290 MORRIS LOEB 



We can now calculate the ion's velocity at 0, from the values 

 of n at that temperature for argentic nitrate and pseudo- 

 cumol-sulphonate (Table II), and [120] the values of X re- 

 duced from Table III by the ratio ^~ (Table IV). 



A 25 



1242 X 0.461 X 0.548 = 314 



842 X 0.727 X 0.517 = 317 Mean = 315.5. 



We again call attention to the smallness of the deviation. 

 Now, since \=u+v, we can obtain the velocities of all our 

 negative ions, by subtracting the velocity of the silver from 

 the respective conducting powers. In Table VI we find on the 

 first line the velocities at 25; on the second, those at 0; on 

 the third, the coefficient of temperature between 25 and 0, 

 as calculated from the formula v = v^[l+a(t-^5)], a being 

 the coefficient: all the values must be multiplied by the 

 factor given in the final column. 



x io- 8 



X IO- 8 



x 10-* 



X 10-8 

 X 10-8 



x 10-* 



The close agreement of the velocities shown in line I with 

 the observations of the rates of transfer and of the conduc- 

 tion, is proved by the fact that the calculated and observed 

 values agree within one half per cent in all cases. A glance at 

 Table VI, in which the velocities are arranged according to 

 magnitude, will bring out a striking relation of the coefficients 

 of temperature. The coefficient of temperature decreases when 

 the velocity increases. We may add that the coefficients of 

 temperature for the monovalent ions OH and H, which 



