That primitive economics bear the impress of sex cleavage is borne out by every class of evidence, and it is in this circumstance that we first come upon societies distinguished by containing two of the most important social elements, exogamy and totemism. Before, however, examining examples of societies containing the two elements of exogamy and totemism, it will be necessary to say something by way of preliminaries on these two elements themselves. They have rightly been made the subject of important special inquiry by anthropological scholars, as being in fact the key to the question of social evolution, and we shall clear the ground considerably by first of all turning to the principal authorities on the subject, and ascertaining the present position of the inquiry.

I must however note, in the first place, that as I have stated the case, exogamy and totemism appear as two separate and distinct elements, whereas it is usual to consider exogamy as an essential part of totemism. I cannot, however, see that this is so. In advanced totemism, it is true, they are found as inseparable parts of one system, but they may well have started separately and coalesced later. In point of fact, all the evidence points in this direction, and if we cease to consider exogamy as a necessary element of totemism, we can advance investigation more rapidly and with greater accuracy.

We come very quickly upon what may be termed natural exogamy. Male working with male outside the groups formed by women and the younger offspring would produce a natural exogamy, which would have followed upon the exogamy produced by hostile capture of women, and two streams of influence would thus tell in favour of the evolution of a system of formal exogamy, and Dr. Westermarck's theory of a natural avoidance of housemates, with all its wealth of evidence, helps us at this point.

The position is not so clear as to totemism. If we begin, however, with a clear understanding that it is not a part of the machinery of exogamous grouping, but an independent growth of its own, we shall have gained an important point, for the contrary opinion has very often obscured the issue and prevented research in the right direction.

It will be advisable to have before us the principal theories as to the origin of totemism. There are practically three—Mr. Frazer's, Mr. Lang's, and Mr. Baldwin Spencer's. Mr. Frazer considers totemism to be "in its essence nothing more or less than an early theory of conception, which presented itself to savage man at a time when he was still ignorant of the true cause of the propagation of the species." Mr. Frazer explains this theory further by saying that "naturally enough, when she is first aware of the mysterious movement within her, the mother fancies that something has that very moment passed into her body, and it is equally natural that in her attempt to ascertain what the thing is, she should fix upon some object that happened to be near her, or to engage her attention at the critical moment."[356]

Mr. Lang rejects Mr. Frazer's theory in toto, and propounds his own as due to the naming of savage societies, and to a sort of natural exogamy produced by practically the same set of conditions as I have already described. Mr. Lang's totemism began in the primary groups, and began with exogamy as a necessary part of it. "Unessential to my system," says Mr. Lang, "is the question how the groups got animal names, as long as they got them, and did not remember how they got them, and as long as the names according to their way of thinking indicated an essential and mystic rapport between each group and its name-giving animal. No more than these three things—a group animal name of unknown origin; belief in a transcendental connection between all bearers human and bestial of the same name; and belief in the blood superstitions (the mystically sacred quality of the blood as life)—was needed to give rise to all the totemic creeds and practices including exogamy," and further, "we guess that for the sake of distinction, groups gave each other animal and plant names. These became stereotyped we conjecture, and their origin was forgotten. The belief that there must necessarily be some connection between animals and men of the same names led to speculation about the nature of the connection. The usual reply to the question was that the men and animals of the same name were akin by blood. The kinship with animals being particularly mysterious was peculiarly sacred. From these ideas arose tabus, and among others that of totemic exogamy."[357]

Mr. Baldwin Spencer, and with him Dr. Haddon, consider totemism to have arisen from economic conditions. Primitive human groups, says Dr. Haddon, "could never have been large, and the individuals comprising each group must have been closely related. In favourable areas each group would have a tendency to occupy a restricted range, owing to the disagreeable results which arose from encroaching on the territory over which another group wandered. Thus, it would inevitably come about that a certain animal or plant, or group of animals or plants, would be more abundant in the territory of one group than in that of another."[358]

These theories are not necessarily mutually destructive, though they seem to me even collectively not to contain the full case for totemism. Mr. Frazer does not account for woman's isolation at the time of conceptual quickening, for the closeness of her observation of local phenomena, and for the separateness of her ideas from the actual facts of procreation. Mr. Lang overloads his case. He is accounting not for the origin of totemism, but for the origin of all, or almost all, that totemism contains in its most developed forms—"all the totemic creeds and practices including exogamy" as he says. He postulates a name-giving process by drawing upon the conceptions as to names by advanced savage thought, and he does not account for the fact that according to his theory, animals and plants must not only have been named, but named upon some sort of system known to a wide area of peoples, before totemistic names for the groups could have been given to them. Mr. Spencer's and Dr. Haddon's theory is perhaps open to the doubts caused by Mr. Lang's criticism of it that there is only one case of a known economic cause for totemism—an Australian case where two totem kins are said to have been so called "from having in former times principally subsisted on a small fish and a very small opossum;"[359] but on the other hand it does supply a vera causa, the actual evidence for which may well have passed away with the development of totemism, without leaving survivals.

All these theories, however, are the result of considerable research and experience, and it is more than probable that they may each contain fragments of the truth which need the touch of combination to show how they stand in relation to the problem which they are propounded to solve. There are features of totemism which are not noticed by any of these distinguished authorities. By using the hitherto unnoticed features, I think it possible to produce a theory as to the origin of totemism, which will contain the essential features of those theories now prominently before the world.

I will set down the order in which the problem can be approached from the standpoint already reached, and we may afterwards try to ascertain what proof is to be derived from totemic societies of the rudest type.

Now totemism is essentially a system of social grouping, whose chief characteristic is that it is kinless—that is to say, the tie of totemism is not the tie of blood kinship, but the artificially created association with natural objects or animals. It takes no count of fatherhood, and only reckons with the physical fact of motherhood. It is not the actual fatherhood or the actual motherhood which is the fundamental basis of totemism, but the association with animal, plant, or other natural object. This is evidently the fact, whatever view is taken of totemism, and that totemism is, in its origin and principle, a kinless, not a kinship system, is the first fact of importance to bear in mind throughout all inquiry. Thus Messrs. Spencer and Gillen say "the identity of the human individual is often sunk in that of the animal or plant from which he is supposed to have originated."[360]

The next fact of importance is that as it commences at birth time, it must be closely associated with the mother and her actions as mother. This leads us to the observation that it is through the agency of the mother that the totem name is conferred upon their children, and to the necessary antecedent fact that women must have themselves possessed the name they conferred—possessed, that is, either the name as a personal attribute and valued as such, or else the power of evolving the name and the capacity of using it with totemic significance. I conclude from this, therefore, that the search for the origin of totemism must be made from the women's side of the social group. Such a search would lead straight to the industrialism of early woman, from which originated the domestication of animals, the cultivation of fruits and cereals, and the appropriation of such trees and shrubs as were necessary to primitive economics.[361] The close and intimate relationship with human life which such animals, plants, and trees would assume under the social conditions which have been postulated as belonging to this earliest stage of evolution, and the aid which these friendly and always present companions would render at all times and under most circumstances, would generate and develop many of those savage conceptions which have become known to research. As human friends they would become part of humanity, just as Livingstone notes of an African people that they did not eat the beef which he offered to them because "they looked upon cattle as human and living at home like men,"[362] an idea which is also the basis of the custom in India not to taste fruit of a newly planted mangrove tree until it is formally "married" to some other tree.[363] These are but the fortunate instances where definite record in set terms has been made. At the back of them lies a whole collection of anthropomorphic conceptions, indulged in by man at all stages of his career.[364] As superhuman agencies for pregnancy and birth, they would do what the human father in the society we are contemplating could not be expected to do, for he would be seldom present during the long period of pregnancy; he would have shared with other males the privileges of sexual intercourse, and he would therefore not be so closely in companionship with the women of the local groups as the friendly animal, plant, or tree who did so much for the mothers. There would thus be formed the groundwork for the fashioning of that most incredible of all beliefs, well founded, as Mr. Hartland has proved both from tradition and belief,[365] that the human father was not father, and that other agencies were responsible for the birth of children.

Gathering up the several threads of this argument, it seems to me that there is within this sphere of primitive thought and within these conditions of primitive life, ample room for the growth of all the main conceptions belonging to totemism; and it will be seen how necessary it is to separate totemism at its beginning from totemism in its most advanced stages. Totemism has not come to man fully equipped in all its parts. It is like every other human institution, the result of a long process of development, and the various stages of development are important parts of the evidence as to origins. At the beginning, it was clearly not connected with blood kinship and descent; it was as clearly not connected with any class system of marriage. But its beginnings would allow of these later growths, would perhaps almost engender these later growths.

