There is another feature, however, common to all liquids, which has a most important bearing on our subject. Every liquid is capable of forming a boundary surface of its own; and this surface has the properties of a stretched, elastic membrane. Herein a liquid differs from a gas or vapour, either of which always completely fills the containing vessel. You cannot have a bottle half full of a vapour or gas only; if one-half of that already present be withdrawn, the remaining half immediately expands and distributes itself evenly throughout the bottle, which is thus always filled. But a liquid may be poured to any height in a vessel, because it forms its own boundary at the top. Let us now take a dish containing the commonest of all liquids, and in many ways the most remarkable—water—and examine some of the properties of the upper surface. *Properties of the Surface Skin of Water.*—Here is a flat piece of thin sheet silver, which, volume for volume, is 10½ times as heavy as water, in which it might therefore be expected to sink if placed upon the surface. I lower it gently, by means of a piece of cotton, until it just reaches the top, and then let go the cotton. Instead of sinking, the piece of silver floats on the surface; and moreover, a certain amount of pressure may be applied to it without causing it to fall to the bottom of the water. By alternately applying and relaxing the pressure we are able, within small limits, to make the sheet of silver bob up and down as if it were a piece of cork. If we look closely, we notice that the water beneath the silver is at a lower level than the rest of the surface, the dimple thus formed being visible at the edge of the floating sheet (Fig. 1). If now I apply a greater pressure, the piece of silver breaks through the surface and sinks rapidly to the bottom of the vessel. Or, if instead I place a thick piece of silver, such as a shilling, on the surface of the water, we find that this will not float, but sinks immediately. All these results are in agreement with the supposition that the surface layer of water possesses the properties of a very thin elastic sheet. If we could obtain an extremely fine sheet of stretched rubber, which would merely form a depression under the weight of the thin piece of silver, but would break under the application of a further pressure or the weight of a heavier sheet, the condition of the water surface would then be realized. We may note in passing that a sheet of metal resting on the surface of water is a phenomenon quite distinct from the floating of an iron ship, or hollow metal vessel, which sinks until it has displaced an amount of water equal in weight to itself. [Illustration: __Fig._ 1.—Silver sheet floating on water._]