IP, TIIF. ABSORPTION AND EMISSION OF AIR. 



another with their tleviatioii, and are shaded off toward the red. ( tuiijilete absorji 

 tion is foiiiid with the most lefraugible of them. It is this which makes the air 

 onat[ue for all lays beyond 185 /'/'• Tlie presence of a mcxlerate amount of ozone 

 tlid not alter the absorption at all. 



The emission-spectrum of oxygen is compounded of three continuous maxima, 

 of which the most refrangible is the strongest. It lies at about 185 H^*^ The ob- 

 servation of these maxima is attended with considerable difficulty, on account of 

 their slisrht photographic efficiency and because of the bauds of oxide of carbon 

 which appear with them. It succeeds best with a Paalzow-Vogel tube and 10 to 

 l(> mm. of meicury pressure. With much smaller pressure even these tubes, 

 though completely guai'ded against fat and mercury, gave dominating bands of 

 carbonic oxide which hai'dl}' allowed the oxygen maxima to appear. It is surpris- 

 ing that I have in this region never obtained lines of oxycren. True, traces of them 

 sometimes emei'ged upon the maximum near 185 MM, but as soon as I tried to bring 

 them out by longer exposure, the bands of carbonic oxide appeared and covered 

 the supposed lines, so as to make them invisible. Nor was it possible, notwith- 

 standing numerous attempts, to bring out more refrangilde rays than those of the 

 maximum at al)out 185 /'/'■ It may well be assumed that the stratum of oxygen 

 between the capillary opening and the window of the Geissler tube contributed to 

 the ill success of this search by its defective transpai-ency. This stratum might be 

 considerably reduced in thickness, were it not that the deposit on the window 

 would then become worse. Now this inconvenience seems to come on earlier with 

 oxygen than with other fjases. For that reason, I have not reduced the thickness 

 of the stratum of this gas below 2 cm., though with others I have not hesitated to 

 go down to 1 cm. The deposit on the window takes place very easily with the 

 Paalzow-Vogel tubes, especially if they have an outei- covei'ing, and is a very incon- 

 venient adjunct of them. In this case, in a few hours' use, it often goes so far as 

 to produce complete opacity. The opening and reluting of the window thus neces- 

 sitated is not only a waste of time, but, on account of the filling witli sulphuric 

 acid, a veiy dubious ])iece of work. 



Carbon dioxide. — This gas was pi'epared by the action of hydrochloric acid 

 upon calcium carbonate. The gas was twice washed w'ith a solution of sodium 

 l)icarbonate in water in order to eliminate the hydrochloric-acid vapoi'. It was 

 dried with phosphorus pentoxide. 



The absorption-sitecti'um of cai'bon dioxide is similar to that of oxygen. Here, 

 too, there are indications of a rhythmical series in the shape of inverted groups of 

 lines. But the end of this series is consideiably more refrangible than that of 

 oxygen. Accordingly, total absorption begins at a shorter wavelength. 



The emission-spectrum is overlaid with the bands of carbon monoxide, the 

 uinvelcome attendants of all my spectra. Its photographic action is uncommonly 

 strong, and it extends far V)eyond 162 /</< into the region of the shortest wave-lengths. 

 I doubt not that, were the tubes sufficiently transparent, it could be photographed 

 as far as the hydrogen spectrum extends. For its wealth of lines it stands unrival- 

 led. I have not been able to make suie whether these bands really belong to car- 



