The amount of pure caustic alkali in a sample of alkali is best determined by Fresenius’s method, as follows:——The total amount of pure alkali, both caustic and carbonated, expressed in per-cents. of carbonate of soda or carbonate of potassa, is ascertained by any of the usual methods. The apparent quantity of alkali per cent. is then determined, without previous treatment of the sample with carbonate of ammonia, by the method of Will and Fresenius (p. 86). The difference between the results indicates the per-centage of dry caustic alkali present; or if the volumetric method be in use, it can be often fairly estimated by adding the first portions of the test-acid very gradually to the sample, carefully observing the effect. When the effervescence at length commences the weight or measure of the test-liquor expended shows the quantity of pure caustic alkali under treatment (nearly). The result depends upon the fact, that little or no carbonic-acid gas is expelled from the liquid on the addition of the test-acid, until the caustic portion is very nearly neutralised. The quantity of WATER or MOISTURE, per cent., present in an alkaline carbonate, is indicated by the loss of weight which 100 gr. suffer on gentle ignition in a loosely-covered iron dish or platinum crucible. So also with samples containing caustic alkali, except that here the water of hydration (= 1 equiv. = 9) is not expelled from the ‘caustic’ portion, and must therefore be determined by calculation. Other matters deserving the serious attention of the operator are——hitting the exact point of neutralisation, and——preparing the test-acids of the proper strength. The method of effecting the former correctly has been already referred to in this article, and is also fully noticed under ACETIMETRY and ACIDIMETRY.