mr Frazer in his erudite and exhaustive work Totemism. In the Kaitish, he writes, " we may detect the first stage in the transition from promiscuous marriage and fortuitous descent of the totem to strict exogamy of the totem clans and strict heredity of the totems in the paternal line." 6 By " promiscuous marriage," marriage within or without the totem, at pleasure, is obviously intended, for the Arunta do not marry " promiscuously " - do not marry their nearest kin.

How, on Mr Frazer's theory, was the transition from the condition of the Arunta to that of the Kaitish made? If the Kaitish were once in the actual Arunta stage of totemism, how did their totems come now to be much more strictly divided between the two moieties, though " the division is not so absolute as amongst the Urabunna in the south and the tribes farther north.. ."? How did this occur? The Kaitish have not made totems hereditary by law; they are acquired by local accident. They have not made a rule that all totems should, as among the more northern neighbours of the Arunta, be regimented so that no totem occurs in more than one division: to this rule there are exceptions. A man " is not actually forbidden " to marry a woman of his own totem provided she be of " the right division," but it is clear that he " does not usually do so." This we can explain as the result of a survival in manners of the old absolute universal prohibition.

Meanwhile our view of the facts makes all the phenomena seem natural and intelligible in accordance with the statement of the observers, Messrs Spencer and Gillen, that the cause of the unique non-hereditary non-exogamous totems of the Arunta is the combination of the churinga spiritual belief with the belief in spiritual conception. This cause, though now present among 3 Northern Tribes, pp. 1 53, 1 54, 1 75.4 Ibid. p. 152.

5 Ibid. p. 175. ° Totemism, i. 244.

the Kaitish, has, so far, operated but faintly. We have been explicit on these points because on them the whole problem of the original form of totemism hinges. In our view, for the reasons stated, the Arunta system of non-exogamous non-hereditary totemism is a peculiarity of comparatively recent institution. But Mr Frazer, and the chief observer of the phenomena, Mr Spencer, consider the Arunta system, non-exogamous and nonhereditary, to be the most archaic form of totemism extant.

As to non-hereditary, we find another report of the facts in Die Aranda and Loritja Steimme, by the Rev. Mr Strehlow, who has a colloquial and philological knowledge of the language of these tribes. As he reports, among other things, that the Aranda (Arunta) in his district inherit their mother's totems, in addition to their " local totems," they appear to retain an archaic feature from which their local totem system and marriage rules are a departure.' The hereditary maternal totem is, in Mr Strehlow's region, the, protective being (altjira) of each Arunta individual.

Are the Arunta " Primitive" or not ? - In the whole totemic controversy the question as to whether the non-exogamous non-hereditary totemism of the Arunta or the hereditary and exogamous totemism of the rest of Australia and of totemic mankind, be the earlier, is crucial.

That Arunta totemism is a freak or " sport," it is argued, is made probable first by the fact that the Arunta inherit all things hereditable in the male line, whereas inheritance in the female descent is earlier. (To this question we return; see below, Male and Female Lines of Descent.) M. Van Gennep argues that tribes in contact, one set having female, the other male, descent, " like the Arunta have combined the systems." 2 But several northern tribes with male descent of the totem which are not in contact with tribes of female descent show much stronger traces of the " combination " than the Arunta, who intermarry freely with a tribe of female descent, the Urabunna; while the Urabunna, though intermarrying with the Arunta who inherit property and tribal office in the male line, show no traces of " combination." Thus the effects occur where the alleged causes are not present; and the alleged causes, in the case of the Urabunna and Arunta, do not produce the effects.

Next the Arunta have no names for their main exogamous divisions, these names being a very archaic feature which in many tribes with sub-classes tend to disappear. In absence of phratry names the Arunta are remote from the primitive. M. Van Gennep replies that perhaps the Arunta have not yet made the names, or have not yet borrowed them. This is also the view of Mr Frazer. As he says, the Southern Arunta lived under the rule of eight classes, but of these four were anonymous, till the names for them were borrowed from the north. The people can thus have anonymous exogamous divisions; the two main divisions, or phratries, of the Arunta may, therefore, from the first, have been anonymous.

To this the reply is that people borrow, if they can, what they need. The Arunta found names for their four hitherto anonymous classes to be convenient, so they borrowed them. But when once class-names did, as they do, all that is necessary, the Arunta had no longer any use for the names of the two primary main divisions: these were forgotten; there is nothing to be got by borrowing that; while four Arunta " sub-classes " are gaining their names, the " classes " (phratries or main divisions) have lost them. It is perfectly logical to hold that while things useful, but hitherto anonymous, are gaining names, other things, now totally useless, are losing their names. One process is as natural as the other. In all Australia tribes with two main divisions and no sub-classes, the names of the two main divisions are found, because the names are useful. In several tribes with named sub-classes, which now do the work previously thrown on the main divisions, the names of the main divisions are unknown: the main divisions being now useless, and superseded by the subclasses. The absence of names of the two main divisions in the Arunta is merely a result, often found, of the rise of the sub 1 Strehlow, ii. 57 (1908).

2 Mythes et legendes d'Australie, p. xxxii.

classes, which, as Mr Frazer declares, are not primitive, but the result of successive later legislative acts of division.3 Manifestly on this point the Arunta are at the farthest point from the earliest organization: their loss of phratry names is the consequence of this great advance from the " primitive." All Arunta society rests on a theory of reincarnated spirits, a theory minutely elaborated. M Van Gennep asks " why should this belief not be primitive? " Surely neither the belief in spirits, nor the elaborate working out of the belief connecting spirits with manufactured stone amulets, can have been primitive. Nobody will say that peculiar stone amulets and the Arunta belief about spirits associated with them are primitive. To this M Van Gennep makes no reply.4 The Arunta belief that children are spirit-children (ratapa) incarnated is very common in the other central and northern tribes, and, according to Mrs Bates, in Western Australia; Dr Roth reports the same for parts of Queensland. It is alleged by Messrs Spencer and Gillen that the tribes holding this belief deny any connexion between sexual unions and procreation. Mr Strehlow, oil the other hand, says that in his region the older Arunta men understand the part of the male in procreation; and that even the children of the Loritja and Arunta understand, in the case of animals. 5 (Here corroboration is desirable and European influence may be asserted.) Dr Roth says that the Tully River blacks of Queensland admit procreation for all other animals, which have no Koi or soul, but not for men, who have souls. (Their theory of human birth, therefore, merely aims primarily ataccounting for the spiritual part of man.)6 According to Mrs Bates, some tribes in the north of South Australia, tribes with the same " class " names as the Arunta, hold that to have children a man must possess two spirits (ranee). If he has but one, he remains childless. If he has two, he can dream of an animal, or other object, which then passes into his wife, and is born as a child, the animal thus becoming the child's totem. This belief does not appear to apply to reproduction in the lower animals. It is a spiritual theory of the begetting of a soul incarnated. If a man has but one spirit, he cannot give one to a child, therefore he is childless.

