72 VAPORIZATION. 



made use of a jar which had a slight fissure in it. He was 

 surprised to find that the water of the pneumatic trough rose 

 into this jar, one and a half inches in 12 hours ; and that after 

 24 hours, the height of the water was two inches two thirds 

 above the level of that in the trough. During the experiment, 

 neither the height of the barometer, nor the temperature of 

 the place had sensibly altered.* He ascribed the phenomenon 

 to capillary action, and supposed that hydrogen only is at- 

 tracted by the fissures, and escapes through them on account 

 of the extreme smallness of its atoms. It is unnecessary to 

 examine this explanation, as Dobereiner did not observe the 

 whole phenomenon. On repeating the experiment, and varying 

 the circumstances, it appeared to me that hydrogen never es- 

 capes outwards by the fissure without a certain portion of air 

 penetrating at the same time inwards, amounting to between 

 one-fourth and one-fifth of the volume of the hydrogen which 

 leaves the receiver. It was found by an instrument, which 

 admits of much greater precision than the fissured jar, that 

 when hydrogen gas communicates with air through such a 

 chink, the air and hydrogen exhibit a powerful disposition to 

 exchange places with each other ; a particle of air, however, 

 does not exchange with a particle of hydrogen of the same 

 magnitude, but of 3.83 times its magnitude. We may adopt 

 the word diffusion-volume, to express this diversity of disposi- 

 tion in gases to interchange particles, and say that the diffusion- 

 volume of air being 1, that of hydrogen gas is 3.83. Now 

 every gas has a diffusion- volume peculiar to itself, and depend- 

 ing upon its specific gravity. Of those gases which are lighter 

 than air, the diffusion-volume is greater than 1, and of those 

 which are heavier, the diffusion- volume is less than l.f 



Exact results are obtained by means of a simple instru- 

 ment, which may be called a diffusion tube, and which is con- 

 structed as follows. A glass tube, open at both ends, is selected, 

 half an inch in diameter, and from six to fourteen inches in 

 length. A cylinder of wood, somewhat less in diameter, is in- 



* Annales de Chiniie et de Physique, 1825. 



f The mathematical relation which subsists between the diffusion-volume, and 

 the density of a gas is expressed thus : 



Diffusion-volume = - - 



I 



where d represents the specific gravity of the gas. 



