ON THE BOILING POINTS OF AQUEOUS SOLUTIONS OF ELECTROLYTES. 197 



tions being noted. It was found essential to the success of an experiment to allow at 

 least twenty minutes for the diffusion of the salt. 



Corresponding to the successive readings of the thermometer, successive readings of 

 the barometer were made, in order that any change in atmospheric pressure might be 

 allowed for. The barometer gave readings to an amount corresponding to an elevation 

 of 0*004° Centigrade, but could be estimated to the equivalent of 0'002° Centigrade. 



The salts used were supplied by Messrs Merck & Co. as specially pure, and were 

 tested analytically and spectroscopically. 



In the calculation of results ionization coefficients were necessary at high temperatures, 

 as near the boiling point as possible. These were obtained in several instances from 

 conductivity values as given by Krannhals # (later from my own observations), as 

 follows. Krannhals gives the molecular conductivity at 99*4° Centigrade for solutions 

 containing one, one-half, one-fourth, one-eighth, etc., and one-thousandth gramme 

 equivalents per litre. The ionization at any of these was obtained by dividing the 

 conductivity value, as given by Krannhals, by the value at one-thousandth of a 

 gramme equivalent. The concentration ionization curve was drawn for the range of 

 a series of experiments. From the amount of salt added to the solvent the percentage 

 composition was obtained, and from this, by the aid of tables,+ the concentration in 

 gramme equivalents per litre ; the ionization was then determined graphically from 

 the above-mentioned concentration ionization curves. Krannhals claims to work with 

 an error limit of from 2 to 3 per cent. Having repeated many of his experiments 

 I found this claim well justified. 



Schaller and Lyle, and Hosking haA^e also done some conductivity work at 

 99° or 99 "4° Centigrade. Krannhals' values were chosen because they were best suited 

 to series of experiments. 



Lyle and Hosking| deal chiefly with sodium chloride solutions. Schaller § 

 worked principally at 256, 512, and 1024 litres per gramme equivalent. 



Values up to 80° Centigrade have been given by Trotsch,|| and Campetti and Nazari,1 

 which would be too low a temperature for my purpose. Those who have given con- 

 ductivity values to 99° or 99*4° have only given them to about one-thousandth gramme 

 equivalent per litre, which dilution could scarcely be supposed to give the molecular 

 conductivity at infinite dilution for every salt. 



I have used Krannhals' conductivity values for KC1, KBr (dilute solutions), 

 NaN0 3 , KN0 3 , and NaCl (dilute solutions). 



AVhen calculating the results obtained by experiment, at first, total elevations above 

 the boiling point of the solvent were used. The calculations were made from the formula 



ot.W.E 

 C = — 1 



(1 + M - la)iv 



* Zdt. fur phys. Ghent., 5, 250 (1890). 



t B. A. Report on the Present State of our Knowledge of Electrolysis and Electro-chemistry (1893). 



X Phil. Mag. (6), 3, 487 (1902). § Zeit.fur phys. Chem., 25, 497. 



|| Wied. Ann., 41, 259 (1890). If Ace. Sci. Torino, 40, Nos. 2 and 3, pp. 155, 163 (1904-5). 