It is clear that this, and all other systems in which reproduction is explained in spiritual terms, can only arise among peoples whose whole mode of thinking is intensely " animistic." It is also plain that all such myths answer two questions - (i) How does a being of flesh and spirit acquire its spiritual part? - (2) How is it that every human being is in mystical ro p port with an animal, plant, or other object, the totem? Manifestly the second question could not arise and need answer before mankind were actually totemists. It may be added that in the south of Western Australia the name for the mythical " Father of All " (a being not there worshipped, though images of him are made and receive some cult at certain licentious festivals) and the name for " father-stock " is maman, which Mrs Bates finds to be the native term for membrum virile. All this appears to be proof of understanding of the male part in reproduction, though that understanding is now obscured by speculation about spirits.

The question arises then, is the ignorance of procreation, where that ignorance exists, " primitive," and is the Arunta totemism also " primitive," being conditioned, as we are told it is, by the unique belief in some churinga? Or is the ignorance due to attempts of native thinkers to account for the spirit in man as a pre-existing entity that has been from the beginning? The former view is that of Messrs Spencer and Gillen, and Mr Frazer. For the latter see Lang, Anthropological Essays presented to E. B. Tylor, pp. 210-218. We can hardly call people primitive because they have struggled with the problem " how has material man an indwelling spirit? " Theories of the Origin of Totemic Exogamy. - Since the word " exogamy " as a name for the marriage systems connected (as a rule) with totemism was used by J. F. McLennan in his 3 Totemism, i. 282, 283.4 Van Gennep, pp. xxxiii - xxxv. Loritja Steimme, p. 52, note 7.

6 Roth, Bulletin, No. 5, pp. 57, 22, 65, 81.

Primitive Marriage (1866), theories of the origin of exogamy have been rife and multifarious. All, without exception, are purely conjectural. One set of disputants hold that man (whatever his original condition may have been) was, when he first passed an Act of Exogamy, a member of a tribe. Howitt's term for this tribe was " the undivided commune." It had, according to him, its inspired medicine-man, believed to be in communication with some superior being. It had its probouleutic council of elders or " headmen " and its general assembly. Such was man's political condition.' It is not distinguishable from that of many modern Australian tribes. Other tribes, said by some to be the most primitive, the Arunta and their neighbours, pay no attention to the dictates of a superior being, and the Arunta of Spencer and Gillen seem to know no such entity, though as Atnatu, Tukura, Altjira, and " the Great Ulthaana of the heavens," he exists in a dwindled form among the Kaitish, Loritja and outlying portions of the Arunta tribe. In religion Howitt's early men were already in advance of Mr Spencer's Arunta. Socially, man, at this date, according to Howitt, at first left the relations of the sexes wholly unregulated; the nearest kinsfolk by blood coupled at will, though perfectly aware that they were, at least on the maternal side, actual brothers and sisters, parents and children.

Upholders of the first theory, that man lived promiscuously in a tribal state with legislative assemblies and then suddenly reformed promiscuity away, must necessarily differ in their opinion as to the origins of totems and exogamy from the friends of the second theory, who believe that man never was " promiscuous," and given to sexual union with near kin. Why man, on the first theory - familiar as he was with unions of the nearest kin - suddenly abolished them is explained in four or five different ways. Perhaps the most notable view is Mr Frazer's; he easily confutes, in thirty-five pages, the other hypotheses. 2 Man saw, or thought he saw, injurious consequences to the wedded nearrelated couples, and therefore he prohibited, first, unions between mothers and sons, and brothers and sisters. 3 But, in his fourth volume, Mr Frazer sees conclusive objections to this view 4 and prefers another. Some peoples, far above the estate of savagery, believe that human incest blights and sterilizes the crops, women and animals. " If any such belief were entertained by the founders of exogamy, they would clearly have been perfectly sufficient motives for instituting the system, for they would perfectly explain the horror with which incest has been regarded and the extreme severity with which it has been punished." 5 That is to say, people had a horror and hatred of incest because they supposed that it blighted the crops and other things. Mr Frazer had previously written (iv. 108) " It is important to bear steadily in mind that the dislike of certain marriages must always have existed in the minds of the people, or at least of their leaders, before that dislike, so to say, received legal sanction by being embodied in an exogamous rule." Again (iv. i 2) " There had, for some reason unknown to us, been long growing up a strong aversion to consanguineous unions " - before any legislative bar was raised against them. This is insisted on. The prohibition " must have answered to certain general sentiments of what was right and proper " (iv. 121). But here the theorist has to explain the origin of the strong aversion, the general sentiment that unions of near kin are wrong and improper. But Mr Frazer does not seem to explain the point that most needs explanation. That " strong aversion," that " general sentiment," cannot have arisen from a growing belief that unions of close kin spoiled the crops or the natural resources of the country. That superstition could only arise as a consequence of the horror and aversion with which " incest " was regarded. Now no idea corresponding to " incest " could arise before unions of near kin were deemed abominable. When once such unions were thought hateful to gods and men, and an upsetting of the cosmic balance, then, but not till then, they might be regarded as injurious to the crops. All such beliefs are sanctions of ideas already in strong ' N.T.S.E.A. pp. 89, 90.3 Ibid. i. 165.

2 Totemism, Ibid. iv. 1 55, 156.

5 Ibid. iv. 158.

force. The idea that such or such a thing is wrong begets the prohibition; followed by the sanction - the belief that the practice of the thing is injurious in a supernormal way: where that belief exists. We do not know it in Australia, for example.

A belief that close sexual unions were maleficent cosmic influences could not possibly arise previous to, and could not then cause, " the dislike of certain marriages "; " the strong aversion to consanguineous unions " - which existed already. This latest guess of Mr Frazer at the origin of the idea of " incest " - of the abomination of certain unions - is untenable. What he has to explain is the origin of the dislike, the aversion, the horror. Once that has arisen, as he himself observes, the prohibition follows, and then comes the supernormal sanction. Thus no theory of exogamous rules as the result of legislation to prevent the unions of persons closely akin, can produce, or has produced, any reason for the aversion to such unions arising among people to whom, on the theory, they were familiar. Mr Frazer has confuted the guesses of MacLennan, Morgan, Durkheim and others; but his own idea is untenable.

The Supposed Method of Reform

On Mr Frazer's theory the reformers first placed half of the mothers of the tribe, with their children, in division A; and the rest of the mothers, with their children, in division B. The members of each division (phratry) must marry out of it into the other, and thus no man could marry his sister or mother. (The father could marry his daughter, but in tribes with no exogamous explicit rule against the union, he never does.) Later the two divisions were bisected each into a couple of pairs (classes) preventing marriage between father and daughter; and another resegmentation prohibited the unions of more distant relations. These systems, from the simplest division into two phratries, to the more complex with two " sub-classes " in each phratry, and the most elaborate of all with four sub-classes in each phratry, exist in various tribes. Environment and climate have nothing to do with the matter. The Urabunna and the Arunta live in the same climate and environment, and intermarry. The Urabunna have the most primitive, the Arunta have the most advanced of these organizations. While the rules are intended to prevent consanguineous marriages, the names of the " sub-classes " (when translatable, the names of animals) cannot perhaps be explained. They have a totemic appearance.

