( 339 ) 



The difference between the two means is at the utmost 0.1 mm. 



1° 



which corresponds to . Thence we mav conclude that a vapour 



1 sun • ' 



pressure apparatus is extremely suited for the standardizing of tem- 

 peratures. The data obtained here show that in this respect the 

 apparatus surpasses the gas thermometer and probably also the resistance 

 thermometer. For the hydrogen thermometer, namely, the error of 2 

 adjustments amounts to .01 a 0°.()2 comp. § .'! and Comm. N°. 95 e 

 $8 and N°. I01 a §3), for the resistance thermometer to .01 comp. 

 $3 and Comm. N". 101" §4) l ). 



§ 6. The determination of the boilino point of oxygen. 



With the resistances of Pt'i given in § 4 we can by means of the 

 data of table II derive the corresponding temperatures on the scale 

 of our hydrogen thermometer II jjj. To this end we start from the 

 mean of the two data of March 25 and Dec. 18, '07, because 

 these are probably more accurate than those obtained with the thermo- 

 meter 1\. To 34.433 and 34.098 i2 correspond the temperatures 

 — 182°.460 and - 183°. 040 respectively. To the first belongs the 

 vapour pressure 806.40 m.m. (the mean from the series I and 111 

 to the second 759.61 mm. (series II). Thence follows by means of 

 rectilinear interpolation for 760 mm. at Leiden for the temperature 

 on the thermometer l>m : t = — 183 .035 and for the normal boiling 

 point (760 mm. on sea-level and 45 northern latitude) : 



t — — 183°.030 — 0°.007 = — 183. 037 



on the normal hydrogen thermometer and (comp. table XXV of 

 Comm. N°. 101 h Dec. 1907) 



ft = — 183°.()42 -f 0°.056 = — 182°.986 



on the absolute scale. 



If we take into consideration the degree of accuracy of the correction 

 to the absolute scale (comp. Comm. N°. 97 6 March '07) and the results 

 for the control determinations for the measurement of temperature 

 made as well with the same hydrogen thermometer as with different 

 ones, then it follows that this value does not probably deviate from 

 the real value by more than .03. 



') The data <>f table 111 also enable us to judge of the accuracy of the adjust- 

 ment of the resistance. Lei the error in the reading of vapour pressure apparatus 

 = 0, which approximately is permissible according to what precedes, then the 

 difference of the means for the series I and 111 must be ascribed to the error of 



the measurement of the resistance. This then would be .005. 



