The examination of the ice on the wall gave some curious results. The horizontal arrangement of the prisms, which we had found to prevail in vertical columns, was here modified to suit the altered conditions of the case, and the axes of the prisms changed their inclination so as to be always perpendicular to the surface on which the ice lay, as far as could be determined by the eye. Thus, in following the many changes of inclination of the wall, the axes of the prisms stood at many different angles with the vertical, from a horizontal position where the wall chanced to be vertical, to a vertical position on the horizontal ledges of the rock. The extreme edges, too, of the ice, presented a very peculiar appearance. The general thickness, as has been said, varied from a foot to a foot and a half; and this diminished gradually along horizontal lines, till, at the edges of the sheet, where the ice ceased, it became of course nothing. The extreme edge was formed of globular or hemispherical beads of ice, like the freezing of a sweating-stone, lying so loosely on the rock that I could sweep them off in detail with one hand, and catch them with the other as they fell. Passing farther on towards the thicker parts of the ice, these beads stood up higher and higher, losing their roundness, and becoming compressed into prisms of all shapes, in very irregular imitation of the cellular tissue in plants, the axes of the prisms following the generally-observed law. There seems to be nothing in this phenomenon which cannot be accounted for by the supposition of gradual thaw of small amount being applied to a sheet of prismatic ice. One fact was remarkable from its universal appearance. Wherever an incision was made in this sheet of ice, the prisms snapped off at the depth of an inch, and could be mowed down like corn by means of a stout knife. Although they broke naturally at this constant depth, and left a surface of limpid ice without any signs of external or internal division, still the laminae obtained by chiselling this lower surface carefully, broke up regularly into the shapes to be expected in sections of prisms cut at right angles to the axis. The roughness of my instruments made it impossible to discover how far this extended, and whether it ceased to be the case at any given depth in the ice.