DIALTTIC SEPARATION OF GASES BY COLLOID SEPTA. 431 



other volatile body which has been observed to pass, like hydrogen, through a plate of 

 palladium is common ether — and that at the atmospheric temperature, while a passage 

 was denied to hydrogen at the same time. The palladium was in the form of foil. 

 Although thin foil of this metal is generally visibly porous and allows air to pass through 

 like a sieve, a tube diffusiometer, covered with a disk of the selected palladium foil, 

 and standing over mercury, retained a volume of 40*5 millims. of air over a vertical 

 column of 155 millims. of mercury for twenty-four hours without depression of the 

 mercury. The air was dried by sticks of potash, but still it did not penetrate the pal- 

 ladium. Dry hydrogen was then conducted to the upper surface of the palladium disk, 

 but still without any penetration by that gas after several hours. Cotton-wool moistened 

 with ether was now placed upon the disk, when, after eight minutes, the confined air 

 within the tube began to expand ; and in the course of an hour longer, the 40*5 volumes 

 of confined air increased to 90"4 (thermometer 18°-5, barometer 758), when the expan- 

 sion ceased. The increase of volume appeared to be due entirely to ether-vapour, 

 absorbable by a pellet charged with sulphuric acid. Why hydrogen proved to be 

 incapable of penetrating the palladium in such circumstances it is difficult to say. It 

 can only be imagined that the palladium foil may have previously condensed on its 

 surface a minute film of foreign matter, which rendered the palladium inactive to 

 hydrogen but not to ether-vapour. 



On the other hand, the penetrating power of hydrogen, here referred to the liquefac- 

 tion of that gas, appears not to be solely confined to metallic septa. There is reason 

 to suspect that in diffusing through a plate of graphite hydrogen passes in a small 

 proportion as a liquid, without any counterdiffusion of air. Hence the constant excess 

 observed of the diffusive coefficient of hydrogen, which came out 3-876, 3-993, and 4-067 *, 

 instead of the theoretical number 3*8, corresponding to the density of the gas referred 

 to air. Such phenomena of gaseous penetration suggest a progression in the degree of 

 porosity. There appear to be (1) pores through which gases pass under pressure or by 

 capillary transpiration, as in dry wood and many minerals, (2) pores through which gases 

 do not pass under pressure, but pass by their proper molecular movement of diffusion, 

 as in artificial graphite, and (3) pores through which gases pass neither by capillary 

 transpiration nor by their proper diffusive movement, but only after liquefaction, such as 

 the pores of wrought metals and the finest pores of graphite. 



Osmmm-mdium. 



A portion of small grains of osmium-indium, amounting to 2-528 grms., was exposed 

 to hydrogen through all descending temperatures from a red heat, as the preceding 

 metals had been treated. The osmium-iridium was then heated again to redness in 

 the Sprengel vacuum, to extricate any hydrogen that might have been absorbed. But 

 only a bubble or two of gas, too minute to be measured, passed over in fifteen minutes, 

 at a red heat. Osmium-iridium, then, exhibits no absorbent power for hydrogen — a result 

 which is consistent with the crystalline character of the substance. 



* Philosophical Transactions, 1863, p. 404. 



MDCCCLXVI. 3 



