ON SPECTROSCOPY. 9 



terrestrial spectrum. If the spectrum is sufficiently bright, webs or wires 

 may be moved along it by a screw, and the relative positions of the 

 lines obtained. If the spectra are faint, some method of illuminating 

 the wires may be employed, or the continental method of forming by 

 the side of the spectrum the image of an illuminated scale. Zollner 

 has applied the principle of the divided object-glass to the small 

 telescope, with the addition of a prism, by which one spectrum is in- 

 verted relatively to the other, and the distance to be measured doubled. 

 Perhaps the most convenient form of bright line micrometer is one 

 recently invented by Mr. Hilger. 



For the purpose of accurate comparison with terrestrial spectra the 

 use of the ordinary little reflecting prism, before the slit, is not suf- 

 ficiently trustworthy, unless some special arrangements are adopted. 

 It is obvious that, unless the light from the terrestrial source comes upon 

 the slit in precisely the same direction as that from the stars, the 

 method fails in accuracy. Practically, the introduction of the spark 

 or vacuum tube in the axis of the telescope, at a distance of about 

 two feet from the slit, gives sufficient accuracy of identity of posi- 

 tion of the two spectra. The most perfect method consists in 

 causing the image of the spark, by a suitable arrangement of lenses, 

 to fall precisely at the spot where the image of the celestial body is 

 formed. 



It is now time to state in a few words the principal results of spectrum 

 analysis applied to the heavenly bodies other than the sun. 



We have learned that the planets Mars, Jupiter, and Saturn, have' 

 atmospheres not very dissimilar probably from our own. The spectra 

 of the more distant planets Uranus and Neptune indicate atmospheres 

 of a wholly different constitution. 



This analysis has taught us that the fixed stars are truly suns after 

 the order of our own. The spectra of the stars are not precisely similar 

 to the solar spectrum. They differ too, greatly, from each other. 

 Roughly, the spectra of the stars may be arranged in some four or live 

 divisions from the brilliant Sirius to telescopic red stars. These dif- 

 ferences of the spectra point to different degrees of temperature, and it 

 may be, also, to some differences of chemical constitution. It is 

 evident that great differences of temperature would be sufficient to 

 give rise to different chemical conditions of the investing atmospheres 





