Various devices and methods are available for ascertaining the density of liquids—for instance, Westphal’s balance; but, apart from the inconvenience attending such a determination, the density of all liquids is somewhat seriously affected by changes in the temperature, and it is therefore better to make direct comparison with fragments of substances of known specific gravity, which are termed indicators. If of two fragments differing slightly in specific gravity one floats on the surface of a uniform column of liquid and the other lies at the bottom of the tube containing the liquid, we may be certain that the density of the liquid is intermediate between the two specific gravities. Such a precaution is necessary because, if the liquid be a mixture of two distinct liquids, the density would tend to increase owing to the greater volatility of the lighter of them, and in any case the density is affected by change of temperature. The specific gravity of stones is not much altered by variation in the temperature. A more convenient variation of this method is to form a diffusion column, so that the density increases progressively with the depth. If the stone under test floats at a certain level in such a column intermediate between two fragments of known specific gravity, its specific gravity may be found by elementary interpolation. To form a column of this kind the lighter liquid should be poured on to the top of the heavier. Natural diffusion gives the most perfect column, but, being a lengthy process, it may conveniently be quickened by gently shaking the tube, and the column thus formed gives results sufficiently accurate for discriminative purposes.