A fat, or a liquid, wets the glass, _i.e._ comes into intimate contact with it, but if a solid substance be tested in the same way, a film of air would intervene and entirely prevent an observation. To displace it, a drop of some liquid which is more highly refractive than the substance under test must first be applied to the plane surface of the dense glass. The most convenient liquid for the purpose is methylene iodide, CH_{2}I_{2}, which, when pure, has at ordinary room temperatures a refractive index of 1·742. It is almost colourless when fresh, but turns reddish brown on exposure to light. If desired, it may be cleared in the manner described below (p. 66), but the film of liquid actually used is so thin that this precaution is scarcely necessary. If we test a piece of ordinary glass—one of the slips used by microscopists for covering thin sections is very convenient for the purpose—first applying a drop of methylene iodide to the plane surface of the dense glass of the refractometer (Fig. 20), we notice a coloured shadow-edge corresponding to the glass-slip at about 1·530 and another, almost colourless, at 1·742, which corresponds to the liquid. If the solid substance which is tested is more highly refractive than methylene iodide, only the latter of the shadow-edges is visible, and we must utilize some more refractive liquid. We can, however, raise the refractive index of methylene iodide by dissolving sulphur[2] in it; the refractive index of the saturated liquid lies well beyond 1·800 and the shadow-edge corresponding to it, therefore, does not come within the range of the refractometer. The pure and the saturated liquids can be procured with the instrument, the bottles containing them being japanned on the outside to exclude light and fitted with dipping-stoppers, by means of which a drop of the liquid required is easily transferred to the surface of the glass of the instrument. So long as the liquid is more highly refractive than the stone, or whatever may be the substance under examination, its precise refractive index is of no consequence. The facet used in the test must be flat, and must be pressed firmly on the instrument, so that it is truly parallel to the plane surface of the dense glass; for good results, moreover, it must be bright. [Illustration: FIG. 20.—Faceted Stone in Position on the Refractometer.] [Illustration: FIG. 21.—Shadow-edges given by a doubly refractive substance.]