37 MAXWELL'S LAW 87 



The deviation in the case of hydrogen has presumably 

 the same cause as that shown in Graham's experiments. 



38. Thermal Effusion 



If the speed with which a gas issues from a narrow 

 tube is really proportional to the speed of its molecular 

 motion the two must increase in the same ratio when the 

 temperature rises. From this it follows that it must be 

 possible to augment the efflux of a gas by warming it. 

 If we let a gas pass through a porous plate, we shall be 

 able to increase its speed by warming the exit side of the 

 plate. 



A one-sided heating of the porous partition produced 

 in this way is not only able to augment an already existing 

 flow, but it is sufficient of itself, without any difference of 

 pressure, to cause a percolation through the pores of the 

 partition. Just as effusion results from a difference, of pres- 

 sure at the two sides of a porous partition, so can a similar 

 phenomenon be brought about by a difference of temperature 

 of the two sides of a partition ; and the latter phenomenon, 

 according to Maxwell's 1 suggestion, is called thermal 

 effusion. 



The possibility of in this way producing a flow of gas by 

 means of an unequal distribution of temperature was first 

 pointed out by Car! Neumann 2 when he was attempting 

 to explain the production of a thermoelectric current by 

 analogy with a thermal effusion. At Neumann's sugges- 

 tion F edder sen 3 arranged a simple experiment, in which 

 he stuffed into a glass tube a tolerably long plug of spongy 

 platinum and then warmed one of the ends of this plug ; he 

 observed the phenomenon expected, viz. a flow of air 

 through the plug from its colder to its warmer side. He 

 obtained the same result on substituting hydrogen for air 

 and spongy palladium for spongy platinum. He observed 

 the same action, too, when he used other partitions made of 



1 Phil Trans, clxx. 1879, p. 255 ; Scientific Papers, ii. p. 711. 



2 Ber. d. math.-phys. CL d. K. Ges. d. Wiss. zu Leipsig,IS12, p. 49. 

 5 Pogg. Ann. cxlviii. 1873, p. 302. 



