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



applied to wires for which pt° was numerically less than 

 273°. When, therefore, I succeeded in obtaining in 1892 a 

 very pure specimen of wire, with the coefficient c = '00389, 

 pt° =257°, I quite expected to find it behave like iron and 

 tin, w T ith the opposite curvature to the impure platinum, and 

 a negative value for the coefficient d. On testing it at the 

 S.B.P. and also at the Ag.F.P. I found, on the contrary, that 

 it gave a value d=+l*50, and that its scale agreed very 

 closely with that of all the other platinum wires I had tested, 

 at least at temperatures above 0° 0. I sent a specimen to 

 Prof. Fleming shortly afterwards and he used it as the 



mo- thermometer P 9 " in his researches on the thermo- 



c 



electric properties of metals at low temperatures. The test 

 of this wire is given by Fleming in the Phil. Mag. July 

 189 5, p. 101, from which the following details are extracted: — 



c = -003885,jrt° = 257°-4. C0 2 B.P., pt= -81°-3. 

 O.B.P.,^=-193°3. 



Assuming *=-182°-5 at the O.B.P., we have ^=+2'10, 

 which gives £=— 78 c, 4 for the temperature of solid C0 2 . 

 The value of the difference-coefficient, so far from vanishing 

 or changing sign, appears to be actually greater at very low 

 temperatures. According to this formula, the resistance of 

 the wire tends to vanish at a temperature t° = — 240 o, 2, cor- 

 responding to 2)t°= — -257°-4. It seems not unlikely, however, 

 according to the observations of Do war, that the resistance, 

 instead of completely vanishing at this temperature, which is 

 close to the boiling-point of hydrogen, ceases to diminish 

 rapidly just before reaching this point, and remains at a small 

 but nearly constant value, about 2 per cent, of its value at 0° C. 

 Application of the Difference-Formula to the case of other 

 Metals. — The application of the difference-formula is not 

 limited to the case of platinum. It affords a very convenient 

 method of reduction of observations on the resistance-varia- 

 tion of other metals. I employed it for this purpose in the 

 comparison of platinum and iron wires *, as a means of veri- 

 fying the suitability of the parabolic formula for the expres- 

 sion of variation of resistance with temperature. Thus, if the 

 symbol ft stands for the temperature by an iron- wire thermo- 

 meter, defined by formula (1), in exactly the same manner 

 as the platinum temperature, and if d and d' stand for the 

 difference-coefficients of platinum and iron respectively, as- 

 suming that both wires are at the same temperature t, we 

 have clearly the relation 



ft-pt={d-d f )Xp(t). 



* PM1. Trans. A. 1887, p. 227. 