Thus, the primary notion of the totem birth of children would, when blood kinship and descent became a consciously accepted element in social development, easily slide into the belief of a totemic ancestor and kinship with the totem; the protection and assistance afforded by the totem to the women of the primary groups who became the mothers of new generations, would easily grow into a sort of worship of the totem; the adoption of the totem name from the circumstances of birth implying the origin of the name from within the group and not from without would, as aggregation took the place of segregation, give way before the association of groups of persons with common interests; the aggregate totem name would come to the separate local totems as soon as, but not before, aggregation had taken the place of segregation in the formation of the social system, and this was not at the earliest stage; the close association of the totems with groups of mothers who always took the fathers of their children from without the mother group, would readily develop into differentiating the mother totems within the group from the totems of the fathers without the group, and this differentiation would produce a special relationship between the sexes based upon the difference of totems instead of upon the sameness of them; and finally there would be produced first a two-class division founded on sex—all the mothers and all the fathers—and, only in a developed form, a two-class division founded on the accepted totem name.

If this is a probable view of the course of totemic evolution, we may more confidently refer to its final stages for further evidence. Advanced totemic society shows a constant tendency to substitute blood kinship for the association with natural objects: first, blood kinship with the mother, then with the mother and the father, finally recognised through the father only. At this last stage, blood kinship has practically succeeded in expelling totemic association altogether in favour of tribal kinship by blood descent, for totemism with male descent as the basis of the social group is totemism in name only; the names of totemism remain but they are applied to kinship tribes or sections of tribes, and they do duty therefore as a convenient name-system without reference to their origin in definite association with the naming animal or plant; and it is already in position to surrender also the names and outward signs. Blood kinship is therefore the destroyer, not the generator, of totemism, and we are therefore compelled to get at the back of blood kinship if we want to find totem beginnings.

This is an important aspect of the case, and it is one which, I think, cannot be ignored. We have found that rudimentary totemism was the basis of a social system founded on artificial associations with animal or plant, was therefore kinless in character; and we have found that when totemism has been carried on into a society developed upon the recognition of blood kinship, blood kinship became antagonistic to totemism, and ultimately displaced it. These two facts point to the rudimentary kinless system as the true origin of totemism.

III

Now we may test these conclusions by applying the theory they contain to an actual case of totemic society. It would be well to choose for this purpose a people who had specialised their totemic organisation, and there are only two supreme instances of this among the races of the world—the North American Indians and the Australians. Everywhere else, where totemism exists, it is not the dominant feature of the social organisation. In Asia and in Africa totemism is subordinate to, or at all events in close or equal association with, other elements, and we cannot be quite sure that we have in these cases pure totemism. North American totemism is in the most advanced stage. Australian totemism is to a very considerable degree less advanced, and it is therefore to Australian totemism I shall turn for evidence.

But even here it is necessary to bear in mind that primitive as the Australians are, they are not so primitive as to be in the primary stages of totemic society. They have developed, and developed strongly along totemic lines, and we know that such development once started has the capacity to proceed far. What we have to do, therefore, is to attempt to penetrate beneath the range of development, to search for the social group at the farthest from the centre point from which migration started, to discover, if we can, relics of group hostility, hostile capture of women and of kinless society, all of which belong to the primary stage from which totemic development has taken place. If we can do this, we may hope to arrive at the origin of totemism, and we are more likely to accomplish it in the case of the Australians than with any other people. If we cannot, as Mr. Lang alleges, anywhere see "absolutely primitive man and a totemic system in the making,"[366] we may go back along the lines from which totemism has developed in Australian society and see somewhat of the process of the making.

We may commence with evidence of the survival of the most primitive human trait, the condition of hostility among the local groups produced by the struggle for women. "The possession of a girl appears to be connected with all their ideas of fighting ... after a battle the girls do not always follow their fugitive husbands from the field, but frequently go over as a matter of course to the victors, even with young children on their backs."[367] Mr. Curr puts the evidence even more definitely in a primitive setting when he informs us of "the young bachelors of the tribe carrying off some of the girl wives of the grey-beards," leaving the old territory and settling at the first convenient place within thirty or forty miles of the old territory. I call this state of things "survival,"[368] because it is the existence in totemic society of the fundamental basis of pre-totemic society. It is checked in Australian totemic society by rules which show a strong development from the primitive. Thus the successful warrior may not take any of his captives to himself; "if a warrior took to himself a captive who belonged to a forbidden class, he would be hunted down like a wild beast," is the evidence of Mr. Fison, who allows it to be "a strong statement, but it rests upon strong evidence."[369] This is the exogamous class system operating even in the case of conflict, when men have resorted to their primitive instincts and their primitive methods.

This discovery of primitive hostility accompanying the obtaining of wives leads us to look for other survivals of the earliest conditions, and we come upon mother-right groups in which the females in each local group are the sexual companions of males from outside their own social group. This is shown by the Kamilaroi organisation, where "a woman is married to a thousand miles of husbands."[370] This phrase may be textually an exaggeration of actual fact, but it undoubtedly expresses a condition of things which actually existed. Women in Australian society must look outside their class, and in general outside their totem, for their sexual mates, and they must expect to be claimed as rightful sexual mates by men whom they have never seen and who live at great distances. Carry this state of things but a few steps back, and we must come to a condition of localised female groups with males moving from group to group. Surely there is something more here than savage organisation. The something more is the development into a system of one of the results of the enforced migratory conditions of early man, namely, the migratory instincts of the males moving outside the female local groups and thus producing natural exogamy. This is what appears to me to be clearly a distinct element in the Australian system. But there is a new element in juxtaposition with it. The new element is the organisation into marriage classes—not every man from without, but only special men from without, are allowed the sexual companionship.

Now in both these cases, where we have apparently penetrated to the most primitive conditions, we are also brought up abruptly against conditions which are not primitive, namely, the exogamous class system, and we are bound to conclude that this class system thus shows itself to be an intruding force which has not, however, been strong enough to quite obliterate the older forces of hostile marriage-capture and mother-right society.

Our next quest is therefore to find out, if we can, an explanation of these two contrasted elements in Australian totemic society, and for this purpose it is advisable to still further narrow down the range of inquiry to one special section of the Australian peoples. For this purpose I shall take the Arunta. There has been much controversy about this people. Mr. Lang argues that the presence of exogamous classes and male descent shows the Arunta to be more advanced than other Australian peoples;[371] Messrs. Spencer and Gillen that the survival of totem beliefs, which are local and unconnected with the class system, proves them to be the least advanced. In this country Mr. Hartland and Mr. Thomas side with Mr. Lang; Mr. Frazer with Messrs. Spencer and Gillen.

The first point of importance to note about the Arunta people is that they occupy the least favourable districts for food supply.[372] This means that they have been pushed there. They did not choose such a location—in other words, they are among the last units of the migration movements which peopled Australia; they are among the last people to have become stationary as a group, and to have been compelled to resort to the development of social organisation in lieu of constantly swarming off from the centre or from the last stopping place to the ends. This tells for primitive, not advanced, conditions.

The next point is the totem system. Messrs. Spencer and Gillen, describing one special case as an example of the rest, give us the following particulars. The Arunta believe that the most marked features of the district they inhabit, the gaps and the gorges, were formed by their Alcheringa ancestors. These Alcheringa are represented as collected together in companies, each of which consisted of a certain number of individuals belonging to one particular totem. Each of these Alcheringa ancestors carried about with him or her one or more of the sacred stones called churinga. These are the general traditions related by the Arunta of to-day to explain their own customs, and let it be noted that the explanation does not necessarily lead us to the primitive conceptions of the Arunta people, but to their present conceptions as to unknown facts. The local example is found close to Alice Springs, where there are deposited a large number of churinga carried by the witchetty grub men and women. A large number of prominent rocks and boulders, and certain ancient gum trees, are the nanja trees and rocks of these spirits. If a woman conceives a child after having been near to this gap, it is one of these spirit individuals which has entered her body, and when born must of necessity be of the witchetty grub totem; "it is, in fact, nothing else but the reincarnation of one of the witchetty grub people of the Alcheringa;" the nanja tree, or stone, ever afterwards is the nanja of the child, and there is special connection between it and the child, injury to the nanja object meaning injury to the nanja man.[373] There is evidence that the reincarnation theory is not admissible,[374] and, indeed, it does not seem warranted on the facts presented by the authors. With this unnecessary element out of the way, then, there is left a system of local totemism, arising at birth and depending upon the mother, without reference in any way to the father, associated with natural features, rocks and trees, and showing in a special way a curious system of sex cleavage by the men of the group being the exclusive guardians of the sacred churinga, and the women the active power by which the churinga becomes connected with the newly-born member of the totem group.[375]

