198 Prof. C. Olszewski on the 



wave-lengths 628, 577, 535, and 480. Messrs. Liveing and 

 Dewar *, who at a later date examined the absorption spectrum 

 of gaseous oxygen in a long steel tube under a strong- 

 pressure, found the same four absorption-bands in the visible 

 part of the spectrum, and, besides them, in the utmost red two 

 others, corresponding to the lines A and B of the solar 

 spectrum, which were also observed by EgorofF and Janssen. 



The apparatus I have described enabled me to repeat my 

 former experiments and to examine more exactly the absorption 

 spectrum of a thicker layer of liquid oxygen in the utmost 

 red. 



The liquid oxygen was poured out of the liquefying appa- 

 ratus into a thin-walled glass tube, the lower end of which 

 was soldered and closely fixed into three glass vessels one 

 outside of the other, to preserve it from external heat. The 

 column of oxygen was 30 millim. thick and about 50 millim. 

 deep. In this glass tube the oxygen remained at its boiling- 

 temperature ( — 181°'4 C.) under atmospheric pressure, in 

 sufficient quantity for the experiment, during more than half 

 an hour, though it was strongly heated by a Drummond's 

 lime-light, concentrated on it by means of a collecting-lens : 

 this light was used to produce the absorption spectrum. In 

 examination of the spectrum I used a universal spectroscope 

 of Kriiss, with a prism of Rutherford. Besides the four 

 known absorption- bands, the experiment also proved the 

 existence of a fifth band, corresponding to the solar line A : 

 it is somewhat blurred, but can be seen distinctly enough if 

 a red glass is put between the source of light and the slit in 

 the spectroscope. This band appeared feebler than the 

 absorption-bands which correspond to the wave-lengths 628, 

 577, and 480, but stronger than the absorption-band at 553. 

 With this relatively slight dispersion, the band A could of 

 course not be decomposed into lines. And this time too I 

 was unable to perceive any absorption corresponding to the 

 solar B. 



The experiments in 1883 made out liquid oxygen to be 

 a colourless fluid, for but small quantities of it were then 

 obtained. Since then I have several times observed that 

 oxygen, when liquefied in wider tubes about 15 mm. thick, 

 appears of a bluish colour. During my experiments already 

 alluded to, in which for the first time a relatively considerable 

 quantity of liquid oxygen was collected in a glass vessel, its 

 bluish colour appeared quite distinctly. The oxygen was 

 prepared from potassium chlorate and manganese dioxide ; to 

 ascertain that it contains no traces of ozone from which the 

 * Phil. Mag. [5] xxvi. p. 286 (1888.) 



