432 DETERMINATION OF FREEZING-POINT DEPRESSIONS, ETC. 



TABLE II. 



If now we plot ionization coefficients as ordinates, against 

 equivalent depressions as abscissae, it is generally assumed that 

 at great dilution we shouid get a straight line. My results are 

 too erratic to lie on a straight line, but the general course of 

 them is no doubt a straight line, and does not tend to either the 

 right or left, as do the curves of others at least up to the con- 

 centration .01. Above this it seems to have a slight rightward 

 tendency, but not nearly as great as Jones'. Of all the observer's 

 results to which I have access, and this includes Loomis, Jones, 

 Raoult, Abegg, Ponsot and Wildermann,* there are none which 

 give a curve as high or higher than mine. Jones' curve at the 

 lower part seems to coincide with mine, but from the concentra- 

 tion of about .08 to .007 it goes to the left of mine, and from 

 this on it passes away to the right. Loomis' curve is to the left 

 of both Jones' and mine and has the leftward tendency, but looks 

 as if it would pass off to the right, if dilution were carried far 

 enough. Abegg's curve is to the left of Loomis'. It starts at a 

 concentration of .07, runs parallel to mine for a space and then 

 passes off to the right. Wildermann's curve has the leftward 

 tendency, while Raoult's seems to be inclined towards the right. 

 In plotting all the above curves I have used Whetham's 

 coefficients. 



Hence it appears to rne that my results have bourne out at 

 least to a large degree what Archibald's and Barnes' results 

 seem to imply. 



These data are all takcu from MacGregor's paper cited above. 