Now at this point we may surely refer back to the custom and belief of the Semang people of the Malay Peninsula, and I suggest that we have the closest parallel between Semang belief and custom and Arunta totemism, not quite the same formula perhaps, but assuredly the same fundamental conception of every child at birth being in intimate association with objects of nature, and this association being the determining force of the newly-born man's social status and class, lasting all through life. In each case the kinless basis of totemism is thus fully shown. The totem names given by women, or assumed on account of the conditions attachable to women as mothers, did not extend to the human fathers. The fathers may be known or unknown to the mothers, but they did not become associated with the totems which the mothers associated with their children. To the extent of fatherhood, therefore, totemism of this type was clearly not based upon the natural fact of blood kinship, but upon the conscious adoption of a non-kinship form of society. To the extent of motherhood also it was not based upon blood kinship, for it was the local totem, not the mother's totem, which became the totem of the newly-born member of the group. We thus have an entirely non-kinship form of society to deal with, a kinless society, "where there is no necessary relationship of any kind between that of children and parents."[376] Primitive man consciously adapted certain of his observations of nature to his social needs, and among these observations the fact of actual blood kinship with father and mother played no part. It would appear therefore that totemism at its foundation was based upon a theoretical conception of relationship between man and animal or plant. Place of birth, association with natural objects, not motherhood and not fatherhood, are the determining factors.

We may proceed to inquire as to the social form which has become evolved from this kinless system.

In the case of the Semangs we have the kinless totemic belief and custom existing within a kinless society. In the case of the Arunta we have the kinless totemism existing in a society based on a kinless organisation still, but containing also full recognition of motherhood,[377] and perhaps recognition of physical fatherhood.[378] There is, therefore, an important distinction in the social position of the two parallel systems. Among the Semang people, their totemic belief and custom do not carry with them a superstructure of society. They form the substantive cult of the scattered social groups, which are kinless groups dependent upon ties local in character and derived from the conscious use of the facts of nature surrounding them. Among the Arunta people, on the contrary, the totem belief and custom are contained within a social system of extraordinary dimensions and proportions. Of course, the obvious questions to raise are—have the Semang people lost a once existing social system connected with their totemic cult? Have the Arunta people had imposed upon them a social system which has not destroyed their primitive totemic cult?

To answer these questions I can only deal with the Semang evidence as it appears in researches of great authority and weight, and there is undoubtedly in all the evidence produced by Messrs. Skeat and Blagden, and the authorities they use, nothing whatever to suggest that Semang totemism once possessed above it an elaborate social organisation of the usual totemic type. There is indeed, the myth which points to a two-class exogamous division for marital purposes,[379] but there is more than myth for the unrestricted intercourse of the sexes both before and after marital rights.[380] In every other direction we get simple groups fashioned on no larger basis than nomadic roaming and journeying to fresh food grounds. On the other hand, there is much to suggest that the Arunta have a dual system of organisation; one, in which the primitive types are still surviving, the second, a more advanced type which covers but does not crush out the first. If this is so, it is clear that the parallel between Semang and Arunta totemism is considerably closer than at first appears.

It will be necessary, therefore, to deal with the two principal signs of alleged Arunta progress, male descent and the exogamous classes. I see no evidence whatever of male descent; male ascendancy, a very different thing, appears, but there cannot strictly be male descent where fatherhood is unrecognised. And here I would interpose the remark that the use of the term descent, male descent and female descent, in these studies is far too indiscriminate.[381] Descent means succession by blood kinship by acknowledged sons or daughters, and this is exactly what does not always occur. Sonship and daughtership in our sense of the term are not always known to savagery. They were not known to the Arunta males, for fatherhood was not recognised by them and motherhood was not definitely used in the social sense. All that the Arunta can be said to have developed is a mother-right society with male ascendancy in the group.[382] Group sons succeeded to group fathers, but individual descent from father to son there is not.

There remain the exogamous classes. In the first place, it is necessary to get rid of a difficulty raised by Mr. Lang. "In no tribe with female descent can a district have its local totem as among the Arunta.... This can only occur under male reckoning of descent."[383] But surely so acute an observer as Mr. Lang would see that with female descent right through, as it exists among the Khasia and Kocch people of Assam, local totem centres are just as possible as with male descent. Mr. Lang is conscious of some discrepancy here, for a little later on he repeats the statement that local totem centres "can only occur and exist under male reckoning of descent," but adds the significant qualification "in cases where the husbands do not go to the wives' region of abode."[384] This is the whole point. Where husbands do go to the wives' region of abode, as they do among the Khasis and the Kocch, female descent would allow of the formation of local totem centres. This is not far from the position of the Arunta. They are mother-right societies. The mother secures the totem name. The father, de facto, is not father according to the ideas of the Arunta people, is at best only one of a group of possible fathers according to the practices of the Arunta people. Therefore, the local totem centre is formed out of a system which may be called a mother-right system for the purpose of scientific description, but which is not even a mother-right system to the natives, because motherhood is not the foundation of the local group.

Secondly, we have the important fact, which Mr. Lang has duly noted, though he does not apparently see its significance in the argument as to origins, that the class system "arose in a given centre and was propagated by emigrants and was borrowed by distant tribes."[385] Messrs. Spencer and Gillen distinctly affirm that the "division into eight has been adopted (or rather the names for the four new divisions have been) in recent times by the Arunta tribe from the Ilpirra tribe which adjoins the former on the north, and the use of them is at the present time spreading southwards."[386] This view is supported by the widespread organisation of eaglehawk and crow, and by the general homogeneity of Australian social forms. It is clear, therefore, that room is made for the external organisation of the class system and the consequent production of the dual characteristics of the Arunta—the joint product of the fossilisation of mother-right society at the end of the migration movement, and the superimposing upon this fossilisation, with its tendency towards the class system, of the fully organised class system. The two systems are not now fully welded in the Arunta group. Whatever view is taken of these, whether they be considered advanced or primal, the undoubted dualism has to be accounted for, and the best way of accounting for this dualism is, I submit, that of differential evolution. Further study of Messrs. Spencer and Gillen's work, together with the criticisms of various scholars, Mr. Lang, Mr. Hartland, Mr. Frazer, Mr. Thomas, and others, convinces me that the extreme artificiality of the class system is due partly to a want of understanding of the entire facts, and partly to the ad hoc adoption by the natives themselves of new plans to meet difficulties which must arise out of a too close adhesion to their rules. Mr. Lang has allowed me to see a manuscript note of his, in which he points out that the inevitable result of the one totem to the one totem rule of marital relationship,—that is, totem A always intermarrying with totem B, males and females from both totems, and with no others,—is the consanguineous relationship of all the members of the two totems. The rule for non-consanguineous marriage has therefore broken down, and when it breaks down the Australian introduces a new rule which satisfies immediate necessities. When this in turn breaks down a further new rule is made, and this is the way I think the differing rules resulted. They represent, therefore, not varying degrees of culture progress, but only varying degrees of artificial social changes, and they spring from the oldest conditions of all where there is no class system at all.[387] Arunta society is not a "sport" under this view, but a product—a product to be accounted for and explained by anthropological rules, derived not only from Australian society but from the general facts of human society which have remained for observation by the science of to-day. The parallel between Semang and Arunta, therefore, helps us in two ways. It enables us to go back to Semang totemism as an example of primitive kinless society, and forward to Arunta totemism as an example of early development therefrom. We have, in point of fact, discovered the datum line of totemism. Upon this may be constructed the various examples according to their degrees of development, and we may thus see in detail the commencing elements of totemism as well as the means by which we may proceed from the commencing elements to the more advanced elements, and finally to the last stages of totemic society where blood kinship is fully recognised and used, where, in fact, totemic tribes as distinct from totemic peoples take their place in the world's history.

