Dissociation of Compounds in Water. 495 



the value found by Bunsen* being 79*7. This would diminish 



2T 2 



the constant as calculated from the formula tt^tt? • Using 



100L & 



the above value, 79*7, the constant as calculated would be 

 1*87. The value found is in reality divided by 10 for the 

 sake of direct comparison. 



The agreement, then, is very satisfactory between the 

 constant as calculated and that found from the above three 

 compounds. 



Very recently Mr. C. Dietericif has suggested some points 

 which he thinks would slightly affect the accuracy of my 

 results as obtained from this improved freezing-point method,, 

 The very sensitive thermometer, he thinks, would detect the 

 changes in the freezing-point of the solution produced by the 

 changes in the barometric pressure. A change of an entire 

 atmosphere would alter the freezing-point only about 0*0075°. 

 The maximum time between the determination of the freezing- 

 point of the water and of the last solution in which that 

 water was used was six hours, since a new litre of water 

 was employed and its freezing-point determined at the 

 beginning of each day's work. The freezing-point of the 

 pure water was determined about 12 noon, and the last deter- 

 mination with this water as solvent was made about 6 p.m. 

 The barometer is read at 2 p.m. and 8 p.m. daily at the 

 meteorological station, near the laboratory, so I am enabled 

 to calculate the magnitude of the error from this source. 

 From the data which have been kindly furnished me at the 

 station, the greatest change in the barometer between 2 p.m. 

 and 8 p.m., from the 1st of November, 1892, to the 1st of 

 August, 1893, the time during which my measurements were 

 made, was less than 10 millim. Hg (7*8 millim.). A change 

 of 10 millim. Hg would alter the freezing-point only 0'0001°, 

 and this is greater than the maximum change even for one 

 day. The average change during this same time was less 

 than 2 millim. Hg, and the mean error from this cause was 

 therefore less than 0'00002°. 



The second point, that different layers of the solution have 

 slightly different freezing-points due to hydrostatic pressure, 

 has no application whatsoever to my work, because the 

 thermometer remained always at the same depth in the 

 solution. The thermometer was firmly fixed in cork c (Zeits. 

 phys. Chem. xi. p. 532, fig. 1) , and was not removed from it 

 during the entire time in which this work was in progress. 



* Bunsen, Pog°> Ann. cxli. p. 1. 

 t Dieterici, Wied. Ann. 1. p. 77. J 



