THE MEASUREMENT OF TEMPERATURE. 71 



alterations were so small ! If constancy is to be secured the 

 protection from the furnace gases must be complete, and the 

 material on which the wire is wound selected with great 

 care. It is a significant fact that the committee pointed 

 out that the resistance of the coil which was placed in 

 a platinum, instead of iron, sheath remained prac- 

 tically unchanged, and it is probable that what small 

 change was noticed therein was a consequence of the in- 

 sufficient annealing of the wire. Again, the relation between 

 the resistance and the temperature of the wire does not 

 appear to have been determined by Sir William Siemens 

 with sufficient accuracy. The formula suggested by him, 

 R = oT^ + jlT + -y, is deduced from insufficient data. 



I extract the following" from Callendar's paper ((9), p. 1 10) : 

 " It appears from Sir William Siemens's account of his 

 experiments that they were undertaken rather with a view 

 of graduating a commercial pyrometer than of investigating 

 the law of change of electrical resistance. Temperatures up 

 to 350 C. were determined by mercury thermometers in an 

 air- or oil-bath, and it does not appear that any corrections 

 were applied to their readings. The individual observa- 

 tions are somewhat irregular and often show divergencies 

 amounting to two per cent, and over. Only three observa- 

 tions at higher temperatures are given ; they show a mean 

 deviation of about 3O C. A copper ball-pyrometer was 

 used to determine the temperatures, which are given as 

 8io°, S35 and 854°C. ; the corresponding temperatures de- 

 duced by Siemens's formula from the observed resistances 

 of the platinum pyrometer were 772 , 81 1' and 882°C. . . . 

 The resistances apparently were only measured to about 

 one per cent, in most cases, and the temperatures are only 

 given to the nearest degree." 



2. Another objection occasionally urged is the change 

 of resistance found in platinum filaments when used in in- 

 candescent lamps. A little reflection, however, will show 

 that the conditions are in no way similar. Callendar ob- 

 serves on this point ((9), p. 105) : " The sudden heating and 

 cooling of the wire when the current is turned on or off, and 

 the intense radiation which keeps the surface at a lower 



