298 SCIENCE PROGRESS. 



Loomis were not quite pure. Pickering also criticised 

 Jones' method, and showed that the character of the 

 results indicated that the thermometer used had not been 

 properly calibrated. In answer, Jones has stated that his 

 thermometer was calibrated, and that he considered the 

 results of different observers were not the same because 

 each obtained but an approximate value for the freezing- 

 point, since the actual temperature at which solid solvent 

 and solution are in equilibrium cannot be determined by the 

 ordinary method. 



To Nernst and Abegg (107) belongs the credit of first 

 obtaining definite evidence as to how the discrepancies arose. 

 They show theoretically and practically that in ordinary 

 cases the observed freezing-point is a function of the 

 apparatus used, the rate of stirring, the temperature of the 

 cooling bath, etc. By keeping the temperature of the cool- 

 ing bath constant and stirring at a constant rate they show 

 how to find approximately the corrections necessary to 

 obtain the true freezing-point. Their experiments prove 

 that by altering the conditions of experiment, values varying 

 between r6 and 2'i, that is, by 25 per cent., could be 

 obtained for the molecular depression of a 1 per cent, 

 solution of sugar, but on applying the corrections deduced 

 by their method all these results gave practically the value 

 demanded by theory, namely, 1 "87. They also make it clear 

 that even when the corrections are used it is quite useless to 

 attempt to measure the depressions nearer than the l/ioooth 

 of a degree. 



A new method applicable to the case of very dilute 

 solutions has been described by Lewis (108). By its use 

 Wildermann (109) has verified van't Hoff's law of the 

 depression of the freezing-point in the case of aqueous 

 solutions of sugar, urea, and alcohol ; he also shows that the 

 degree of electrolytic dissociation calculated from the freez- 

 ing-point of solutions of potassium chloride and sulphuric, 

 trichloracetic, dichloracetic, and nitro-benzoic acids agrees 

 with that calculated from the conductivity ; and finds that 

 the dissociation of the last two acids obeys Ostwald's dilu- 

 tion law. 



