VARIABLE STARS. 109 



tlie liy])othesis. We may go further; taking into account the fact that 

 the whole "eclipse" lasts for about eight hours, and that the star loses 

 about two- thirds of its light at the time of maximum obscuration (when 

 the brightness remains without change for about twenty minutes), it 

 can be safely inferred that the dark attendant is just about four-fifths 

 the diameter of the bright star and about half as heavy. Also that 

 the distance between the two is about 3,250,000 miles, the diameters of 

 the two bodies being respectively about 1,060,000 and 840,000 miles. 

 Moreover, the combined mass of the two is about two thirds that of the 

 sun, and their density only one-flfth as great, not much exceeding that 

 of cork. 



Nor is this all ; since the discovery of the variability of the star in 

 1782, a slight but distinct variation of the period has been observed, of 

 such a character as to set the times of eclipse alternately backAvard and 

 forward by about two hours and a half in the interval between 1804 and 

 1809. From this, combined with certain minute irregularities in the 

 " proper motion " of the star, Mr. Chandler two years ago showed that 

 Algol and its companion must be moving together in a great orbit, 

 nearly as large as that of the planet Uranus, completing the revolution 

 in a period of about one hundred and thirty years; at least, such a revo- 

 lution would account for the facts, while no other hypothesis yet proposed 

 will do so. And if this revolution is real, it must be around some neigh- 

 boring body, so laintly luminous that it has not yet been detected. The 

 story of Algol opens up to' us a new universe of "dark stars" — bodies 

 invisible, but massive and powerful. Presumably the other stars of the 

 Algol class admit of a similar explanation, although as yet they have not 

 been spectroscopically studied, being all of them small a-nd too faint for 

 investigation with the instruments at present available. When the great 

 Yerkes telescope, nownearing completion, is mounted at Lake Geneva it 

 is expected that one of its earliest aj)plications will be to investigations 

 of this sort. It should be added also that certain peculiarities in the 

 light curves of some of these stars seem to indicate that we have not 

 quite reached the bottom of the matter; that in addition to the eclipse, 

 and perhaps caused by it, there are real variations in the light-giving 

 power of the star. 



As to the other classes of variables, there is no such satisfactory 

 explanation nor any generally received theory. In the case of Beta 

 Lyra>, and its congeners, we have apparently mechanical and other 

 causes combined — the axial revolution of a star, perhaps, or possibly 

 the whirling of a group of stars, accompanied by tidal changes in the 

 brightness and spectroscopic characteristics of their light, for while 

 the minima of the Algol variables are not accompanied by any altera- 

 tion of the character of the lines in their spectra, it is not so with the 

 other variables; their spectra are almost all of them characterized by 

 bright lines, which become especially brilliant at the time of maximum 

 brightness. Beta Lynv presents conspicuously in its spectrum both 