IV

I do not propose in this chapter to proceed further with this inquiry. It will not advance my object, nor is it absolutely necessary. Totemism in the full has been described adequately by Mr. Frazer in his valuable abstract of the evidence supplied from all parts of the world, and there is not much in dispute among the authorities when once the stage of origin is passed. There is danger, however, at the other extreme, namely, the attempt to discover totemism in impossible places in civilisation. Mr. Morgan has shown us totemic society in its highest form of development, untouched by other influences of sufficient consequence to divert its natural evolution. This, I think, is the merit of Mr. Morgan's great work, and not his attempt, his futile attempt as I think, to apply the principles of totemic society to the elucidation of societies that have long passed the stage of totemism. In particular, the great European civilisations are not totemic, nor are they to be seen passing from totemism. It is true that Mr. Lang, Mr. Grant Allen, and others have attempted to trace in certain features of Greek ritual and belief, and in certain tribal formations discoverable in Anglo-Saxon Britain, the relics of a living totemism in the civilised races of Europe;[388] but I do not believe either of these scholars would have endorsed his early conclusions in later studies. Mr. Grant Allen did not, so far as I know, repeat this theory after its first publication, and Mr. Lang has given many signs of being willing to withdraw it. The fact is, there is no necessity to think of Greek or English totem society because in Greece and England there are traces of totem beliefs. We may disengage them from their national position and put them back to the position they occupied before the coming of Greek or Englishman into the countries they have made their own.

In that position there may well have been totemic peoples in Britain of the type we have been considering from Australia. I have already indicated that totemic survivals in folklore have been the subject of a special study of my own which still in the main stands good, and for which I have collected very many additional illustrations and proofs. I discovered that folklore contained some remarkably perfect examples of totemic belief and custom, and also a considerable array of scattered belief and custom connected with animals and plants which, unclassified, seemed to lead to no definite stage of culture history, yet when classified, undoubtedly led to totemism. The result was somewhat remarkable. At many points there are direct parallels to savage totemism, and the whole associated group of customs received adequate explanation only on the theory that it represented the detritus of a once existing totemic system of belief.

The present study enables me to take the parallel to primitive totemism much closer. One of the perfect examples was of a local character. This was found in Ossory. Giraldus Cambrensis tells an extraordinary legend to the following effect: "A priest benighted in a wood on the borders of Meath was confronted by a wolf, who after some preliminary explanations gave this account of himself: There are two of us, a man and a woman, natives of Ossory, who through the curse of one Natalis, saint and abbot, are compelled every seven years to put off the human form and depart from the dwellings of men. Quitting entirely the human form, we assume that of wolves. At the end of the seven years, if they chance to survive, two others being substituted in their places, they return to their country and their former shape."[389] Here is a saintly legend introduced to explain the current tradition of the men of Ossory, that they periodically turned into wolves. Fynes Moryson, in 1603, ridiculed the beliefs of "some Irish who will be believed as men of credit," that men in Ossory were "yearly turned into wolves."[390] But an ancient Irish MS. puts the matter much more clearly in the statement that the "descendants of the wolf are in Ossory,"[391] while the evidence of Spenser and Camden explains the popular beliefs upon even more exact lines. Spenser says "that some of the Irish doe use to make the wolf their gossip;"[392] and Camden adds that they term them "Chari Christi, praying for them and wishing them well, and having contracted this intimacy, professed to have no fear from their four-footed allies." Fynes Moryson expressly mentions the popular dislike to killing wolves, and they were not extirpated until the eighteenth century.[393] Aubrey adds that "in Ireland they value the fang-tooth of an wolfe, which they set in silver and gold as we doe ye Coralls;"[394] and Camden notes the similar use of a bit of wolf's skin.[395]

In the local superstitions of Ossory, therefore, we have several of the cardinal features of savage totemism, the descent from the totem-animal, the ascription to the totem of a sacred character, the belief in its protection, and a taboo against killing it. I will venture to suggest, however, that to these important features there is to be added a parallel in survival to the Semang and Arunta features where the local circumstances of birth are the determining forces which supply the totem name, for the relationship of "gossip," "god-sib," is clearly of the same character as that of the soul-tree of the Semang and the alcheringa of the Australian.[396] The condition of survival has altered the detail of the parallel, but the parallel is on the same plane.

The wolf as gossip to the men of Ossory leads us on to inquire whether any other animal had such close connections with human beings. In Erris, a part of Connaught, "the people consider that foxes perfectly understand human language, that they can be propitiated by kindness, and even moved by flattery. They not only make mittens for Reynard's feet to keep him warm in winter, and deposit these articles carefully near their holes, but they make them sponsors for their children, supposing that under the close and long-established relationship of Gossipred they will be induced to befriend them."[397] Thus it appears that the selfsame conception which the men of Ossory had in the thirteenth century for the wolf, the men of Erris had for the fox in the nineteenth century. No explanation from the dry details of the natural history of these animals is sufficient to account for this curious parallel, and we must turn to ancient beliefs for the explanation.

The general attitude of the men of Erris towards the fox is confirmed as an attribute of totemism when we come to examine a special local form of it. This we can do by turning to Galway. The Claddagh fishermen in Galway would not go out to fish if they saw a fox: their rivals of a neighbouring village, not believing in the fox, do all they can to introduce a fox into the Claddagh village.[398] These people are peculiar in many respects, and are distinctively clannish. They retain their old clan-dress—blue cloaks and red petticoats—which distinguishes them from the rest of the county of Galway, and it may be conjectured that the present-day custom of naming from the names of fish—thus, Jack the hake, Bill the cod, Joe the eel, Pat the trout, Mat the turbot, etc.[399]—may be a remnant of the mental attitude of the folk towards that belief in kinship between men and animals which is at the basis of totemism. But, returning to the fox, we have in the belief that meeting this animal would prevent them from going out to fish, a parallel to the prohibition against looking at the totem which is to be found among savage people, and we have in the neighbours' disbelief in the fox and a corresponding belief in the hare,[400] that local distribution of different totems which is also found in savagery. But all these particulars about the relationship of the fox to the Claddagh fishermen receive unexpected light when we inquire into the biography of their local saint, named MacDara. This saint is the patron saint of the fishermen who, when passing MacDara's island, always dip their sails thrice to avoid being shipwrecked. But then, in the folk-belief, we have this remarkable fact, that MacDara's real name was Sinach, a fox[401]—an instance, it would seem, of a totem cult being transferred to a Christian saint. Thus, then, in the superstitions of these Claddagh fisherfolk we can trace the elements of totemism, the root of which is contained, first, in the nominal worship of a Christian saint, and second, in the actual worship of an animal, the fox.

These examples of local totemism may be followed by a remarkable example of tribal or kinship totemism. It was noted by Mr. G. H. Kinahan in his researches for Irish folklore, and is mentioned quite incidentally among other items, the collector himself not fully perceiving the importance of his "find." This really enhances the value of the evidence, because it destroys any possibility of an objection to its validity—a really important matter, considering the remarkable character of this survival of totem-stocks in Western Europe. The exact words of Mr. Kinahan are as follows:—

"In very ancient times some of the clan Coneely, one of the early septs of the county, were changed by 'art magick' into seals; since then no Coneely can kill a seal without afterwards having bad luck. Seals are called Coneelys, and on this account many of the name changed it to Connolly."[402] The same local tradition is mentioned by Hardiman in one of his notes to O'Flaherty's Description of West or H-iar Connaught,[403] but the note is equally significant of genuineness from the fact that the tradition is styled "a ridiculous story." It strengthens Mr. Kinahan's note in the following passage: "In some places the story has its believers, who would no more kill a seal, or eat of a slaughtered one, than they would of a human Coneely."

The clan Coneely is mentioned both by Mr. Kinahan and by Mr. Hardiman as one of the oldest Irish septs; and that it is widely spread, and not congregated into one locality, is to be inferred from the description of the tradition as prevalent in Connaught, especially from Mr. Hardiman's words, describing that "in some places" the story has its believers now; and hence we may conclude that wherever the clan Coneely are situated there would exist this totem belief.

The full significance of these facts may best be tested by reference to the conditions laid down by Dr. Robertson Smith for the discovery of the survivals of totemism among the Semitic races. These conditions are as follows:—

"'(1) The existence of stocks named after plants and animals'—such stocks, it is necessary to add, being scattered through many local tribes; (2) the prevalence of the conception that the members of the stock are of the blood of the eponym animal, or are sprung from a plant of the species chosen as totem; (3) the ascription to the totem of a sacred character which may result in its being regarded as the god of the stock, but at any rate makes it be regarded with veneration, so that, for example, a totem animal is not used as ordinary food. If we can find all these things together in the same tribe, the proof of totemism is complete; but even when this cannot be done, the proof may be morally complete if all the three marks of totemism are found well developed within the same race. In many cases, however, we can hardly expect to find all the marks of totemism in its primitive form; the totem, for example, may have become first an animal god, and then an anthropomorphic god, with animal attributes or associations merely."[404]

Now in the Irish case all three of these conditions are found together in the same tribe, the clan Coneely, and it is impossible to overlook the importance of such a discovery. It proves from survivals in folklore that totemistic people once lived in ancient Ireland, just as the corresponding evidence proved that the ancient Semitic stock possessed the totemic organisation.

