Bergman estimates the quantity of water in this salt at 25, and the acid at 43 per cent; but his real acid was not so concentrated as Kirwan’s, consequently the nitric acid in nitrated soda should contain more water than my true acid. Wenzel, from an experiment on the composition of nitrated soda, concludes that it contains 37,48 of alkali, and 62,52 of nitrous acid; and 1000 of this acid, from Kirwan’s calculation, contain 812,6 of his real acid; consequently, 100 parts of my aëriform acid should contain 93,28 of Wenzel’s acid, and 6,72 of water. I saturated with potash 54 grains of solution of nitric acid, of specific gravity 1,301. Evaporated at about 212°, it produced 66 grains of nitre. This nitre exposed to a higher temperature, and kept in fusion for some time, was reduced to 60 grains. Now from the table, 54 of 1,301, should contain 26,5 of true acid. But according to Kirwan’s estimation, 100 parts of dry nitre contain 44[56] of his real acid, with 4 water; consequently 60 should contain 26,4. [56] Additional Observations, page 70. Again, 90 grains of acid, of specific gravity 1,504, saturated with potash, and treated in the same manner, gave 173 grains of dry nitre. Consequently, 100 parts of it should contain 47,3 grains of true acid. Now Lavoisier[57] allows about 51 of dry acid to 100 grains of nitre, and Wenzel 52. From Berthollet’s[58] experiments, 100 grains of nitre, in their decomposition by heat, give out nearly 49 grains of gas.[59] Hence it appears that the aëriform acid, that is, the true acid of my table, contains rather less water than the acid supposed to exist in nitre. [57] Elements, pag. 103, Kerr’s Translation. [58] Mem. Acad. 1787. [59] As well as oxygene and nitrogene, Mr. Watt’s experiments prove that much phlogisticated nitrous acid is produced. DIVISION II. _EXPERIMENTS and OBSERVATIONS on the composition of AMMONIAC and on its combinations with WATER and NITRIC ACID._ I. _Analysis of AMMONIAC or VOLATILE ALKALI._ The formation and decomposition of volatile alkali in many processes, was observed by Priestley, Scheele, Bergman, Kirwan, and Higgins; but to Berthollet we owe the discovery of its constituent parts, and their proportions to each other. These proportions this excellent philosopher deduced from an experiment on the decomposition of aëriform ammoniac by the electric spark:[60] a process in which no apparent source of error exists. [60] Journal de Physique, 1786. Tom. 2, pag. 176. Since, however, his estimations have been made, the proportions of oxygene and hydrogene in water have been more accurately determined. This circumstance, as well as the conviction of the impossibility of too minutely scrutinizing facts, fundamental to a great mass of reasoning, induced me to make the following experiments.