Totems in Relation to Exogamy

So far, in this theory nothing has been said of totems, though it is an all but universal rule that people of the same totem may not intermarry, even if the lovers belong to tribes separated by the breadth of the continent. In fact, according to the hypothesis which has been set forth, totems, though now exogamous, played no original part in the evolution of exogamy. They came in by accident, not by design, and dropped into their place in a system carefully devised.

Originally, on this theory, a totem came to a child, not as is usual now, by inheritance, but by pure accident; the mother supposing that any object which caught her attention at the moment when she first felt the life of her child, or any article of food which she had recently eaten, became incarnate in her, so that the emu (say) which she saw, or had eaten of, was her child. He or she was an Emu man or woman, by totem was an Emu.

Certain localities, later, were somehow associated each with one given object - cat, kangaroo, grub, or anything else, and now " local totems " (if the phrase may be used) took the place of " conceptional totems," as among the Arunta. The child inevitably was of the local totem and its supposed place of conception.

Finally all tribes except the Arunta " nation " made the totem hereditary, either from mother or father; and as the mother or father, an Emu, was in division A, so was the child, and he or she must marry out of that division into the other, B.6 The objections taken to this theory are now to be stated: Frazer, Totemism, i. 157-167.

(1.) The theory can by no possibility apply to tribes with three or more main exogamous divisions cr phratries, such as we find in North America. In a three-phratry tribe we are reduced to suppose that there were three sexes, or resort to some other solution not perhaps compatible with the theory. (ii.) We have no evidence that any totemic people, except the Navajoes, think the closest sexual unions injurious to the parties or their offspring. The theory is thus Nerdy extracted from the facts - certain unions are forbidden, therefore they must have been deemed injurious. Now, even if they were generally thought injurious, the belief would be a mere inference from the fact that they were forbidden. (iii.) The supposed original legislative exogamous division produced a very different effect than that said to be aimed at, namely, the prohibition of marriage between brothers and sisters. It forbade to every man marriage with half the women of his tribe, most of whom were not, even in the wide native use of the term, his " tribal " sisters, that is, women in a man's phratry of the same status as his own sisters. Such relationships, of course, could not exist before they were created by the supposed Act of Division. It would have been easy to prohibit marriages of brothers with sisters directly, just as, though no exogamous rule forbids, the father, in tribes of female descent, is directly forbidden to marry his daughters. The natives can take a simple instead of a bewildering path. To this natural objection Mr Frazer replies:' " If we assume, as we have every right to do, that the founders of exogamy in Australia recognized the classificatory system of relationship, and the classificatory system of relationship only, we shall at once perceive that what they intended to prevent was not merely the marriage of a man with his sister, his mother, or his daughter in the physical sense in which we use these terms; their aim was to prevent his marriage with his sister, his mother and his daughter in the classificatory sense of these terms; that is, they intended to place bars to marriage not between individuals merely but between the whole groups of persons who designated their group, not their individual relationships, their social, not their consanguineous ties, by the names of father and mother, brother and sister, son and daughter. And in this intention the founders of exogamy succeeded perfectly." Mr Frazer's theory of the origin of exogamy appears now to waver. It was 2 that the primal bisection of the tribe was " deliberately devised and adopted as a means of preventing the marriage, at first, of brother with sisters... ." Here was the place to say, if it was then intended to say, that the Australians " recognized the classificatory system of relationships only." As a matter of fact they recognize both the consanguine and the classificatory systems. It is not the case that " the savage Australian, it may be said with truth, has no idea of relationships as we understand them, and does not discriminate between his actual father and mother and the men and women who belong to the group, each member of which might have lawfully been either his father or his mother, as the case may be." This statement is made inadvertently and unfortunately by Messrs Spencer and Gillen,' but it is contradicted by their own observations. An Arunta can tell you, if asked, which of all the men whom he calls " father " is his very own father.4 The Dieri have terms for " great " (actual) and " little " (tribal) father, and so for other relationships. In Arunta orgies a woman's " tribal " " fathers " and " brothers " and " sons " are admitted to her embraces; her actual father and brothers and sons are excluded.' Thus, if the prohibition be based on aversion to unions of persons closely akin by blood, as the actual father is excluded, the actual father, among the Arunta, is, or has been, amongst that people, regarded as near of blood to his daughters. The Arunta are ignorant, we are told, of the part of the male in procreation. Be it so, but there has been a time when they were not ignorant, and when the father was recognized as of the nearest kin by blood to his daughters. If 1 Totemism, Ibid. i. 163.

3 Northern Tribes, pp. 95 seq.; Totemism, Central Tribes, p. 57. Ibid. p. 97.

not, and if the prohibition is based on hatred of unions of close kin, why is the father excluded? Nothing, in short, can be more certain than that Australian tribes distinguish between " social " or " tribal " relations on the one hand, and close consanguine relations on the other. Among the Arunta office is inherited by a man from his mother's husband, his father quem nuptiae demonstrant; not from any " tribal " father.6 Mr Frazer ? apparently meant in his earlier statement that brothers and sisters consanguine, and these only, were to be excluded from intermarriage, because he went on to say that science cannot decide as to whether the closest interbreeding is injurious to the offspring of healthy parents, however near in blood; and that very low savages could not discover what is hidden from modern science. He had therefore marriages of consanguine brothers and sisters present to his mind: " the closest interbreeding." Brothers and sisters were finally forbidden, on this theory, to intermarry, not because of any dread of injury to the offspring. " The only alternative open to us seems to be to infer that these unions were forbidden because they were believed to be injurious to the persons engaged in them, even when they were both in perfect health." 8 These " incestuous unions " are between brothers and sisters, mothers and sons. Here brothers and sisters consanguine, children of the same mother in each case, certainly appear to be intended. Who else, indeed, can be intended? But presently 9 we are to assume that the Australians, before they made the first exogamous division of the tribe " recognized the classificatory system of relationship, and the classificatory system only." They meant, now, to bar marriage between " whole groups of persons," related by " social, not consanguineous ties." But this seems to be physically impossible. These " whole groups " never existed, and never could exist, as far as we can see, till they were called into being by the legislative division of the tribe into two exogamous phratries - which had not yet been made. How could a man call a whole group of women " p upa," as at present (the word being applied to his wife and to all women of the opposite phratry to his whom he might legally marry) before the new law had constituted such a group? In what sense, again, were all women of a certain status called my " sisters " (like my actual sisters) before the new law made a new group of them - in regard to marriage as sacred as my own sisters now were to me ? It cannot be said that all women of my status were called, collectively, my " sisters " before the new division of the tribe and new rule arose, because previously, all women of my status in the tribe have been my " sisters." Who else could be collectively my " sisters "? If to marry a " sister " were reckoned dangerous to her and to me, I must have been forbidden to marry all the women of my status in the tribe. How could a law which merely halved the number of my " sisters " remove the unknown danger from half of them? If any women except my actual sisters were, before the new rule, reckoned as socially my sisters, all women in the tribe of a certain status must have been so reckoned. If all dangerous, I must marry none of them. But by the new rule, I may marry half of them! Why have they ceased to be dangerous?

If the theory be that originally only brothers and sisters consanguine were thought dangerous to each other in sexual relations, and the superstition was later extended so as to include all " classificatory " brothers and sisters, who were in these days (before the exogamous division) classificatory brothers and sisters? How and for what reason were some marriageable girls in the tribe classificatory sisters of a young man while others, equally young and marriageable, were not ? The classificatory brothers and sisters must have been all the marriageable youth of both sexes in a generation, in the tribe.

