l^ov. 8, 1877] 



NATURE 



31 



my views about the configuration of the sea-tottom | around Jan Mayen. Tl^ ^gure of the bottom which I at 



present find the most probable I have given in the 

 chart which I send herewith. It will be observed that 

 it is the pait of the sea between Jan Mayen and Ice- 



land which is to be corrected on the small chart which 

 was published in NATURE, vol. xvi. p. 527. 



Christiania, October 23 H. Mohn 



ON THE DIFFUSION OF MATTER IN RELA- 

 TION TO THE SECOND LA W OF THERMO- 

 DYNAMICS 



I. nPHE purpose of this paper is to call attention to a 

 ■^ natural process that appears to constitute an 

 exception to the second law of thermodynamics, and 

 which, if noticed by others, would at least appear from 

 its importance to merit a rr.ore general recognition. The 

 subject may be best dealt with by means of a simple 

 illustration, the principles involved in the action of which 

 are already perfectly well known. 



2. Let the annexed figure represent a cylinder, contain- 

 P 



ing a piston, p ; a suitable (plumbago) porous diaphragm 

 (as used for diffusion experiments) being fitted into the 



piston. The piston can be connected conveniently with 

 any outer arrangement for doing work. Suppose the one 

 half of the cylinder to be filled with oxygen, the other 

 half with hydrogen. Then, as is known, according to 

 the kinetic theory, the molecules of O and H are im- 

 pinging continually against the porous partition or 

 diaphragm, P, and the molecules in their impacts thus 

 occasionally encounter vacant spaces or pores, and so 

 continue their motion on across the diaphragm into the 

 opposite compartment. Owing, however, to the fact that 

 the molecules of hydrogen are moving four times as fast 

 as the molecules of oxygen, they strike the diaphragm 

 correspondingly more frequently, and thus four times as 

 many hydrogen molecules pass through into division O, 

 as oxygen molecules pass through into division H. [The 

 piston is supposed fixed at present, so that no work being 

 done, there is consequently no heating or cooling of the 

 gas.] But on account of the excess of molecules passing 

 into division O, the pressure there will rise. If, then, after 

 the pressure has risen to a certain degree, the piston be 



