1892.] 



Determination of Low Temperatures. 



absent, while the experimental difficulties of the air- thermometer 

 are avoided. 



It is thus evident that a platinum-thermometer, especially 

 of the H pattern*, is a convenient instrument for the determi- 

 nation of very low temperatures where an error of 0°"5 is im- 

 material ; and the above investigation strengthens our confidence 

 in the accuracy of the method adopted for the measurement, 

 by platinum-thermometers, of temperatures over the range of 

 - 273° to + 700°. 



[Note by E. H. G., Feb. 1893.] The constants of all the 

 thermometers referred to in Table I. (with the exception of 

 Callendar's 1887 thermometer) were determined by means of 

 the same resistance box, which was one of the dial pattern, 

 constructed by Messrs Elliott. Mr Glazebrook has, since the 

 reading of the above communication, been so kind as to make 

 a direct comparison of the coils of this box and the B. A. stand- 

 ards ; and although the errors in the individual coils are small 

 their cumulative effect is, for certain values of R, considerable. 



I have applied the corrections thus rendered necessary to the 

 values of R^, R^, R^, and redetermined the consequent values 

 of 8. 



The results are as follows. 



Table III. 



* A full description of the construction of H will be found in the British 

 Association Eeport on Electrical Standards, 1890, reprinted in the Electrical 

 RevieiP, No. 670, p. 363; and a short account on p. 153, Phil. Trans. A., 

 vol. clxxxii., 1891. 



