ANALYSIS OF GASEOUS MIXTURES. 299 



may be determined by the proportions in which these exist 

 in it, and a correct determination of their proportion and 

 constitution will afford a correct and true measure of the 

 quality and value of the gas, and a test of the excellence of 

 the process by which the gas has been formed, as well as of 

 the value of the coal, canuel, or other substance, used as a 

 gas-producing material. 



Gas-makers have been much in the habit of relying on 

 the weight or specific gravity of gas as an indication of its 

 quality. The heavier the gas, and, it is said, the better 

 is its quality. This notion has arisen from observing, that 

 olefiant gas is one of the heaviest of the constituents of coal 

 gas. But the specific gravity is not to be depended upon 

 as a test of the excellence of the gas, and in any case could 

 only give a very crude idea of the general quality of it, 

 without giving any knowledge of the nature and propor- 

 tion of its constituents. The specific gravity of carbonic 

 oxide is 967*8, and 100 cubic inches weigh 29*83 grains, 

 at 60*' Fahrenheit and 30 inches barometer. The specific 

 gravity of olefiant gas is 985*2, and 100 cubic inches weigh 

 30*37 grains. The specific gravity of light carburetted 

 hydrogen is 559*6, and 100 cubic inches weigh 17*25 grains. 

 Now it is evident, that a gas containing much carbonic 

 oxide, very little olefiant gas, and an inferior amount of 

 light carburetted hydrogen, and consequently very poor in 

 illuminating power, may weigh heavier, and seem better, 

 than gas of far higher and better quality, so that the weight 

 is only valuable as an adjunct to analysis. 



The measurement of the light by the eye, whilst, like the 

 above, aflFording no index of the constitution of the gas, is 

 open to many irregularities and fallacies. There is no con- 

 stant means of comparison, and after a few trials the eye fails 

 to appreciate any but large differences. The next method 

 of determining the illuminating power of gas, by estimating 

 the amount of its constituents condensible by chlorine, 



