Pressures of Saturation of Oxygen. 455 



below 20 millim. undetermined. Besides, the points of his 

 curve are often rather far from their places on the ideal line. 

 Finally, Wroblewski started with the assumption that the 

 Irydrogen-thermometer gives too low results at temperatures 

 lower than —193°, and used a thermoelectric junction to 

 measure such temperatures ; in consequence of which, for 

 instance, the temperature — 207 o, 2 of the hydrogen- thermo- 

 meter was only — 200 o, 4 according to the thermoelectric 

 junction ; these differences changed irregularly with the 

 temperature. The measurements by means of a hydrogen- 

 thermometer have not yet been systematically carried out. 



We have no reason to suppose that the hydrogen-ther- 

 mometer fails at very low temperatures "* ; Prof. Olszewski's 

 experiments with thermometers filled with nitrogen, oxygen, 

 and nitric oxide t, as well as Dr. P. Chappuis's with a 

 carbon-dioxide thermometer J, showed that gas-thermometers 

 give temperatures in agreement with true data, even at 

 temperatures which are much lower than the critical tem- 

 perature of the corresponding gas. Moreover a comparison of 

 the hydrogen-thermometer with an electric thermometer based 

 on an increase of electrical conductivity when the temperature 

 is diminished, shows that both are pretty well in agreement ; 

 the line representing the variation of electrical resistance with 

 the variation of temperature is nearly straight §. If we use 

 Wroblewski's determinations (taking his results to — 193° 

 to be in agreement with the hydrogen-thermometer) to 

 draw the line of resistances, on approaching — 200° we see an 

 inflection of the line down towards the axis of temperatures, 

 contrary to its whole course as far as this temperature. 



If we draw by means of the numbers given by Wroblewski 

 the curve of the boiling-points of oxygen, and compare it 

 with the curve we have obtained from measurements with a 

 hydrogen-thermometer, we at once see the very irregular 

 shape of his line, and the disagreement of temperatures. The 

 temperatures which he assumed to be in agreement with the 

 hydrogen-temperatures to —193° are even at the tempera- 

 ture of — 190°*5 (corresponding to the pressure of 90 millim.) 

 lower by 8 C than those of the hydrogen -thermometer. This 



* A-Vitkowski, Rozpraicy (Transactions) of the Crac. Acad. T21 iii. 

 p. 380 (1891). 



t Transact, of the Crac. Acad. xiv. p. 283. Also Wied. Ann. xxxi. 

 p. 69 (1887). 



X Tiav. et Mem. du Bureau Intern, d. Poids et Mesures, vi. pp. 92- 

 104 (1888). 



§ Cf. Dewar and Fleming, < On the Electr. Resist, etc./ Phil. Mag. T51 

 xxxvi. p. 271 (1893). 



