[352] Op. cit. xli. (1885) p. 354. See also his paper in vol. xlii. of the same Journal.
Even from a distance of eight or ten miles, the hills of Rum are seen to be obviously built up of successive nearly horizontal tiers of rock. As the summer tourist is carried past the island, in that wonderful moving panorama revealed to him by the "swift steamer" of modern days, these great dark cones remind him of colossal pyramids, and as the ever-varying lights and shadows reveal more prominently the alternate nearly level bars of crag and stripes of slope, the resemblance to architectural forms stamps these hills with an individuality which strikes his imagination and fixes itself in his memory. If choice or chance should give him a nearer view of the scene, he would not fail to notice that it is among the northern hills of the island that the bedded character is so conspicuous, and that it ceases to be prominent in the southern heights, though here and there, as in the upper part of Scuir na Gillean, it may in certain lights be detected even from a distance. Crossing over from Eigg, he would recognize each of the features represented in the sketch reproduced in [Fig. 339]. Along the shore, red sandstones rise in naked cliffs, from the top of which the ground slopes upward in brown moors to the bare rocky declivities. A deep valley (Glen Dibidil) is seen to run into the heart of the hills, between the bedded group to the north and the structureless group to the south. If the weather is favourable, some eight or more prominent parallel bars of rock may be counted on the two higher cones to the right. These bars are not quite level, but slope gently from right to left. They remind one of the terraced basalts of the plateaux, but present a massiveness and a breadth of intervening bare talus-slope such as are not usual among those rocks.
Fig. 339.—Outline of the Hills of the Island of Rum, sketched from near the Isle of Eigg.
Nor is this impression of regularity and bedded arrangement lessened when we actually climb the slopes of the hills. I had for years been familiar with the outlines of Rum as seen from a distance, and had sketched them from every side, but I shall never forget the surprise and pleasure when my first ascent of the cones revealed to me the meaning of these parallel tiers of rock. I found it to be the structure of the Cuillin Hills repeated, but with some minor differences which are of interest, inasmuch as they enlarge our conceptions of the process by which the gabbro-bosses were formed.
The northern half of the island of Rum consists almost entirely of red sandstone, which, as already stated, is a continuation of the same formation (Torridonian) so well developed in the south-east of Skye, Applecross and Loch Torridon, and traceable between the Archæan gneiss and the Cambrian strata up as far as Cape Wrath. The sandstones, though full of false bedding, show quite distinctly their true stratification, which is inclined with singular persistence towards W.N.W., at angles averaging from 15° to 20°. If they are not repeated by folds or faults, they must reach in this island a thickness of some 10,000 feet. Their red or rather pinkish tint seems mainly to arise from the pink felspar so abundant in them, for in many places they really consist of a kind of arkose. Pebbly bands with rounded pieces of quartz are of common occurrence throughout the whole formation. Dykes and veins of basalt are profusely abundant. Sometimes these run with the bedding, and might at a distance be taken for dark layers among the pink sandstones. They often also strike obliquely up the face of the cliffs like ribbons.
But, notwithstanding their apparent continuity, there can be no doubt that these sandstones have suffered from those powerful terrestrial disturbances which have affected all the older rocks of the North-West Highlands. On the west side, where they plunge steeply into the sea, they have undergone a change into fine laminated rocks, which might at first be mistaken for shales, but which owe their fissility to shearing movements. Along their southern border, from a point on the east coast near Bagh-na-h-Uamha, south of Loch Scresort, to the head of Kilmory Glen, they are abruptly truncated against a group of dark, flaggy and fissile schists and fine quartzites or grits, which in some places are black and massive like basalt, and in others are associated with coarse grey gneiss. That some of these rocks are portions of the Lewisian series can hardly be doubted, and their structure and relations are probably repetitions of those between the Lewisian gneiss and Torridon sandstone of Sleat in Skye. I found also on the northern slopes of Glen Dibidil a patch of much altered grey and white limestone or marble, which reminded me of the Cambrian limestone of Skye. The red sandstones in a more or less altered condition are prolonged to the south-east promontory of the island.
In passing over the zone of these more ancient rocks, we find them to present increasing signs of alteration as they are traced up the slopes towards the great central mass of erupted material. The pink sandstones gradually lose their characteristic tint, and grow much harder and more compact, while the veins and dykes of basalt and sheets of dolerite intersecting them increase in number. The zone of black compact quartzite, which lies to the south of the sandstones, and which at one point reminds us of basalt, at another of the flinty slate of the schistose series, likewise displays increasing induration. Its bedding, not always to be detected, is often vertical and crumpled. But the most remarkable point in its structure is the intercalation in it of bands of breccia. These vary from less than an inch to several yards in diameter; they run mostly with the bedding, but occasionally across it. The stones in them are fragments of the surrounding rock embedded in a matrix of the same material, but also with pieces of a somewhat coarser grit or quartzite. A band of coarse breccia forms the southern limit of this zone along the northern base of Barkeval and Allival. In general character it resembles the thinner seams of the same material just referred to. The matrix so closely agrees with the black flinty quartzite, that but for the included stones it could hardly be distinguished; so greatly has the mass been indurated that the stones seem to shade off into the rest of the rock. But here and there its true brecciated nature is conspicuously revealed by prominent blocks of hardened sandstone. This band of breccia must in some places be 150 or 200 feet broad. It has no distinct bedding, but seems to lie as a highly inclined bed dipping into the hill. It may possibly be a crush-breccia belonging to a period earlier than the volcanic eruptions. It is at once succeeded by a black flinty felsite like that of Mull. The groundmass of this rock, so thickly powdered with magnetite grains as to be almost opaque under the microscope, displays good flow-structure round the turbid crystals of orthoclase and the clear granules of quartz. Further up the hill, the rock becomes lighter in colour and less flinty in texture—a change which is found to arise from more complete devitrification, the groundmass having become a crystalline granular aggregate of quartz and felspar with scattered porphyritic crystals of these minerals (microgranite). In some places, the felsite incloses fragments of other rocks. A specimen of this kind, taken from the head of Coire Dubh, shows under the microscope a brown micro-felsitic groundmass, with crystals of felspar and augite, inclosing a piece of basalt, composed of fine laths of plagioclase, abundant magnetite and a smaller proportion of granules of augite.
Fig. 340.—View of Allival, Rum, sketched from the base of the north-east side of the cone.