armour of a Viking invader, and I believe that relics of a battle between Vikings and natives have been found on the spot. But Mr. MacInnes was no archaeologist, though the armour of the Dhuine Mor was correctly described by him from careful study of the ghost. In this case tradition was not the source of the knowledge of the percipient. There were two percipients, but the other was alarmed and showed no scientific curiosity.

Mr. Gomme calls on psychical research to hand supranormal phenomena over to folklore, but I cannot see that folklorists make anything of them. Folklore does not cross-examine the witnesses and compare adjacent evidence tending towards proof or disproof of the phenomena. Yet this appears to be the only scientific method of dealing with such things.

In other cases where märchen contain vestiges of institutions, the fact has not been "ignored"; thirty years ago I was busy in tracing these vestiges, and the knowledge has been vulgarisé in most treatises on folklore. For example I gave the Frog Prince story, and so does Mr. Gomme, but he says "Frog Prince=totem "It mout be so, or it moutn't," quoth Uncle Remus. Again, a prince coming from a foreign country to win a bride, makes "an exogamous marriage." Now in Australia you make an exogamous marriage with a neighbour in your own tribal territory (save in one or two cases as in that of the Kurnai, where you go to a remote part of tribal territory), but it was not a case of exogamy when his Majesty married a "Sea-King's daughter from over the sea." We must not be in such a hurry to find exogamy! Nausicaa wanted to marry a foreign prince, the Ithacan, but it was legal, and desirable, that she should wed a Phaeacian of her own island.

I really do not quite see what novelty Mr. Gomme thinks he is introducing. If he makes exogamy and foreign marriage coextensive, or thinks that a case of foreign marriage or a dozen cases, are necessarily due to the exogamous prohibition, then the information is rather novel than convincing.

When Mr. Gomme finds fault with historians for demurring to Greek and Roman accounts of the low savage estate of the natives of these islands in the "La Tène" or "Late Celtic" period, does he mean that our Graeco-Roman books are folklore? He says "the terms 'savage' and 'barbarism' indulged in by the Greek and

Roman writers, cannot be rejected by modern historians simply because they are too harsh." "Barbarism" is not harsh; and "savage" must be rejected because archaeology proves beyond possibility of refutal, that, in England, Ireland, and Scotland, the people of Caesar's time were not savages.

The material civilisation and art of the La Tène iron age of late Celtic ornament, arts, and crafts, were no more savage than the art and civilisation of Late Minoan Greece. As to the alleged "community of women," if it existed, the fact was quite out of harmony with the material culture, and quite out of harmony with the traditional evidence (which Mr. Gomme should respect), of the Tain Bo Cualgne, though Queen Maive "wur a bad 'un,— she," as the Northern Farmer says. Mr. Gomme (p. 116) plucks a crow with me about all this, citing my History of Scotland, vol. i. pp. 3-5. I have read vol. i. pp. 3-5, and to these pages I refer the curious inquirer. As it happens, I did not say (or "declare roundly"), that "to found theories upon such evidence as archaeology provides is the province of another science, not of history." I said, "to discuss the race and language of the tribes who incised on the rocks the universal hieroglyphs of early man; who used the polished neolithic weapons; to found theories on the shapes of skulls unearthed from barrows, is the province of another science, not of history." History is not craniology, and history is not philology bombinans in vacuo, studying a language of which we have not a single word on record.

The language of the people who inscribed Scottish rocks with Chiriqui and Arunta decorative designs is not known to us. We cannot prove that they were men of Celtic speech, and, if they were not, they have left no linguistic traces except perhaps in some mysterious monosyllabic river names such as the Spanish Ter and the English Ver and Cher. How can history criticise such names, or read the Cretan inscriptions, not in Greek, but written in Greek characters?

History is primarily concerned with contemporary documents. When these do not exist, she strains her eyes with the aid of old books of history, and gratefully accepts what help she can receive from philology, archaeology, anthropology, folklore, sphragistic

