1889.] Absorption- Spectra of Oxygen and its Compounds. 229 



and that "when they are melted the liquid and superincumbent vapour 

 are. of a very pale yellow colour. As the temperature rises both liquid 

 and vapour become, as is well known, of a deep orange, and finally of 

 a dark, reddish-brown colour. We examined the spectra produced by 

 two thicknesses of liquid and vapour — (1) by that contained in a narrow 

 tube about 1 mm. in diameter, and (2) by that in a tube about 1 cm. in 

 diameter. At 15° to 20° the vapour in the narrow as well as in the wider 

 tube showed the well-known absorption-spectrum of fine, dark lines ; 

 no absorption by the liquid in the narrow tube could be detected, 

 and the liquid in the wide tube showed no fine lines, but several 

 faint, very diffuse bands, unresolvable into lines with a spectroscope 

 of three prisms. These bands had their maxima in places where the fine 

 lines of the vapour were most intense and most closely set, so that it 

 might be inferred that they were due to similar molecules in both 

 cases, but that in the liquid the vibrations of these molecules were no 

 longer sharply defined but modified by the constraint arising from the 

 liquid state. Some parts, however, of the spectrum of the vapour, 

 where the lines were closely set, did not appear to be represented by 

 any definite bands in the liquid. The liquid absorbed a good deal of 

 blue light in a continuous manner, while the vapour only absorbed 

 it selectively. At the red end the limit of the visible spectrum was 

 lower for the liquid than for the gas, that is, there w T as more absorption 

 of red light by the vapour than by the liquid, so much so that below 

 a certain point the absorption by the vapour appeared continuous. 



The narrow tube was next immersed in a wider tube full of glycerine, 

 which was gradually heated. As the temperature rose, the colour of 

 both liquid and vapour deepened, the absorptions of the vapour were 

 stronger, and the liquid gave the same bands as had been before ob- 

 served with the greater thickness. At still higher temperature the 

 absorption of blue light, both by liquid and vapour, diminished sensibly, 

 until at 85° the groups of lines in the blue had pretty well disappeared 

 from the spectrum of the vapour. In fact, at 85° there was no sensi- 

 ble difference between the actions of liquid and vapour on blue light, 

 it seemed only some continuous absorption. At the red end the 

 difference between the liquid and vapour remained quite as strongly 

 marked as at lower temperatures, if anything, more so ; and the ab- 

 sorptions in the orange, yellow, and green were unaltered. At 90° the 

 lines of the vapour in the green began to fade, and at 100° they w r ere 

 very faint ; but those in the orange, as well as the corresponding 

 diffuse bands in the liquid, were as strong as before. There was still 

 considerably more absorption of red light by the vapour than by the 

 liquid, as if there were a strong absorption-band in the red of the 

 vapour which was absent in the liquid. 



As the temperature rose to 110° all the lines in the vapour had 

 become faint, and at 115° they were no longer discernible, and there 