But then if all the youth of a generation, of both sexes, were classificatory brothers and sisters, and if therefore their unions were dangerous to themselves, or to the crops, the danger could not be prevented by dividing them into two sets, and s See Proceedings of British Academy, iii. 4. Lang, "Origin of Terms of Human Relationships." 7 Totemism, 163. Ibid. i. 165. Ibid. i. 288.

allowing each set of brothers to marry each set of sisters. The only way to parry the danger was to force all these brothers and sisters to marry out of the local tribe into another local tribe with the same superstition. When that was done, the two local tribes, exogamous and intermarrying, were constituted into the two phratries of one local tribe. But that is not the theory of observers on the spot: their hypothesis is that a promiscuous and communistic local tribe, for no known or conceivable reason, bisected itself into two exogamous and intermarrying " moieties." On the face of it, it is a fatal objection to the theory that when men dwelt in an undivided commune they recognized no system of relationships but the classificatory, yet were well aware of consanguineous relationships; were determined to prohibit the marriages of people in such relationships; and included in the new prohibition people in no way consanguineous, but merely of classificatory kin. The reformers, by the theory; were perfectly able to distinguish consanguineous kinsfolk, so that they might easily have forbidden them to intermarry; while if all the members of the tribe were not in the classificatory degrees of relationship, who were? How were persons in classificatory relationships with each other discriminated from other members of the tribe who were not? They were easily discriminated as soon as the phratries were instituted, but, we think, not before.

Term of Classificatory Relationships

Here it is necessary to say a few words about " classificatory " terms of relationship. Among many peoples the terms or names which with us denote relationships of consanguinity or affinity, such as Father, Mother, Brother, Sister, Son, Daughter, Husband, Wife, are applied both to the individuals actually consanguineous in these degrees, and also to all the other persons in the speaker's own main exogamous division or phratry who are of the same " age-grade " and social status as the Father, Mother, Brother, Sister, Son, Daughter, Husband, Wife, and so forth. As a man thus calls all the women whom he might legally have married by the same term as he calls his wife, and calls all children of persons of his own " age-grade," class and status by the same name as he calls his own children, many theorists hold this to be a proof of the origin of the nomenclature " in a system of group marriage in which groups of men exercised marital rights over groups of women, and the limitation of one wife to one husband was unknown. Such a system would explain very simply why every man gives the name of wife to a whole group of women, and every woman gives the name of husband to a whole group of men," and so on with all such collective terms of relationship.' Certainly this is a very simple explanation. But if we wished to explain why every Frenchman applies the name which he gives to his " wife " (femme) to every " woman " in the world, it would be rather simpler than satisfactory to say that this nomenclature arose when the French people lived in absolute sexual promiscuity. The same reasoning applies to English " wife," German Weib, meaning " woman," and so on in many languages. Moreover the explanation, though certainly very simple, is not " the only reasonable and probable explanation." Suppose that early man, as in a hypothesis of Darwin's, lived, not in large local tribes with the present polity of such tribes in Australia, but in " cyclopean families," where the sire controlled his female mates and offspring; and suppose that he, from motives of sexual jealousy, and love of a quiet life, forbade amours between his sons and daughters. Suppose such a society to reach the dimensions of a tribe. The rules that applied to brothers and sisters, mothers and sons, would persist, and the original names for persons in such relationships in the family would be extended, in the tribe, to all persons of the same status: new terms being adopted, or old terms 'extended, to cover new social relationships created by social laws in a wider society.

Another Theory of the Origin of Totemism and Exogamy

How this would happen may be seen in studying the other hypothesis ' Totemism, i. 304.

of exogamy and totemism. 2 Man was at first, as Darwin supposed, a jealous brute who expelled his sons from the neighbourhood of his women; he thus secured the internal peace of his fire circle; there were no domestic love-feuds. The sons therefore of necessity married out - were exogamous. As man became more human, a son was permitted to abide among his kin, but he had to capture a mate from another herd (exogamy). The groups received sobriqugts from each other, as Emu, Frog, and so forth, a fact illustrated copiously in the practice of modern and English and ancient Hebrew villages.3 The rule was now that marriage must be outside of the local group-name. Frog may not marry Frog, or Emu, Emu. The usual savage superstition which places all folk in mystic rapport with the object from which their names are derived gradually gave a degree of sanctity to Emu, Frog and the rest. They became totems.

Perhaps the captured women in group Emu retained and bequeathed to their children their own group-names; the children were Grubs, Ants, Snakes, &c. in Emu group. Let two such groups, Emu and Kangaroo, tired of fighting for women, make peace with connubium, then we have two phratries, exogamous and intermarrying, Emu and Kangaroo, with totem kins within them. (Another hypothesis is necessary if the original rule of all was, as among the Urabunna and other tribes, that each totem kin must carry out of itself into only one other totem kin. 4 But we are not sure of the fact of one totem to one totem marriage.) In short, the existence of the two main exogamous divisions in a tribe is the result of an alliance of two groups, already exogamous and intermarrying, not of a deliberate dissection of a promiscuous horde.' The first objection to this system is that it is not held by observers on the spot, such as Mr Howett and Mr Spencer. But while all the observed facts of these observers are accepted (when they db not contradict their own statements, or are not corrected by fresh observations), theorists are not bound to accept the hypotheses of the observers. Every possible respect is paid to facts of observation. Hypotheses as to a stage of society which no man living has observed may be accepted as freely from Darwin as from Howitt, Spencer and L. Morgan.

It is next objected that " the only ground for denying that the elaborate marriage-system " (systems ?) " of the Australian aborigines has been devised by them for the purpose which it actually serves, appears to be a preconceived idea that these savages are incapable of thinking out and putting in practice a series of checks on marriage so intricate that many civilized persons lack either the patience or the ability to understand them. .. The truth is that all attempts to trace the origin and growth of human institutions without the intervention of human intelligence and will are radically vicious and foredoomed to failure." 6 But nobody is denying that the whole set of Australian systems of marriage is the result of human emotions, intelligence and will. Nobody is denying that, in course of time, the aborigines have thought out and by successive steps. have elaborated their systems. The only questions are, what were the human motives and needs which, in the first instance, set human intelligence and will to work in these directions; and how, in the first instance, did they work? The answers given to these questions are purely and inevitably hypothetical, whether given by observers or by cloistered students.

It is objected, as to the origin of totemism, that too much influence is given to accident, too little to design. The answer is that " accident " plays a great part in all evolution, and that, 2 Lang and Atkinson, Social Origins and Primal Law; Lang, Secret of the Totem. Lang, Social Origins and Secret of the Totem. 4 Anthropological Essays, pp. 206-209.

