NOVEMBEB 4, 1904. J 



SCIENCE. 



579 



the soiutions of these salts. And the sec- 

 ond is that the migration velocities of the 

 two ions of a salt vary by the same per- 

 centage amount, if they vary at all, with 

 changes in its concentration. It is scarcely 

 admissible, however, to regard this last fact 

 even as an indication that the hypothesis 

 of constant migration velocities is correct; 

 for any change in the character of the 

 liquid medium might well affect the veloci- 

 ties of different ions not far from equally. 

 Important evidence in regard to this 

 hypothesis and that stating that ions and 

 the un-ionized molecules associated with 

 them have a normal osmotic pressure is, 

 however, furnished by the agreement of the 

 ionization values derived, on the one hand, 

 from the conductivity-ratio, and, on the 

 other, from the properties thermodynamic- 

 ally related to osmotic pressure. Three of 

 these properties have been measured with 

 sufficient accuracy with certain electrolytes 

 to make the results of significance, namely, 

 the freezing-point lowering, the electro- 

 motive force of concentration cells, and the 

 heat of solution in relation to change of 

 solubility with the temperature. ^ Under 

 the assumption that osmotic pressure and 

 gaseous pressure are equal under identical 

 conditions, a relation between each of these 

 properties and the degree of ionization of 

 an electrolyte can be derived with the help 

 of the second law of energetics. Then, 

 either this ionization value may be directly 

 compared with the conductivity ratio ; or, 

 assuming provisionally that the latter is a 

 correct measure of ionization, the magni- 

 tude of the property in question may be 

 calculated, and the result compared with 

 that obtained by direct measurement. In 

 the case of the freezing-point lowering, I 

 have adopted the first of these methods. 

 For the five salts for which both reliable 

 freezing-point determinations and accurate 

 conductivity measurements at 0° exist, the 

 ionization values corresponding to both of 



these properties have been computed. Es- 

 pecial attention was given to the selection 

 of the best value of the freezing-point 

 lowering constant and to the extrapolation 

 of the conductivity for zero concentration, 

 the details of which can not be here de- 

 scribed. The results may be summarized 

 as follows: 



In case of the two uni-univalent salts 

 and the three uni-bivalent salts hitherto 

 carefully investigated, the ionization values 

 derived from freezing-point lowering do 

 not differ from those derived from con- 

 ductivity, between the concentrations of 

 1/200 and 1/4 normal, by more than two 

 or three per cent. 



The five salts referred to are potassium 

 and sodium chlorides, potassium and so- 

 dium sulphates and barium chloride. The 

 two sets of values for potassium chloride, 

 for which an abundant experimental ma- 

 terial exists, exhibit no pronounced or sys- 

 tematic differences; but for the other four 

 salts the freezing point leads to values 

 which are in general from two to three 

 per cent, higher at all concentrations than 

 the conductivity ratio. The fact that these 

 differences do not, as a rule, increase with 

 increasing concentration indicates that they 

 may be due to some constant experimental 

 error, or to an error in the extrapolated 

 conductivity value. 



Accurate measurements have been made 

 by Jahn of the electromotive force of con- 

 centration cells consisting of two silver or 

 mercury electrodes covered with silver 

 chloride or mercurous chloride, one of 

 which is immersed in a weak solution and 

 the other in a strong solution of sodium or 

 potassium chloride. These measured values 

 were compared by him with those calcu- 

 lated from the thermodynamic relation be- 

 tween electromotive force and the concen- 

 trations and degrees of ionization of the 

 salt in the cell. Unfortunately, however, 

 the thermodynamic relation employed in- 



