(2) The percentage of iron oxide present is determined in the following way. A considerable quantity of the sample under examination is dissolved in water, and to the boiling solution a large excess of caustic potash solution is added. The caustic potash precipitates both the alumina and ferric hydrates, the former, however, redissolves. The solution is filtered, and the precipitate of ferric hydrate washed, dissolved in dilute hydrochloric acid, and reprecipitated with excess of caustic potash. It is again thrown on to a filter, washed, redissolved in hydrochloric acid, and again reprecipitated with ammonia. In this way it is obtained free from alumina.[17] [Footnote 17: For a method of estimating minute quantities of iron in alums see ‘Journ. Soc. Chem. Ind.,’ April 1887, p. 276.] (3) _Free Acid._—This can be estimated indirectly, i.e. by estimating the alumina and ferric oxide, calculating the amount of sulphuric acid necessary to combine with them, and deducting this from the total amount of sulphuric acid obtained. In the case of alum it is of course necessary to make due allowance for the sulphate of potash or ammonia present. The sulphuric acid is estimated as barium sulphate in the usual way. It should be borne in mind that sulphates of alumina are sometimes met with containing uncombined alumina. A convenient method of estimating free acid consists in digesting a weighed quantity of the finely powdered sample in strong alcohol. The alcohol dissolves away the free acid, which can be estimated in the solution by means of a standard solution of alkali. Free acid can be detected by means of a solution of {208} Congo-red, which is turned blue with free acid, but not with pure sulphate of alumina. _Antichlor, Sodium Thiosulphate, Sodium Sulphite, &c._—Sodium thiosulphate can be estimated by means of a standard solution of iodine in potassium iodide. The operation should be performed in dilute solution (see p. 111). The same solution serves also for the estimation of sulphurous acid in sodium and other sulphites. _Starches._—After having made a careful microscopical examination of the sample, it should be examined for water by drying at 100°C., and for mineral matter by igniting in a platinum crucible and weighing the residue. The water should not exceed 18 per cent., and the ash ·50 per cent. The samples should be carefully examined for insoluble matter by dissolving in water and filtering. It is also useful in comparing different samples of starch to convert the specimens into pastes, under exactly similar conditions. When perfectly cold, the “stiffness” of the pastes should be compared. This may be done by noting the relative resistance to weights placed upon their surface. Information on this subject is contained in a paper by W. Thomson in the ‘Journ. Soc. Chem. Ind.,’ March 1886.