THE MEASUREMENT OF RESISTANCE 225 



furnace gases which at high temperatures penetrated the iron 

 tubes then employed to protect the resistance coils. IH view 

 of these facts, the coils are now wound on very light frames of 

 mica, which touch each turn of the wire at only four points; the 

 areas of the surfaces of contact are thus reduced to a minimum. 

 The coils are now protected by porcelain tubes glazed externally. 



In order to make the relation of resistance to temperature 

 comparatively simple and one which may be determined by 

 measurements at three known temperatures, pure platinum 

 must be employed. 



If the resistance of the coil be measured at and at 100C., 

 the average change per degree will be 



Rioo RO 



and if R be the resistance at some other temperature, 



R-Ro 



100 



R\OQ /to 



will be the corresponding temperature, on the assumption that 

 the change of resistance is linear. A temperature so defined is 

 called the "platinum temperature," so 



/ ._ inn R - RQ 



lp 1UU D r> 



/1 100 /t<) 



Callendar showed that for high temperature measurements it 

 was not correct to assume a linear variation of the resistance, 

 but that a correction must be applied in order to give the proper 

 temperature on the scale of the gas thermometer. He found that 

 if t be the temperature on that scale, the difference of t and t p 

 could be expressed by the following empirical relation: 



where if pure platinum be used 5 is a constant, its value being 

 about 1.505. This form of expression holds with great exactness 

 for pure platinum, but is not general; for instance, it fails in the 

 case of palladium. For impure platinum 5 must be expressed 

 thus: 5 = a + bt. 



15 



