THERMAL CONDUCTIVITIES OF SOME ELECTRICAL INSULATORS. . 451 

 " Fundamental coefficient"* 



for AC coil = -r = -003775, 

 ' 



for CB coil = ^|^ = -003780. 

 " Fundamental zero " 



for AC coil = . QQ3 1 775 = 264-9 pt. degrees, 



for CB coil = . 00 3 786 = 264-6 pt. degrees. 



These results show that the two coils are practically identical in their pro]>ertie8. 

 Taking the means of the constants of the two coils we have temperature t f on 



"R 1 '"'I'll 



platinum scale corresponding to resistance R = - Hence temperature of 



liquid air = ' 3265 ~ / }'f 491 = -200'6 pt. degrees, which corresponds to -186'5 C., or 



86'5 absolute. Hence for the 8 of CALLENDAR'S formula 



-- _, 



100 / 100 



where t is the temperature on the hydrogen, t f that on the platinum scale, we have 

 8 = 14-l/(2-865x 1-865) = 14*1/5-35 = 2'64. 



Hence tlie platinum wire agrees closely in its properties with that used hy 

 DEWAK.t and the platinum temperature scale given hy it will not ditt'er materially 

 from the normal platinum scale. 



It was originally proposed to reduce all temperatures to the normal scale of the 

 hydrogen thermometer, but since the 8 of CALLENDAK'S formula in the case of pure 

 platinum comjmred with the hydrogen thermometer at 0, 100, and the boiling- 

 point of sulphur, appears to have the value 1'5, and when compared at 100, 0, and 

 the boiling-point of liquid oxygen to have the value 2'G, it is still somewhat uncertain 

 what the relation between the platinum and hydrogen thermometers is at low 

 temperatures. This fact seemed to render it advisable to give the results in terms of 

 the platinum scale, and along with them the results of conversion to the hydrogen 

 scale, on the assumption that CALLENDAR'S 8 is constant. When the connection 

 between the two scales at low temperatures is more definitely known, the results can 

 then be more accurately calculated in terms of the hydrogen scale. 



* For pure platinum this should Iw -00389. 

 t 'Roy. Soc. Proc.,' vol. 68, p. 363, 1901. 

 3 If 2 



