( S4i ) 



the other observations where the temperature of a bath of oxygen 

 boiling under atmospheric pressure was determined. For then impurities, 

 nitrogen as well as less volatile substances are unavoidable. It seems 

 that the influence of the latter is paramount; all these results for 

 the boiling point are too high by 0°.3 or more. The first mentioned 

 determinations yield for the boiling point of oxygen ls2°. ( .»:> on 



the normal hydrogen scale. For the pressure 7(iO ram. is given 

 without indication of a further reduction. Our value for 760 nun. 

 mercury (at ()") i^ on the normal hydrogen scale — 183°. 030, differing 

 by 710 from the value mentioned above. One of the last deter- 

 minations is that of Grünmach '). He finds L82°.23. With the 

 correction derived by Hoffmann and Rothe") for the pentane thermo- 

 meter ( — 0°.42) this becomes — 182°. <->(>, a result which after being 

 corrected is still much too high. 



If we compare the two observations at lower pressure with those 

 of Travers, Senter and Jaquerod then it appears thai both our 

 temperatures are lower by 0M3. Hence it is clear that a systematic 

 difference exists between the two series. 



i) Berliner Sitz. Ber. 1906. 



*) Zeitschr. f. Instrkd. 27. 1807. 



(November 26, 1908). 



