274 Profs. Dewar and Fleming on the Electrical 



^ 



resistance of pure platinum. H. L. Callendar* and Callendar 

 and Griffiths have carefully examined this matter for tempe- 

 ratures from about zero Centigrade upwards to about 1000° C. 

 Callendar has indicated the possible causes of the discrep- 

 ances in the results obtained by previous observers in the use 

 of the platinum resistance-thermometer, as originally sug- 

 gested by Siemens ; and he has shown that a properly 

 arranged pure platinum wire may be employed to measure 

 temperatures, and may have its fixed points determined. We 

 have in like manner found that a wire of pure soft annealed 

 platinum may be cooled again and again to the lowest attain- 

 able temperatures, and heated up to high temperatures, without 

 suffering any permanent change in its specific resistance after 

 the first careful annealing has been made. This property is 

 not, however, confined to pure platinum, but is equally 

 possessed by other properly annealed metals. 



There are, of course, good reasons for giving preference to 

 platinum in the construction of resistance-thermometers : but, 

 on the other hand, it is possible that for low temperatures 

 chemically pure nickel, prepared by the process discovered 

 by Mr. Ludwig Mond from nickel carbonyl, will prove to be 

 of great utility on account of its very large mean coefficient 

 of variation of resistance with temperature. 



It is, however, clear that a wire of pure platinum which 

 has been carefully annealed possesses a definite electrical 

 resistance at each temperature to which it is exposed, and 

 may be made use of as a thermometric body with great con- 

 venience. Callendar has introduced the useful term platinum 

 temperature to denote temperatures measured with the pla- 

 tinum resistance-thermometer. On this scale the temperature 

 is defined as follows : — Let E be the electrical resistance of 

 a given wire of pure platinum at zero Centigrade, that is 

 when immersed in melting ice, and let R 100 be the resistance 

 of the same wire at 100° Centigrade. Then an increment of 

 temperature, which causes a change in the resistance of the 



platinum equal to ] l n ^ ° , is called one degree of platinum- 

 temperature. If pt denote any temperature measured on this 

 platinum-scale, and R be the resistance of the wire at that 

 temperature, then 



Hence, if we take any straight line as a temperature-axis, 



* Phil. Trans. Roy. Soc. vol. clxxiii. (1887) A. pp. 161-230, " On the 

 Practical Measurement of Temperature." 





