The tubes must be heated in some form of water-bath; an ordinary glass beaker serves the purpose satisfactorily. The pure salt should never be diluted; but the density of the contents of tubes (_e_) and (_f_) may be varied at will, water being added in order to lower the density, and concentration by means of evaporation or addition of the nitrate being employed in order to increase it. To avoid the discoloration of the skin, rubber finger-stalls may be used, and the stones should not be handled until after they have been washed in warm water. The staining may be minimized if the hands be well washed in hot water before being exposed to sunlight. It is advisable to warm the stone to be tested in a tube containing water beforehand lest the sudden heating develop cracks. A piece of platinum, or, failing that, copper wire is of service for removing stones from the tubes; a glass rod, spoon-shaped at one end, does equally well. It must be noted that although Retgers’s salt is absolutely harmless to the ordinary gem-stones—with the exception of opal and turquoise, which, as has already been stated, being to some extent porous, should not be immersed in liquids—it attacks certain substances, for instance, sulphides and cannot be applied indiscriminately to minerals. The procedure described above is intended only as a suggestion; the method may be varied to any extent at will, depending upon the particular requirements. If such tests are made only occasionally, a smaller number of tubes may be used. Thus one tube may be substituted for the two marked _a_ and _b_, the liquid contained in it being diluted as required, and a series of indicators may be kept apart in small glass tubes. On the other hand, any one having constantly to test stones might increase the number of tubes with advantage, and might find it useful to have at hand fragments of all the principal species in order to make direct comparison. (2) DIRECT WEIGHING The balance which is necessary in both the methods described under this head should be capable of giving results accurate to milligrams, _i.e._ the thousandth part of a gram, and consistent with that restriction the beam may be as short as possible so as to give rapid swings and thus shorten the time taken in the observations. A good assay balance answers the purpose admirably. Of course, it is never necessary to wait till the balance has come to rest. The mean of the extreme readings of the pointer attached to the beam will give the position in which it would ultimately come to rest. Thus, if the pointer just touches the eighth division on the right-hand side and the second on the other, the mean position is the third division on the right-hand side (½(8 − 2) = 3). Instead of the ordinary form of chemical balance, Westphal’s form or Joly’s spring-balance may be employed. Weighings are made more quickly, but are not so accurate.