496 Dr. Harry C. Jones on the 



It then passed loosely through cork Gr, on which c rested, and 

 was never raised or lowered to the extent of 5 millim. The 

 freezing-point of the water was thus determined at a certain 

 depth, and that of the solution at the same depth. The 

 lowering was the difference between these two determinations 

 made under the same hydrostatic pressure. The difference 

 between the specific gravity of the water and that of the very 

 dilute solutions employed producing slight differences in 

 pressures, is almost too small to consider. The difference 

 e.g. for a 0*2 normal solution of cane-sugar at 20 centim. 

 beneath the surface would be about \ millim. Hg, and this 

 would alter the freezing-point 0*000005°. 



The third point, as to whether the thermometer had reached 

 the same temperature as the solution, can only be decided by 

 the results. Several solutions were determined several times 

 each to test this point. In some of these determinations twice 

 as much time was allowed as in others, and the maximum 

 difference between the readings was only a few ten-thousandths 

 of a degree. While I could not wait an unlimited time to 

 make the readings, because the solution became more con- 

 centrated as more and more ice separated, yet Mr. Dieterici 

 must remember that in the apparatus which was employed the 

 litre of solution cooled very slowly, and the time at my dis- 

 posal to make the readings seems to have been quite sufficient, 

 as shown oj the results. 



Thus the three points suggested by Mr. Dieterici, as 

 probably introducing slight errors into my results, are almost 

 entirely of a theoretical character, as shown by the calculation 

 of the exact amount of error introduced by each. From his 

 final statements the author himself does not seem to attach 

 any great significance to them. 



I suggested the error 0*0002° "* simply as the half of the 

 maximum difference between the results in a series of deter- 

 minations of the same solution. Of course any error either 

 of weight or measurement in the preparation of the solutions, 

 or any traces of impurity in the compounds, which always 

 exist to some slight degree, would not be included in this maxi- 

 mum error in the determination of the freezing-point. 



To briefly summarize : — 



The dissociation of the electrolytes investigated, as calcu- 

 lated from the lowering of the freezing-point, agrees in 

 general with that calculated from the conductivity. The 



* Jones, Zeits. phys. Chem. xi. p. 112. 



