~Effect of Varying Ammonic Salts.~--The quantities of copper, water, and ammonia were the same as in the last mentioned set of experiments, but no nitric acid was used. Ammonic nitrate added 1 gram 5 grams 10 grams 20 grams Cyanide required 21.2 c.c. 22.1 c.c. 23.1 c.c. 24.1 c.c. These show that combined ammonia seriously affects the titration, and that the principle sometimes recommended of neutralising the acid with ammonia, and then adding a constant quantity of ammonia, is not a good one, because there is then an interference both by the ammonia and by the variable quantity of ammonic salts. The same quantity of combined ammonia has the same effect, whether it is present as sulphate, nitrate, chloride, or acetate, as the following experiments show. Four lots of 20 c.c. of "copper nitrate" were taken, and 20 c.c. of dilute ammonia added to each. These were carefully neutralised with the respective acids, rendered alkaline with 30 c.c. more of ammonia, cooled, diluted to bulk, and titrated. The results were:-- With sulphuric acid 22.5 c.c. of cyanide " nitric acid 22.6 " " " hydrochloric acid 22.6 " " " acetic acid 22.5 " " ~Effect of Foreign Salts.~--Sulphates, nitrates and chlorides of sodium or potassium have no action, whilst the hydrates, carbonates, bicarbonates, sulphites, and nitrites have an important effect. The interference of ammonic salts has already been shown. Salts of silver, zinc, and nickel react with cyanide just as copper does, and consequently interfere. Ferrous salts are sure to be absent, and ferric salts yield ferric hydrate with the ammonia, which is not acted on by the cyanide, but, owing to its bulkiness, it settles slowly; this lengthens the time required for titration, and so modifies the manner of working. _An assay should not be worked with ferric hydrate present, unless the standard contains about the same amount of it._ On mines it is often inconvenient to separate the copper by means of sulphuretted hydrogen; hence it is customary to titrate without previous separation. In this case, instead of standardising the cyanide with electrotype copper, a standard ore should be used. This should be an ore (of the same kind as those being assayed) in which the copper has been carefully determined. ~Effect of Varying Copper.~--In these experiments 10 c.c. of nitric acid, 30 c.c. of ammonia, and water to 200 c.c. were used. Copper nitrate present 1.0 c.c. 10.0 c.c. 20.0 c.c. 50.0 c.c. 100.0 c.c. Cyanide required 0.7 " 11.2 " 21.7 " 54.5 " 108.1 " These results show that under the conditions laid down the various causes of disturbance nearly neutralise one another, and the results within a fair range are practically proportional.