PAPERS ON PHYSICS 



303 



tos and cooled directly to -180° C by means of liquid air. 

 The temperature of A was allowed to fall until the tone 

 emitted was just maintained. This by repeated trials was 

 found to be at 1 C. Thus the temperature difference when 

 B was cooled to -181°C, for this particular tube, was found 

 to be 182 -C. In observations 2 and 3 the part B was placed 

 in a special copper tube designed by the author^ some years 

 back for the determination of intermediate temperatures. 

 Observation 4 was for B at room temperature (note that 

 it was now necessary to heat A to 448^ C), while in No. 5 

 the part B was warmed up to 57 °C, and the tip heated elec- 

 trically to 527 °C before the tube responded. 



The absolute temperatures of the part B are given in 

 the second last column, while the corresponding vibrations 

 per second are listed in the last column. These data are 

 represented graphically in Fig. 2, in which the total tem- 

 perature differences as ordinates are plotted against abso- 

 lute temperatures. The relation is strictly linear except 

 possibly for the last reading at 330" absolute temperature. 

 By extending the straight line to the left we are able to 

 determine the temperature difference that should maintain 



c 



IP 



40 BO 120 >«0 ZK IK its iU 



Deoreei flbsoluto v^cale 



FIGURE 2. 



the tone when B is cooled to absolute zero. For this par- 

 ticular tube the graph shows this temperature difference to 

 be about 80 = C. 



1. Phys. Rev., Vol. XV, p. 125. 



