388 Mr. T. C. Fitzpatrick on the Action of 



and at —19° the tube was filled with a soft solid, which had 

 a considerable conductivity. The conductivity of these solu- 

 tions were measured also at 0°, and the conductivity for the 

 more dilute solution was much smaller than the temperature- 

 coefficient at higher temperature would lead one to expect. 

 I conclude that the definite compound, the cryohydrate, 

 existed at this temperature in the solution, though it did not 

 separate out till —7°. 



To further experiment on the existence of these molecular 

 groups, I have determined the conductivity of two solvents ; 

 of the solvents mixed ; of solutions of equivalent quantities of 

 the same salt in the two solvents ; and in the mixed solvents. 

 The results are given below. 



250 cubic centim. of the solution were in each case intro- 

 duced into the cell. 



Conductivity of alcohol '0000082. 



„ water '000091. 



Conductivity of 250 cubic centim. °f^ .aaaaoq 



mixed solvents J 



Calculated value '0000248. 



Conductivity of CaCl 2 alcoholic solution '00506. 

 „ „ water „ '0775. 



Conductivity of CaCl 2 solution in mixed 1 .A990 

 alcohol and water . . . . j 

 Calculated value '0206. 



There is a considerable contraction of volume on mixing 

 the alcohol and water ; for the mixed solutions of calcium 

 chloride the volume was made up to 500 cubic centim. by the 

 addition of a mixture of the solvents. 



The conductivity- values obtained differ by about 10 per 

 cent, from the calculated values, both for the mixed solvents 

 and for the mixed salt solutions. I conclude, therefore, that 

 the alcohol reacts with the salt associated with the water, and 

 vice versa ; it might be possible to obtain two solutions of 

 different strength, such that the mixture had a conductivity 

 the mean of the two, similar to Arrhenius'' isohydric solutions. 



All these experimental facts point in the same direction, to 

 the action of the solvent being a chemical one ; the salt being 

 decomposed, and molecular groups formed in the solution. 



Quite recently a paper has been published in the Annates 

 de Chimie by Foussereau"*, on the decomposition of metallic 

 chlorides in solution ; in which to measure the change in the 

 character of the solution, due to such decomposition, he deter- 



* Annates de Chimie, July 1887. 



