16 Mr. W. Sutherland on Ionization in Solutions 



Solutions," although the formula? were developed by means 

 of approximations applicable only to fairly dilute solutions, 

 yet these could be carried up to strengths of some gram- 

 equivalents per litre, and there appeared in the whole in- 

 vestigation no necessity to distinguish between ionized and 

 non-ionized solute. Complete ionization was tacitly assumed 

 throughout, and led to no result in conflict with experience. 



We can test (13) as to the relation between conductivity 

 and viscosity at different temperatures by much appropriate 

 experimental material, especially the determinations of Kohl- 

 rausch over a small range of temperatures for a large number 

 of solutions at infinite dilution (Sitz. d. K. Preuss. Akad. 

 1901, p. 1026), those of Lyle and Hosking for NaCl up to 

 100° 0. (Phil. Mag. [6] iii. p. 487), those of Hosking for 

 LiCl up to 100° (J. (vii. p. 469), and those of Noyes and 

 Coolidge (he. cit.) forNaCl and KC1 up to 306° C. Accord- 

 ing to (11) and (12) \ r) Q f° r a given solution ought to be 

 constant at all temperatures, unless some ionic parameter on 

 the right-hand side of (12) varies with temperature. That 

 some parameter does vary with temperature is shown by the 

 result of Lyle and Hosking for NaCl up to 100°, namely 

 Ao?7 = 1-189(1- 0-00174*). Hosking's result for LiCl is 

 similar. The work of Kohlrausch gives the temperature 

 coefficient for a large number of ions. From 0° to 36° he 

 writes 



t \=^ [l + «(t-l$)+!3(t-lSy-], . . (16) 

 and finds 



/3 = 0-0163(<2-0'Q174) or 0*0177(a-0*0177). (17) 



He finds, moreover, that just as X for a solute is separable 

 into two parts, Aoi and A02, each characteristic of an ion, 

 A i for each ion has its own characteristic coefficient a. As 

 regards the variation of r] with temperature we can write 

 between 0° and 40° 



i/pi =[i+v(t-i*)+s(t-im/i*vo. • • (is) 



Bousfield (Proc. Hoy. Soc. lxxi. p. 47) obtains 7 = 0*0251 

 and 8 = 0-000115. From (16) and (18) 



A o ^ o =i8X ol 8^o[l+(«-7)(^-18)H-(/3-S-a 7 + ^)0-18) 2 ] 



. . . (19) 



The values found by Kohlrausch for a are 0*0154 for H, 

 0*0174 for OH, and then values ranging from 0*0203 for 

 N0 3 to 0*0269 for C0 3 . Now I have sought to show in " The 

 Molecular Constitution of Aqueous Solutions " that the H 



