from the first; they did not get so through the denudations caused by time. But the general rule is to find with these circles other structural forms. Mr Fergusson, who has written a most able, though prejudiced, work on "Rude Stone Monuments," considers the mound as architecturally the first step in the development of these monuments, and, for the mere explanation of terms, we may accept his order of exposition. The mound would require a row of stones round its base to keep it together; hence arose the circle of close-fitting stones, which so often accompanies mounds and cairns, and which sometimes also appear alone, though not so often. These mounds may be of earth or stone, and their purpose is, as a rule, for burial, though cairns might have also been raised for the sake of "remembrance," 66 witness," and boundary marks. The burial mound or barrow may have a cist in it-that is four slabs of stone set in box form, and with another slab super-imposed. These cists were intended to receive the bodies or the urns. The cists when exposed, that is when the earth of the mound is all cleared away, appear as a box of large slabs, with a slab covering it; and this description, with the addition that the stones are large, or megalithic, is true of a perfect "dolmen " or "cromlech." The dolmen consists of at least two supporting stones and one covering stone, but it usually has side stones as well as end stones. The true dolmens are found unattended by any trace of tumulus, which shows that they were erected independent of any mound or cairn. Nevertheless, the best antiquarians are of opinion that they were intended for burial purposes, and not for altars of worship, as the "Druid" theorists have held. Whether they were a development from the stone cists is, perhaps, an open question. Another feature of certain barrows is the internal chamber. This chamber is generally circular, and built by overlapping the stones at a certain height, and thus gradually narrowing the circle until at last the apex of the chamber can be closed with one slab. The chambered cairns at Clava will illustrate the principle of this construction. Here there is a foundation laid of very large stones-some three feet high, and on this a course or two of stones is laid, not perpendicularly as is more usual, but with a backward inclination. At from four to five feet high the stones begin to overlap all round until, at last, at a height probably of eleven or twelve feet, the circle could be closed by a single stone, thus forming a domed chamber of a dozen feet in diameter and height. Leading to this chamber there may be a passage made of two walls of stone, with slabs across. The foundation-stones of the passage are usually large-megalithic, in

fact. Now, two points are to be noticed here, according to the theory of the architectural development of these structures maintained by Mr Fergusson: If the cairn or mound is removed, leaving only its megalithic foundations, there remain an interior circle of closely-placed stones, with an avenue" leading into it. Hence the avenue or alignment of stones may be regarded as a development of the passage into the chamber of the mound, at least from an architectural point of view. These avenues of stone are common in France, but we have a good specimen of them in Lews, where, at Callernish, we find an avenue of megalithic pillars-stones six or seven feet high on the average, leading to an interior circle of 42 feet diameter. The peculiarity at Callernish is that there branch off from the circle three lines of stones, making, with the avenue, a kind of cruciform groundplan. The great chambered mound at New Grange, in Ireland, is entered by a long passage three feet wide and some six feet high, the sides of which are composed of megalithic pillars covered over with slabs. The chamber has branches running off right and left, and a third in continuation of the passage. The general resemblance of this groundplan to that of Callernish caused Dr Stuart, in his "Sculptured Stones of Scotland," to say: "If the cairns at New Grange were removed, the pillars would form another Callernish." But Callernish was never covered with a mound; it was, indeed, threatened to be covered with peat, accreted through countless centuries. The avenue is too broad-eight feet broad-and the stones too pointed to be covered with cross slabs, while they stand apart from one another at a distance of some six feet, and not close together, as such a theory would require. Besides, where would the mound material be taken to in such a place? Callernish, from these and other considerations, was never even intended to be covered with a mound or cairn. These avenues attain their highest development when unattended with any other structures or superstructures in the way of circles or of mounds, as at Carnac, in France. Another accompaniment of the stone circle may be a single standing stone or "menhir," placed either interior or exterior to it. And, lastly, we may mention the existence of a ditch or trench exterior to a circle or to a mound. Such ditches have passages leading across them- -a fact which throws some light on the passages of the Clava middle cairn. Specimens of these are to be seen in Derby and Cumberland. To sum up, we find circles connected with mounds, either inside the mound, round its base, or at a distance outside; we find circles connected with or surrounding dolmens, whether mound-covered or