This theory, already suggested by the Rev. J. Mathew, and Mr Daniel McLennan, occurred independently to M. Van Gennep, who, in Mythes et le'gendes d'Australie, suppressed his chapter on it, after reading The Secret of the Totem. The conclusions were almost identical with those of that work (Op. cit. pp. vi. xxxiv.). The details of the evolution, which are many, may be found in Social Origins and Primal Law, and revised in The Secret of the Totem. 6 Totemism, i. 280, 282.

in the opposed theory, the existence and actual exogamous function of totems is also accidental, arising from ignorance and a peculiar superstition. It is urged that no men would accept a nickname given from without by hostile groups. This is answered by many examples of cases in which tribes, clans, political parties, and, of course, individuals, have accepted sobriquets from without, and even when these were hostile and derisive.' It is asked, Why, on this theory, are there but two exogamous divisions in the tribe? The reply is that in America there may be three or more: that in the Urabunna there are as many exogamous divisions (dual) as there are totems, and that these, like the main exogamous divisions, go in pairs, because marriage is between two contracting parties.2 It is maintained in this theory that Australian blacks, who are reflective and by no means illogical men, have long ago observed that certain marriages are rigorously barred by their social system, for no obvious reason. Thus a man learns that he must not marry in his own main exogamous division, say Eagle Hawk. He must choose a wife from the opposite division, Crow. She must belong to a certain set of women in Crow, whose tribal status is precisely that, in Crow, of his own sisters, and his " little sisters " (the women of his sister's status) in Eagle Hawk. The reflective tribesman does not know why these rules exist. But he perceives that the marriageable women in his own main division bear the same title as his sisters by blood. He therefore comes to the conclusion that they are all what his own sisters manifestly are, " too near flesh," as the natives say in English; and that the purpose of the rule is to bar marriage to him with all the women who bear the name " sisters " that denotes close consanguinity. Presently he thinks that other kinsfolk, actual, or bearing the same collective title as actual kinsfolk of his, are also " too near flesh," and he goes on to bar them till he reaches the eight class model; or like some south-eastern tribes, drops the whole cumbrous scheme in favour of one much like our own.

The reflective savage, in short, acts exactly as the Church did when she extended to cousins the pre-existing Greek and Roman prohibitions against the marriages of very near kin; and, again, extended them still further, to exclude persons not consanguineous at all but called by the same title as real consanguines, " father," " mother " and " child " in " gossipred " - godfather, godmother, godchild. The savage and ecclesiastical processes are parallel and illustrate each other. Probably when a tribe with two main exogamous and intermarrying divisions came into existence in the way which we have indicated, the names used in families for father, mother, daughter, son, husband, wife, brother, sister, were simply extended so as to include, in each case, all persons in the tribe who were now of the same status, socially, with the same rights, restrictions and duties, as had been theirs in the fire-circle before the tribe was made a tribe by the union of two exogamous and previously hostile intermarrying local groups; or two sets of such groups. The process is natural; the wide extension now given to old names of relationships saved the trouble of making new names. Thus we have found a reasonable and probable way of accounting for classificatory terminology without adopting the hypothesis that it arose out of " groupmarriage " and asking " But how did group-marriage arise?" There is no accident here, all is deliberate and reflective design, beginning with the purely selfish and peace-loving design of the jealous sire. Meanwhile the totemic prohibition, " no marriage in the same totem name," has been retained and expanded even beyond the tribe, and " however remote the hunting grounds " of two persons, they may not intermarry if their totem name be the same.

Such are the two chief opposed theories of the origins of exogamy, and of the connexions of exogamy with totemism. The second does not enjoy the benefit of notice and criticism in Mr Frazer's Totemism. 1 The Secret of the Totem, pp. 128, 134.

For other arguments explaining the duality of the divisions see Van Gennep, ut supra, p. xxxiv. and note 1.

Relations of the Social and Religious Aspects of Totemism

It is a curious fact (if it be accepted as a fact) that the social aspect of totemism - the prohibition to marry a person of the same hereditary totem name - is sometimes strongest where the " religious " prohibition against killing or eating the totem is weakest; while the highest regard is paid to the totem, or to the god which is supposed to inhabit the totem species, where there is no prohibition on marrying within the totem name. Thus in Australia, where (except in the centre, among the Arunta) almost all tribes prohibit marriages within the totem name, it is scarcely possible to find an instance in which irreligious treatment of the totem, killing or eating it, is (as among many other totemic peoples) thought to be automatically or " religiously " punished by illness, death or miscarriage. Religion, in these cases, does not hold that the injured majesty of the totem avenges itself on the malefactor. On the other hand the Samoans, who pay no regard to the sacred animal of each community in the matter of not marrying within his name, believe that he will inflict death if one of his species be eaten - and if no expiatory rite be performed. 3 In Samoa, we saw, the so-called totem is the vehicle of a God; in Australia no such idea is found.

Meanwhile the offence of marrying within the totem name is nowhere automatically punished in any way except among the American Navajos, where, to make certain, the totem kin also inflicts secular penalties; 4 and it is part of the magic of the Intichiuma rites for the behoof of the totem that his kin should eat of him sparingly, as on all occasions they may do. In all other quarters, where marriage within the totem kin is forbidden, the penalty of a breach of law. has been death or tribal excommunication. The offence is secular. The Euahlayi, who never marry within the totem name, " may and do eat their hereditary totems with no ill effects to themselves." s This is very common in South Australia. As a rule, however, in Australia some respect is paid to the actual plant or animal, and some Northern tribes who inherit the paternal totem respect it almost as much as the maternal totem. As they also inherit property in the maternal line, it seems clear that they have passed from female to male descent, as regards the totem, but not as regards inheritance.6 Male and Female Descent of the Totem. - It was the almost universal opinion of anthropologists that, in the earliest totemic societies, the totem was inherited from the mother, and that inheritance from the father was a later development. But when the peculiar totemism of the Arunta was discovered, and it was desired to prove that this non-exogamous totemism was the most primitive extant, it was felt to be a difficulty that the Arunta reckon descent of everything hereditable in the male, not the female line. If then, the Arunta were not primitive but advanced, in this matter as well: as in their eight sub-classes and ceremonies, how could their totemism be primitive? It would have been easy to reply that a people might be " primitive " in some details though advanced in others - the fact is notorious. But to escape from the dilemma the idea was proposed that neither male nor female descent was more primitive than the other. One tribe might begin with male, one with female descent. Nobody can prove that it was not so, but " whereas evidence of the passage from female to male reckoning may be observed, there is virtually none of a change in the opposite direction."' Thus the Worgaia and Northern neighbours of the Arunta, with male descent, have certainly passed through a system of female descent of the totem, and actually inherit property in the female line, while Strehlow's Aranda or Arunta inherit their mothers' totems. Moreover Howitt shows us at least one tribe ' Turner, Samoa, p. 31, sqq.

4 Bourke, Snake Dance of the Moquis, p. 279.

Mrs Langloh Parker, The Euahlayi Tribe, p. 279.

6 See for Worgaia and Warramunga reverence of the mother's totem, though they inherit the father's, Spencer and Gillen, Northern Tribes, p. 166. That these tribes, though reckoning descent in the paternal line, inherit property in the maternal is certain, see PP. 5 2 3, 524.

