76 



ROYAL SOCIETY OF CANADA 



It Avill be seen from the above tuble that in the case of the six stronger 

 solutions the differences are for the most part considerably beyond the 

 limit of the error of an observation of the eonduotivity, which tvould be 

 about 0^25 per cent, and are for the most part also of the same sign. 

 This result was to be expected, as the values of the ionization coefficients 

 were taken to be the ratios of the s])ecific molecular conductivity to 

 the specific molecular conductivity at infinite dilution, which is rigorously 

 admissible only for infinitely dilute solutions. Also, the several values ot 

 the specific molecular conductivity at infinite dilution for the electrolytes 

 in the mixture were taken to be the same as the values found by obsei". 

 vations on the simple solutions, and these values would only hold rigor- 

 ously for infinitely dilute mixtures. 



In the case of the weaker solutions, the difierences are either within 

 or^but little beyond the limit of experimental error, and arc nearly 

 equally distributed as to sign. Considering the many sources of error in 

 obtaining the data required for the calculations, this agreement between 

 observed and calculated values appears to be very satisfactory, and to 



