CONSTANT FOR ELECTROLYTES—HEBB 413 
The data, as I have indicated, are taken from the second of 
the papers referred to above. Before, however, plotting the 
ionization-equivalent depression curve, I plotted in each case the 
equivalent depression against concentiation, and drew in the 
smooth curve which best represented the results, so that approxi- 
mately as many points fell on one side of the curve as on the 
other. In drawing in this curve, however, I did not use many 
of the observations—only those of the stronger concentrations: 
This was done in order to get rid, as much as possible, of the 
error due to the rightward or leftward tendency of the observa- 
tions ; for, had I plotted all the points and then drawn in the 
smooth curve which best represented them, these latter erratic 
points would have given a rightward or leftward tendency even 
to the part of the curve corresponding to the stronger concentra- 
tions. Hence, as a general thing, I discarded all but three or 
four of the observations on the stronger concentrations. In 
cases where I have kept more it is due to one of three things : 
(1) because the points were close together; (2) because there was 
very little rightward or leftward tendency to the ionization- 
equivalent depression curve, or (3) because the ionization-equi- 
valent depression curve for these stronger concentrations had a 
rapid natural bend to it. The information in (2) and (3) was 
ascertained by a preliminary plotting of ionization coefficients 
against equivalent depressions. Having thus drawn in the 
smooth concentration-equivalent depression curve, I read off 
the values of the equivalent depressions corresponding to the 
given concentrations, and have given them in brackets alongside 
of the observations themselves. 
1 then plotted these corrected equivalent depressions against 
their corresponding ionization coefficients, and finding by inspec- 
tion the portion of the curve thus obtained, which was straight, 
I drew in the straight line which I thought best represented the 
results. In drawing in this straight ‘ine I generally gave more 
weight to points corresponding to solutions of greater concen- 
tration. This straight line was then treated as pointed out 
above, and k and J were thus obtained. 
