36 Prof. A. W. Rticker on a Suggestion 



gases the specific heat at constant pressure varies considerably 

 with the temperature. 



More recently still, Wiillner (Wied. Ann. Bd. iv. p. 321, 

 1878), using Kundt and Warburg's method, has determined 

 the ratio of the specific heats of air, carbonic oxide, carbonic 

 acid, nitrous acid, ethylene, and ammonia at 0° and 100° C. 

 He finds that for gases which obey Boyle's law, c p and c v are 

 constant, but that in the case of less perfectly gaseous bodies 

 they increase with the temperature. The difference between 

 them, however, is always very approximately constant and 

 equal to the theoretical number — thus justifying the applica- 

 tion of the first of the above equations to imperfect gases, and 

 proving that the observed increase in the specific heats is due 

 to work done within the molecules, and not against the inter- 

 molecular forces, which must therefore be negligible. 



On the whole, then, the result of these researches is to show 

 that m + e can be calculated very approximately from the above 

 equations if c p is given, and that Regnault's values of this 

 quantity are probably trustworthy to 6 per cent. 



One of the chief difficulties of the thermodynamic theory of 

 gases has been to attribute to m and e values which would at 

 once lead to the observed ratios of c p to c v , and satisfy any 

 rational supposition as to the interior mechanism of a mole- 

 cule. Kundt and Warburg proved that for mercury vapour 



— =1*666, which is consistent only with the supposition that 



the atoms of that substance are smooth rigid spheres. Boltz- 

 mann (Pogg. Ann. Bd. clx. p. 175, 1877) and Bosanquet (Phil. 

 Mag. April 1877) have since drawn attention to the fact that, 

 for a smooth rigid surface of revolution, m = 5 and e=0, which 



would make -^ = 1*4. The fact, therefore, that this number 



Cv 



agrees very closely with those given by experiment for a large 

 number of gases (air, 0, N, H, CO, and NO) would be accounted 

 for by supposing their molecules to be surfaces of revolution. 

 This condition would be fulfilled, as is pointed out by Mr. 

 Bosanquet, by two spheres rigidly united, and would thus 

 accord well with our conception of the atomic constitution of 

 the above gases. It would perhaps be better to regard the 

 spheres not as rigidly united, but as bound together by forces 

 which prevent the separation of their surfaces, while leaving 

 them otherwise free to move. The required five degrees of 

 freedom would thus be obtained, and the hypothesis would 

 better coincide with the supposition of the union of two smooth 

 spheres to form the ultimate particles of the gases. 



For a discussion of the difficulties offered to any such theory 



