THE ELECTROLYTIC DISSOCIATION THEORY 395 



Ag(NH 3 >/^Ag- + 2NH3, and by the law of mass action the 



expression A" .mu^ 3 must be constant. This was tested 

 CA g (JN H 3 ) 2 



by varying the concentrations of the ammonia and of the silver 



ions, and good constants were obtained, so that the validity 



of the original assumption is rendered probable. Complex 



ions containing ammonia, which have been investigated by 



Werner, Jorgensen, and others, have already been referred to 



(P- 387). ' 



An interesting light is thrown upon the question of com- 

 plex formation by the theory of electro-affinity of Abegg and 

 Bodlander, 1 a short account of which will now be given. 

 According to their views, complex compounds are formed by 

 the combination of ions with neutral molecules ; they only differ 

 in degree from double salts, in that the latter are more com- 

 pletely split up into their components in solution. The tendency 

 shown by the different ions to enter into complex compounds is 

 so much the greater the smaller their electro-affinity; in other 

 words, the ions which have the least affinity for electricity 

 have the greatest tendency to enter into complexes. The 

 electro-affinity is measured by the "decomposition potential" 

 (" Zersetzungsspannung ") of the ion ; it is the equal and 

 opposite potential which must be applied to discharge it, and 

 the order of the electro-affinity of the elements is the same as 

 their order in the " tension series." 2 The alkali metals have the 

 greatest affinity for positive electricity, while such metals as 

 mercury, platinum, and gold have very little. We now find 

 that, whilst the alkali metals occur almost exclusively as simple 

 ions in solution — that is, show no tendency to complex forma- 

 tion — such metals as platinum and mercury enter so readily into 

 complex ions that they are scarcely known in the free condition. 

 This is just what is to be expected on Abegg and Bodlander's 

 theory, so that the latter affords us a convenient method of 

 representing the facts, and has been of great service in the 

 experimental investigation of the question. It should, however, 

 be mentioned that the decomposition potential is no exact 

 measure of the electro-affinity, as is already evident from 

 the fact that the former alters with dilution. The difficulty 

 of obtaining an exact measure of the affinity between an ion 



1 Zeit. Anorg. Chem. 1899, 20, 453. 



2 Van 't Hoff, Lectures, iii. 90. 