We have now examined the most archaic forms of the survival of totemism in Britain. If we pass on to inquire whether we can detect the more scattered and decayed remnants of totem beliefs and customs, we turn to Mr. Frazer as our guide. From Mr. Frazer's review of the beliefs and customs incidental to the totemistic organisation of savage people, it is possible to extract a formula for ascertaining the classification of savage beliefs and practices incidental to totemism. This formula appears to me to properly fall into the following groups:—

(a) Descent from the totem.

(b) Restrictions against injuring the totem.

(c) Restrictions against using the totem for food.

(d) The petting and preservation of totems.

(e) The mourning for and burying of totems.

(f) Penalties for non-respect of totem.

(g) Assistance by the totem to his kin.

(h) Assumption of totem marks.

(i) Assumption of totem dress.

(j) Assumption of totem names.

My suggestion is that if a reasonable proportion of the superstitions and customs attaching to animals and plants, preserved to us as folklore, can be classified under these heads this is exactly what might be expected if the origin of such superstitions and customs is to be sought for in a primitive system of totemism which prevailed amongst the people once occupying these islands. The clan Coneely and the Ossory wolves are proofs that such a system existed, and if such perfect survivals have been able to descend to modern times, in spite of the influences of civilisation, there is no prima facie reason why the beliefs and customs incidental to such a system should not have survived, even though they are no longer to be identified with special clans. When once a primitive belief or custom becomes separated from its original surroundings, it would be liable to change. Thus, when the wolf totem of Ossory passes into a local cultus, we meet with the belief that human beings may be transformed into animal forms, as the derivative from the totem belief in descent from the wolf. Fortunately, the process by which this change took place is discernible in the Ossory example; but it will not be so in other examples, and we may therefore assume that the Ossory example represents the transitional form and apply it as a key to the origin of similar beliefs elsewhere.

Again, if we endeavour to discover how the associated totem-beliefs of the clan Coneely would appear in folklore supposing they had been scattered by the influences of civilisation, we can see that at the various places where members of the clan had resided for some time there would be preserved fragments of the once perfect totem-belief. Thus, one place would retain traditions about a fabulous animal who could change into human form; another place would preserve beliefs about its being unlucky to kill a seal (or some other animal specially connected with the locality); another place would preserve a superstitious regard for the seal (or some other local animal) as an augury; and thus the process of transference of beliefs into folklore, from one form into other related forms, from one particular object connected with the clan to several objects connected with the localities, would go on from time to time, until the difficulty of tracing the original of the scattered beliefs and customs would be well-nigh insurmountable without some key. But having once proved the existence of such examples as the clan Coneely and the Ossory wolves, this difficulty, though still great, is very much lessened. Our method would be as follows. We first of all postulate that totem peoples did actually exist in ancient Britain, or whence such extraordinary survivals? We next examine and classify the beliefs and customs which are incidental to totemism in savage society, and having set these forth by the aid of Mr. Frazer's admirable study on the subject, we ascertain what parallels to these beliefs and customs may be found in the folklore of Britain. And then our position seems to be very clearly defined. We prove that in folklore certain customs and superstitions are identical, or nearly so, with the beliefs and customs of totemism among savage tribes, and we conclude that this identity in form proves an identity in origin, and therefore that this section of folklore originated from the totemistic people of early Britain.

I shall not take up all these points on the present occasion, especially as they have in all essentials appeared in the study to which I have referred; but as an example of the scattering of totem beliefs I will refer to the well-known passage in Caesar (lib. v. cap. xii.), from which we learn that certain people in Britain were forbidden to eat the hare, the cock, or the goose, and see whether this does not receive its only explanation by reference to the totemic restriction against using the totem for food. Mr. Elton, with this passage in his mind, notices that "there were certain restrictions among the Britons and ancient Irish, by which particular nations or tribes were forbidden to kill or eat certain kinds of animals;" and he goes on to suggest that "it seems reasonable to connect the rule of abstaining from certain kinds of food with the superstitious belief that the tribes were descended from the animals from which their names and crests or badges were derived."[405]

Let us see whether this reasonable conjecture holds good. The most famous example is that of Cuchulainn, the celebrated Irish chieftain, whose name means the hound of Culain. It is said that he might not eat of the flesh of the dog, and he came by his death after transgressing this totemistic taboo. The words of the manuscript known as the Book of Leinster are singularly significant in their illustration of this view. "And one of the things that Cuchulainn was bound not to do was going to a cooking hearth and consuming the food [i.e. the dog]; and another of the things that he must not do was eating his namesake's flesh."[406] Diarmaid, whose name seems to be continued in the current popular Irish name for pig (Darby), was intimately associated with that animal, and his life depended on the life of the boar.[407] These examples are so much to the point that we may examine the cases mentioned by Caesar from the same standard.

Mr. Frazer points out that even among existing totem-tribes the respect for the totem has lessened or disappeared, and among the results of this he notes instances where, if any one kills his totem, he apologises to the animal. Under such an interpretation as this, we may surely classify a "memorandum" made by Bishop White-Kennett about the hare, the first of the British totems mentioned by Caesar: "When one keepes a hare alive and feedeth him till he have occasion to eat him, if he telles before he kills him that he will doe so, the hare will thereupon be found dead, having killed himself."[408] But respect for the hare, in accordance with totem ideas, was carried further than this at Biddenham, where, on the 22nd September, a little procession of villagers carried a white rabbit [a substitute for hare] decorated with scarlet ribbons through the village, singing a hymn in honour of St. Agatha. All the young unmarried women who chanced to meet the procession extended the first two fingers of the left hand pointing towards the rabbit, at the same time repeating the following doggerel:—

Gustin, Gustin, lacks a bier, Maidens, maidens, bury him here.[409]

This points to a very ancient custom, not yet fully explained, but which clearly had for its object the reverential burying of a rabbit or hare. It is characteristic of the totem animal that it serves as an omen to its clansmen, and we find that the hare is an omen in Britain. Boudicca is said to have drawn an augury from a hare, taken from her bosom, and which when released pursued a course that was deemed fortunate for her attack upon the Roman army;[410] and in modern south Northamptonshire the running of a hare along the street or mainway of a village portends fire to some house in the immediate vicinity.[411] In 1648 Sir Thomas Browne tells us that in his time there were few above three-score years that were not perplexed when a hare crossed their path.[412] In Wilts and in Scotland it was unlucky to meet a hare, but the evil influence did not extend after the next meal had been taken.[413] Then, too, the prohibition against naming the totem object is found in north-east Scotland attached to the hare, whose name may not be pronounced at sea, and Mr. Gregor adds the significant fact that some animal names and certain family names were never pronounced by the inhabitants of some of the villages, each village having an aversion to one or more of the words.[414] A classification of the beliefs and customs connected with the hare takes us, indeed, to almost every phase of totemistic belief, and it is impossible to reject such a mass of cumulative evidence.

Of the second of the British food taboos mentioned by Caesar we have the most perfect illustration in the instance of the Irish chieftain, Conaire, who, descended from a fowl, was interdicted from eating its flesh.[415]

Turning next to the goose, we find that at Great Crosby, in Lancashire, there is held an annual festival which is called the "Goose Fair," and although it is accompanied by great feasting, the singular fact remains that the goose itself, in whose honour the feast seems to have been held, is considered too sacred to eat, and is never touched by the villagers.[416] In Scotland also the goose was never eaten, being too sacred for food.[417]

Thus the hare, the fowl, and the goose have retained their sacred character in a special manner in various parts of the country, and I may add a further note of more general significance. In Scotland there exists a prejudice against eating hares and cocks and hens.[418] In the south-western parts of England the peasant would not eat hares, rabbits, wild-fowl, or poultry, and when asked whence this dislike proceeds, he asserts that it was derived from his father[419]—the traditional sanction which is so essential to folklore.[420]

The ideas surrounding these three special animals might be easily extended to others, but I will only observe that Mr. Elton, noting both the classical and modern accounts of certain districts in Scotland and Ireland where fish, though abundant, is tabooed as food, quotes with approval a modern suggestion that this abstinence was a religious observance.[421] That fish are carved on numerous stones is a curious commentary on this assertion, while another point to be noted is that the inhabitants of the various islands have each their peculiar notions as to what fish are good for food. Some will eat skate, some dog-fish, some eat limpets and razor-fish, and as a matter of course, says Miss Gordon Cumming, those who do not, despise those who do.[422] A prejudice also existed against white cows in Scotland, and Dalyell ventures upon the acute supposition that this was on account of the unlawfulness of consuming the product of a consecrated animal.[423] These are not stray notes of inexperienced observers, and with two centuries between them it must be that they contain the essence of the people's conception—a conception which leads us back to totemism for its explanation.

