1920] on The Blue Sky and the Optical Properties of Air 159* 



spectra of the sun and stars in the extreme ultra-violet region, using 

 for the purpose a reflecting telescope, and prisms and lenses made of 

 quartz or Iceland spar. In this way the absorption of a glass objective 

 was avoided. He noticed in 1890 that the spectrum of Sirius showed a 

 number of bands near the extreme limit of atmospheric transmission, 

 the bands tailing off into complete absorption. 



These bands were observed and discussed by other authors, but 

 no definite conclusion was reached as to their origin until 1917, when 

 the matter was taken up by my colleague, Prof. Fowler, and myself. 

 Our interest was stimulated by an excellent photograph of the bands,. 

 taken at Edinburgh Observatory under Prof. Sampson's direction, 

 which I show on the screen. We found that the same bands were 

 present in the solar spectrum. It may seem strange that this had 

 not been observed long ago, considering how closely the solar spectrum 

 has been scrutinised for more than a generation. As a matter of 

 fact this is one of the cases where a powerful instrument is a positive 

 disadvantage. The bands are diffuse, and under high dispersion 

 they are unrecognisable. In any case they are less conspicuous than 

 in the spectrum of Sirius, because in the sun numerous metallic lines 

 are superposed upon them and distract the eye. 



Xow the position and general aspect of these bands suggested 

 that they were connected with the absorption which terminates 

 the spectrum. This led us to suspect that they were due to ozone, 

 and the suspicion was readily confirmed by experiment. Burning 

 magnesium ribbon gives a convenient source of continuous spectrum 

 in the ultra-violet region. Interposing a long tube containing ozone 

 between the burning magnesium and the slit, a series of bands was 

 photographed which exactly corresponded to those photographed in 

 the solar spectrum with the same instrument, as you will see in the 

 slide shown. 



Absence of Ozone Near the Ground. 



AVe are driven then to the conclusion that the absence of short 

 waves from the spectra of the sun and stars is due to absorption by 

 terrestrial ozone. But it was not thought desirable to let the matter 

 rest there. It is true that many attempts had been made to determine 

 the (no doubt very small) quantity of ozone in air by chemical means, 

 but with very conflicting results, because other constituents of air, 

 such as oxides of nitrogen, are liable to produce reactions not unlike 

 those of ozone. It seemed more satisfactory to test the absorbing- 

 power of air near the ground for ultra-violet rays, to which ozone is 

 so opaque. I used for this purpose a mercury vapour lamp in a 

 quartz vessel, which is a powerful source of ultra-violet rays, and 

 observed its spectrum four miles away, so that the mass of air 

 intervening was as great as that between the midday summer sun and 

 the top of the Peak of Teneriffe, from which observations of the 

 extent of the solar spectrum have been made. The result was to 



