PROGRESS IN PHYSICAL CHEMISTRY. 229 



finds a plausible explanation in the supposition that these 

 liquids contain molecular aggregates. 



We now come to the question, Can the degree of com- 

 plexity be estimated ? and in the first place it must be stated 

 that the method does not rest upon a firm theoretical basis. 

 Ramsay and Shields have subjected Eotvos's theoretical views 

 to adverse criticism, and although this is affected by a mis- 

 conception (19) due to the fact that Eotvos does not use the 

 term "corresponding temperature" in the sense employed 

 by van der Waals, it must be admitted that Eotvos's proof 

 is not rigorous. Ramsay and Shields do not attempt to 

 give a proof, but strive to show that the mode of estimating 

 the molecular weight of a liquid from the rate of change of 

 its surface energy is strictly analogous to the mode of 

 estimating the molecular weight of a gas from the rate of 

 change of its volume energy. Now the latter method pre- 

 supposes that we can calculate volumes of different gases, 

 the volume of energy of which is due in each case to the 

 same number of molecules, and if in a similar way we could 

 ascertain for different liquids areas of liquid surface, the 

 surface energy of which is due in each case to the same 

 number of molecules, then the analogy might be valid. In 

 the present state of the theory of liquids, however, it seems 

 to me that we do not know how to estimate areas of dif- 

 ferent liquid surfaces satisfying the above condition. The area 

 which has been used is the molecular area, or an area upon 

 which there lies on the average in the interior of the liquid 

 the same number of molecules. But although this is so, it 

 does not follow that the surface energy possessed by the 

 molecular area, in the case of different liquids, is due to the 

 aoencv of the same number of molecules. For, in the first 

 place, it is certain that the distribution of molecules on the 

 surface will not be quite the same as on a plane in the in- 

 terior of the liquid, and in the second place, how can we tell, 

 speaking roughly, that in the case of one liquid one layer of 

 molecules, while for another two layers, etc., are not operative 

 in producing the surface energy? The number of molecules 

 participating in the surface energy depends not only on the 

 area of the liquid surface, but also on the sphere of action of 



