276 SCIENCE PROGRESS. 



carbon tetrachloride, etc., may be taken as proof that the 

 liquid contains complex molecules. Evidence of a like 

 nature has also been brought forward by Thorpe and Rodger 

 during their investigation of the viscosity of liquids. 



This must suffice to show that between normal liquids, 

 such as the esters, on the one hand, and the fatty alcohols 

 and acids on the other, there exists an important difference. 

 The characteristic difference between normal and associated 

 liquids is, that in the case of the former the molecular 

 surface energy is a linear function of the temperature, whilst 

 in the latter case it is not so. 



It has been assumed that the molecular weight of the 

 normal liquids in the liquid state is the same as that of their 

 gases. There is no positive evidence to show that the mole- 

 cules of such liquids as follow the law do not associate into 

 twos or threes on assuming the liquid state. One thing 

 however is certain : if they do, they all associate to an equal 

 extent, and the degree of association is not altered by rise 

 of temperature. It is unlikely that mere liquefaction should 

 produce in almost all cases equal association, and it is 

 unlikely that a rise of temperature should not cause the 

 dissociation of an associated body. 



It is evident, however, that the assumption that the 

 molecular weights of such liquids as the esters are not 

 changed on assuming the liquid state is a mere hypothesis ; 

 but it is one for which there is a great deal of probability, 

 and probability of the same kind as that which led to the 

 adoption of the usually received molecular formulae for 

 gases. 



It is now pretty generally known that the molecular 

 weight of liquids is not the same thing as the molecular 

 weight of liquids in liquid solution. Van't Hoff has 

 shown that dissolved substances behave as gases, in obeying 

 the well-known gas laws, if we only substitute osmotic for 

 ordinary gas pressure. The molecular weight of a liquid as 

 determined by the cryoscopic or ebullioscopic method, for 

 example, is therefore not to be considered the same as 

 the molecular weight of the liquid per se, but rather the 

 same as the molecular weight of the substance in the gaseous 



