Practical Thermometry Standard. 



525 



galvanometer with telescope and scale for reading the E.M.F. 

 He also proved, by the direct experiment of placing a silver 

 wire in a porcelain tube in boiling zinc, that the F.P. of silver- 

 was certainly higher than the B.P. of zinc. It is interesting 

 to observe that even at this early date the thermoelectric 

 pyrometer was capable of giving results which were far 

 superior to the air-thermometer in point of consistency. 



Table I. — Comparison of Gas-Thermometers at 950° C. * 



Observer. 



Date. 



Pouillet 



1836 



Deville & 1 

 T roost ] 



1860 





1863 



„ 



1864 



Weinhold 



1873 



Deville & 



T1880 





1 » 





I „ 



Violle 



1882 

 1894 

 1899 







0., H., &N. ... 



1895 



Thermometer. 



Pore. Air. 



Pore. Iodine. 



Plat. Air. 

 Pore. Air. 

 Pore. Air. 

 Pore. Air. 



,, Hydrogen, 



„ CO,. 

 Pore. Air. 

 Pore. Air. 

 Pore. Nitrogen. 

 Plat. Res. 



B.P. of Zinc. 



1040 



(932) 

 883-897 



1035 



928-954 



916-924 



1037-1079 



930 



(930) 

 910-930 



(916) 



F.P. of 



Silver. 



1000 



917-960 

 (916) 



(954) 

 (986) 

 (968) 

 (961) 



The striking discrepancy between the results of Becquerel 

 and those of Deville and Troost was the origin of a keen and 

 instructive controversy. Deville and Troost pointed out that, 

 the air contained in Becquerel's platinum bulb had diminished 

 by nearly 20 per cent, in the course of one series of observations, 

 and showed by a conclusive experiment that this was due to 

 the leakage of hydrogen into the platinum bulb at high 

 temperatures, and the subsequent absorption of the aqueous 

 vapour produced. It is impossible, for this reason, to use 

 hydrogen as the standard material in metallic bulbs, but 

 other gases, such as nitrogen or argon, can be used, if the 

 metal is not exposed to reducing gases externally. Becquerel, 

 in reply, maintained that this could not have produced an 



* The numbers enclosed in brackets were referred to the gas-thermo- 

 meter indirectly by auxiliary methods. The limits of divergence of the 

 several observations are given, wherever accessible, as an indication of 

 the order of consistency attained. 



