§ 100. =Separation of Ammonia.=--Ammonia is separated in all cases by distillation, and if the organic or other liquid is already alkaline, it is at once placed in a retort and distilled. If neutral or acid, a little burnt magnesia may be added until the reaction is alkaline. It is generally laid down that the contents of the stomach in a putrid condition cannot be examined for ammonia, because ammonia is already present as a product of decomposition; but even under these circumstances it is possible to give an opinion whether ammonia _in excess_ is present. For if, after carefully mixing the whole contents of the stomach, and then drying a portion and reckoning from that weight the total nitrogen (considering, for this purpose, the contents to consist wholly of albumen, which yields about 16 per cent. of nitrogen)--under these conditions, the contents of the stomach yield more than 16 per cent. of nitrogen as ammonia reckoned on the dry substance, it is tolerably certain that ammonia not derived from the food or the tissues is present. If, also, there is a sufficient evolution of ammonia to cause white fumes, when a rod moistened with hydrochloric acid is brought near to the liquid, this is an effect never noticed with a normal decomposition, and renders the presence of extrinsic ammonia probable. An alkaline-reacting distillate, which gives a brown colour with the “nessler” reagent, and which, when carefully neutralised with sulphuric acid, on evaporation to dryness by the careful heat of a water-bath, leaves a crystalline mass that gives a copious precipitate with platinic chloride, but is hardly at all soluble in absolute alcohol, can be no other substance than ammonia. § 101. =Estimation.=--Ammonia is most quickly estimated by distilling, receiving the distillate in decinormal acid, and then titrating back. It may also be estimated as the double chloride of ammonium and platinum (NH₄Cl)₂PtCl₄. The distillate is exactly neutralised by HCl, evaporated to near dryness, and an alcoholic solution of platinic chloride added in sufficient quantity to be always in slight excess, as shown by the yellow colour of the supernatant fluid. The precipitate is collected, washed with a little alcohol, dried, and weighed on a tared filter; 100 parts of the salt are equal to 7·6 of NH₃. VI.--Caustic Potash and Soda. § 102. There is so little difference in the local effects produced by potash and soda respectively, that it will be convenient to treat them together. =Potash= (=potassa caustica=).--Hydrate of potassium (KHO), atomic weight 56, specific gravity 2·1.