4 PHILOSOPHICAL TRANSACTIONS. [ANNO J 7 8 J . 



specific gravity. But, in fact, the specific gravity of compounds, found by 

 actual experiment, seldom agrees with that found by calculation, but is often 

 greater without any diminution of the lighter ingredient. This increase of 

 density must then arise from a closer union of the component parts to each 

 other than either had separately with its own integrant parts ; and this more inti- 

 mate union must proceed from the attraction or affinity of these parts to each 

 other : I therefore imagined that this attraction might be estimated by the in- 

 crease of density or specific gravity, and that it 

 was proportionable to it, but was soon undeceived. Cubic inch „f common air .... 0.385 



I must also premise, that the absolute weights Fixed air 0.570 



r r ■ , , . , , . Marine Air 0.65+ 



or many sorts 01 air nave been accurately deter- jsjitrous air 300 



mined by Mr. Fontana, the thermometer being Vitriolic air 0.77s 



o , ,, , , Alkaline air 0.2 



at 55 , and the barometer at 2Q-' T inches, or Inflammab le air 0.035 



nearly so. Their weights were as annexed : 



Of' spirit of salt. — From the time I first read in Dr Priestley's Experiments 

 on Air (that inexhaustible source of future discoveries) of the exhibition of 

 marine acid in the form of air, free from water ; and that this air, reunited with 

 water, formed an acid liquor in all respects the same as common spirit of salt ; I 

 conceived the possibility of discovering the exact quantity of acid in spirit of salt 

 of any given specific gravity, and by means of this the exact proportion of acid 

 in all other acid liquors ; for if a given quantity of pure fixed alkali were satu- 

 rated, first by a certain quantity of spirit of salt, and then by determined quan- 

 tities of the other acids, I concluded, that each of these quantities of acid liquor 

 must contain the same quantity of acid ; and this being known, the remainder 

 being the aqueous part, this also must be known ; but this conclusion entirely 

 rested on the supposition that the same quantity of all the acids was requisite for 

 the saturation of a given quantity of fixed alkali ; for if such given quantity of 

 fixed alkali might be saturated by a smaller quantity of one acid than of another, 

 the conclusion fell to the ground. This point might indeed be in some measure 

 determined by weighing the neutral salts, formed by these acids, when thoroughly 

 dry ; but still a source of inaccuracy remained : for if they were exposed to a 

 considerable heat, part of the acid would necessarily be expelled, and more of 

 one acid than of another; and if the heat were not considerable, much of the 

 water of crystallization would remain ; so that, if the weights were found to be 

 equal, this equality could not be ascribed to equal quantities of acid, but might 

 perhaps arise from a smaller proportion of acid in one of them, and a larger 

 proportion of water, and in another from a larger proportion of acid and a 

 smaller proportion of water ; and if the weights were unequal, no certain con- 

 clusion could he drawn. To obviate this difficulty, I used the following expe- 

 dient. 1st. I supposed the quantities of nitrous and vitriolic acids, necessary to 



