come very near to what I regard as the solution, I do not think that either can be accepted. Whatever theory we form must be inclusive, and this canon seems absolutely fatal to the whole of these suggestions as they stand. Account for all these remains, or you account for none. The notion of a processional path, for example, whatever form it takes, could only be accepted if all the rows were double and definitely parallel. To say that there are such things as avenues of approach to structures elsewhere is in itself no argument, but simply a begging of the question. Where there is but one row there can be no path; where there are more there are too many paths-labyrinths, in short. The accident that attention was first called to Avebury and Merivale is entirely responsible for this avenue or processional idea.

The suggestion that the rows were race-courses is, as we have seen, open to all the "processional path" objections, with the addition of its own inherent absurdity.

The serpent hypothesis was admittedly evolved from Dr. Stukeley's "inner consciousness," and has few backers now.

Mr. Fergusson's twin proposals are mutually destructive. The same thing is made to represent two wholly different ideas. One might be true, but not both; and I do not know of a single instance in which these rows are not, or may not have been, associated with circles or tumuli. Certainly that is the case with those at Merivale. And the battle theory is wholly inconsistent with the frequency of single rows. The "army" must have been drawn up on parade merely in such a case-clearly there was no enemy.

Mr. Spence Bate threw out the idea that the long row in the valley of the Erme was intended to guide the inhabitants from the "sacred circle" to the village where they mostly dwelt, in foggy weather. But his "sacred circle" is clearly sepulchral, and the village is a kistvaen!

Mr. Bate's general view, however, was at one time that the "avenues" were sepulchral, though later on he adopted Mr. Fergusson's battle-plan hypothesis. In 1871 his conclusion. was that a kistvaen was generally connected with the rows; while there was always a cairn detached and not very distant, and he proceeded :—

"I take it that the kistvaen within the circle held the remains of the honoured dead, their priests, their counsellors, their successful warriors; that whenever they opened the kist to receive their ashes, they planted a single stone or pair of stones, according to the custom of the tribe, to commemorate. . . . The larger stones,

though not invariably, stand near the circle. I have thought that the important stones may commemorate the deaths of the first interred or founders of the tribe, and that the others are evidence of the merits of the individual. The length of the avenue, therefore, records the number of individuals interred, and to a certain extent the duration of the tribe. The large cairn that stands apart is the burial place of the many. Their bodies were buried, and with each interment each mourner added his portion of stones to the common heap." 4

Unfortunately for Mr. Bate's suggestion, there is no evidence that a kistvaen ever contained the remains of more than one body—it is only a stone coffin; and all analogy is against the assumption. Nor is it possible that the rows could have been customarily erected piecemeal-the indications of structural design are too clear. The second suggestion, touching the cairns, has this much in its favour, that we have ample proof of joint burial in such stone heaps and barrows elsewhere; and clearly the people so buried would be of less importance than those for whom the kistvaen was formed and the circle reared, to leave for the moment the rows out of the question.

I do not feel certain, however, that all these cairns were burial places. One under Shell Top, to which Mr. Spence Bate referred as a chambered cairn, is neither more nor less than a hut cluster-a group of chambers formed for residence in a pile of stones, with a long common passage; surely could not have stood alone.

GENERAL FEATURES OF THE ROWS.

and it

In the absence then of any accepted and satisfactory solu tion of our problem, we proceed to inquire what are the essential features and common characteristics of the "rows" We find:

A. That the universal principle of these monuments is the existence of a row of isolated stones, arranged with some approximation to regularity of position and commonly of

size.

:

B. That the number of rows associated is not essentialthere may be one, two, three, or any number up to three and twenty the variation not being in style or purpose, but apparently either in the direction of the importance of the structure; or signifying additional interments-on which latter head the fact that there are several groups of these monu• Prehistoric Antiquities of Dartmoor. Op. cit.

ments on Dartmoor, and that some have certainly been the subject of additions, is noteworthy.

C. That the bulk of the stones is also a non-essential. They are of all sizes from twenty or more feet above the ground to a foot, and even less. The same desire for appearance commonly shown in the regularity of the spacing is, however, commonly though not invariably indicated in the proximity of those approaching the same proportions. The best seem, as a rule, to have been used first.

D. That the spacing is non-essential, since on Dartmoor alone it ranges generally between two and six feet, though fairly regular in the individual examples.

E. That while commonly or approximately straight, the rows occasionally branch; and while parallel when double, when in more numerous groups frequently diverge, as at Caithness, or converge, as at Carnac. These variations are therefore non-essentials.

at

F. The direction of the rows is also a non-essential, but one which seems to have a meaning, seeing that something like system is apparent. On Dartmoor they have a more general easterly and westerly bearing in the southern quarter, and a more general northerly and southerly in the northern, though with exceptions in both. Still the distinct agreements can hardly be accidental. The Kennet "avenue Avebury runs nearly north-west and south-east; those at Stanton Drew east and west; the line at Aylesford northeast and south-west; the Ashdown row north and south; the Caithness ditto; the Carnac east and west. The point to be noted here is that agreeing directions are, as a rule, characteristic of special localities-the bearings were clearly not taken haphazard, but with a definite object, whatever that object may have been.

