Laio of Molecular Force. 307 



Combining equations (1) and (2), we get 



Ti ~ Ta ^ ^ 



Now, according to Van der Waals, if two bodies are in corre- 

 sponding states at temperatures T^ and Tg, they will continue 

 to be so at temperatures (l + n)Ti and (l+n)T2 j ^^ ^he case 

 where n is an infinitely small fraction we get from (4), 



^ d f ci^Vl^ \ rj, ^ /<W_\ . 



i. e. 



~T~ ("i^'i^) = ;;- («2^2*) = constant. 

 OX at 



Thus, according to this argument the last expression is not 

 only constant for a given body at all temperatures, but is the 

 same for all bodies. Eotvos has verified the formula from 

 0°-190° for ethyl oxide^ and has found the mean value of the 

 constant for a number of substances to be "227. 



The exceptional behaviour of water, the alcohols, and the 

 fatty acids is easily understood when we remember the large 

 amount of experimental evidence there is pointing to irregu- 

 larity in the molecular structure of water and ethyl alcohol, 

 while the well-known anomalous vapour-density of acetic acid 

 and of some of the higher acids of the series helps to explain 

 their exceptional behaviour. 



Robert Schiff'^s discovery (Annalen der Chemie, ccxxiii.) 

 was the result of a comprehensive experimental determination 

 of the capillary constants of a large number of organic com- 

 pounds, and consisted in the unfolding of a definite relation 

 loetween the number of molecules of different liquids, which 

 rise in a given capillary tube, and their chemical structure. 

 He showed that the number of molecules raised could be ex- 

 pressed as a number of hydrogen atoms ; thus the presence of 

 a atom in a molecule has as much influence in determining 

 how many molecules will rise in a tube as two H atoms, an 

 atom as much as three H, and so on. Calling the capil- 

 lary equivalent of a molecule thus expressed in terms of H 

 atoms h, and n 1000 times the number of molecules raised, 



a^d 

 or 1000 -x — , where a^ is the height to which the substance 



rises in a millimetre tube, d its density, and m its molecular 

 weight, Schitf^s empirical law is 



logion = 2-8155 --00728 A- logioA. 



We shall not have occasion to use this result of SchifF's ; but 



X2 



