SCIENCE 



[N. S. Vol. XXX. No. 757 



exhibits other properties which are found 

 in all chloride solutions. In other words, 

 the solution has no properties which are 

 not included in one of two distinct sets, 

 one possessed by all acids, and one by all 

 chlorides. So it seems natural to regard 

 this solution as a mixture of two sub- 

 stances, hydrogen ion, which is present in 

 all acids, and chloride ion, which is present 

 in all chlorides. To illustrate this additive 

 property let me perform this simple and 

 familiar experiment. Here are two solu- 

 tions in alcohol, one containing cobalt ni- 

 trate, the other an equivalent amount of 

 cobalt chloride. One is red, the other a 

 brilliant blue. On pouring these two into 

 equal large volumes of water, the difference 

 in color disappears and both assume the 

 pink color which is typical of aqueous 

 cobalt solutions. Here again are three cop- 

 per salts, nitrate, chloride and bromide, 

 which in alcohol are respectively blue, 

 green and dark brown, but when poured 

 into equal volumes of water all show the 

 same blue color of the cupric ion. 



Finally, the third main argument in 

 favor of ionization is derived from the 

 electrical properties of solutions. Some 

 fifty years ago Clausius believed that the 

 conduction of electricity in electrolytes 

 affords sufficient evidence to show a 

 dissociation into ions. Whether or not 

 we accept this proof, which is perhaps a 

 little metaphysical, other experimental 

 facts such as Kohlrausch's law of the addi- 

 tivity of conductivities at infinite dilutions, 

 the agreement between conclusions drawn 

 from conductivity and transference experi- 

 ments, and the coincidence in the degree of 

 dissociation calculated from conductivity 

 and from freezing point, all give strong 

 support to the theory of ionic dissociation. 



Unfortunately, some over-enthusiastic 

 advocates of the ionic theory, not content 

 with this solid evidence, have superimposed 

 on the theory other extraneous speculations 



which, when later discredited, have in some 

 quarters brought the parent theory into 

 disrepute. For example, I may mention 

 the dictum that all chemical reactions in 

 aqueous solutions are ionic in character, a 

 notion which not only is intrinsically im- 

 probable upon theoretical grounds, but has 

 been refuted experimentally by the experi- 

 ments of Kahlenberg and others. Some too 

 zealous ionists have applied the theory to 

 highly concentrated solutions, without 

 making allowances for the deviations from 

 the laws of the perfect solution which are 

 to be expected there. It would be absurd 

 to class as a dilute solution one which is 

 five or ten times molal, yet we see attempts 

 to apply to a pair such as sodium and 

 potassium nitrates the law of solubility 

 lowering which has been obtained for ideal 

 solutions. 



In justice to the author of the ionic theo- 

 ory, it should be noted that he has had no 

 part in these attempts to stretch the theory 

 beyond its elastic limit. In his papers on 

 this subject, brief as they have been, Ar- 

 rhenius has with great fairness and ex- 

 traordinary acumen stated, as far as our 

 present limited knowledge permits, the 

 truth, the whole truth, and nothing but 

 the truth about ionic dissociation. 



Against the excess of zeal in some advo- 

 cates of the theory may be balanced the 

 dogmatism of others who for a priori rea- 

 sons have declared it absurd that a sub- 

 stance like potassium chloride, bound to- 

 gether presumably by an enormous affinity, 

 could break spontaneously into its constitu- 

 ent parts. Instead of attempting to refute 

 such circular reasoning, let us return rather 

 to the consideration of that experimental 

 material upon which are based the three 

 main arguments for the ionic theory. Here 

 a careful scrutiny reveals facts which, dis- 

 concerting as they may be, no fair advocate 

 of ionic dissociation can afford to ignore. 



