NO. 1865 OUR KNOWLEDGE OE THE SUN — HALE 345 



temperature, a red star, assumed to have the same composition as 

 the Sun, might be expected to give a spectrum resembhng that of a 

 Sun-spot, if its temperature were the same. 



In order to test this question with sufficient precision, the spectra 

 of Arcturus, an incipient red star, and of a Ononis, a conspicuous 

 red star in the constellation of Orion, were photographed with a 

 very powerful spectrograph. Here, again, the principle of using a 

 high dispersion spectrograph, mounted on a massive stone pier in 

 a constant temperature chamber, was substituted for the ordinary 

 method of attaching a small spectrograph to the tube of a moving 

 telescope. The Snow telescope provided a fixed image of the star, 

 and it was only necessary to maintain this upon the slit of the spec- 

 trograph during an exposure long enough to permit the greatly dis- 

 persed light to impress itself upon the photographic plate. With 

 the comparatively small aperture of the Snow telescope, exposures 

 of from fifteen to twenty hours, carried on through several suc- 

 cessive nights, were required. The great amount of light which 

 will be collected by our 6o-inch reflector will reduce these exposures 

 and will also permit fainter stars to be photographed with high dis- 

 persion. 



A study of the plates thus obtained showed an interesting parallel- 

 ism between the relative intensities of the lines in the spectra of 

 these stars and those of Sun-spots. Many of the lines that are 

 strengthened in spots are strengthened in these stars, and many of 

 the lines that are weakened in spots are weakened in these stars. 

 There are some important points of difference, probably due to the 

 fact that the relative intensities of the lines in spots and stars are 

 not determined solely by temperature condition. In general, how- 

 ever, the agreement is sufficiently close to indicate the probability 

 that a common cause — reduced temperature — is at work in both 

 cases. If any doubts remained as to the resemblance between the 

 spectra of red stars and Sun-spots, they were removed when the 

 titanium oxide bands were discovered in our photographs of spot 

 spectra. These bands are the characteristic feature of one of the 

 two great classes of red stars, their spectra showing them in all de- 

 grees of intensity, from the comparative faintness which delayed 

 their discovery in Sun-spots to the blackness observed in such deep 

 red stars as aHerctilis and Antares. The absence of these bands in 

 the other great class of red stars, in whose spectra the bands of 

 carbon (not found in Sun-spots) predominate, suggests interesting 

 possibilities in future work on the Sun's stellar relationships. 



These results leave unanswered scores of questions involved in 



