216 Prof. L. Kahlenbero- on the 



& 



molecular-weight determinations on solutions. The only 

 objection which the adherents of the dissociation theory have 

 made against using the data obtained as an argument against 

 the dissociation hypothesis is that the solutions employed 

 were not dilute enough, since the dissociation theory and the 

 theory of van't Hoff* hold only for infinitely dilute solutions. 

 Now the work was performed with apparatus like that used 

 by Beckmann, and the lowest concentrations examined were 

 such as could very well be tested by means of apparatus of 

 that kind without incurring too large an experimental error ; 

 i. <?., the most dilute solutions tested were so chosen that the 

 rise of the boiling-point or lowering of the freezing-point 

 would as a rule be about one-tenth of a degree. But stronger 

 solutions, in some cases almost up to the point of saturation, 

 were also examined in order to observe the changes of boiling- 

 point and freezing-point with increase of concentration. The 

 objection that all the solutions employed were too concen- 

 trated to enable one to use the data to confute the dissociation 

 theory, may be met by the statement that the adherents of 

 the theory, among them Arrhenius himself, have repeatedly 

 used solutions as strong as and even stronger than the most 

 dilute solutions I employed in the investigation cited; for it 

 is a fact known to all that the bulk of cryoscopic and ebullio- 

 scopic data gathered to support Arrhenius's claim of the 

 relation between molecular conductivity and boiling-points 

 and freezing-points have been obtained with the Beckmann 

 apparatus, and hence the solutions used were as a rule not 

 more dilute than those of lesser concentration employed by 

 my co-workers and me. Not all the adherents of the theory 

 of electrolytic dissociation would be ready to admit with 

 Whetham* that the only satisfactory cryoscopic measure- 

 ments at the extreme dilutions necessary to test the theory 

 are those of Mr. E. H. Griffiths for cane-sugar and potassium 

 chloride. It certainly is demonstrated by Griffiths's work that 

 the so-called molecular lowering of the freezing-point of 

 potassium chloride is within very small limits of experimental 

 error double that of cane-sugar. This is merely one isolated 

 experimental fact, and it has not yet been shown that the 

 curve representing the change of the molecular conductivity 

 with the concentration runs parallel with the curve indicating 

 1 he change of the freezing-point with the concentration in the 

 case of potassium chloride within the range investigated by 

 Griffiths. Nor has this parallelism between freezing-point 

 and molecular conductivity been rigidly demonstrated for the 



* Paper read betore the Faraday Society. ' Electro-Cheuiist,' iii. July 

 1903, p. 16. 