7 Thomas, ut supra, p. 15.

with female descent, the Dieri, actually in the process of diverging from female to male descent of the totem. " A step further is when a man gives his totem name to his son, who then has those of both father and mother. This has been done 'even in the Dieri tribe," which appears to mean that it is also done in other tribes.1 A difficult case in marriage law is explained by saying that " possibly some man, as is sometimes the case, gave his Murdu (totem) to his son, who was then of two Murdus, and so could not marry a girl of one of his two totems." 2 We thus see how the change from female to male descent of the totem is " directly led to," as Mr Howitt says, 3 by a man's mere fatherly desire to have his son made a member of his own totem kin. On the other hand, we never read that with male descent of the totem a mother gives hers to son or daughter. All these facts make it hard to doubt (though absolute proof is necessarily impossible) that female everywhere preceded male descent of the totem.

Proof of transition from female to male descent of the totem appears to be positive in some tribes of the south of South Australia. Among them each person inherits his mother's totem, and may not marry a woman of the same. But he also inherits his father's totem, which " takes precedence," and gives its name to the local group. No person, as apparently among the Dieri when a father has " given his totem " to a son, may marry into either his father's or his mother's totem kin (Mrs Bates) .

Thus we have a consecutive series of evolutions: (a) All inherit the maternal totem only, and must not marry within it. This is the rule in tribes of south-east Australia with female descent. (b) Some fathers in this society give their totems to sons, who already inherit their maternal totems. Such sons can marry into neither the paternal nor maternal totems. This was a nascent rule among the Dieri. (c) All inherit both the paternal and the maternal totem, and may marry into neither (southern South Australia). (d) All inherit the religious regard for the maternal totem, but may marry within it, while they may not marry within the paternal totem (Worgaia and Warramunga of north central Australia). (e) The paternal totem alone is religiously regarded, and alone is exogamous (tribes of southeast Australia with male descent). (f) The totem is neither hereditary on either side nor exogamous (Spencer's Arunta). (g) The maternal totem is hereditary and sacred, but not exogamous (Strehlow's Arunta).

In this scheme we give the degrees by which inheritance of the totem from the mother shades into inheritance of the totem from both parents (Dieri), thence to inheritance of both the maternal and paternal totem while the paternal alone regulates marriage (Worgaia and Warramunga), thence to exclusive inheritance of the paternal, without any regard paid to the maternal totem (some tribes of South Australia), and so on.

Meanwhile we hear of no tribe with paternal descent of the totem in which mothers are giving their own totems also to their children. We cannot expect to find more powerful presumptions in favour of the opinion that tribes having originally only maternal have advanced by degrees to only paternal descent of the totem. Mr Frazer says, " So far as I am aware, there is no evidence that any Australian tribe has exchanged maternal .for paternal descent, and until such evidence is forthcoming we are justified in assuming that those tribes which now trace descent from the father formerly traced it from the mother."4 We have now provided, however, the evidence for various transitional stages from maternal to paternal descent, but have found no traces of the contrary process, nor more than one way of interpreting the facts. It is admitted by Mr Frazer that in several North American tribes the change from female to male descent has to all appearance been made. 5 Among the Delawares the initial process was much akin to that of the Dieri, who, in a tribe of female descent, " gives " his own totem to his sons. " The Delawares had a practice of sometimes naming a child into its father's clan," and a son thus became a member of his father's 1 N.T.S.E.A. p. 284.2 Ibid. p. 167.

Ibid. p. 284. Totemism, Ibid. iii. 42, 58, 72, 80.

clan. This " may very well have served to initiate a change of descent from the female to the male line." s Howitt says precisely the same thing about the paternal practice of the Dieri. Thus there is no reason for denying that the change from female to male descent can be made by Australian as readily as by American tribes. We have given evidence for every step in the transition. The opposite opinion arose merely in an attempt to save the primitiveness of the Arunta, some of whom actually still make the maternal totem hereditary.

The change to male descent is socially very important. The totem kin of a man, for example, takes up his blood feud. Where the descent is female a " man may probably have some (totemic) kinsmen in the same group, but equally a considerable number of members of other totem kins." But it is clear that the rule of male descent gives far greater security to the members of a local group; for they are surrounded by kinsmen, local totem groups only occurring where male descent of the totem prevails, or is predominant? The change from female to male descent of the totem, or the adoption of male descent from the first (if it ever occurred) is thus a great social advantage.

The Ways out of Totemism

While Howitt believed (though later he wavered in his opinion) that female had always preceded male descent of the totem, he also observed that with male descent came in abnormal developments. One of these is that the people of a district with male descent are often known by the name of the region, or of some noted object therein (say wild cherries). 8 They may even regard (or white observers suppose that they regard) some object as their " local totem," yet they marry within that so-called totem. But they take to marrying, not out of the hereditary totem kin, which becomes obsolescent, but out of their own region into some other given locality. Thus in the Kurnai tribe there were no inevitable hereditary totems, but thundung were given by the fathers to lads" when about ten years old or at initiation." 9 The animal thundung(elder brother) was to protect the boy, or girl (the girl's thundung was called banung). The names of the creatures, in each case, appear to have been given to their human brothers and sisters; the thundung name descended to a man's sons. " The names are perpetuated " (under male descent) " from generation to generation in the same locality." 10 Thus it appears that when a Kurnai wishes to marry he goes to a locality where he finds girls of banung names into which he may lawfully wed. So far he seems, in fact, to practise totemic exogamy; that he has to travel to a particular locality is merely an accident. Though the thundung and banung names are not inherited at birth by the children, they are given by the father when the child is old enough to need them." On the whole, we seem to see, in tribes where male descent is of old standing, that the exogamous function of the totem becomes obsolete, but a shadow of him, as thundung, retains a sort of " religious " aspect and even an unappreciated influence in marriage law.

In Fiji and Samoa, in Melanesia 12 and British New Guinea, many types of contaminated and variegated survivals of totemism may be studied. In the Torres Islands 13 hero-worship blends with totemic survivals. As in parts of South Africa, where a tribe, not a kin, has a sacred animal, as in Fiji, he seems to be the one survivor of many totems, the totem of some dominant local 6 Totemism, 7 Except among the Arunta, where, though totems come by change, local groups are usual. See Spencer and Gillen, Central Tribes, p. 9. How this occurs we can only guess. See Folk Lore, vol. xx., No. 2, pp. 229-231. Here it is conjectured that adults of the totem congregate for the purpose of convenience in performing Intich'uma, or magical services for the propagation of the totem as an article of food. For the nature of these rites, common in the central and northern but unknown to the south-eastern tribes, see Central Tribes, pp. 167-212, and Northern Tribes, pp. 283-320. The Arunta totem aggregates are magical local societies.

8 Central Tribes, pp. 8, 9.9 N.T.S.E.A. p. 146.

"° Ibid. p. 046.11 Cf. Howitt, ibid. pp. 270-279.

