466 Van Name and Brown — Ionization and 



fraction" at this concentration is only about 20 per cent."^ Qy) 

 The molecular conductivity of the solution as calculated for 

 18° from the above composition is 22 5, f or less than half the 

 value actually measured, which is 46*7. (e) The value of (F) 

 calculated by McBain is much lower than that given by the 

 electromotive force method. 



In a similar manner it is possible to calculate the approxi- 

 mate composition of cadmium iodide solutions from the experi- 

 mental data furnished by our measurements of electromotive 

 force, and by our previous study of the iodine-cadmium iodide 

 equilibrium. These calculations will be confined to the O'Ol 

 and 0*125 molar solutions, for which the data are presumably 

 most accurate. It will be assumed that Cdlg^ is the only com- 

 plex ion present in significant amount, and that the degree of 

 ionization of (Cdl2)3 is of about the same order of magnitude 

 as that of the average uni-bivalent electrolyte. 



Allowing for the effect of the excess of Cd"*"^ ions, which is 

 much larger in the 0*01 molar than in the 0*125 molar solution, we 

 may assume that the degree of ionization is 80 per cent in the 

 former and 75 per cent in the latter. An error of a few per 

 cent in the degree of ionization assumed will not greatly 

 change the results. Using these degrees of ionization, the 

 " active fractions " given on page 453 and the values of (F) 

 from Table 1, we obtain the results recorded in Table YI. 

 That these values differ greatly from those of McBain is evi- 



Table VI. 



(Cdl3') (Cd++) 



0-01 molar 2-4 6'45 



0-125 molar, 52- 39-'7 



dent. Interpolation of these results for O'l molar concentra- 

 tion gives, approximately, ^Cdlg^) = 0*045 and (T) = 0*021, 

 figures which are respectively 5 and 1*7 times those of McBain. 

 Since the values in Table YI depend upon neither freezing 

 point nor conductivity measurements, their correctness may 

 properly be tested by calculating the van't Hoff' coefiicient i 

 and the molecular conductivity for each solution. For i we 

 obtain 0*0199/0*01 = 1*90, and 0*135/0*125 = 1*08 respec- 

 tively, while the measured values (see Table Y) are 1*96 and 

 1*19.' 



* This low power to unite with iodine cannot be explained by the low ion- 

 ization of the Cdls molecules unless it is assumed that the degree of ioniza- 

 tion of the cadmium tri-iodide is about nine times larger. This is obviously 

 very improbable. 



t For the method of calculation see p. 467. The ionic conductivities here 

 used were those employed by McBain, namely, ^Cd = 51, Cdls' = 41, and 

 V — 66 '4. If the values given on p. 467 be employed the result is 21 "1 instead 

 of 22-5. 



(I') 



(Cdl.) 



(Cdl.)= 



10-5 



0-25 



0*3 



2'7-5 



7*2 



8-6 



