348 JOHN W. SHIVE 



fluence of the freezing mixture ceases and that of the external 

 temperature conditions begins, and a correction for the variation 

 in the mercury column can be only roughly approximated. The 

 thermometer employed in the present work exposed, to temper- 

 ature conditions external to the freezing chamber, a mercury 

 column somewhat more than 25 cm. in height. 



Since the influence of external air temperature upon the read- 

 ing of the Beckmann thermometer seems not to have been em- 

 phasized in connection with this sort of work, the data of table I 

 are here presented to show the nature of this influence, these data 

 being representative of a much larger number of determinations. 

 The value of the freezing point of the solvent, as used in the 

 computations of the table, was determined with a room tem- 

 perature of 20°C., and the corresponding values for the solu- 

 tions were determined with varying air temperatures, the ap- 

 parent depressions being, in each case, the difference between 

 the freezing point of the pure solvent (air temperature 20°C.) 

 and that of the solution. Similar variations of the apparent 

 from the true depression must, of course, result in every case 

 where the air temperature at the testing of the solvent differs 

 from that at the testing of the solution. When pure water is 

 tested with two different air temperatures an apparent depres- 

 sion of its freezing point is evident, and this well emphasizes the 

 point here under discussion. It should be noted that none of 

 the diffusion tensions given in the last column of table I are cor- 

 rect, since none of them were determined from freezing point 

 determinations with air temperature at 20°, the air temperature 

 prevailing at the determination of the freezing point of pure water 

 here used in calculation. These numbers represent only apparent 

 diffusion tensions at the freezing point. 



From table I it is clear that air temperature is not to be neg- 

 lected in such determinations as are here undertaken. It was 

 impracticable to maintain the room temperature constant, and 

 the difficulty was obviated by reading the freezing point of pure 

 water at the beginning of each day's operations and repeating 

 this determination at intervals of an hour or two, according to 

 fluctuations of the air temperature. The depression was deter- 



