this line absent, its form and proportion would be nearly that of the Roman cross or common crucifix. The longest line of this cross, which may be considered as the general bearing of the work, lies in a direction 24° west of the meridian. The total length of this line is at present 588 feet, but there are stones to be found in the same direction for upwards of 90 feet further, which have apparently been a continuation of it, but which having fallen, like others, through different parts of the building, have sometimes been overwhelmed with vegetation, leaving blanks that impair its present continuity. The whole length may therefore with little hesitation be taken at 700 feet. The cross line, intersecting that now described at right angles, measures 204 feet, but as it is longer on one side than the other, its true measure is probably also greater, although I was not able to discover any fallen stones at the extremities; the progress of cultivation having here interfered with the integrity of the work. The diameter of the circle which occupies the centre of the cross is sixty-three feet, the lines ceasing where they meet the circumference. The stone which marks the centre is twelve feet in height. The heights of the other stones which are used in the construction are various, but they rarely reach beyond four feet: a few of seven or eight feet are to be found, and one reaching to thirteen is seen near the extremity of the long line. The additional line already mentioned, extends northwards from the outer part of the circle, on the eastern side. It is however very defective, a great number of the stones being absent towards its northern extremity; although there is apparent evidence of their former continuity, in one which remains erect and in others which have fallen from their places. I could not discover any traces of a line parallel to this on the western side; but as some inclosures have been made in the immediate vicinity, it is possible that such might have originally existed; notwithstanding the superstitious reverence with which the Scots in general regard these remains, and the care with which, in their agricultural operations, they commonly avoid committing any injury to them. The intervals between the stones vary from two to ten feet or more, but it is probable that the larger spaces have resulted from the falling of the less firmly rooted pillars which occupied those places. The number of stones in the circle is thirteen, independently of the central one; and the number in the whole building, either erect or recently fallen, is forty-seven.

The aspect of this work is very striking, as it occupies the highest situation on a gentle swelling eminence of moor land; there being no object, not even a rock or stone, to divert the attention and diminish the impression which it makes. The circles found in the vicinity are less perfect, and present no linear appendages their average diameter varies from forty to fifty feet, and one of them contains four uprights placed in a quadrangular form within its area. I may add to this general account, that solitary stones, apparently of a monumental nature, are found in this neighbourhood, as well as in the island of Bernera, and in other parts of Lewis.

• The

The cruciform shape of the structure described above is a remarkable, and I believe a solitary circumstance. It has not at least been noticed among the numerous descriptions of these erections which antiquaries have given to the public. It is true, that in some of the cromlechs or smaller monuments, a disposition of the stones resembling that of a cross has sometimes been remarked, but it seems in all these cases to have been the result either of accident or necessity. No monuments in which that form is obviously intended, have been traced higher than the period of the introduction of Christianity; nor was it indeed till a later age, that of Constantine, that the cross became a general object of veneration. From that time its use is common; and it is frequently found applied under a great variety of structures and forms, to numerous objects, civil and military, as well as ecclesiastical. Those cases in which the figure of the cross has been found marked or carved on stones of higher antiquity, which had served either for the purposes of sepulchral memorials or Druidcal worship, appear to have resulted from the attempts of the early catholics to convert the supposed monuments of ancient superstition to their own ends; either from economical motives or from feelings of a religious nature. But such attempts cannot be supposed to have given rise to the peculiar figure of the structure here described. The whole is too consistent and too much of one age to admit of such a supposition; while, at the same time, it could not under any circumstances have been applicable to a Christian worship. Its essential part, the circular area, and the number of similar structures found in the vicinity, equally bespeak its ancient origin. It must therefore be concluded that the cruciform shape was given by the original contrivers of the fabric, and it will afford an object of speculation to antiquaries, who, if they are sometimes accused of heaping additional obscurity on the records of antiquity, must also be allowed the frequent merit of eliciting light from darkness. To them I willingly consign all further speculations concerning it.'

Along the coast are innumerable examples of the passage of granite veins, and of the contortions of gneiss. Not far from Oreby is seen a detached and incurvated rock of this material, forty feet in height, the surrounding parts having fallen away.

Enough, however, of these gneiss islands; in which the hoary monotony of the geology seems to have palled even on the Doctor's patient spirit of research: but his observations concerning them bespeak both originality of statement and correctness of induction; while the corollaries deducible from them may be regarded as important contributions to the history of a rock, which forms an extensive ingredient in the composition of many mountainous tracts of the globe.

The Trap islands are so denominated from the prevailing rock, although in a few of them no masses of that sub

4

stance

stance are to be found; yet their geological features, being in other respects analogous to the rest of the groupe, will justify an apparent deficiency of arrangement. The non-conformity of their outline to the general bearings of the coasts, and their secondary characters, at once proclaim a difference from the preceding division. They abound, too, in trap-veins, which may generally be traced to large masses of the parentrock. These islands, in the order most conducive to their elucidation, are, Rasay, Flodda, Sky, Soa, Longa, Scalpa, the Shiants, Canna, Sandy Isle, Rum, Egg, Muck, Inch Kenneth, Mull, Ulva, Colonsay, and a few others of subordinate dimensions.

