456 Thaddens Estreicber on the 



difference afterwards decreases, and amounts to 7° at the ther- 

 moelectrical temperature — 200 o, 4 ; Wroblewski did not go 

 further. Besides this irregularity, Dr. Ch.-Ed. Guillaume * 

 proved that we have no reason to distrust the hydrogen-ther- 

 mometer at the lowest temperatures hitherto reached, and that 

 if we were obliged to extrapolate temperatures, it would be 

 better to do so by measurements of the variation of resistance 

 rather than by the electromotive force of the contact ; that the 

 formula by means of which Wroblewski calculated the tem- 

 peratures was badly deduced, and, as a necessary sequence, 

 that his results do not agree with fact. 



For these reasons, Prof. Olszewski directed me to measure 

 the temperatures of saturated vapour of oxygen by means of 

 a hydrogen-thermometer. In the present state of our know- 

 ledge we might hope to obtain more positive results, were it 

 but for this reason, that the manipulation of liquefied gases in 

 quantities incomparably larger and under conditions much 

 more convenient than was possible a few years ago, now 

 presents neither great difficulties nor dangers in practice. 



Oxygen was liquefied in the apparatus of Prof. Olszewski's 

 construction f , which served to obtain large quantities of 

 liquefied gases under atmospheric pressure. From the steel 

 cylinder, in which oxygen was liquefied, it was poured out 

 into the glass vessel a, surrounded by three others in 

 order to isolate liquid oxygen from the influence of ex- 

 ternal heat, and to prevent the vessel from freezing on the 

 outside, which would render it impossible to observe the 

 liquid oxygen within ; for the same reason there was at the 

 bottom of the outermost vessel b a layer of phosphoric anhy- 

 dride. The second vessel had a brass setting above, to which 

 a cover c, also of brass, to be fixed on it by means of four 

 screws s/, was closely adapted. In this cover, four openings 

 were made : the central one, which was the largest, was 

 destined to receive a little hydrogen-thermometer t ; round 

 about it were three openings : the first served to let in a 

 narrow glass tube h, admitting a very slow stream of hydrogen, 

 which was used to aid the oxygen in boiling when the 

 pressure was very much reduced. Through the second 

 the copper tube o passed from the liquefying apparatus ; 

 this tube, closed at the end, had four lateral openings 

 nearly at the end, in order to prevent the liquid from being 

 thrown out of the vessel ; but to prevent liquid oxygen from 



* Archives d. Sc. Phys. et Natur. 1888 ; afterwards EInditstrie Electr. 

 i. p. 77. 



t Transactions of the Crac. Acad. [2] iii. p. 385 ; Bullet. Intern. 1890, 

 p. 177. Also Phil. Mag. [o] xxxix. p. 103. 



