164 Intelligence and Miscellaneous Articles. 



experimental values, instead of the present " smoothed" values, are 

 published. The most remarkable part of the conclusions which he 

 draws from them is that non-electrolytes in very weak solutions 

 give an abnormally large depression similar to that given by 

 electrolytes, but which according to the dissociation theory should 

 be shown by electrolytes only. It is still more remarkable, how- 

 ever (especially when we remember that these results emanate 

 from Prof. Ostwald's laboratory), that Mr. Jones should have pub- 

 lished them without auy reference to the fact that my own results 

 (although they differ somewhat from his in individual cases) had 

 established the existence of this same phenomenon, not only as 

 regards aqueous solutions, but still more conspicuously as regards 

 benzene solutions (Ber. xxiv. pp. 1469, 3329 ; xxv. pp. 1854, 2011, 

 2518, 3434). 



Mr. Jones suggests a possible explanation of this excess in the 

 case of non-electrolytes, without noticing, however, that it must 

 apply equally to electrolytes also, and thus upset the very conclusions 

 which he has drawn from the rest of his work, namely, that the excess 

 is due to dissociation and agrees accurately with the electric con- 

 ductivity. Perhaps I may help him a little out of his difficulty by 

 suggesting that, alter all, part of this 25 per cent, excess may be 

 due to experimental error. When a student whose work is 

 hitherto unknown publishes results which claim an accuracy ten 

 times greater than any previously attained, it is but reasonable to 

 expect fuller details than those given by Mr. Jones. A mercurial 

 thermometer reading with certainty to the ten thousandth of a 

 degree is an instrument unknown in England, and any systematic 

 investigation of its capabilities would add confidence to our esti- 

 mate of the results obtained. Independent of errors due directly 

 to the thermometer, I should like to ask if Mr. Jones has satisfied 

 himself that he gets the same value for the freezing-point of water 

 (on which all his conclusions depend) whatever method he uses for 

 determining it. In his apparatus the sides and bottom of the vessel 

 are being constantly cooled by the surrounding medium, the top 

 is being constantly heated by partial contact with the air, and the 

 contents are being heated by the friction of the stirrer. When 

 freezing begins, the liquid can only be kept at the same temperature 

 as the ice if this latter is disseminated throughout it in con- 

 siderable quantity and in minute particles ; but, according to my 

 experience, this is almost impossible when dealing with a large 

 bulk of pure water, even when violently stirred : the ice forms 

 and agglomerates on the sides of the vessel, leaving the liquid in 

 the middle to become superheated and give too high a reading for 

 the freezing-point. A similar phenomenon occurs to a less extent 

 with very weak solutions, and gradually becomes negligible as the 

 strength increases, and as the ice which separates forms in smaller 

 crystals which do not agglomerate so easily. An error due to 

 this cause would account for the very high values obtained by 

 Mr. Jones for weak solutions. 



