5f)2 PHILOSOPHICAL TRANSACTIONS. [ANNO 1784. 



tion saturated the orange-coloured precipitate, and all the acid liquor obtained 

 from the process: therefore we have the proportion as 352 : 120 :: 1395 : 475, 

 from which it appears, that all the acid employed was recovered again in the 

 form of acid, excepting only 5 grs. ; a smaller quantity than what might reason- 

 ably be supposed to be lost in the process by the extreme volatility of the nitrous 

 air. In order to ascertain the exact point of saturation, slips of paper, stained 

 by the juice of the petals of the scarlet rose, were employed, which were the 

 nicest test I could procure, as litmus will not show the point of saturation of any 

 liquor containing much phlogisticated nitrous acid, or even fixed air, but will 

 turn red, and show it to be acid, when the test of those leaves, violets, and 

 some other of the like kind, will turn green in the same liquor, and show it to 

 be alkaline. But the exact point of saturation of so dilute a liquor is so very 

 difficult to ascertain, that an error might easily be committed, notwithstanding 

 the attention bestowed on it. Supposing this experiment to be unexceptionable, 

 the conclusions which may be drawn from it are very favourable to the hypothesis 

 I endeavour to support. Thirty-six oz. measures of dephlogisticated air were 

 obtained, and only 5 grs. of a weak nitrous acid were lost in the process. 218 

 grs. of mercury out of 240 were revived, and all the dephlogisticated nitrous 

 acid employed is found to be highly phlogisticated in the process. It appears, 

 that the nitrous acid does not enter into the composition of dephlogisticated air ; 

 it seems only to serve to absorb phlogiston from the watery part of the mercurial 

 nitre. 



1 1 . As this last process proved very tedious and complicated, on account of 

 the necessity of ascertaining the quantity of acid in the receiving water, by 

 means of an alkali which afforded a double source of error in the point of satu- 

 ration, I resolved to try the distillation of dephlogisticated air from cubic nitre 

 in a glass vessel, and to draw from it only such a quantity of air as it would yield 

 without acting much on the retort, which latter circumstance is essentially neces 

 sary to be attended to. An ounce of the crystals of mineral alkali were dissolved 

 in nitrous acid, and the mixture brought to an exact saturation by the test of 

 litmus ; 30 oz. measures of air were distilled from it, which, during the latter 

 part of the process, was accompanied with slightly yellow fumes ; the receiving 

 water was found to be acid, and the residuum alkaline. The residuum being 

 dissolved in the receiving water, the solution was neutral, or very nearly so, 

 by every test ; for in this case litmus might be used, as the acid was very 

 slightly phlogisticated. On adding a few drops of a very dilute nitrous acid, the 

 tests showed the liquor to be acid. 



12. Encouraged by the success of this experiment, I took an ounce = 480 

 grs. of pure common nitre, and put it into a flint-glass retort, coated, which 

 was placed in a furnace. It began to give air about the time it became red-hot, 