G. We now come to an essential of the first importance. The rows of Dartmoor (and, so far as I am aware, those elsewhere) are invariably connected or associated with stone circles, kistvaens, barrows, cromlechs, cairns, and other indications of sepulture.

H. Another general rule is that Dartmoor rows begin with a circle or barrow on higher ground and then trend downward. (Coryndon Ball is the only example I know that follows along a hillside.) The point of the compass for the commencement seems immaterial. Thus with the east and west rows of the Plym, Meavy, and Walkham valleys, the head of the row is on the east or west, according to the level. The circle of the great row in the Erme valley is the highest point

at that end, though the kistvaen at the other is higher still. Here, however, we have a double slope to the river, and such a result was inevitable.

Deduction G clearly establishes the character of the rows. If they are always associated with definite indications of sepulture they must in some way be sepulchral likewise. It has, indeed, been customary to call the larger stone circles with which they are not infrequently connected "sacred circles"-simply, I take it, because they are specially important. Certainly not because there is any authority for such assignation, or any evidence of such purpose. But Dartmoor alone supplies instances of undoubted sepulchral circles of considerable size. For example, that at the head of the great row in the Erme valley, which is 45 feet across and encloses a distinct barrow, while they range down to five or six feet in diameter about a kistvaen. Minus the barrow, the Stalldon Moor peristalith would have been dubbed a sacred circle without doubt-indeed, it has been. The well-known circle at Penrith had also traces of sepulture, and with the pyramids in view, we need not be in a hurry to exclude the most gigantic from this category.5

So at Avebury, Stanton Drew, Shap, and Callernish, and partially at Carnac, the rows are connected with circles; at Caithness and elsewhere with cairns and tumuli. Circles, no doubt, once existed where they cannot now be found; but it is absolutely certain that these rows have always sepulchral relations of one kind or another: and that in the immense majority of cases the association with circles-some of which still do, and the rest probably did, enclose barrows or cairns and kistvaens-is clearly evident.

There might thus be something to be said in favour of Mr. Fergusson's idea that the rows connected with circles represented externally the passages in tumuli which led to the central chamber, if the double row was invariable. We have seen that it is not.

I do not think, however, our quest for the purpose of these rows, if we accept their sepulchral origin, is at all hopeless.

5 So WESTROPP (Prehistoric Phases) cites Colonel Ross-King as stating that "the Todas, a wild and rude tribe of the Nilgiri Mountains," at the present day burn the remains of their dead within a circle of stones, and afterwards bury them there. And the late Mr. W. J. Henwood, F.R.S., described the existence at Agur and other localities in the Northwest Provinces of India, of a people who rear cromlechs, some flat and some inclined, on the first of which flowers are often placed; while within the enclosure formed by the latter lamps are burnt, clearly in honour of the dead, whether ancestor worship is actually involved or not.

We must bear in mind that they are a distinctive form of interment that in the immediate vicinity of the barrows and kistvaens and circles connected with them we find barrows and cairns and kistvaens-even circles, presumably of much the same period-which are not. The rows clearly indicate special elaboration, and therefore special honour; and by all analogy they should mark the burial places of persons of importance-such as the head of a family or a village, or the chief of a tribe.

So far as we can judge, in its complete form this class of "rude stone monument" consists of a "row connecting a circle and a menhir. At least this is the case in so many of our Dartmoor instances that we are fairly entitled to suggest the absence of circle or menhir as probably due to accident or mutilation. If so, and the perfect structure consists of circle, row, and menhir, the similarity of the arrangement to the Egyptian "Key of Life" is equally striking and suggestive, as indicating a possible association with the widespread worship of the productive powers of Nature, which in forms more or less disguised finds place, in symbol if not in doctrine, in all the elder religions. How far this cult was present to the rearers of our Dartmoor rows is, however, a very difficult question; and the rows would rather seem to represent an engrafted expansion of the original idea. Circles and menhirs are very commonly associated where the connecting row is not found. Moreover, it is one of the commonest experiences of humanity that customs and symbols long outlive their meaning. The urn retained its place as a funeral symbol, while cremation had been abandoned for centuries; the obelisk and the headstone are the old menhir; and in parts of Yorkshire you may see miniature modern cromlechs by the score in parish churchyards, as, for example, at Skipton and Bolton Abbey-slabs supported above the ground by two or four supports. But we do not think of associating these with the original idea of the cromlech builders, and it is quite possible the circle and menhir on Dartmoor may equally have reached the habit stage.

In Professor Max Muller's Gifford Lectures on Anthropological Religion, recently published, we get accounts of Vedic funeral ceremonies taken from the Aranyakas (about 600 B.C.) and the Sûtras. In case of death, says the Sûtra, let someone have a piece of land dug up, south-east or southwest (of the village), inclining towards the south or the 6 pp. 241 et seq.