1819.] and on the Atomic Weight for Iron. 297 



from it by an excess of nearly half a cubic inch in their total 

 quantity ; the true quantity, 1 believe, lies just half way between 

 them ; that is to say, it should be precisely two cubic inches ; 

 nevertheless I shall not here make any corrections in the expe- 

 rimental quantities above obtained, but draw my conclusions 

 strictly from them, leaving it ultimately to the atomic theory to 

 polish off the rough product of experiments ; taking, therefore, 

 1-79 cubic inches of carbonic acid gas as representing 0*2264 of 

 a grain of carbon, and 0*45 of a cubic inch of azote gas as equal 

 to 0-1332 of a grain of azote, we obtain the numbers 22*64 and 

 13*32 for the quantities of carbon and of azote in 100 of the salt 

 in question. 



In the experiments made by burning the salt with peroxide of 

 copper, I was for some time puzzled to account for the necessity 

 of employing so much of the oxide as I found requisite in order 

 to obtain the maximum quantity of gases ; but I subsequently 

 discovered that the oxide which I employed contained a minute 

 quantity of silica ; and that when a sufficiency of this last was 

 not introduced by means of the oxide, the potash of the salt 

 retained behind with it a portion of carbonic acid, which, in the 

 opposite case, was expelled as gas when the silica combined 

 with that alkali at a red heat. Hence it would seem adviseable, 

 in all future attempts, to ascertain the composition of a combus- 

 tible acid by burning it united to an alkaline base, to add to the 

 peroxide of copper employed a proportion of silica sufficient to 

 saturate the alkali, and prevent its remaining behind in the state 

 of subcarbonate, or of nitrite, by converting it into a silicate. 



I estimated the quantity of hydrogen by ascertaining that of 

 the peroxide decomposed in the above experiments beyond what 

 was employed in converting the carbon into carbonic acid. My 

 mode of doing this was to act upon the residuum of the combus- 

 tion by diluted sulphuric acid, which dissolved all the oxide that 

 had not been decomposed, and left the remainder in the metallic 

 state. From the weight of this reduced copper, it was easy to 

 infer that of the oxygen expended, and by subtracting therefrom 

 what combined with the carbon, to deduce the proportion of 

 hydrogen which united with the residue. The quantity of 

 hydrogen which, by this means, I computed to exist in 100 gr. 

 of the salt was 0*8 of a grain. 



I may here remark, that this mode of computing the hydrogen 

 appears to be susceptible of greater precision than that of weigh- 

 ing the water which may be collected after the experiment ; for 

 to say nothing of the difficulty of such collection, it must some- 

 times be a matter of uncertainty whether part of that water did 

 not exist as such in the substance operated upon ; besides, one 

 part of hydrogen furnishes only nine times its weight of water, 

 but it decomposes 40 times its weight of peroxide of copper. A 

 minute quantity of hydrogen, therefore, is more readily and 



