Jtjlt 2, 1909] 



SCIENCE 



The first of these unpleasant facts is that 

 the values for the degree of dissociation of 

 strong electrolytes calculated on the one 

 hand from freezing points, and on the other 

 from conductivities, while usually fairly 

 concordant, frequently differ by an amount 

 far greater than the experimental error. 

 For half normal solutions of lithium chlo- 

 ride, magnesium chloride and calcium fer- 

 rocyanide, the degrees of dissociation cal- 

 culated from the freezing points are 94 

 per cent., 99 per cent, and 2 per cent., 

 while from conductivities we calculate 71 

 per cent., 62 per cent, and 20 per cent., 

 respectively. Of course these are moder- 

 ately concentrated solutions and at higher 

 dilutions the discrepancies become less. 

 Moreover, it is not unlikely that the at- 

 tempts to explain such facts, by assump- 

 tions of hydration, association, and the 

 like, may ultimately be successful, but in 

 the meantime these facts can not be neg- 

 lected. 



In the second place, the additivity of the 

 properties of electrolytic solutions, striking 

 as it is, seems to prove too much. If it is 

 an- argument for the dissociation of elec- 

 trolytes, it seems to be an argument for 

 complete dissociation. Why should the 

 properties, of a normal solution of potas- 

 sium chloride be simply those of potassium 

 and chloride ions if, as measurements of 

 conductivity show, it is 25 per cent, undis- 

 sociated? Why should the undissociated 

 part have no individual properties of its 

 own? It is easy to see why completely 

 dissociated acids and bases should give the 

 same heat of neutralization, since we regard 

 this heat as simply due to the union of 

 hydrogen and hydroxide ions, but half- 

 normal potassium and sodium hydroxides 

 give essentially the same heat of neutraliza- 

 tion with an acid, although they are 20 per 

 cent, undissociated. Half-normal barium 

 hydroxide gives the same, although 40 per 



cent, undissociated. Copper sulphate as 

 dilute as one tenth normal is stiU more 

 than half undissociated, but its color is 

 nearly the pure color of cupric ion. In- 

 deed in all the strong electrolytes the par- 

 tial volume, heat capacity, internal energy, 

 viscosity, refractive index, rotary power, in 

 fact practically all the significant physical 

 properties of the undissociated part of the 

 electrolyte, seem practically identical with 

 the properties of the constituent ions. If 

 we had no other criterion for the degree of 

 dissociation, these facts would undoubtedly 

 lead us to regard salts, up to a concentra- 

 tion of normal or half normal, as com- 

 pletely dissociated. 



Finally, the phenomena of electrical con- 

 duction present several puzzling, and as yet 

 unexplained, features. For example, at- 

 tention has recently been called to the in- 

 teresting fact that the two ions which in 

 aqueous solution possess by far the greatest 

 mobility, are the ions of water itself, hydro- 

 gen and hydroxide. This might possibly be 

 regarded as chance if it had not also been 

 found that in other solvents a similar con- 

 dition exists. Thus in methyl alcohol, the 

 methylate ion moves with unusual velocity. 

 To explain this curious fact, it has been 

 suggested that the ions of the solvent have 

 a mode of progress different from that of 

 other ions, due to their ability to pass vir- 

 tually through the molecules of solvent. 

 This view, in a certain sense, requires a 

 return to a modified Grotthiis theory, and 

 if accepted, necessitates the conclusion that 

 the process of conduction is not quite so 

 simple as it may have seemed to the orig- 

 inal advocates of the ionic theory. 



Perhaps the most vulnerable point in the 

 whole armor of the ionist is reached when 

 we attempt to apply the mass law to the 

 dissociation of strong electrolytes. The 

 mass law derived rigorously only for the 

 perfect solution could hardly be expected 



