64 Mr. Rollo Appleyard on a 



The Platinum Coils, 



There were six platinum coils, forming six thermometers, 

 each having a resistance of about 7 ohms. Each consisted of 

 about 2 metres of pure platinum wire, of 8 mils diameter, 

 wound in double spirals upon an ebonite rod. Thick copper 

 leads were soldered to the ends of the coils. To determine a. 

 and R 32 the six coils were separately tested in melting ice, and 

 then at various temperatures in heated water, side by side with 

 a mercurial thermometer. These measurements were made 

 with the coils submerged in a large tank, the temperatures 

 being controlled by a gas-regulator. The resistances were 

 measured on an ordinary "dial" bridge having german- 

 silver coils. This part of the work occupied much time : a 

 was derived from the mean of the best of 30 determinations ; 

 it was somewhat tedious, but the values R 32 and a. proved, in 

 the result, very accurate. The six coils of the six thermo- 

 meters were afterwards connected in series, and regarded as 

 one thermometer. The mean values were : — 



R 32 =4005 ohms (six coils, in series). 

 a= -002097. 



The six platinum coils, still connected in series, were then 

 imbedded within their respective dielectrics. Holloiv metal 

 tubes, for mercurial thermometers, were inserted at the same 

 time within the dielectrics, for check-readings. The dielec- 

 trics were then immersed together into one tank of water. 



Measurement of R, by " Dial " bridge. 



The apparatus was now complete, and tentative tests were 

 made for R t at different temperatures. The results, when 

 compared with the mean of the readings of the mercurial 

 thermometers in the tubes, were rather discouraging. The 

 temperatures could not be depended upon within about 2° 

 Fahr. These errors were due either to the german-silver 

 coils, or the plugs, of the dial-bridge. Slightly better results 

 were obtained by reducing the battery to a single cell. It 

 was impossible to determine the bridge-temperature, at any 

 time, with accuracy; and, moreover, even if that temperature 

 could have been found *, the temperature-coefficient for dif- 

 ferent samples of german-silver varies so widely with different 

 samples, that no dependence whatever could have been placed 

 upon values " corrected " by text-book coefficients. The 



* A platinum coil interwound with one of the resistance-bobbins 

 would probably be the best method under these circumstances. 



