These, of course, preclude a method of the kind aimed at, and at the same time emphasise the importance of uniformity of work in the ordinary process. In the determination of chlorides in sea-water, Dittmar used a combined method: precipitating the bulk of the silver as chloride, and after filtering, determining the small excess of silver by sulphocyanate. This modification answers admirably when applied to the assay of bullion. In the ordinary Gay-Lussac method, the precipitation of the bulk of the silver by the 100 c.c. of salt solution leaves nothing to be desired, either as to ease in working or accuracy of result; the silver precipitate settles quickly, and leaves a clear liquor admirably fitted for the determination of the few milligrams of silver remaining in solution. But the method of determining this residual silver by adding successive small quantities of salt so long as they continue to give a precipitate is unsatisfactory, and, judged on its own merits apart from the rest of the process, could hardly escape condemnation. It is clumsy in practice, for the continued adding of small portions of salt solution is laborious and becomes impossible with more than a few milligrams of silver in solution. The proposed modification is simple; having precipitated the silver with the 100 c.c. of salt solution, as described under Gay-Lussac's method (page 120), shake till the liquor clears, and filter into a flask, washing with a little distilled water. Add 2 c.c. of "ferric indicator" to the filtrate and titrate with a standard "sulphocyanate solution" made by diluting the ordinary standard solution to such an extent that 100 c.c. after diluting shall be equivalent to 0.1 gram of silver.[18] Calculate the weight of silver found by "sulphocyanate" and add it to the weight which 100 c.c. of the salt solution will precipitate. An advantage of this modification is that an excess of 15 milligrams may be determined as easily and exactly as 5. In standardising the salt solution, then, weigh up, say 1.0150 gram of pure silver, dissolve and titrate. Suppose 13.5 c.c. of "sulphocyanate" required; then these are equivalent to .0135 gram of silver, (100 c.c. = .1); the silver precipitated by the salt is 1.0150-.0135--_i.e._, 1.0015 gram, which is the standard. ~Application of the Method to Assays for Arsenic.~--If silver nitrate be added to a neutral solution of an arsenate of one of the alkali metals, silver arsenate (Ag_{3}AsO_{4}), is thrown down as a dark-red precipitate. If, after adding excess of silver nitrate to insure a complete precipitation, the arsenate of silver be filtered off, the weight of the arsenic could be estimated from the weight of silver arsenate formed. But this may be done much more conveniently by dissolving the precipitate in nitric acid, and titrating with sulphocyanate; the silver found will be to the arsenic present as 324 (108×3) is to 75.