grand and imposing effect; and from the idea of lofty height which the distant mountains suggest, they convey a stronger impression of gigantic proportions than is produced even by the first sight of the giant monoliths of Salisbury Plain.

The most remarkable of the Hebridean groups is that of Callernish, near Loch Roag, in the Lewis, of which an accurate view is given in the frontispiece to this volume, from a sketch by my friend, George Harvey. It occupies the summit of a ridge of hilly ground, and embraces a cruciform group of monoliths attached to a central circle about forty feet in diameter. In the centre is

a column measuring nearly seventeen feet in height, around which the circle is formed of flat columnar

blocks of gneiss. From this an avenue of similar stones stretches two hundred and seventy feet to the north, while single rows placed towards the other cardinal points complete the cruciform arrangement of the whole. Its greatest length is stated by Logan as 558 feet, and by Macculloch as about 680 feet; but its present actual measurement, from the most southern stone to the northern end of the avenue, is barely 380 feet. Attention has been recently directed to an interesting fact, which seems to confirm the idea that this megalithic group has been expressly arranged with refer

ence to the cardinal points by astronomical observation. Mr. Henry Callender remarks, in a communication on this subject: "That the position was chosen and laid down from astronomical observation, can easily be demonstrated by visiting the spot on a clear night, when it will be found that by bringing the upper part of the single line of stones extending to the south to bear upon the top of the large stone in the centre of the circle, the apex of that stone coincides exactly with the pole-star; this is more readily done from the south line being on sloping ground, so that looking along the line upwards to the higher level of the centre stone is very much the same as taking an observation through the incline of a telescope." The peculiar arrangement of the Callernish group, with its northern avenue, and cardinal rows of columns, strongly confirms the conviction, that we have here a memorial of primitive astronomical knowledge; of the observation of that one ever-resting polar star, around which all others seem to revolve; and of the study of the motions of the heavenly bodies in connexion with native rites of worship in prehistoric times. Until recently, many of the stones were completely buried in the moss, and of two other circles lying about a mile to the eastward on a low moor, nothing could be seen but a few grey blocks slightly protruding above the heather and rushes. But since the first edition of this work appeared, the liberal zeal of Sir James Matheson has effected the removal of the superincumbent peat from all the three circles, to a depth of between five and six feet; thereby leading to important discoveries. Beneath the moss surrounding the great Callernish circle, the disclosure of a rough causeway basement, and other equally conclusive proofs, showed that the stones had been founded on the boulder-clay, apparently before the 1 Proceedings of Soc. Antiq. Scot, vol. ii. p. 382.

growth of the peat commenced. This received confirmation from evidence of a still more comprehensive character, by the fact that fallen stones of the smaller circles were uncovered, lying upon the clay, with the whole growth of peat above them; so that the commencement of the peat-forming epoch appears to date subsequent to their desertion and ruin. We have thus a singularly suggestive evidence of their remote antiquity; and a gauge of the lapse of time since the abandonment of those megalithic temples: which, though as yet undefined, only requires some approximate determination of the annual rate of growth of the peat, to enable us to apply it to such purposes of chronology.

But other discoveries rewarded the labours of the explorers. As the excavations at the great temple proceeded, a circular stone building was disclosed on the east side of the central stone, with its diameter equal to the radius of the circle, as shown in the accompanying ground-plan, and containing two chambers, the largest

of which measures six feet nine inches by four feet three inches. Within this, embedded in an unctuous substance, consisting apparently of peaty and animal mat

1 Edin. Phil. Jour. New Series, vol. xv. p. 236.

ter, fragments of human bones were found, which seem to have been subjected to the action of fire. The disclosures are altogether replete with interest; and reveal some novel features in connexion with this class of monuments, which tradition has associated with such unvarying tenacity with the worship of the Druids. But the buried monoliths of the Lewis circles are not the only examples of such change of level in the slow lapse of time. On various parts of the mainland similar megalithic groups remain partially entombed in like manner in the slowly accumulating mosses, the growth of unnumbered centuries. On one of the wildest moors in the parish of Tongland, Kirkcudbrightshire, an example may be seen, consisting of a circle of eleven stones, with a twelfth of larger dimensions in the centre, the summits of the whole just appearing above the moss. Adjoining the group there stands a large cairn with its base doubtless resting on the older soil beneath. With such evidence at command, it is obvious that however vague many of the speculations may be which have aimed at the elucidation of rites and opinions of the Celtic Druids; and have too often substituted mere theory for true archæological induction: we shall run to an opposite error in ascribing to a Scandinavian origin structures manifestly in existence long prior to the earliest Norwegian or Danish, or even perhaps Celtic, descent on our coasts.

One other remarkable class of works, the Rocking Stones, remains to be noted. These are found among the ancient monuments of England and Ireland, as well as on various parts of the Continent, and are no less frequent in Scotland; and no illustrations of ancient skill, or of primitive superstitious rites, are more calculated to awaken astonishment, and admiration of their singular

1 Proceedings Soc. Antiq. Scot. vol. iii. p. 112.

constructors. So strange a mixture of extreme rudeness and great mechanical skill appears to be combined in these memorials of the remote past, that they excite greater awe in the thoughtful mind than even the imposing masses enclosing the sacred area of Stonehenge or the circle of Stennis. Nor is such an estimate of them unjust; for it would undoubtedly prove a much more complicated problem for the modern engineer to poise the irregular and amorphous mass on its point of equilibrium, than to rear the largest megalithic group that still stands to attest the mechanical power which the old builders could command.

It has indeed been supposed by some that the origin of Rocking Stones may be traced entirely to natural causes; and this opinion is adopted by Worsaae and other Danish and Norwegian antiquaries.1 Such a theory, however, seems to stand still more in need of proof than that which regards them as stones of ordeal, by which Druid or Scandinavian priests were wont to test the guilt or innocence of the accused. Apollonius Rhodius speaks of rocking-stones placed on the apex of tumuli, and Mr. Akerman refers, in his Archæological Index, to the famous Agglestone Barrow, in the island of Purbeck, as having been similarly surmounted. One such undoubted example would abundantly suffice to overthrow this geological theory of natural formation. It is a less conclusive, though not altogether valueless argument, that some of the most remarkable logan stones. of Scotland are found in the immediate vicinity of other undoubted primitive stone-works. The great rockingstone in the parish of Kirkmichael, Perthshire, for example, has already been referred to as one of a large group of circles, cairns, and other monuments of the same class. Its form is that of a rhombus, of which

Primeral Antiquities of Denmark, p. 110.