466 



SCIENCE. 



[N. S. Vol. IV. No. 92. 



in proportion to the amount of the electro- 

 lyte present, for dilute than for concen- 

 trated solutions, and for an infinite dilution 

 the molecules would, presumably, become 

 all ' active.' Arrheneus pointed out, also, 

 that there is a close connection between the 

 number of ' active ' molecules as deter- 

 mined by the electrical conductivity of so- 

 lutions and the ' avidity ' of acids as deter- 

 mined by the thermo-chemical researches of 

 Thomsen. His first explanation of the cause 

 of the difference between the ' active ' and 

 '■ inactive ' molecules, was, however, un- 

 satisfactory and was not well received. 



Shortly after, in his first development of 

 his theory of solutions, van't Hoflf was com- 

 pelled to admit that many substances in 

 aqueous solutions cause a depression of the 

 freezing point much greater than they 

 should in proportion to their molecular 

 weights. He expressed the deviation by 

 use of a factor, ' i,' which is, for electrolytes, 

 always greater than unity and expresses the 

 number of times the depression exceeds the 

 theoretical depression as calculated from the 

 molecular weight. This factor was at first 

 considered to be a constant, but it is now 

 known that it is variable and that it in- 

 creases with the dilution. The obvious 

 meaning of this factor is that the molecules 

 of electrolytes are separated into two or 

 more parts when dissolved in water, or other 

 liquids which have a similar effect in caus- 

 ing electrical conductivity. But, just as 

 chemists were very slow to see that the ab- 

 normal densities of ammonium chloride and 

 of many other substances are due to dissocia- 

 tion, so van't Hoflf did not draw a conclu- 

 sion which seemed to be so contradictory to 

 all preconceived notions about the bodies in 

 question. Arrhenius, however, saw the 

 logical conclusion, and his studies had pre- 

 pared him for its acceptance. As a result, 

 he proposed, in 1887, his theory of electro- 

 lytic dissociation." 



This theory, which seemed at first very 



improbable, has shown itself capable of co- 

 ordinating the facts of many diverse fields 

 of work and has proved more valuable in 

 the incentive which it has given to re- 

 search and more prolific of results than any 

 other theory proposed during the last dec- 

 ade. According to the theory, an electro- 

 l3^te when dissolved in water, and sometimes 

 when dissolved in other solvents, is sepa- 

 rated more or less completely into its ions. 



The empirical basis for the theory lies in 

 the correspondence between electrolytic con- 

 ductivity and the divergence from the nor- 

 mal depression of the freezing point and 

 lowering of the vapor pressure ; in the cor- 

 respondence of both with the ' avidity ' of 

 acids which has already been referred to ^ 

 in the quantitative connection between each 

 of these and the chemical efiect of acids as 

 shown in the inversion of cane sugar and 

 saponification of methyl acetate ; in the 

 satisfactory explanation which it gives for 

 the independent migration of ions during 

 electrolyses as established by the work of 

 Hittorf, Kohlrauch and others ; in the fact 

 that an electrolyte obeys the same law for 

 dissociation with increasing dilution as a 

 gas under diminishing pressure, first pointed 

 out theoretically by Ostwald and Planck,^* 

 and then experimentally established by 

 Ostwald, Wildermann,^^ Loomis,^^ and 

 others ; and in general by the fact that the 

 properties of a dilute solution of an elec- 

 trolyte are dependent on the sum of the 

 properties of the ions present rather than 

 on the properties of the chemical compound 

 which those ions may combine to produce. 

 It would take me too far to illustrate this 

 last statement as shown to be true of the 

 density, color and other properties of solu- 

 tion. 



The theory has thrown light upon many 

 chemical riddles and has placed the chemist 

 in a position to predict phenomena which 

 could formerly be known only as the re- 

 sult of . experiment. It suggests at once 



