July 2, 1909] 



SCIENCE 



tained in different ways are in complete 

 accord; the properties are additive only 

 just in as far as the electrolytes are ion- 

 ized ; the electrical properties seem entirely 

 normal. Here, then, is an enormous field 

 in which the whole theory of Arrhenius 

 may be quantitatively applied, with perfect 

 safety, to a wide variety of problems. 



Again, in the ease of strong electrolytes 

 at high dilution, the theory of ionic disso- 

 ciation is completely in accord with all 

 known facts. The agreement between 

 transference numbers measured directly 

 and those calculated from the conductivi- 

 ties at infinite dilution is eminently satis- 

 factory. There can hardly be any question 

 that with increasing dilution the ratio of 

 the molal concentration to that calculated 

 when no dissociation is assumed approaches 

 just two for binary, and just three for ter- 

 nary salts. Is there any other hypothesis 

 which will account for this cardinal fact? 



Any valid criticism of the ionic theory 

 must, therefore, be based upon its applica- 

 tion to solutions of strong electrolytes at 

 moderately high concentrations. Here the 

 problem is unquestionably one of great dif- 

 ficulty. Even the simple question of de- 

 termining the true degree of dissociation is" 

 one which still permits much divergence of 

 opinion. One of the suggestions which has 

 been made to account for some of the 

 anomalies of the strong electrolytes is that, 

 owing to the change in the mobility of the 

 ions with the concentration, the conduct- 

 ivity is not a correct measure of the degree 

 of dissociation. In some cases, especially 

 in the case of the hydrogen ion, this sup- 

 position has indeed been definitely verified. 

 Yet in the majority of cases it seems very 

 unlikely that we may thus explain the ex- 

 traordinary discrepancies such as those that 

 I have pointed out in the application of the 

 mass law to strong electrolytes. I have 

 shown in another place that while conduct- 

 ivity may not indeed be an absolutely reli- 



able measure of the concentration of the 

 ions, it furnishes, nevertheless, the only 

 way that we have of determining them. 

 Every other method which has been used 

 gives a measure, not of the real concentra- 

 tion, but rather of the escaping tendency, 

 or activity, which may not in all cases be 

 proportional to the concentration. I be- 

 lieve that we shall make no great error in 

 assuming that the degree of dissociation 

 calculated from the conductivities is in 

 most cases substantially correct and that the 

 lack of fulfilment of the requirements of 

 the mass law for strong electrolytes is due 

 to deviation of one or more of the dissolved 

 substances from the laws of the perfect 

 solution. This assumption is by no means 

 inherently improbable. In fact, in the case 

 of many non-electrolytes we find marked 

 deviations from the ideal laws even at 

 small concentrations. 



If we must conclude that one of the sub- 

 stances present in a solution of an elec- 

 trolyte is abnormal in its behavior, we are 

 inclined at first to suspect the ions, which 

 on account of their peculiar electrical con- 

 dition might be expected to differ ma- 

 terially from ordinary substances. But 

 this view proves to be untenable. As far 

 as we can judge, the ions seem to obey, at 

 lea^t to a fair degree of approximation, the 

 laws of the perfect solution, and thus we 

 are forced to place the responsibility for 

 the observed anomalies chiefly upon the un- 

 dissociated part of the electrolyte. 



The correctness of this idea may be tested 

 by applying the simple laws of solutions to 

 those cases where the properties of the 

 undissociated substances may be eliminated 

 from consideration. Every such test, which 

 our present experimental material permits, 

 substantiates this hypothesis. Thus, for 

 example, when we consider the equilibrium 

 between a dissolved salt and the ions in its 

 saturated solution we need not consider the 

 undissociated portion. Assuming that the 



