If copper is present in the alloy as well as silver, it is necessary to add about the same quantity of copper to the checks as is supposed, or known, to be present in the assays. If the substance to be assayed is an alloy of silver and copper, first cupel 0.5 gram of it, with, say, 10 grams of lead, and weigh the resulting button of silver, in order to get an approximate knowledge of its composition. Suppose the button weighs 0.3935 gram. We know that this is below the truth: for the sake of round numbers take it as 0.4, and assume that the rest of the alloy (0.1 gram) was copper. Two check pieces are then weighed out, each containing 0.4 gram silver and 0.1 gram of copper wrapped in 5 grams of lead. Of course the silver must be pure. And there is also weighed out two (or better, four) assay pieces each containing half a gram of the alloy wrapped in 5 grams of lead. The whole lot are then cupelled as nearly as possible under the same conditions. With four assay pieces, the cupels should be placed close together in two rows of three across the muffle; the two check pieces are put in the middle cupels. Suppose the buttons of silver got weighed as follows:-- Check pieces I. 0.3940 II. 0.3945 Assay pieces I. 0.3905 II. 0.3912 III. 0.3910 IV. 0.3909 The average loss on the two check pieces is 5.7 milligrams, and the average result of the four assay pieces is 0.3909. Add the average loss to the average result, and there is got the corrected result, 0.3966. And if 0.5 gram of alloy contain 0.3966 of silver, 1000 will contain 793.2 of silver, and this is the degree of fineness. A correction for the loss in cupellation is always made in this way when rich alloys are being assayed; and in the case of rich ores it may be done after the manner of the first of the above illustrations. There is another method of working which relies more on experiment. This is to smelt the cupel as described further on (p. 114), and to again cupel the resulting button of lead. The button of silver got in this second cupellation is added to that first obtained. It will sometimes, but not often, happen that the two buttons together will slightly exceed in weight the silver which was actually present. This is because of the retention in the buttons of a small quantity of lead. It has been stated that the proportion of lead thus retained may be as much as 1% of the silver present; this, however, can only be under exceptional conditions. A determination of the actual silver in the buttons got in the series of cupellations quoted on pages 102, 103, gave an average percentage of 99.85, so that even with the larger buttons the effect of the retained lead would be only to increase the weight by about 1 milligram. In the method of working with checks, the retained lead has no disturbing influence.