12 Rivers, " Totemism in Polynesia and Melanesia," Journ. Anthrop. Inst. vol. xxxix.

13 Haddon, Cambridge Expedition, vol. v.

totem group, before which the other totems have fled, or but dimly appear, or are vehicles of gods, or, in Africa, of ancestral spirits. (These African tribal sacred animals are called Sibokol.) Some tribes explain that the Siboko originated in an animal sobrique, as ape, crocodile, given from without? Sibokoism, the presence of a sacred animal in a local tribe, can hardly be called totemism, though it is probable that the totem of the leading totem kin, among several such totem kins in a tribe, has become dominant, while the others have become obsolete. On the Gold Coast of Africa as long ago as 1819, Bowdich 3 found twelve " families," as he called them, of which most were called by the name of an animal, plant or other object, more or less sacred to them. They might not marry a person of the same kindred name, and there can be little doubt that totemism, with exogamy, had been the rule. But now the rules are broken down, especially in the peoples of the coast. The survivals and other information may be found in the Journal of the Anthropological Institute (1906) xxxvi. 178, 188.

There are fainter traces of totemism in the Awemba between Lake Tanganyika and Lake Bangweolo. 4 A somewhat vague account of Bantu totems in British East Africa, by Mr C. W. Hobley, indicates that among exogamous " clans " a certain animal is forbidden as food to each " clan." 5 The largest collection of facts about African totemism, from fresh and original sources, is to be found in Mr Frazer's book. For totemism in British Columbia the writings of Mr Hill Tout may be consulted. 6 The Thlinkit tribes have the institution in what appears to be its earliest known form, with two exogamous phratries and female descent. Among the Salish tribes " personal " totems are much more prominent. Mr Hill Tout, with Professor F. Boas, considers the hereditary exogamous totem to have its origin in the non-exogamous personal totem, which is acquired in a variety of ways. The Salish are not exogamous, and have considerable property and marked distinctions of rank. It does not, therefore, appear probable that their system of badges or crests and personal totems is more primitive than the totemic rules of the less civilized Thlinkits, who follow the form of the south-east Australian tribes.?

Other very curious examples of what we take to be aberrant and decadant totemism in New Guinea are given by Mr Seligmann (Man, 1908, No. 89), and by Dr Rivers for Fiji (Man, 1908, No. 75). Mr Seligmann (Man, 1908, No. lo)) added to the information and elucidated his previous statements. The " clans " in British south-east New Guinea usually bear geographical names, but some are named after one of the totems in the " clan." " Every individual in the clan has the same linked totems," of which a bird, in each case, and a fish seem to be predominant and may not be eaten. " The clans are exogamous ... and descent is in the female line." It appears, then, that a man, having several totems, all the totems in his " clan," must marry a woman of another " clan " who has all the totems of her " clan." Similar multiplicity of totems, each individual having a number of totems, is described in Western Australia (Mrs Bates). In this case the word " totem " seems to be used rather vaguely and the facts require elucidation and verification. In this part of Australia, as in Fiji 8 " pour la naissance ... l'apparition du totem-animal avait toujours lieu." In Fiji the mother sees the animal, which does not affect conception, and " is merely an omen for the child already conceived." But in Western Australia, as we have seen, the husband dreams of an animal, which is supposed to follow him home, and to be the next child borne by his wife. If it is correctly stated that when the husband has dreamed of no animal, while nevertheless his wife has a baby, the husband spears the man whom he suspects of having dreamed of an animal, the marital jealousy 1 Frazer, "Totemism, South Africa," Man (1901), No. iii.

2 See Secret of the Totem, pp. 25, 26. a Mission to Ashanti. 4 Journ. Anthrop. Inst. (1906), xxxvi. 154.

6 Ibid. (1903), xxxiii. 346-348. B Ibid. (1903-1904).

See discussion in Secret of the Totem for details and references.

8 Pere Schmidt, Man (3908), No. 84, quoting Pere de Marzan, Anthropos, ii. 400-405. takes an unusual form and human life becomes precarious. But probably the husband has some reason for the direction of his suspicions. He never suspects a woman.

" The Banks' Islanders," says Mr Frazer, " have retained the primitive system of conceptional totemism." 9 On the other hand Dr Rivers, who is here our authority, writes " totemism is absent " from " the northern New Hebrides, the Banks' and the Terres groups." 1° In a place where totemism is absent it does not prima facie seem likely that we shall discover " the primitive system of conceptional totemism." The Banks' Islanders have no totemism at all. But they have a certain superstition applying to certain cases, and that superstition resembles Arunta and Loritja beliefs, in which Mr Frazer finds the germs of totemism. The superstition, however, has not produced any kind of totemism in the Banks' group of isles, at least, no totemism is found.." There are," writes Dr Rivers, " beliefs which would seem to furnish the most natural starting-point for totemism, beliefs which Dr Frazer has been led by the Australian evidence " (by part of the Australian evidence, we must .say) " to regard as the origin of the institution." Thus, in Banks' Islands we have the starting-point of the institution, without the institution itself, and in many Australian tribes we have the institution - without the facts which are " the most natural starting-point." As far as they go these circumstances look as if " the most natural " were not the actual starting-point. The facts are these: in the Isle of Mota, Banks' group, " many individuals " are under a tabu not to eat, in each case, a certain animal or fruit, or to touch certain trees, because, in each case, " the person is believed to be the animal or fruit in question." This tabu does not, as in totemism, apply to every individual; but only to those whose mothers, before the birth of the individuals, " find an animal or fruit in their loin-cloths." This, at least, " is usually " the case. No other cases are given. The women, in each case, are informed that their child " will have the qualities of the animal " (or fruit) " or even, it appeared would be himself or herself the animal " (or fruit). A coco-nut or a crocodile, a flying fox or a brush turkey, could not get inside a loin-cloth; the animal and fruits must be of exiguous dimensions. When the animal (or fruit) disappears " it is believed that it is because the animal has at the time of its disappearance entered into the woman. It seemed quite clear that there was no belief in physical impregnation on the part of the animal nor of the entry of a material object in the form of the animal. ., but, so far as I could gather, an animal found in this way was regarded as more or less supernatural, a spirit animal and not one material, from the beginning." " There was no ignorance of the physical role of the human father, and the father played the same part in conception as in cases unaccompanied by an animal appearance." The part played by the animal or fruit is limited to producing a tabu against the child eating it, in each case, and some community of nature with the animal or fruit. Nothing here is hereditary. The superstition resembles some of those of the Arunta, Loritja and Euahlayi. Among the Euahlayi the superstition has no influence; normal totemism prevails; among the Arunta nation it is considered to be, and Dr Rivers seems to think that it is, likely to have been the origin of totemism. In Mota, however, it either did not produce totemism, or it did; and, where the germ has survived in certain cases, the institution has disappeared - while the germinal facts have vanished in the great majority of totemic societies. Dr Rivers does not explain how a brush turkey, a sea snake or a flying fox can get into a woman's loin-cloth, yet these animals, also crabs, are among those tabued in this way. Perhaps they have struck the woman's fancy without getting into her loin-cloth.

It is scarcely correct to say that " the Banks' Islanders have retained the primitive system of conceptional totemism." They only present, in certain instances, features like those which are supposed to be the germs of a system of conceptional 9 Man, iv. 128.

to " Totemism in Polynesia and Melanesia," Journ. Anthrop. Inst. xxxix. 173, sqq.

totemism. In the case of the Arunta we have demonstrated that hereditary and exogamous totemism of the normal type preceded the actual conceptional method of acquiring, by local accident, " personal totems." If the Banks' Islanders were ever totemists they have ceased to be so, and merely retain, in cases, a superstition analogous to that which, among the Arunta, with the aid of the stone churinga, has produced the present unique and abnormal state of affairs totemic.

