~Effect of varying the quantity of Soda.~--There were two series of experiments, one with 2 c.c. of nickel solution (= .0044 gram of nickel), the other with 20 c.c. The conditions were as before, except that the quantity of soda was varied. Soda added 5 c.c. 15 c.c. 30 c.c. Nickel found, 1st series .0037 .0042 .0045 " " 2nd series .0444 .0444 .0442 These show that the presence of much soda, though it has only a small effect, is beneficial rather than otherwise. Ammonia has a bad effect, if present in anything like the same quantities. ~Effect of varying the Nickel.~--In experiments with 10, 20, and 40 c.c. of the nickel solution, the results were:-- Nickel present .0222 .0445 .0890 Nickel found .0220 .0442 .0884 ~Effect of Zinc.~--In these experiments 20 c.c. of nickel solution (= .0445 gram of nickel), 10 c.c. of soda, 6 drops of potassium iodide and water to 100 c.c. were used. The excess of cyanide was purposely kept at from 10 to 15 c.c., which is hardly sufficient. Zinc added 0 .25 gram. .5 gram. Nickel found .0442 .0440 .0407 On increasing the excess of cyanide to over 20 c.c. and doubling the quantity of soda, the experiment with 0.5 gram of zinc gave 0.441 gram of nickel. Hence the titration is satisfactory in the presence of zinc provided that not fewer than 20 or 30 c.c. of soda are used, and that the excess of cyanide is such that not fewer than 20 or 30 c.c. of silver nitrate are required in the titration. Moreover, these precautions should be taken whether zinc is present or not. ~Effect of other Metals.~--If metals of the first and second groups are present they should be removed by passing sulphuretted hydrogen and filtering. If _iron_ is present it must be removed, since ferrous salts use up much cyanide, forming ferrocyanides, and ferric salts yield ferric hydrate, which obscures the end reaction. Hence the sulphuretted hydrogen must be boiled off and the iron removed as basic ferric acetate by the method described on p. 233. If the precipitate is bulky it should be dissolved in a little dilute acid, neutralised and again precipitated as basic acetate. The nickel will be in the two filtrates. In the absence of manganese and cobalt the titration may be made without further separation.