The most convenient procedure is to have at hand three glass tubes, fitted with stoppers or corks, to contain liquids of different densities— (_a_) Methylene iodide reduced to 2·7; using as indicators orthoclase 2·55, quartz 2·66, and beryl 2·74. (_b_) Methylene iodide reduced to 3·1; indicators, beryl 2·74 and tourmaline 3·10. (_c_) Methylene iodide, undiluted, 3·32. The pure liquid in the last tube should on no account be diluted; but the density of the other two liquids may be varied slightly, either by adding benzol in order to lower it, or by allowing benzol, which has far greater volatility than methylene iodide, to evaporate, or by adding methylene iodide, in order to increase it. The density of the liquids may be ascertained approximately from the indicators. A glance at the table of specific gravities shows that as regards the gem-stones methylene iodide is restricted in its application, since it can be used to test only moonstone, quartz, beryl, tourmaline, and spodumene; opal and turquoise, being amorphous and more or less porous, should not be immersed in liquids, lest the appearance of the stone be irretrievably injured. Methylene iodide readily serves to distinguish the yellow quartz from the true topaz, with which jewellers often confuse it, the latter stone sinking in the liquid; again aquamarine floats, but the blue topaz, which is often very similar to it, sinks in methylene iodide. By saturating methylene iodide with iodine and iodoform, we have a liquid (_d_) of density 3·6; a fragment of topaz, 3·55, may be used to indicate whether the liquid has the requisite density. Unfortunately this saturated solution is so dark as to be almost opaque, and is, moreover, very viscous. Its principal use is to distinguish diamond, 3·535, from the brilliant colourless zircon, with which, apart from a test for hardness, it may easily be confused. It is easy to see whether the stone floats, as it would do if a diamond. To recover a stone which has sunk, the only course is to pour off the liquid into another tube, because it is far too dark for the position of the stone to be seen.