20 Clarke, The Atomic Theory. 



lengths of the spectral lines are related to the atomic 

 weights of the several elements, as has been shown by the 

 researches of Runge and his colleagues, of Rummel,* and 

 of Marshall Watts,-f- If we \.xy to study the specific 

 gravity of solids or liquids, the only clues to regularity 

 are furnished by the atomic ratios. Atomic and molecular 

 volumes give us the only approximations to anything like 

 order. Similarly, we speak of atomic and molecular 

 refraction, of molecular rotation for polarised light, of 

 molecular conductivity, and the like. In Trouton's law, 

 the latent heat of vaporisation of any liquid becomes a 

 function of the molecular weight. And, finally, all thermo- 

 chemical measurements are meaningless until they have 

 been stated in terms of gramme molecular weights ; then 

 system begins to appear. Chaos rules until the atomic or 

 molecular weight is taken into account ; with that con- 

 sidered, the reign of order begins. 



Even to the study of solutions the same conditions 

 apply. Substances in solution exert pressure, and in this 

 respect they closely resemble gases van't Hoff has 

 shown that equal volumes of solutions, having under like 

 conditions equal osmotic pressures, contain equal numbers 

 of molecules, and thus Avogadro's gas law is curiously 

 paralleled. The two laws are even equivalent in their 

 anomalies. The abnormal density of a gas is explained 

 by its dissociation, and the variations from van't Hoff's 

 law are explicable in the same way. The theory of ionic 

 or electrolytic dissociation, proposed by Arrhenius, shows 

 that certain substances, when dissolved, are split up into 

 their ions, and through this conception the analogy 

 between gases and solutions is made absolutely complete. 

 The ions, however, are atoms or groups of atoms ; and 



* Proc. Roy. Soc. Victoria, vol. lo, part I., p. 75. 

 t I'hii. ATaq. (6), 5, 203. 



