304 Mt\ Dahon^s Analysis of [Oct. 



Article XIII. 



On the Analysis of Atmospheric Air hif Hydrogen* 

 By John Dultoii, Esq. FRS.&c. 



(To llichard Phillips, Esq.) 



RESPECTED FRIEND, Manchester y Sept. 17, 1825. 



According to my promise I transmit the results of some 

 late experiments on the analysis of atmospheric air by hydrogen. 

 My chief object was to find under what circumstances the 

 union of the oxygen and hydrogen, by the electric spark, is 

 complete, that is, so that one or both of the gases are entirely 

 consumed ; and in what cases either no union takes place or a 

 partial one, leaving portions of both gases still in mixture in the 

 residue. 



From a memoir of M. M. Humboldt and Gay-Lussac (Ann. 

 de Chimie, 53, 1805,) we learn that one volume of hydrogen, 

 mixed with two or nine volumes of oxygen, gives the same loss 

 by electricity, namely, 1-46; but if mixed with 9*5 oxygen, the 

 loss is only '68 ; and this loss diminishes rapidly till the oxygen 

 becomes 16, when there is no loss at all. They found that if the 

 surplus gas was azote or carbonic acid, the loss was not much 

 difterent ; but they do not seem to have ascertained this with 

 precision. 



It is right to observe that the hydrogen I used was obtained 

 in the usual way from zinc and dilute sulphuric acid, and was 

 received in bottles filled with as pure rain water as I could 

 procure ; the bottles were filled with the gas, and not more than 

 one-third of the gas of each bottle was used; the hydrogen 

 was free from atmospheric air, except what was expelled from 

 the water by the hydrogen bubbling into the bottle : this quan- 

 tity of atmospheric air, however, must be something ; yet, on 

 firing 10 measures of hydrogen with oxygen, the diminution is 

 usually 14*6 to 15. 



The mixtures of gases fired at once were commonly about 

 150 measures, each measure being the volume of one grain of 

 water. The eudiometer has six inches in length, correspondent 

 to 150 measures; and all the experiments were made over 

 water. 



The atmospheric air I mostly used was procured in the 

 country, and was found by frequent trials to contain almost 

 exactly 21 per cent, of oxygen. This is not the case at all 

 times. I once found the oxygen as hi^h as 21*15 per cent, 

 from an average of many experiments; it was on the 8th of 

 January last, when the barometer was 309, wind N.E. and very 

 moderate, after three days of calm and gentle frost. But the 

 general state of the atmosphere yields only 20*7 or 20*8 per 



