JANUAKT 22, 1915] 



SCIENCE 



113 



time thait Michell showed that on the doe- 

 trine of probabilities double stars were 

 too numerous to be fortuitous groupings 

 in all eases, so that binary stars were in a 

 sense discovered by a mathematician and 

 not by an astronomer. Twenty years 

 later Hersehel proved at the telescope that 

 some double stars are real binaries, and 

 that they revolve around each other by rea- 

 son of their mutual attractions. 



In 1880 Pickering showed that Algol's 

 changes in light conform well with the 

 eclipse explanation, and he suggested that 

 the matter might be settled by the spectro- 

 scope. He argued that the orbital velocity 

 of Algol due to the attraction of the dark 

 companion should be considerable, and 

 should change its sign according as the 

 observations are made before or after the 

 time of minimum light. The spectroscope 

 was not quite ready at that time to handle 

 problems of such delicacy, but a few years 

 later Vogel succeeded in greatly increas- 

 ing its accuracy for the determinations of 

 velocities, by substituting the photographic 

 plate for the human eye. Algol was 

 among the first stars to be tested by Vogel, 

 and his observations indicate precisely such 

 velocities as the eclipse explanation im- 

 plies. This explanation has been ac- 

 cepted without reserve since that time, 

 and has been extended to all the numerous 

 variables of the same kind that have in the 

 mean time been discovered. 



It was early noticed by Argelander and 

 others that the period of Algol, the time 

 between two successive light minima, is 

 not constant. Attempts were made to 

 represent these inequalities by formulae 

 involving the second and higher powers of 

 the time, but the star refused to conform 

 to such equations. In 1888 Chandler ex- 

 amined this question with great thorough- 

 ness; he showed that by the introduction 

 of periodic terms aU the observations up 



to that time could be well represented. 

 The most important of these terms has a 

 coefficient of 173 minutes and a period be- 

 tween 130 and 140 years. To account for 

 this Chandler supposed that the system 

 contains a third body, and that Algol and 

 its eclipsing companion revolve around 

 the common center of gravity of all three 

 bodies in this long period. The dimen- 

 sions of this orbit were supposed to be 

 such that the light equation in it for an 

 observer on the earth would be 173 min- 

 utes, and thus the eclipses would be ad- 

 vanced or delayed by this amount, ac- 

 cording as they occur on the nearer or the 

 farther side of this vast orbit. Chandler 

 was quick to see that this explanation en- 

 tails irregularities in the proper motion 

 of Algol, and that these might be large 

 enough to be unearthed from meridian ob- 

 servations. An examination of all the ma- 

 terial of this kind then available con- 

 vinced him that such an effect is really 

 present, the coefficient of the oscillation 

 coming out l."3, and its period 131 years. 

 This result was apparently confirmed in a 

 general way by Searle at Harvard Observa- 

 tory, making use of additional observa- 

 tions secured for this express purpose. 

 Baushinger, however, after applying to 

 the catalogue positions the best available 

 systematic corrections, concludes that there 

 is no evidence whatever of a periodic term 

 in Algol's proper motion. In the follow- 

 ing year, Boss overhauled the same obser- 

 vations once more and decided that the 

 probabilities were in favor of the presence 

 of a term with a period of 131 years, but 

 with a coefficient much smaller than that 

 found by Chandler, 0."5 against l."3. In 

 later years Boss seems to have changed his 

 mind as to the reality of this term; for in 

 his Preliminary General Catalogue, pub- 

 lished in 1910, he treats Algol as though 

 its motion were uniform, although in the 



