302 Dr. Henry on the Analysis of some of [Oct. 



In this case, the carhonic acid was 40 — 21*5 = 18*5, and 

 therefore exceeded in volume the theoretical proportion (18*4) 

 by only -pi-, a deviation much within the limits of possible 

 errors, arising either from the difficulty of measuring small quan- 

 tities, or of ascertaining the purity of nitrous oxide. We 

 may, also, from this mode of operating, deduce the quantity of 

 nitrogen, which exists as an element of nitrous oxide*; for since 

 1 volume of carbonic acid results from the combustion of 1 vo- 

 lume of carbonic oxide, the residuary 21 - 5 measures must have 

 contained 21 — 18-5 = 2*5 of carbonic oxide and nitrogen in- 

 troduced by that gas, +19 measures of nitrogen disengaged 

 from the nitrous oxide. 



The results of the experiments with an excess of carbonic 

 oxide suggested to me a ready and correct method of testing 

 the nitrous oxide, which had hitherto been a desideratum. The 

 only test, before applicable to this purpose, was the amount, to 

 which the gas is absorbed, when agitated with well boiled 

 water. But besides the uncertainty whether all the nitrous 

 oxide be in this case condensed, the proportion of the unab- 

 solved residuum is subject to variation, from the quantity of 

 other gases extricated from the water itself. Reduced to the 

 form of a rule, the new method may be stated as follows : Let 

 a given volume of nitrous oxide be exploded with a slight ex- 

 cess of carbonic oxide of known purity; for example, 110 or 

 115 measures of the latter to 100 of the former. Now as each 

 volume of real nitrous oxide gives, under these circumstances, 

 an equal volume of carbonic acid, we may impute whatever 

 carbonic acid is deficient of that proportion to the mixture of 

 so much nitrogen with the nitrous oxide. If, for example, 

 using an excess of carbonic oxide, there should result, from 

 100 measures of nitrous oxide, only 95 of carbonic acid, we 

 may safely consider the nitrous oxide to be contaminated with 

 5 per cent, of nitrogen gas. A proportion of nitrous gas may, 

 1 am aware, be occasionally mixed with the nitrous oxide, but 

 this may be easily discovered, and previously separated, by so- 

 lution of green sulphate of iron. 



Having determined the application of carbonic oxide to the 

 analysis of nitrous oxide to be so easy and satisfactory, I had 

 hoped that the same agent might have been employed in the 

 analysis of nitrous gas, which, as is well known, does not form 

 a combustible mixture with simple hydrogen gas.* But on 

 trial, I could not, by any variation which I made in the propor- 

 tions of the two gases, obtain a mixture combustible by electri- 

 city. I had recourse, therefore, to olefiant gas ; but had nearly 

 abandoned this method also as impracticable, on finding that 

 the mixture could not be set on fire by a spark from the prime 



* Davy's Researches, p. 136. 



