Not only so, but the prisms could be detached with great ease, by using no instrument more violent than the fingers; while the point of a thin knife entered freely at any of the surface lines, and split the ice neatly down the sides of the prisms.

When one or two of the sides of a prism were exposed, at the edge of the piece of ice, the prism could be pushed out entire, like a knot from the edge of a piece of wood.

In some cases there seemed to be capillary fissures coincident with the lines where several sides of prisms met.

Considering the shape of the whole column, it is clear that the two ends of each prism could not be parallel; neither was one of the ends perfectly symmetrical with the other, and I do not think that the prisms were of the nature of truncated pyramids.

On descending again, I found that the columns were without exception formed of this prismatic ice, either in whole, as in the clear column, or in part, as where limpid prisms existed among the white ice which ran in veins down the cascades.