Platinum Temperatures. 447 



Professor Callendar gave (Call.) the formula, convenient 

 for some cases, at 



K=rv n ^, (5) 



where a and (3 are constants: and Mr. E. H. Griffiths used 

 (Gt.) the more elaborate formula 



t — TX = fl-57 -1- hvS 1 -f CGT Z + dltT*, .... (6) 



where a, b, c, d are constants. 



Formula (1) is, in the light of later and more accurate 

 experiments, only a first approximation : and Professor Cal- 

 lendar (Call.) found that (2) was unsuitable for the results he 

 obtained in his experiments. 



Formulae (3) and (4), on eliminating •cr, take the form 

 (given hv Professor Callendar as well as others) of 



where a and ft are positive constants. This is the equation 

 of a parabola whose axis is perpendicular to the axis of t ; 

 and (1) it leads to a maximum value for R at t = a/2/3, and (2) 

 at lower temperatures any given resistance corresponds to tiuo 

 temperatures. Both of these statements indicate physical 

 conditions which we have no reason to suppose exist ; and 

 therefore, however closely the formulae for any particular set 

 of experiments correspond with observations within the limits 

 of the experiments, it is obvious that they can only be an 

 approximation. Professor Callendar gives the values (Call. 

 p. 220) 



*=-003,448,0; /3= -000,000,533 ; 

 whence 



*=^=3234°-5, and R=6-576E . 



No experiments have been published over a range of tem- 

 perature above the freezing-point of water so great as to 

 increase the resistance to this extent. However, in a series 

 of experiments by Messrs. Holborn and Wien extending to 

 1610° C, in the course of which a part of the apparatus was 

 cracked, thus leading to a smaller increase of resistance as the 

 temperature rose than was to be expected, they found a 

 resistance nearly equal to 6Pi , while throughout the range of 

 the 1600° of their experiments the temperature-coefficient 

 had varied very slowly and steadily and only to the amount 

 of some 20 or 25 per cent, of its initial value. It is therefore 



2 L2 



