Broadly considered, two classes of facts are presented to the glacier-observer; the one suggestive of viscosity, and the other of the reverse. The former are seen where _pressure_ comes into play, the latter where _tension_ is operative. By pressure ice can be moulded to any shape, while the same ice snaps sharply asunder if subjected to tension. Were the result worth the labour, ice might be moulded into vases or statuettes, bent into spiral bars, and, I doubt not, by the proper application of pressure, a _rope_ of ice might be formed and coiled into a _knot_. But not one of these experiments, though they might be a thousandfold more striking than any ever made upon a glacier, would in the least demonstrate that ice is really a viscous body. [Illustration: Fig. 30. Moulds used in experiments with ice.] I have here stated what I believe to be feasible. Let me now refer to the experiments which have been actually made in illustration of this point. Two pieces of seasoned box-wood had corresponding cavities hollowed in them, so that, when one was placed upon the other, a lenticular space was enclosed. A and B, Fig. 30, represent the pieces of box-wood with the cavities in plan: C represents their section when they are placed upon each other. [Sidenote: ACTUAL MOULDING OF ICE.] A _sphere_ of ice rather more than sufficient to fill the lenticular space was placed between the pieces of wood and subjected to the action of a small hydraulic press. The ice was crushed, but the crushed fragments soon reattached themselves, and, in a few seconds, a lens of compact ice was taken from the mould. [Illustration: Fig. 31. Moulds used in experiments with ice.] This lens was placed in a cylindrical cavity hollowed out in another piece of box-wood, and represented at C, Fig. 31; and a flat piece of the wood was placed over the lens as a cover, as at D. On subjecting the whole to pressure, the lens broke, as the sphere had done, but the crushed mass soon re-established its continuity, and in less than half a minute a compact cake of ice was taken from the mould. [Illustration: Fig. 32. Moulds used in experiments with ice.] In the following experiment the ice was subjected to a still severer test:--A hemispherical cavity was formed in one block of box-wood, and upon a second block a hemispherical protuberance was turned, smaller than the cavity, so that, when the latter was placed in the former, a space of a quarter of an inch existed between the two. Fig. 32 represents a section of the two pieces of box-wood; the brass pins _a_, _b_, fixed in the slab G H, and entering suitable apertures in the mould I K, being intended to keep the two surfaces concentric. A lump of ice being placed in the cavity, the protuberance was brought down upon it, and the mould subjected to hydraulic pressure: after a short interval the ice was taken from the mould as a smooth compact _cup_, its crushed particles having reunited, and established their continuity.