138 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



picnometer used was of the form recommended by Ostwald and Luther 

 and the setting of the meniscus in the capillary was made in a thermo- 

 stat kept constant at 25*^ C. For each solution, at least two weighings 

 were made, the picnometer being replaced in the thermostat and the 

 meniscus reset between weighings. The weighings were made imme- 

 diately after removing the picnometer from the thermostat and wiping 

 off the surface. The successive weighings of the same solution usually 

 agreed within two tenths of a milligram. Proper corrections were 

 made for the buoyancy of the air. The iodine was prepared by re- 

 subliming with potassium iodide. 



The results are presented in the following table : 



and this was 



The last column gives the value An/Am between a solution con- 

 taining no iodine and one of the concentration indicated. No evi- 

 dence is present of a systematic variation of Av/Am with the concen- 

 tration. Omitting the value of Av/Am for solution no. 2', the average 



^^r.^. [ average dev.' 



value IS Ay/ Am = 0.2376 ± 0.0004 ( = y= — 



\ yo 



taken as the partial specific volume of iodine at great dilutions in 

 potassium iodide solution. The specific volume of pure iodine from 

 the data in Landolt and Bornstein is 0.202. 



In the following tables the results are given for some calculations 

 of Av/Am for potassium, sodium, lithium and hydrogen iodides in aque- 

 ous solution at 20° C. from the data of Wegner and of Perkin given in 

 Landolt and Bornstein. 



