Another portion is treated with tartaric acid: a white granular precipitate is produced. Hydrofluosilicic acid is added to a third portion of the solution: the formation of a gelatinous precipitate is a further indication of the presence of potassa. If the preceding tests have given negative results, and a white precipitate is formed by the addition of antimonate of potassa to another portion of the solution, _soda_ is present. In both cases, it is necessary to confirm the results by means of the spectroscope. The above reactions are distinctive only in the absence of metals precipitated by sulphuretted hydrogen, sulphide of ammonium or carbonate of soda, and small portions of the solution should be tested with these reagents. In order to detect baryta, strontia and lime, the residue, insoluble in alcohol is dissolved in dilute nitric acid, and an excess of carbonate of ammonia added to the solution: the three bases, if present, are precipitated as carbonates. The precipitate formed is separated from the solution by filtration, dissolved on the filter in dilute hydrochloric acid, and the solution then filtered and divided into two parts: sulphuric acid is added to one, the fluid filtered from the precipitate of sulphate of baryta formed, and the filtrate treated with ammonia and oxalate of ammonia. If _lime_ be present,--although its sulphate is not easily soluble--sufficient will be contained in the filtrate to give a white precipitate of oxalate of lime. The remaining portion of the solution is evaporated to dryness, and the residue treated with absolute alcohol. Chloride of strontium goes into solution, chloride of barium remaining undissolved. If upon evaporating the alcoholic solution a residue is obtained which, when dissolved in water, produces turbidity in a solution of sulphate of lime, _strontia_ is present. The residue, insoluble in alcohol, is dissolved in water. If a precipitate is produced by the addition of sulphuric acid or hydrofluosilicic acid to the solution, _baryta_ is present. The latter reaction distinguishes baryta from strontia, which is not precipitated by hydrofluosilicic acid. Should the tests mentioned above fail to give affirmative results, and poisoning by means of baryta and strontia be nevertheless suspected, these compounds may possibly have remained in the materials contained in the alimentary canal, in the state of insoluble sulphates. To effect their detection under these circumstances the organic substances must be decomposed by means of sulphuric acid. The carbonaceous residue is calcined in a crucible at an elevated temperature, and the remaining mass treated with water. In this way, a solution of sulphides of barium and strontium is obtained, which is then tested as directed above. DETECTION OF CHLORINE, BROMINE, AND IODINE. CHLORINE AND BLEACHING CHLORIDES.