I do not think we could get closer to totemic beliefs and ideas than this, nor could we have a better example of the necessity of examining early historical data by anthropological tests and by folklore parallels. Caesar's words are unimportant by themselves. They convey nothing of any significance to the modern reader—a mere dietetic peculiarity which means nothing and counts for nothing. And yet it might be considered certain that Caesar knew that the details he recorded were of importance in the historical sense. He did not indicate what the importance was, probably because he was not aware of it; but because he was conscious that among the influences which counted with these people were the food taboos, he rightly recorded the facts. They have remained unconsidered trifles until now, when anthropology has brought them within the range of scientific observation, and they are now to be reckoned with as part of the material which tells of the culture conditions of a section of the early British peoples.

I must here interpose a remark with reference to this grouping of the evidence. Apart from the significance of the superstitions as they are recorded in their bare condition among the peasantry, there is the additional fact to note that the superstition against eating or killing certain animals or birds, or against looking at them or naming them, etc., is not universal. It obtains in one place and not in another. If the injunction not to kill, injure, or eat a certain animal were simply the reflection of a universal practice, such a practice might originate in some attribute of the animal itself which characteristically would produce or tend to produce superstition. But the spread of this class of superstition in certain districts, and not in others, is indicative of an ancient origin, and it is exactly what might be expected to have been produced from totem-peoples. Unfortunately, neither the negative evidence of superstitious beliefs nor the local distribution of superstitious beliefs has ever been considered worthy of attention. But some little evidence is incidentally forthcoming, and I would submit that this may be taken as indicative of what might be obtained more fully by further research into this neglected aspect of folklore. I drew Miss Burne's attention to this subject, and she has noted some particulars in her valuable Shropshire Folklore.[424] But for the most part this portion of our evidence wants picking out by a long and tedious process from the mass of badly recorded facts about popular superstitions. I do not believe in the generally stated opinion that certain superstitions are universally believed or practised. It is difficult to prove a negative, and such evidence is not absolutely scientific, but when it comes in direct antithesis to positive, there does not seem any harm in accepting it. Every class of superstition wants tracing out geographically, and local variants want careful noting. I cannot doubt if this were properly done that many so-called universal superstitions would be found to be distinctly local. In the meantime, it is not with universal superstitions that we have to deal. It is primarily with those local variants which show us side by side the differences of belief. It is thus that we can afford evidence of that intermixture of totem-objects which is to be expected from the known facts of totem-beliefs and customs. Indeed, Mr. McLennan has laid it down that "we might expect that while here and there perhaps a tribe might appear with a single animal god, as a general rule tribes and nations should have as many animal and vegetable gods as there were distinct stocks in the population ... we should not expect to find the same animal dominant in all quarters, or worshipped even everywhere within the same nation."[425]

It is important that we should thoroughly understand what these survivals of totemism in the British isles really mean. On the extreme west coast of Ireland, farthest away from the centres of civilisation, there are found these unique examples of a savage institution. The argument that they might have been transplanted thither by travellers from the far west, where totemism has developed to its highest form, cannot seriously be advanced. The argument that they might be the accidental form into which some merely superstitious fancies of ignorant peasants happened to have ultimately shaped themselves, is met by the mathematical demonstration that the ratio of chance against such a development would be well-nigh incalculable. The remaining argument is that they indicate the last outpost, or perhaps one of the last outposts, of a primitive savage organisation which once existed throughout these lands. This is the view that appears to me to be the only possible one to meet all the conditions of the case; one proof in support of this view being the discovery of evidence in other parts of the country which shows that totemism has left its stamp in more or less perfect form upon the traditional beliefs and practices of the nation. Though we are not able to identify further complete examples of the same type as the seal clan of Western Ireland, or the wolf people of Ossory, we should be able, if the explanation I have advanced of their origin be the correct one, to produce examples of the varying forms which such an institution as totemism must have assumed when it had been broken up by the advance of civilising influences. If the seal clan, or the wolf clan, is in truth the last outpost of a savage organisation, there will be in the lands less remote from the centres of civilisation some evidences of the break-up of savagery as it has been driven westward. Somewhere in tradition, somewhere in local observances of beliefs or superstition, there must still be echoes, more or less faint, but still echoes, from totemism. Having discovered these undoubted examples of totemism, the argument shifts its ground. We can no longer say that the theory of totemism may possibly explain some of the customs and traditions of the people. We are, by the logic of the position, compelled to say that custom and tradition must have preserved many relics of totemism, and that so far from seeking to explain custom and tradition by the theory of totemism, we must seek to explain the survival of totemism by custom and tradition. I lay stress on this view of the case because it is hard to combat the views of those who look upon "mere superstition" as no explanation of primitive originals. To us of the present day the beliefs of the peasantry are no doubt properly definable as "mere superstition." But when we examine it as folklore we are seeking for its origin, not for its modern aspect; we are asking how "mere superstition" first arose, and in what forms, not how it exists; we are pushing back the inquiry from to-day when it exists side by side with a philosophical and moral religion to the time when it existed as the sole substitute for philosophy and morals. Even if it is "mere superstition" it has a dateless history. It is not conceivable that it suddenly arose at a particular period before which "mere superstition" did not exist, and all, both peasant and chief, were philosophical and moral. It is not conceivable that the mere superstition of to-day has replaced bodily the mere superstition of other ages. Every succeeding age of progress has influenced it, no doubt, but not eradicated it, and hence the mere superstition of to-day has just such an unbroken continuity of history as language or institutions. That we are able to pick out from among its items undoubted forms of totemism, and that we may add to these complete examples a classified grouping of customs and beliefs in survival parallel to the customs and beliefs of savage totemism, affords proof that at least we may carry back that history to the era of totemism, at whatever point that era may cross the line of, or come into contact with, political history.

This is the definite conclusion to be drawn from the anthropological interpretation of the presence of totemic beliefs among the survivals of folklore. The study of the anthropological conditions has occupied a wide range of thought and inquiry, but it leads us back to a safe basis for research, for it brings definitely within touch of that realm of man which lies outside the civilisation wherein folklore is embedded, the peoples who have made, and the peoples who are dominated by, that civilisation. The savage of Britain cannot with this evidence before us be considered as the mere product of the literature of Greece and Rome. He is part and parcel of the savagery of the human race. Anthropology has shown us that savagery reached the land we now call Britain as part of the general movement of people which has caused the whole earth to become a dwelling-place for man, and now that we know this we must appeal to anthropology whenever we find that the problems of folklore take us out of the culture period of a civilisation known to history.[426]

APPENDIX

I append a synopsis of the culture-structure of the Semangs of the Malay Peninsula (references are to Skeat and Blagden's Pagan Races of the Malay Peninsula where not otherwise specified), in order that the position claimed for the one section of totemic belief may be tested by the remaining characteristics of Semang culture. I claim that there is nothing that remains which is inconsistent with the interpretation given of the totemic items.

Physical:—

(a). Live exclusively in the forest surrounded by hostile fauna (i. 13).

(b). Food consists of such wild vegetable food as may happen to fall from time to time in season (i. 109, 341, 525), together with small mammals and birds (i. 112), fish (i. 113).

(c). As soon as they have exhausted the sources of food in one neighbourhood they move on to the next (i. 109).

(d). Fire obtained by friction (i. 111, 113), but meat is eaten raw (i. 112).

(e). Nudity is alleged (Journ. Indian Archipelago, i. 252; ii. 258); no satisfactory proof (i. 137); do not use skins of animals nor feathers of birds (i. 138); a girdle of fungus string (i. 138, 142, 380); fringe of leaves suspended from a string (i. 139, 142); necklaces and ligatures of jungle fibre (i. 144, 145); women wear a comb made of bamboo as a charm against diseases (i. 149).

(f). Habitations are rock shelters (i. 173), tree shelters afforded by branches of trees improved by construction of a weather screen (i. 174); ground screen of palm leaves (i. 175).

(g). Hunt successfully the largest animals, escaping easily up the trees (i. 202-204).

