258 Mr GRAHAM on the Diffusion of Gases. 



responding to its individual diffusiveness. The same separation 

 of mixed gases occurred in diffusion through the pores of stucco, 

 or the fissure of a cracked jar. 



" Which volumes are not necessarily of equal magnitude, be- 

 ing, in the case of each gas, inversely proportional to the square 

 root of the density of that gas." This may be demonstrated, 

 when different gases communicate by very narrow channels, or by 

 very small apertures, and when inequality of pressure is guarded 

 against. In the case of a gas communicating with the air by a 

 wide aperture, on the other hand, although the diffusion or inter- 

 mixture takes place precisely in the same way, still the result is 

 different ; for where a contraction takes place from the process 

 of diffusion, the air flows in mechanically through the aperture, 

 wholly unresisted, and makes up the deficiency. A gas, how- 

 ever, of large diffusion- volume escapes, in these circumstances, in 

 a shorter time than a gas of small diffusion- volume. Indeed, it 

 was the conclusion of the former paper, that gases diffuse more 

 or less rapidly according to some function of their densities, " ap- 

 parently inversely as the square root of their densities." The 

 advantage, in illustrating the process of diffusion, of minute 

 apertures or channels of communication, such as we have in the 

 stucco-plug, depends upon the circumstance, that when a con- 

 traction or expansion takes place in the gaseous contents of a 

 diffusion-instrument, any current in an outward or inward direc- 

 tion is prevented by frictional resistance ; so that the simple re- 

 sult of diffusion is exhibited, not complicated by the effect of any 

 other force. 



The law at which we have arrived (which is merely a descrip- 

 tion of the appearances, and involves, I believe, nothing hypo- 

 thetic), is certainly not provided for in the corpuscular philoso- 

 phy of the day, and is altogether so extraordinary, that I may 

 be excused for not speculating farther upon its cause, till its 

 various bearings, and certain collateral subjects, be fully investi- 

 gated. 



