720 ANNUAL EEPORT SMITHSONIAN INSTITUTION, 192 8 



manner. It is impossible to deny the complete analogy of this 

 phenomenon with that occurring on the dilution of ammonia or 

 acetic acid, although it takes place much more rapidly." He sums 

 up in the following statement : " The aqueous solution of any hydrate 

 [i. e., acid or base] is composed in addition to the water, of two parts, 

 one active (electrolytic), the other inactive (nonelectrolytic). These 

 three substances (viz, water, active hydrate, and inactive hydrate) 

 are in chemical equilibrium, so that on dilution the active part 

 increases and the inactive part diminishes." 



Arrhenius gives nowhere in this memoir a precise account of the 

 nature of the active and inactive portions. He indicates and illus- 

 trates what they might be, but that is all; he does not define. The 

 most important special feature of the paragraph is the statement that 

 the active part increases on dilution. He continues : " In what respect 

 these two parts differ remains to be elucidated. Probably the active 

 part (as with ammonia) is a compound of the inactive part and the 

 solvent. Or possibly inactivity may be caused by the formation of 

 molecular complexes. Or again the difference between the active and 

 inactive parts may be purely physical. The same statement applies 

 to bases, and we may also speak of the inactivity of dissolved sajts, 

 in which case the notions of inactivity and complexity completely 

 coincide." 



With regard to solutions of normal salts he makes the following 

 statements: (1) "Aqueous solutions of all electrolytes contain the 

 dissolved electrolyte at least in part in the form of molecular com- 

 plexes. (2) If the attenuation of the solution of a normal salt is 

 continued, the complexity approaches asymptotically an inferior 

 limit. (3) The limit to which the complexity of a normal salt at 

 extreme dilution tends to approach is of the same degree for all 

 normal salts. Probably this limit wijl not be attained before all 

 the salts are split up into simple molecules, represented by the 

 chemical molecular formula, 



" To fix our ideas, I have introduced the notion of coefficient of 

 activity defined as f oljows : The coefficient of activity of an 

 electrolyte is the number expressing the ratio of the number of 

 ions actually contained in the electrolyte to the number of ions it 

 would contain if the electrolyte were completely transformed into 

 simple electrolytic molecules. 



" Before going on we shall describe more precisely the notion of 

 coefficient of activity by the aid of the hypothesis of Williamson 

 and Clausius. According to section 2 this coefficient is defined by 

 the number of ions present in a solution. But to each pair of ions 

 there corresponds an electrolytic molecule which can take part in 

 the production of a circular current; that is to say, its ions are 

 endowed with the movement assumed by the hypothesis. If, now,. 



