SVANTE ARRHENIUS WALKER 721 



an electrolyte is constituted in such a manner t^at only a certain 

 fraction 1/n can at the same time take part in such a movement, 

 it is evident that its coefficient of activity is I/71. It is not neces- 

 sary, however, that a chemical difference should exist between the 

 active and inactive parts. For greater clearness we choose an ammo- 

 niacal solution as example. In this solution there are two different 

 parts, one active NH^OH, the other inactive NH3. If the latter is 

 transformed into the former, the sum of the molecules of the two 

 species is not augmented. Thus if -ni and n are the numbers of 

 molecules of NH^OH and NII3, the first factor of the coefficient of 



activity will be — , — Now several of the NH.OH molecules may 



be associated with each other, so that the number of physical mole- 

 cules of NH.OH is p, of (NH.OH). q, of (NH.OH) 3 r, etc., where 

 evidenth'^ p + 2q + ^?'+ . . . = m. Again of the molecules NH.OH 

 only a fraction 1/A presents a simultaneous movement of ions. The 

 corresponding numbers for (NH.OH) o and (NH.OH) 3 are 1/jx and 

 1/v. In this case the coefficient of activity of the ammonia will be 

 equal to 



m 



m + n\m\ 



m/i mv / m + n\\ iJL V J 



It is interesting to compare with this coefficient of activity the 

 " dissociation ratio " of Lodge, which is defined in the report from 

 which I have already quoted. Lodge writes : " inn^ is the number 

 of grams of the electrolyzed or dissociated substance in a unit cube, 

 and this we may write iVyu, where N stands for the number of monad 

 gi'am-equivalents of the really electrolyzed substance per cubic centi- 

 meter and /A is its molecular weight compared with hydrogen." Con- 

 sidering the case of two electrolytes dissolved in the same solution, he 

 proceeds : " tliere will be N -^ and N^ to represent the amount of dis- 

 sociated substance present, reckoned in gram-equivalents per cubic 

 centimeter of solution. We come to the conclusion that we do not 

 know the absolute velocity of any ion, and can not know it without 

 further information regarding the dissociation ratio (that is, NJN' 

 or NJN') of each substance present, where N' is the total number 

 of monad gram-equivalents of the dissolved substance in a cubic 

 centimeter of solution." 



To Lodge the " dissociation ratio " is in all probability small. 

 Arrhenius, on the other hand, contemplates the variability of the 

 " coefficient of activity " with dilution and the likelihood of its being 

 large in very dilute solutions. 



So far the considerations are purely theoretical ; now comes an 

 important step, their union with experimental data. Kohlrausch 

 had shown that the molecular conductivity of an electrolyte was 



