of accurate consideration, as they not only determine the con stitution of the most northern part of Europe, but furnish us with information which can hardly be found either so full or so distinct in the interior of mountains. The cliffs, and small islets on the shore, are not composed of gneiss; for the whole exterior of these rocks, their fine slaty structure, and the earthiness of their cross fracture, pro nounce with distinctness that we are entering on clay-slate; but the shining surface of this rock resembles mica-slate : the scales of mica, however, although abundant, are only scattered on the continuous base of this slate. These strata contain beds of massive brown quartz, in which delicate fissures are frequently coated with chlorite, with occasional imbedded crystals of felspar. Large folia of talc, and small greenish- grey splintery cones, resembling serpentine, are not of uncom mon occurrence. Ascending Kielvigs-Eid we find small-grained granite, con taining insulated folia of black mica and a great deal of horn blende : the line of separation of these rocks may be followed for a considerable length, and it may be clearly seen that the clay-slate is continued beneath the granite. This granite frequently changes itself into straight slaty gneiss, in which large and beautiful garnets often occur. In some places, however, this granite contains diallage and but little quartz ; and by the former mineral gradually increasing, whilst the quartz and mica continue to decrease till they disappear, the granite becomes a fine granular greenstone without any visible separation. This change in the nature of the rock is be trayed by weathering ; for the clove-brown diallage becomes tile-red and frequently of a garnet colour. Towards the mountains, this greenstone passes at last into a coarse granular rock, in which the ingredients of felspar and diallage are beautifully distinct, like the gabbro of the Italians. The felspar of this variety more easily disintegrates than the dial lage, although the latter loses its colour sooner. The loose blocks are therefore rough, the diallage projecting when the