212 Prof. H. L. Callendar on Platinum Thermometry. 



assumed tentatively a much lower yalue £=945° for the Ag. 

 F.P., giving a result £=1037° for the Au. F.P., which 

 naturally does not agree with the results of subsequent work. 

 These results have since been misquoted in a manner which 

 has the effect of suggesting that the platinum thermometer 

 gives very capricious results at high temperatures. Holborn 

 and Wien, for instance, quote my value 981°'6 for the Ag. 

 F.P., obtained with the impure wire by the S.B.P. method 

 of reduction, and at the same time quote the value 1037° 

 for the Au. F.P., which was obtained by assuming the value 

 945° for the Ag. F.P. Comparing these with the values 

 obtained by Heycock and Neville with the pure w T ire, one 

 might naturally conclude, in the absence of information as to 

 the manner in which the two results were calculated, that 

 different wires gave very inconsistent results. The truth is, on 

 the contrary, that very different wires agree with remarkable 

 uniformity in giving approximately the same platinum-seal e, 

 and that they also give consistent values of t provided that the 

 reduction is effected in a consistent manner. But, although 

 it is evident that this method may be made to give consistent 

 results in the case of impure wires, it is in all cases preferable 

 to use pure wire of uniform quality. If, forinstancej a pyro- 

 meter gives a value of c less than '0035, or a value of d 

 greater than 1*70, it would be safer to reject it, although it 

 may possibly give very consistent results. Values of d greater 

 than 2*00 at the S.B.P. sometimes occur, but may generally 

 be taken as implying that the wire is contaminated. Such 

 instruments as a rule deteriorate rapidly, and do not give 

 consistent results at high temperatures. 



The Difference- Formula at Loiv Temperatures. — The suita- 

 bility of the Platinum thermometer as an instrument for low- 

 temperature research is shown by the work of Dewar and 

 Fleming, and Olszewski. It has also been adopted by Holborn 

 and Wien, in spite of their original prejudice against the 

 instrument. The first verification of the platinum scale at 

 very low temperatures was given by Dewar and Fleming, 

 whose researches by this method are the most extensive and 

 important. They found that two different specimens of wire 

 with fundamental coefficients c = "00353, and c = "00367 

 respectively, agreed very closely in giving the same values of 

 the platinum temperature down to —220°. The values of 

 the difference-coefficients for these wires, calculated by 

 assuming t= — 182°*5 for the boiling-point of liquid oxygen, 

 are d = 2'15 * and d = 2'72, respectively. The first of these 

 refers to the particular wire which Dewar and Fleming 

 selected as their standard. 



* See below, p. 219, middle, and footnote. 