heraldry, and so forth. But what Caesar says is evidence (not folklore evidence), and, if he were well informed, then the social conditions of the people of these islands, in his time, were much out of harmony with their degree of material culture. I do not discuss Mr. Gomme's theory of kinlessness and totemism among the Semangs, because Mr. Skeat, as he kindly informs me, could find no certain vestige of totemism in that people, and his book proves that they recognise kinship. I do not follow Mr. Gomme into the relations of race and folklore, because about race we know, at least I know, very little; and our own, and Greek, and ancient Egyptian folklore contain many elements common to Arunta or Dieri folklore. Folklore seems, on the whole, universally human. I do not quite understand what Mr. Gomme means when he says that "the Teutonic people, and their Celtic predecessors, came to Britain with a tribal, not an agricultural constitution." Does he mean that these tribes had no agriculture in their old homes? That would be hard to prove, and if they had agriculture it must have been under rules, "an agricultural constitution." Is there any reason why a tribal people should not occupy villages? The word "tribal" is very vague, but the tribes of the north-west coast of America live in villages, and the Celts and Teutons may have done so too. The people of Attica were tribal, living in villages, before the synoecismus. I really do not see that the fact can only be explained by ethnological differences, and perhaps Mr. Gomme does not mean that. All my observations must be taken with the reserve that I may have failed to interpret his meaning with precision.

A. LANG.

THE CAMBRIDGE HISTORY OF ENGLISH LITERATURE.

Vol. i.

FROM THE BEGINNINGS TO THE CYCLES OF ROMANCE.
Cambridge University Press, 1907.

'THE history of a nation's literature cannot be divorced from some consideration of its political, religious, and social life, including its manners as well as its phases of sentiment and fashion, its trivial thoughts no less than its serious moments.' I read this sentence in the preface of the new History of English Literature, and, so far as it goes, I find the principle therein enunciated good. True, as a folk-lorist, I should have liked some recognition of those obscurer, more instinctive, more primitive manifestations of the racial or, as I should prefer to say (the term racial being prejudiced by its pseudo-scientific associations), national, consciousness which are in so large a measure the objects of our study. Still here the principle is asserted, that literature must be regarded, not as a fortuitous succession of individual manifestations of talent unrelated to each other and to the life out of which they spring but as the revelation of the intellectual, aesthetic and spiritual ethos of a community, which ethos has been in the past and is in the present dependent upon conditions which we can analyse and determine. As a folk-lorist my concern with the present volume is chiefly to consider in how far it has carried out this principle, and I necessarily disregard nearly all that most reviewers, dealing with it from the standpoint of literary history as commonly understood, make the subject of their comments. should, however, like to express my appreciation of the utility of the work; it contains a great deal of information accurate in itself and lucidly presented; its merits in this respect are such as to entitle it to a place on the shelves of all students of our early national past; a reprint will doubtless be called for before long and in anticipation thereof a few remarks upon its plan, and upon the way in which that plan has been carried out, may be thought not out of place.

The plan is that of the Cambridge Modern History: special sections of the subject are dealt with by different writers, the various contributions being dovetailed into and more or less harmonised

with each other by the editor, who provides birds'-eye surveys of the main periods and tendencies. I cannot but think the plan a mistaken one. It necessarily leads to overlapping and duplication, serious drawbacks when the literature of well-nigh a thousand years has to be surveyed in 400 pages. A flagrant instance is supplied by chapters ix. and x: Latin Chroniclers from the eleventh to the thirteenth centuries, by Prof. Jones; Latin Literature of England from John of Salisbury to Richard of Bury, by Dr. Sanday. It would have been far better to allow Prof. Jones to treat the whole of this literature, upon which, despite the variations of subject matter, the form and historic antecedents of the language impress a character that distinguishes it from work in the vernacular, whether French or English. The worst fault is undoubtedly the assignation of the Metrical Romances (1200-1500) to two writers, Prof. Ker and Mr. J. W. H. Atkins. It is of less consequence that the two not infrequently differ as that on the whole they go over the same ground without either subjecting it to that exhaustive and penetrating survey we have a right to expect. Thus despite the really excellent 'linking and harmonising' chapters of the editor, Mr. Waller, the reader obtains no clear broad impressions; he must himself in large measure supply his own synthesis, and he is handicapped by the fact that the materials therefor reach him in vertical instead of horizontal sections and thereby hinder a clear insight into the evolution of the literature as a whole.

The capital error of the book remains to be noticed; it is one which defies the excellent principle I quoted at the outset of this review, and one of which the disastrous nature is especially apparent to the folklorist reader. An attempt has been made to treat in one volume two markedly distinct periods in the growth of our national literature, and thereby facts, which were to exercise a most potent influence upon the whole subsequent development of English letters, are slurred over or distorted. In a period of 350 years, 700-1050 in round figures, English literature exhibits a development which it is comparatively easy to trace, characteristics which are comparatively simple and homogeneous, a formal body and an animating spirit which are obviously the product and expression of a comparatively harmonious and genuinely national