~The titration.~--The solution, containing not much more than 0.1 gram of nickel, and free from the interfering metals, must be cooled. It is next neutralised and then made strongly alkaline with a solution of soda (NaHO); an excess of 20 or 30 c.c. suffices. This will produce a precipitate. The cyanide solution is now run in from a burette until the solution clears, after which an excess of about 20 c.c. is added. It is well to use some round number of c.c. to simplify the calculation. Add a few drops of potassium iodide solution, and run in the standard solution of silver nitrate from a burette. This should be done a little at a time, though somewhat rapidly, and with constant shaking, till a permanent yellow precipitate appears. If the addition of the cyanide did not result in a perfectly clear solution, this is because something besides nickel is present. The residue may be filtered off, though with a little practice the finishing-point may be detected with certainty in the presence of a small precipitate. If the student has the slightest doubt about a finish he should run in another 5 c.c. of the cyanide and again finish with silver nitrate. The second result will be the same as the first. For example, if 40 c.c. of cyanide and 30 c.c. of silver nitrate were required at the first titration, then the 45 c.c. of cyanide in the second titration will require 35 c.c. of silver nitrate. The difference between the quantities of the two solutions used in each case will be 10 c.c. It is this difference in the readings of the two burettes which measures the quantity of nickel present. Each c.c. of the difference is equal to .0025 gram of nickel. But if the cyanide solution is not exactly equal in strength to the silver nitrate, the quantity of cyanide used should be calculated to its equivalent in silver nitrate before making the subtraction. The following experimental results illustrate the accuracy of the assay and the effect upon it of varying conditions. A solution containing 1 gram of nickel sulphate (NiSO_{4}.6H_{2}O) in 100 c.c. was used. By a separate assay the sulphate was found to contain 22.25 per cent. of nickel. For the sake of simplicity the results of the experiments are stated in weights of nickel in grams. ~Effect of varying excess of Cyanide Solution.~--In each experiment there was 20 c.c. of the nickel solution, equal to .0445 gram of nickel. There were also 10 c.c. of soda solution, 3 or 4 drops of potassium iodide and sufficient water to bring the bulk to 100 c.c. before titrating. Cyanide in excess 6 c.c. 4 c.c. 8 c.c. 12 c.c. 25 c.c. Nickel found .0434 .0436 .0440 .0442 .0444 Although the difference between the highest and lowest of these results is only 1 milligram, their meaning is quite obvious. The excess of cyanide should not be less than 20 c.c.