18 Mr. W. Sutherland on Ionization in Solutions 



justifying extrapolation for approximate purposes. To check 

 the results thus obtained, I have used another empirical 

 formula, namely : 



log v = 567/(^ + 273) -1-239. 

 In the following table are given the values of \ at the 

 different temperatures from Noyes and Coolidge for NaOl 

 and KC1, along with the experimental values of rj at 1 8° and 

 140° and the two estimated values of tj q at the higher 

 temperatures, and their mean. The last row contains the 

 products \v ' 



Table V. 



Temp 18°. 140°. 218°. 281°. 306°. 



Exper. t] -0105 '00196 



(1) ri ... -00131 -00103 -00095 



(2) j 7o ... -00122 -00091 -00081 



Mean Vo ... -00126 '00097 "00088 



NaCl. 



X 110-3 512 782 9S4 1078 



\ rj 1-16 1-00 0-99 0-95 0-95 



„ calc 1-16 1-02 0-98 0-95 0-94 



KC1. 



X 131-4 572 845 1041 1125 



\ ri 1-38 1-12 106 101 099 



„ calc 1-38 1-12 1-055 l'Ol 99 



For both solutes \ tj falls with increasing temperature. 

 But the approach to constancy in this product is sufficient to 

 show that viscosity is the chief cause of the change in X with 

 temperature. 



A glance shows that over these large ranges of temperature 

 the variation of \r) Q with temperature is not expressible by the 

 simple linear relation holding between 0° and 36°. Yet 1 have 

 found the following comparatively simple relation to hold : 



\9. = « + &{('.-*)/(« + 39)}*, . . . (21) 

 in which t e the critical temperature is 365°, and the 39 is that 

 temperature at which Kohlrausch and Lyle & Hoskmg 

 have found that water would yield dilute solutions of zero 

 conductivity and infinite viscosity, if it changed continuously 

 instead of freezing at 0°. For NaCl a = 0*90 and for KOI 

 0'91, while the values of b are 0'106 and 0*192. The values 

 of \ r) Q calc. in Table V. were obtained by (21) . This empirical 

 formula gives us the rather important result that in (11) 

 and (12) the temperature variation of the viscosity terms, 