(h). Knives made of bamboo, flakes and chips of stone, knives of bone (i. 249, 269); bow and arrow (i. 251, 255); not sufficiently advanced to have produced neolithic implements (i. 268); wooden spear (i. 270).

(i). Ignorant of pottery, vessels made from big stems of bamboo (i. 383).

Social:—

(j). Chief of the group is the principal medicine man, but is on an equal footing with his men, no caste and property is in common (i. 497, 499).

(k). Marriage rights are secured by the presentation of a jungle knife to the bride's parents and a girdle to the bride, and the bride never lets the girdle part from her for fear of its being used to her prejudice in some magic ceremony; adultery is punishable by death (ii. 58, 59) [but this information was not obtained from the most primitive of the Semang people].

(l). Semang women are common to all men (Newbold, Political and Stat. Acc. of Settlements in Straits of Malacca, ii. 379). Great ante-nuptial freedom (ii. 56, 218); "Of the Semang I have not had an opportunity of personally judging" (ii. 377, Newbold).



(m). Eat dead kindred except head (Newbold, ii. 379); burial takes place in the ground, and the older practice was exposure in trees; the Semang have no dread of ghosts of the deceased (ii. 89, 91).

(n). No sacred shrines or places (ii. 197).

(o). Avoidance of mother-in-law (ii. 204).

(p). Myth of the ringdove informing the children of the first woman that they had married within prohibited degrees of consanguinity, and advising them to separate and marry "other people" (ii. 218).

(q). Myth as to ignorance of cause of birth being dispelled by the cocoanut monkey informing the first man and woman (ii. 218).

(r). The Semang are almost ineradicably nomadic, have no fixed habitation, and rove about like the beasts of the forest (i. 172; ii. 470).

(s). Women and girls are not allowed to eat until the men and boys have finished their repast (i. 116); the men do most of the hunting and trapping, and the women take a large share in the collecting of roots and fruits; all the cooking is performed by the women and girls (i. 375).

(t). They are split up into a large number of dialects, each of which is confined to a relatively small area, and it often happens that a little [clan] or even a single family uses a form of speech which is differentiated from other dialects to be practically unintelligible to all except the members of the little community itself (ii. 379).

(u). Natural segregation of the [tribes] into small [clans] to some extent cut off from one another and surrounded by settled Malay communities (ii. 379).

(v). The most thoroughly wild and uncivilised members of our race, regarded by the Malays as little better than brute beasts, with no recorded history (ii. 384).

(w). Nomadic life of the Semang leads them over a considerable tract of country (ii. 388).

Psychical:—

(x). Decorative patterns on quivers representing natural objects, and possessing magical virtue to bring down various species of monkeys and apes and other small mammals (i. 417), and as charms for the men (i. 423).

(y). Decorative pattern on magic comb worn by women to serve as a charm against venomous reptiles and insects, similar design for similar reason sometimes painted on the breast (i. 41, 420-436).

(z). Child's name is taken from some tree which stands near the prospective birthplace of the child. As soon as the child is born this name is shouted aloud by the sage femme, who then hands over the child to another woman, who buries the afterbirth underneath the birth-tree or name-tree of the child. As soon as this is done the father cuts a series of notches in the tree, starting from the ground and terminating at the height of the breast. The cutting of these notches is intended to signalise the arrival on earth of a new human being, since it thus shows that Kari registers the souls that he has sent forth by notching the tree against which he leans. Trees thus "blazed" are never felled. The child must not in later life injure any tree which belongs to the species of his tree; for him all such trees are taboo, and he must not even eat their fruit, the only exception being when an expectant mother revisits her birth-tree. Every tree of its species is regarded as identical with the birth-tree (ii. 3, 4). When an East Semang dies his birth-tree dies too (ii. 5).

(aa). The child's soul is conveyed in a bird, which always inhabits a tree of the species to which the birth-tree belongs. It flies from one tree of the species to another, following the as yet unborn body. The souls of first-born children are always young birds newly hatched, the offspring of the bird which contained the soul of the mother. If the mother does not eat the soul-bird during her accouchement the child will be stillborn or will die shortly after birth (ii. 4, 192, 194, 216). She keeps the soul-bird within the birth-bamboo, and does not eat it all at once, but piecemeal (ii. 6). All human souls grow upon a soul-tree in the other world, whence they are fetched by a bird which was killed and eaten by the expectant mother (ii. 194).

(bb). Semang religion, in spite of its recognition of a thunder-god (Kari) and certain minor deities (so called), has very little indeed in the way of ceremonial, and appears to consist mainly of mythology and legend. It shows remarkably few traces of demon worship, very little fear of ghosts of the deceased, and still less of any sort of animistic beliefs (ii. 174). [As the Kari is the deity common to the Semang and the people higher in culture than the Semang, it is difficult to trace out the primitive idea. The myths also show a common impress, "which is probably mainly due to the same savage Malay element" (ii. 183).]

(cc). During a storm of thunder and lightning the Semang draw a few drops of blood from the region of the shin bone, mix it with a little water in a bamboo receptacle, and throw it up to the angry skies (ii. 204).

(dd). Pretend entire ignorance of a supreme being, but on pressure confessed to a very powerful yet benevolent being, the maker of the world (ii. 209).

FOOTNOTES:

[284] Beddoe, Races of Britain, cap. ii., and Journ. Anthrop. Inst., xxxv. 236-7; Boyd-Dawkins, Early Man in Britain, cap. vii. viii. and ix.; Ripley, Races of Europe, cap. xii.

[285] Rhys, Celtic Britain, 271; Rhys, Celtic Heathendom, passim; Rhys and Jones, Welsh People, cap. i. and Appendix B on "Pre-Aryan Syntax in Insular Celtic," by Professor Morris Jones.

[286] Barrows, mounds, tumuli, stone circles, monoliths are generally admitted to belong to the Stone Age people before the Celts arrived, and when they are adequately investigated, as Mr. Arthur Evans has investigated Stonehenge (Archaeological Review, vol. ii. pp. 312-330), and the Rollright Stones (Folklore, vol. vi. pp. 5-51), the evidence of a prehistoric origin is unquestioned.

[287] I have worked out the evidence for this in the Archaeological Review, vol. iii. pp. 217-242, 350-375, and though I do not endorse all I have written there, the main points are still, I think, good.

[288] Wallace, Darwinism, cap. xv.

[289] Spencer and Gillen, Central Tribes of Australia, 12, 272, 324, 368, 420.

[290] Descent of Man, i. cap. vii. 176.

[291] Cf. Topinard's Anthropology, part iii., "On the Origin of Man," pp. 515-535, for the details of the various authorities ranged on the sides of monogenists and polygenists.

[292] Keane, Man, Past and Present, discusses the important evidence obtained by Dr. Dubois from Java, and Dr. Noetling from Upper Burma, pp. 5-8. It is only fair to that brilliant scholar, Dr. Latham, to point out that without the evidence before him to prove the point, he came to the same conclusion that the original home of man was "somewhere in intra-tropical Asia, and that it was the single locality of a single pair."—Latham, Man and his Migrations, 248.

[293] The most recent example of this is Mr. Thomas's extraordinary treatment of the evidence of migration in Australia. It produces in his mind "novel conditions," but has effects which he cannot neglect, but which he strangely misinterprets. N. W. Thomas, Kinship Organisations in Australia, 27-28.

[294] Spencer, Principles of Sociology, i. 18.

[295] Lord Avebury, Prehistoric Times, 586.

[296] Man, Past and Present, pp. 1, 8.

[297] Latham, Man and his Migrations, 155-6.

[298] The ethnographic movement is a very definite fact in anthropological evidence, though it has been little noted. Thus "the Coles are evidently a good pioneering race, fond of new clearings and the luxuriant and easily raised crops of the virgin soil, and have constitutions that thrive on malaria, so it is perhaps in the best interest of humanity and cause of civilisation that they be kept moving by continued Aryan propulsion. Ever armed with bow, arrows, and pole-axe, they are prepared to do battle with the beasts of the forest, holding even the king of the forest, the 'Bun Rajah,' that is, the tiger, in little fear."—Col. Dalton in Journ. Asiatic Soc., Bengal, xxxiv. 9.

