IN TORSIONAL OSCILLATION 119 



FIGURE 2 Showing relation of (n) and (a) with Temperature. 



270 290 310" 330 350 370 380" 410 



TIME EFFECT 



430 



450 



470 490 



In these temperature experiments, the plan adopted was 

 to raise the brass wire up to the temperature indicated, and 

 to remove it from the furnace at once, as it was found by 

 previous tests that the change was a sudden one. Wires 

 were introduced for \ hour, \ hour, 5 minutes, and 1 minute, 

 respectively, at a temperature of 400 C., and in each case 

 the same values of n got. 



LOWERING OF TEMPERATURE 



A length of original l wire was allowed to remain in liquid 

 air for over half an hour and subsequently tested, but this 

 seemed to have no effect upon the value of n ; and this result 



1 This term will be used in subsequent pages to denote a length of wire cut from 

 a coil as supplied by the dealer. 



