174 SCIENCE PROGRESS. 



from the study of salt-solutions. The evidence given by 

 them was, however, anything but satisfactory ; indeed, it 

 had previously led Raoult to assume 3H0O as the formula 

 of liquid water. Eventually he concluded that the constitu- 

 tion of saline solutions was exceptional, and he adopted 20*3, 

 the value of F given by solutions of indifferent organic sub- 

 stances, as the true value of the molecular lowering. The 

 molecular formula of water had now to be assumed to be 2H 2 

 in order to get a value of F/M agreeing with the rule. The 

 assumption that the molecular complexity of a solvent 

 exercises a definite effect upon the freezing-point of its 

 solutions, is open to the objection that other correlated 

 properties give no indication of being influenced in a similar 

 way. Moreover, any independent testimony which may 

 be taken to prove that the molecule of liquid water is 

 complex, also points to the conclusion that the same is true 

 for formic acid and acetic acid. Hence the molecular 

 lowerings of these solvents as well as that of water should 

 be exceptional if the freezing-point is affected by molecular 

 complexity. The preceding table shows, however, that the 

 acids exhibit no peculiarity. For these and other reasons 

 there is little question that Raoult's explanation of the 

 behaviour of water is incorrect ; water is therefore a marked 

 exception to his rule, and in the light of more recent work 

 it is almost certain that the rule is erroneous. What 

 appears to be the correct value of the molecular lowering 

 was deduced thermodynamically by van't Hoff in 1886. 

 He showed that F is not proportional to the molecular 

 weight of the solvent, but is given by the expression 

 F = •02T 2 /W, where T is the freezing-point of the 

 solvent on the absolute scale, and W is its heat of 

 fusion in gram-calories. 



Although the observations of Fabinyi (1889) on solutions 

 in naphthalene and thymol support Raoult's empirical rule, 

 the work of Hentschel (1888), and in particular that of 

 Eykman (1888-89), wno nas carried out a careful investiga- 

 tion of the question, have decided in favour of van't HofFs 

 theoretical value of the molecular lowering. The data 

 which substantiate this conclusion are collected in the fol- 



