82 Mr. Graham's Experimental Researches 



the hydrogen gas been greatly impeded, 1st. From the direc- 

 tion in which it took place, downwards ; and, &dly. From the 

 density of the medium into which it diffused. 



Had the mixture of olefiant gas and hydrogen been allowed 

 to diffuse upwards, and into an atmosphere of specific gravity 

 intermediate between that of its constituent gases — into steam 

 or ammoniacai gas, for instance, circumstances would have 

 been most conducive to the unequal diffusion and separation 

 of the mixed gases. 



(2.) Hydrogen gas, in a tall receiver, is expanded by sul- 

 phuric ether, I find, four times more rapidly than common 

 air. Mr. Leslie had already observed, that ice evaporates 

 twice as rapidly in hydrogen gas as in common air ; and he 

 and Mr. Dalton found the cooling powers, or mobility of the 

 different gases to be inversely as their density. 



(3.) Gases permeate with increased facility in both direc- 

 tions through the pores of porcelain tubes at high temperatures 

 (Priestley), because, I believe, their tendency to diffusion, 

 which is inversely as their density, is vastly increased by their 

 rarefaction, and not from any dilatation of the pores of the por- 

 celain, which must be utterly trivial in the most intense heat. 



(4.) A tall receiver was jths filled with a mixture of 2 hy- 

 drogen + 1 oxygen, which had remained mixed for three 

 weeks, but was found sensibly pure before the experiment. A 

 little ether being thrown up into the receiver, the experimental 

 mixture rapidly expanded. The first bubble projected from 

 the receiver by the expansion was received, deprived of all 

 ether-vapour by washing, and being exploded, left half its bulk 

 of pure hydrogen gas. 



(5.) The vapour of water appears, from the following expe- 

 riment, to be more diffusive than the vapour of alcohol, as 

 might be expected from the densities of these vapours. Of 

 dilute alcohol (0*964), three ounces were exposed to spon- 

 taneous evaporation in a cylindrical jar two inches deep, and 

 the same quantity in a jar six inches deep, but otherwise 

 similar, the mouths of both- jars being loosely covered with 

 paper. When each of the vessels had lost half an ounce by 

 evaporation, the remaining liquor was examined and found to 

 contain sensibly more alcohol in the case of the deep than of 



