CHEMICAL SCIENCE. 171 



interstitial spaces, or channels, appear to be sufficiently small to extin- 

 guish transpiration, or the passage of masses entirely. The graphite 

 becomes a molecular sieve, allowing molecules only to pass through. 

 With a plate of stucco, the penetration of gases under pressure is very 

 rapid, and the volumes of air and hydrogen passing in equal times are 

 as 1 to 2.891, which is a number for hydrogen intermediate between 

 its transpiration volume 2.04 and diffusion volume 3.8 ; showing that 

 the passage through stucco is a mixed result. The rate of passage of 

 gas through graphite appeared also to be closely proportional to the 

 pressure. Further, hydrogen was found to penetrate through a graph- 

 ite plate into a vacuum, with sensibly the same absolute velocity as it 

 diffused into air ; establishing the important fact, that the impelling 

 force is the same in both movements. The molecular mobility may 

 therefore be spoken of as the diffusive movement of gases ; the passage 

 of gas through a porous plate into vacuum, as diffusion in one direction 

 or single diffusion ; and ordinary diffusion, or the passage of two gases 

 in opposite directions, as double, compound or reciprocal diffusion. 



Atmol i/sis. A partial separation of mixed gases and vapors of un- 

 equal diffusibility can be effected by allowing the mixture to permeate 

 through a graphite plate into a vacuum, as was to be expected from, 

 the preceding views. As this method of analysis has a practical char- 

 acter and admits of wide application, it may be convenient to- distin- 

 guish it by a peculiar name. The amount of the separation is in pro- 

 portion to the pressure, and attains its maximum when the gases pass 

 into a nearly perfect vacuum. A variety of experiments were made 

 on this subject, of which perhaps the most interesting were those upon 

 the concentration of the oxygen in atmospheric air. When a portion 

 of air confined in ajar is allowed to penetrate into a vacuum, through 

 graphite or unglazed earthenware, the nitrogen should pass more rapid- 

 ly than the oxygen in the proportion of 1.0668 to 1 ; and the propor- 

 tion of oxygen be proportionally increased in the air left behind in the 

 jar. The increase in the oxygen actually observed when the air in 

 the jar was reduced from 1 volume to 0.5 volume, was 0.48 per cent. ; 

 to 0.25 volume, was 0.98 percent. ; to 0.125 volume, was 1.54 per cent. ; 

 to 0.0625 volume, was 2.02 per cent. ; or, the oxygen increased from. 

 21 to 23.02 per cent, in the last sixteenth part of air left behind in the 

 jar. The most remarkable effects of separation are produced by means 

 of the tube-atmolyser. This is simply a narrow tube of unglazed earth- 

 enware, such as a tobacco-pipe stem two feet in length, which is placed 

 within a shorter tube of glass and secured in its position by corks, so as 

 to appear like a Liebig's condenser. The glass tube is placed in com- 

 munication with an air-pump, and the annular space between the two 

 tubes is maintained as nearly vacuous as possible. Air or any other 

 mixed gas is then allowed to flow in a stream along the clay tube, and 

 collected as it issues. The gas so atmolysed is, of course, reduced in 

 volume, much gas penetrating through the pores of the clay tube into 

 the air-pump vacuum ; and the slower the gas is collected, the greater 

 the proportional loss. In the gas collected, the denser constituent of 

 the mixture is thus concentrated in an arithmetical ratio, while the 

 volume of the gas is reduced in a geometrical ratio. In one experi- 

 ment, the proportion of oxygen in the air after traversing the atmolyser 

 was increased to 24 50 per cent., or 16.7 upon 100 oxygen originally 