[299] Traditions of great migrations exist among most primitive races. Some of these contain unexpected corroboration from actual discoveries. Thus the natives of New Zealand had a tradition that their ancestors, when they arrived in their canoes some four centuries ago, buried some sacred things under a large tree. It is said that the tree was blown down in recent times and that the sacred things were discovered. Taplin records "a good specimen of the kind of migration which has taken place among the aborigines all over the continent" (The Narrinyeri, p. 4); and similar evidence could be produced in almost every direction. Mr. Mathew in Eaglehawk and Crow deals with "the argument from mythology and tradition" as to the origin of the Australians in a very suggestive fashion (pp. 14-22). Stanley has preserved an African native tradition of local groups spreading out from the parent home (Through the Dark Continent, i. 346).

[300] I am aware this is disputed by O. Peschel—Races of Man, 137 et seq.—but I think the evidence is sufficient; and it must be remembered that there is direct evidence of the most backward races not using the fire they possess for cooking, but always eating their animal food raw, as, for instance, the Semang people of the Malay Peninsula. (See Skeat and Blagden, Pagan Races of the Malay Peninsula, i. 112.) The Andaman Islanders could not make fire, though they possessed and kept it alive. This shows that they must have borrowed it and did not previously possess it.—Quatrefages, The Pygmies, 108. Tylor, Early History of Mankind, cap. ix., should be consulted.

[301] The term political is, I confess, a little awkward, owing to its specially modern use, but it is the only term which, in its early sense, expresses the stage of social development represented by a polity as distinct from a mere localisation.

[302] It was one of the first efforts of the science of language to endeavour to trace out the original home of the so-called Aryas and their subsequent migrations. "Emigration," said Bunsen, "is the great agent in forming nations and languages" (Philosophy of Hist., i. 56); and Niebuhr, who has traced out most of the migrations of the Greek tribes, observes that "this migration of nations was formerly not mentioned anywhere" (Anc. Hist., ii. 212). Quite recently, Professor Flinders Petrie has worked at the question of European migrations in the Huxley lecture of 1907 (Journ. Anthrop. Inst., xxxvi. 189-232), his valuable maps showing "the movements of twenty of the principal peoples that entered Europe during the centuries of great movements that are best known to us" (204). In the meantime, the folklorist has much to do in this direction, and up to the present he has almost entirely ignored or misread the evidence. I do not know whether Mr. Nutt would still adhere to his conclusion that the myth embodied in the Celtic expulsion-and-return formula is undoubtedly solar (Folklore Record, iv. 42), but a restatement of Mr. Nutt's careful and elaborate analysis would lead me to trace the myth to the migration period of Aryan history, just as I agree with von Ihering that the ver sacrum of the Romans is a rite continued from the migration period to express in religious formulae, and on emergency to again carry out, the ancient practice of sending forth from an overstocked centre sufficient of the tribesmen and tribeswomen to leave those who remained economically well-conditioned (The Evolution of the Aryan, 249-290). Pheidon's law at Corinth, alluded to by Aristotle (Pol., ii. cap. vi.), could only be carried out by a sending out of the surplus. See also Aristotle, Pol., ii. cap. xii.; and Newman's note to the first reference, quoting similar laws elsewhere. Both the "junior-right" traditions and customs take us back to the same conditions. The occupation of fresh territories is an observable feature of the Russian mir (Wallace, Russia, i. 255; Laveleye Primitive Property, 34), and Mr. Chadwick has recently called attention to the corresponding Scandinavian evidence (Origin of the English Nation, 334).

[303] Mr. J. R. Logan long ago pointed out that "the further we go back, we find ethnic characteristics more uniform," and further concluded that certain facts observed by himself "lead to the inference that the Archaic world was connected."—Journ. Indian Archipelago, iv. 290, 291.

[304] Descent of Man, pp. 590, 591.

[305] Studies in Ancient History, i. 84.

[306] History of Human Marriage, cap. ii.

[307] Ancient Society, p. 10.

[308] Secret of the Totem, p. 32.

[309] N. W. Thomas, Kinship Organisation in Australia, 4.

[310] Folklore, xii. 232.

[311] Both Dr. Haddon and myself made the same point on a criticism of Mr. Fraser's Golden Bough, mine being from the Aricia rites, and Dr. Haddon's from the savage parallels thereto. See Folklore, xii. 223, 224, 232.

[312] Sproat's Scenes and Studies of Savage Life, 19. The use of the term "tribe" in this quotation is, of course, descriptive only. There is no tribal constitution among the Ahts, and "group" would have been the preferable term.

[313] Dr. W. H. Rivers' recently published work on the Todas is the best authority.

[314] Rivers, op. cit., 432, 455.

[315] Rivers, op. cit., cap. xxi. 504, 517.

[316] Rivers, op. cit., 452-456.

[317] Latham, Descriptive Ethnology, ii, 137.

[318] Bucher, Industrial Evolution, 56.

[319] Rev. George Taplin, The Narrinyeri; South Australian Aborigines, 40. Cf. Howitt, Native Tribes of South-east Australia, 710-720; Grierson, The Silent Trade, 22.

[320] Cf. Skeat and Blagden, Pagan Tribes of Malay Peninsula, i, 10.

[321] Graham, Bheel Tribes of Khandesh, 3.

[322] Herodotos, iv. 180.

[323] Journ. Asiatic Soc., Bengal, xiii. 625.

[324] Major Gurdon, The Khasis, 76, 82.

[325] N. W. Thomas, Kinship Organisations in Australia, 124.

[326] Fustel de Coulange's Cite Antique, cap. xiv. and xv., is, however, the most exaggerated example of this point of view.

[327] Lang, Social Origins, 1. The latest exponent of anthropological principles affirms that "the family which exists in the lower stages of culture, though it is overshadowed by the other social phenomena, has persisted through all the manifold revolutions of society."—N. W. Thomas, Kinship Organisations in Australia, 1.

[328] Jevons' Introd. to Hist. of Religion, 195.

[329] See also Prof. Geikie in Scottish Geographical Mag. (Sept. 1897).

[330] Early Hist. of Mankind, 303; MacCulloch, Childhood of Fiction, 396; Gould, Mythical Monsters.

[331] Mr. Westermarck has collected excellent evidence as to the economic influences upon savage society (Hist. of Human Marriage, 39-49), and we may quite properly assume the same conditions for earliest man.

[332] A very good summary of the pygmy peoples in all parts of the world is given by Mr. W. A. Reed in his useful Negritos of Zambales, 13-22. Cf. Keane, Man, Past and Present, 118-121; Keane, Ethnology, 246-248; and Sir W. H. Flower, Essays on Museums, cap. xix.

[333] Latham, Man and his Migrations, 55, 56. Dr. Beke was a most cautious observer, and I have consulted all his contributions to the Journal of the Geographical Society (vol. xiii.) and have found no sign of his retraction of the evidence. His correspondence in the Literary Gazette of 1843, p. 852, discusses the question of the Dokos being pygmies, but he adheres to his information as to the absence of social structure being correct.

[334] Lib. ii. 32, 8; cf. Quatrefages, The Pygmies, cap. 1, "The Pygmies of the Ancients."

[335] Lieut.-Col. Sutherland, Memoir respecting the Kaffirs, Hottentots, and Bosjemans, i. 67 (Cape Town, 1846).

[336] Burrows, The Land of Pygmies, 182.

[337] Mr. A. B. Lloyd's volume In Dwarfland and Cannibal Country, p. 96, is the most recent evidence.

[338] It is worth noting here that the Chinese traditions of the pygmies are exceedingly suggestive and curious. See Moseley, Notes by a Naturalist, 369.

[339] Skeat and Blagden, Malay Peninsula, ii. 443.

[340] Journ. Indian Archipelago, iv. 425-427; cf. Journ. Anthrop. Inst., xvi. 228; Wallace, Malay Archipelago, 452.

[341] Clifford, In Court and Kampong, 171-181.

[342] Skeat and Blagden, Pagan Races of Malay Peninsula, i. 13.

[343] Op. cit., i. 53-4, 139, 169, 172, 341.

[344] Op. cit., i. 170.

[345] Op. cit., i. 243-248, 268.

[346] Op. cit., i. 494; ii. 56, 218.

[347] Op. cit., ii. 3. Compare Journ. Indian Archipelago, iv. 427, "they are called after particular trees, that is, if a child is born under or near a cocoanut or durian, or any particular tree in the forest, it is named accordingly," and John Anderson, Considerations relative to Malayan Peninsula, 1824, p. xli.

[348] Op. cit., ii. 4, 192, 194.

[349] Op. cit., ii. 174, 209.

[350] Archaeological Review, i. 13, from an official report published in a Government Blue Book.

[351] Brinton, The American Race; Curtin, Creation Myths of Primitive America.