For totemism in India, see Dalton, Descriptive Ethnology of Bengal; for the north of Asia, Strahlenberg's Description, &c. (1738); and in all instances Mr Frazer's book.

Myths of Totem Origins

The myths of savages about the origin of totemism are of no historical value. Not worshipping ancestral spirits, an Australian will not, like an ancestorworshipping African, explain his totem as an ancestral spirit. But where, as in the north and centre, he has an elaborate philosophy of spirits, there the primal totems exude spirits which are incarnated in women.

In their myths as to the origin of totemism, savages vary as much as the civilized makers of modern hypotheses. Some claim descent from the totem object; others believe that an original race of animals peopled the world; animals human in character, but bestial, vegetable, astral or what not, in form. These became men, while retaining the rapport with their original species; or their spirits are continually reincarnated in women and are born again (Arunta of Messrs Spencer and Gillen); or spirits emanating from the primal forms, or from objects in nature, as trees or rocks, connected with them, enter women and are reincarnated (Arunta of Mr Strehlow and some Australian north-western tribes, studied by Mrs Bates). Other Australians believe that the All-Father, Baiame, gave totems and totemic laws to men.' There are many other explanatory myths wherever totemism, or vestiges thereof, is found in Australia, Africa, America and Asia.

All the myths of savages, except mere romantic Mdrehen, and most of the myths of peoples who, like the Greeks, later became civilized, are " aetiological," that is, are fanciful hypotheses made to account for everything, from the universe, the skies, the sun, the moon, the stars, fire, rites and ceremonies, to the habits and markings of animals. It is granted that almost all of these fables are historically valueless, but an exception has been made, by scholars who believe that society was deliberately reformed by an act bisecting a tribe into two exogamous divisions, for savage myths which hit on the same explanation. We might as well accept the savage myths which hit on other explanations, for example the theory that Sibokoism arose from animal sobriquets. Exceptions are also made for Arunta myths in which the primal ancestors are said to feed habitually if not exclusively on their own totems. But as many totems, fruit, flowers, grubs, and so on are only procurable for no longer than the season of the May-fly or the March-brown, these myths are manifestly fabulous.

Again the Arunta primal ancestors are said to have cohabited habitually with women of their own totem, though without prejudice against women of other totems whom they encountered in their wanderings. These myths are determined by the belief in oknanikilla, or spots haunted by spirits all of one totem, which, again, determine the totem of every Arunta.. The idea being that the fabled primal ancestors male and female in each wandering group of miracle-workers were always all of one totem, it follows that, if not celibate, which these savages never are, they must have cohabited with women of their own totem, and, by the existing Arunta system, there is no reason why they should not have done so. In no other field of research is historical value attributed to savage legends about the inscrutable past that lies behind existing institutions.

We are thus confronted by an institution of great importance socially where it regulates marriages and the blood-feud, or where it is a bond of social union between kinsmen in the totem or members of a society which does magic for the behoof ' Mrs Langloh Parker, The Euahlayi Tribe. of its totem (central and north-western Australia), and is of some " religious " and mythical importance when, as in Samoa, the sacred animal is regarded as the vehicle of a god. Of the origin of these beliefs, which have practical effects in the evolution of society and religion, much, we saw, is conjectured, but as we know no race in the act of becoming totemic - as in all peoples which we can study totemism is an old institution, and in most is manifestly decaying or being transmuted - we can only form the guesses of which examples have been given. Others may be found in the works of Herbert Spencer and Lord Avebury, and criticisms of all of them may be read in A. Lang's Social Origins. Whether or not survivals of totems are to be found in the animal worship of ancient Egypt, in the animal attendants of Greek gods, in Greek post-Homeric legends of descent from gods in various bestial disguises, and in certain ancient Irish legends, it is impossible to be certain, especially as so many gods are now explained as spirits of vegetation, to which folk-lore assigns carnal forms of birds and beasts.

Other Things called Totems

As has been said, the name " totem " is applied by scholars to many things in nature which are not hereditary and exogamous totems.' The " local totem " (so called) has been mentioned, also " linked totems." Personal Totems. - This is the phrase for any animal or other object which has been " given " to a person as a protective familiar, whether by a sorcerer 2 or by a father, or by a congress of spaewives at birth; or whether the person selects it for himself, by the monition of a dream or by caprice. The Euahlayi call the personal totem Yunbeai, the true totem they style Dhe. They may eat their real but not their personal totems, which answer to the hares and black cats of our witches.

Three or four other examples of tribes in which " personal totems " are " given " to lads at initiation are recorded by Howitt. 3 The custom appears to be less common in Australia than in America and Africa (except in South Australia, where people may have a number of "personal totems "). In one case the " personal totem " came to a man in a dream, as in North America. 4 Here it may be noted that the simplest and apparently the easiest theory of the origin of totemism is merely to suppose that a man, or with female descent a woman, made his or her personal totem hereditary for ever in his or her descendants. But nobody has explained how it happened that while all had evanescent personal totems those of a few individuals only become stereotyped and hereditary for ever.

Sex-Totems

The so-called " sex totem " is only reported in Australia. Each sex is supposed by some tribes to have its patron animal, usually a bird, and to injure the creature is to injure the sex. When lovers are backward the women occasionally kill the animal patron of the men, which produces horseplay, and " a sort of jolly fight," like sky-larking and flirtation.' The old English " jolly kind of fight," between girls as partisans of ivy, and men as of the holly " sex-totem," is a near analogue. It need not be added that " sex-totems " are exogamous, in the nature of things.

Sub-Totems

This is the name of what are also styled " multiplex totems," that is, numerous objects claimed for their own by totem kins in various Australian regions. The Emu totem kin, among the Euahlayi tribe, claims as its own twenty-three animals and the north-west wind.' The whole universe, including mankind, was apparently divided between the totem kins. Therefore the list of sub-totems might be extended indefinitely.' These " sub-totems " are a savage effort at universal classification.

Conclusion

We have now covered the whole field of controversy as to the causes and origins of totemic institutions. Australia, with North America, provides the examples of those institutions which seem to be " nearest to the beginning," and in Australia the phenomena have been most carefully and 2 The Euahlayi Tribe, p. 21. N.T.S.E.A. pp. 144-148.

4 Ibid. p. 154. Ibid. pp. 148-151. ' The Euahlayi Tribe, p. 15.

7 N.T.S.E.A. p. 454.

elaborately observed among peoples the least sophisticated. In North America most that we know of many great tribes, Iroquois, Hurons, Delawares and others, was collected long ago, and when precision was less esteemed, while the tribes have been much contaminated by our civilization. It has been unavoidably necessary to criticize, at almost every stage, the conclusions and hypotheses of the one monumental collection of facts and theories, Mr Frazer's Totemism (1910). Persons who would pursue the subject further may consult the books mentioned in the text, and they will find a copious, perhaps an exhaustive bibliography in the references of Mr Frazer's most erudite volume's, with their minute descriptive account not only of the totemism, but of the environment and general culture of hundreds of human races, in Savagery and in the Lower and Higher Barbarism. (A. L.)