6 Sosman — Platinum- Ji hod i uni Thermoelement. 



own elements, on which we have complete data, usine; for this 

 purpose the portion of the curve from 1100° to 1550°. A 

 parahola passed tlirough the meltinj^ points of copper (1082"6°), 

 diopside (1391'2°) and palladium (1549-2°), gives, in the case of 

 the various 10 per cent elements, values for platinum from 

 1748° to 1753°; the 1 per cent alloy gives 1750°-1755° (low 

 sensitiveness); the 5 per cent, 1752°; and the 15 per cent, 

 1755°. The mean is 1752°. 



In the preceding paper it was pointed out* that the avail- 

 able optical determinations of the melting points of palladium 

 and platinum agree very well as to the difference between 

 these two points, althougli disagreeing as to their absolute 

 value. The mean value of this difference is 206°. Having 

 fixed the palladium point on the nitrogen thermometer, we 

 were able therefore to give tlie melting point of platinum as 

 1755° with an estimated accuracy of 5°. The extrapolated 

 value in the preceding paragraph agrees with this figure, 

 1755°, within the estimated limit of accuracy. 



A parallel case to this is found in the values obtained by 

 extrapolating the resistance thermometer above 500°, which 

 come remarkably close to those obtained by the nitrogen 

 thermometer, but which should not on that account be given 

 equal weight with a real measurement. There is no reason to 

 expect that such extrapolation will give true results, for the 

 downward extrapolation of the thermoelement equation, and 

 likewise of the resistance thermometer,! goes far astray. The 

 extrapolated values given above for platinum should therefore 

 be considered as merely confirmatory, not as having any 

 independent weight, and we have accordingly, in calculating 

 the interpolation curve for the thermoelement from 1550° to 

 1750°, used our original value of 1765° for platinum.;}: 



The summarized temperature scale adopted for present use 

 in this laboratorj'- for the calibration of thermoelements is as 

 follows: 



Ice, m. p. 0° 



Water, b. p. 100-0°+ 0-037 (p-760) 



Naphthalene, b. p. 217-7° + 0-057 (p-760) 



Benzophenon, b. p. 305-4°+ 0-063 (p-760) 



Cadmium, m. p. 3J0*2^ 



Zinc, m. p. 418-2° 



Antimony, m. p. (in CO) 629-2° 



Silver, m. p. (in CO) 960-0° 



Gold, m. p. 1062-4° 



Copper, m. p. (in CO) 1082-6° 



*Day and Sosman, loc. cit.. p. 160. 



f Travers and Gwyer, Zeitschr. pliys. Chem., Hi, 437, 1905. 



X In view of these facts, it is evident that Harker's figure, 1710°, obtained 

 by extrapolating the curves of a number of platinum-rhodium and platin- 

 iridium thermoelements from 1100° up to their melting points, has little or 

 no value as an estimate of the melting point of platinum. 



