rocks. It is most easily seen at that side of the island which faces Iona, and on the summit of the cliffs of a semicircular bay opening in that direction. The bed is here broken at the edge of the cliff, so as to expose its whole thickness for a considerable extent. But the same appearance may also be observed immediately above the ordinary landing place, where the bed has also been broken. The stones which it contains are all rounded, and of various, often considerable dimensions, and they exhibit specimens of granite, gneiss, micaceous schistus, quartz, and red sandstone. Together with these, are some rolled pieces of basalt.
Here then is a circumstance in the mineral history of Staffa, adventitious it is true, but involving difficulties of no small importance. If we cast our eyes on the map, (Pl. 35.) we shall perceive that it is embayed in a large sinuosity, formed in the island of Mull, and nearly enclosed on the opposite side by Iona, and the Treshanish islands. Beyond the latter, a second line is drawn by Tirey, and Coll; while to the north, but at a greater distance, are placed the islands of Muck, Rum, Egg, Canna, and Sky. The whole island of Mull, with the exception of the Ross, is of a trap formation, containing however some partial tracts of sandstone and other rocks which I need not notice. The islands of Ulva and the Treshanish, with their dependent rocks, are also of trap formation. So are the islands which lie to the north, and which I have enumerated above. Iona however, together with Coll and Tirey, consists principally of gneiss and mica slate traversed by granite veins, rocks which also form the chief parts of the coasts of Lorn, Appin, Morven, and Ardnamurchan.
It is to the former then, that we must look for the origin of the rolled stones which cover Staffa, if, limiting the great operations of nature by our own narrow views, and the ages which have con-
3 s 2