ATMOSPHERES OF THE PLANETS — RUSSELL 155 



The celestial body which we can observe in far the greatest detail 

 tells quite another story. The moon, viewed telcscopically, shows no 

 more atmosphere — whether in the artist's or the physicist's sense — 

 than a bare plaster cast illuminated by a powerful searchlight. 

 Far more delicate tests are possible here than in other instances, and 

 neither refraction nor twilight is present to the minutest degree. Our 

 satellite is naked rock in vacuo. Mercury, too, shows little evidence 

 of atmosphere, though Antoniadi reports occasional obscuration of 

 dark markings, which ho attributes to dust-haze. 



The existence of atmospheres on the majority of the planets, 

 though not on all, is thus established by direct telescopic observation. 

 To determine their composition we must, as usual, have recourse 

 to the spectroscope ; but we meet with two difficulties. 



In the first place, many possible atmospheric constituents show no 

 selective absorption whatever in the region accessible to our study. 

 Hydrogen, nitrogen, helium, neon, and argon belong in this group, 

 and are hopelessly beyond the reach of our investigation. Sec- 

 ondly, the other gases of the earth's atmosphere absorb too much for 

 our advantage. The worst by far is ozone. Though present in but 

 small amounts, and mainly in the higher layers, it cuts off the whole 

 spectrum short of 2900 angstroms, and deprives us of any hope of 

 studying the most interesting parts of all celestial spectra. 



Were we working in the infrared, water vapor would be almost as 

 troublesome. There are long stretches of the solar spectrum, within 

 the range of present-day plates, in which we can find out little or 

 nothing about the sun's own spectrum. The great wide lines of the 

 water-vapor bands, often overlapping, hide almost everything else. 

 The band near 11500 A is quite hopeless; that at 18000 would be 

 worse, if our photographs got so far; one near 9600 is still very bad; 

 while in those near 8200 and 7200 the solar lines can be picked out, 

 with care, among their stronger telluric neighbors. 



Oxygen reveals itself by a strong band, with very regularly spaced 

 lines, at A7594 (Fraunhofer's A), the weaker B band near 6867, 

 and the much fainter a band at 6277. The terrestrial origin of all 

 these lines is conclusively settled by two tests: first, their changes 

 with the altitude of the sun (varying the air-path) and, for the 

 water-vapor lines, with weather conditions; second, the absence of 

 the Doppler shift, due to the sun's rotation, when light from the 

 east and west limbs is compared. The absence of even faint com- 

 ponents of solar origin is explained by the high temperature, which 

 dissociates such molecules completely. 



The intensities of these bands are in inverse order of the abundance 

 of the molecules which produce them — an apparent anomaly, ex- 

 plained by the circumstances of their origin. The ozone band is 

 part of the main system of the O3 molecule, and, like all such bands, 