The west side of Rasay has a dull and uninteresting aspect: but on the east we meet with scattered farms, diversified by towering rocks, mural cliffs of sandstone, patches of brushwood, and striking proofs of enormous fractures and dislocations. The first conglomerate, that forms a part of the sandstone beds, presents some detached and elevated rocks of a very remarkable character, on one of which is situated the very whimsical and picturesque structure, Brochel Castle, the antient seat of Macleod of Rasay. This building is so contrived as to cover the whole summit of the sharp eminence on which it stands; its walls being continuous with the precipitous faces of the rock. The projections of these have been so contrived as to form parts of the building; and they are at the same time so like in appearance to the masonry of which it is constructed, that it is often difficult to distinguish between the artificial and the natural wall.'

Owing to various unfavourable circumstances, the precise sequence of the strata in this island is ascertained with doubt and difficulty; and we shall therefore only mention, generally, that, besides gneiss, which geologically connects the island with Rona, are found red and white sandstones, graywacké slate, shale, and coarse limestone. The porphyry, which is not regularly stratified, has for its base compact felspar, usually of pale tints; and the imbedded crystals of felspar are white, or slightly yellow. Different trap-rocks are also found incumbent on the sandstone: but they bear a very small proportion to the porphyry. One, of common basalt, forms the summit of Duncan-hill, and another has for its base a mixture of augit and felspar, and contains prehnite.

The little island of Flodda is included in this series merely for the sake of illustration, as it throws some light on the structure of Rasay and Sky. The magnitude of the latter, and the great variety and intricacy of its geological structure, occasion it to occupy nearly 160 pages of the present work:

but

but we cannot accompany the writer through such a labyrinth of disquisition; and we must be contented to point to a few of the more remarkable passages, whether of a popular or a scientific description.

On the east side of the promontory of Strathaird are a number of caverns, which have been formed in the rock by the wasting and degradation of trap-veins. One of them was inhabited by the Pretender, during his concealment; and another has lately become the cause of great resort to Sky, on account of its stalactitic concretions, being popularly distinguished by the name of the Spar Cave.

Beyond the high cascade of Holme, the uniformity of the mural rock is finely relieved by a series of columnar cliffs, in some respects superior to the ranges of Staffa; for, though not so regular, it is on a scale of five or six times the magnitude, and, in some of its portions, more assimilated to magnificent Grecian architecture. The virgin scenery of Loch Scavig, and of the vale of Cornisk, is portrayed in the Doctor's best style: but he was not permitted to approach this enchanted ground without submitting to an ordeal of patience, arising out of the want of boat-accommodation, in a degree more than equal to that which we have already mentioned as defeating his efforts to reach Barra-head.

The principal rocks which compose the geology of this island are gneiss, the lowest of the series, graywacké slate, red sandstone, quartz rock, shelly and crystalline limestone, shale, calcareous sandstone, and various members of the trap and syenite families, in the usual irregular and overlying position: with partial strata of siliceous schist, chert, and coal. The first of these latter varies in its tints from pale to dark grey, and to an intense black. On weathering, the grey becomes white, while the black retains its original hue, many of the fragments thus resembling striped jasper. Lydian-stone is merelythe black variety of siliceous schist in a hard and brittle state. With respect to the coal, its position is generally as obscure as its quantity is insignificant. In some instances, it is entangled in the trap-rock, and accompanied by bituminous wood: but, in any situation in the island in which it has been hitherto discovered, it affords no prospect of being worked to advantage.'

The principal unstratified substances, of which the positions and relations are here detailed, are syenite, porphyry, claystone, basalt, amygdaloid, iron clay, jasper, greenstone, augit rock, and hypersthene rock. Dr. M. observes: the jasper is rare. I have used this term because I know of no other by which the substance in question can so well be cha

racterized.

racterized. It is yellow or brown, with a lustre approaching to the resinous, and is well known as a product of St. Helena. The specimens of Sky differ in no respect from those of that island, which have sometimes, but improperly, been called pitch-stones. That they are not such, if proof were necessary, would be sufficiently proved here by the regular gradation which they undergo into clay; appearing indeed to be portions of clay which have undergone changes, in consequence of their vicinity to the basalt, resembling those which sandstones experience in similar situations.*'

Dr. M. proposes to give the name of augit rock to an admixture of felspar and augit, which possibly occurs more frequently than mineralogists are aware, as it is often confounded with green-stone; the difficulty of discriminating between augit and hornblend, especially when the particles are small, and intimately combined with felspar, being very great. In the present instance, the finer varieties, when freshly broken, are comparatively tender, with the aspect of serpentine, the felspar being of a green colour, and the augit of a pitchy black: but, in a day or two, this distinction vanishes, the whole acquiring great additional hardness, and a generally grey aspect; so that neither of the ingredients can be accurately discerned. Here, then, we have another example of the close alliance between lavas and traps, since the observations of mineralogists have recently proved that augit and not hornblende forms the dark part of these volcanic products.'

The simplest varieties of the hypersthene rock, which occurs so abundantly in the Cuchullin range of hills, consist of hypersthene, with compact greenish or crystalline felspar; the crystals or concretions of hypersthene being, in some cases, half an inch in length, and in others not exceeding a pin's point. The masses are remarkable for the bareness and integrity of their surfaces, and for the spiry forms of their summits, which emulate the picturesque aspect of the hills of Arran. Dr. Macculloch's able sketch of the numerous trapveins observed in Sky will amply reward the trouble of perusal. The independent minerals which he specifies are, analcyme, chabasite, stilbite, (which is very abundant,) nadelstein, (in various forms,) laumonite, olivine, prehnite, chalcedony, steatite, actynolite, &c. The rarest was hypersthene.

* I have since received similar specimens from Guadaloupe, where they occur among the lavas of that island, adding one more to the numerous analogies already existing between the volcanic rocks and those of the trap family.'

Several