Of THE COSMOS. PERIODICALLY VARIABLE STARS, 165 



in the mean it is 331 days 20 hours, but its fluctuations 

 are as great as a month ; for the shortest time which has 

 been known to elapse from one maximum to the next was 

 only 306, and the longest was 367 days. These ir- 

 regularities become still more striking if we compare the 

 epochs of the actually observed maxima of light with the 

 results which would be obtained by calculating them upon 

 the assumption of a uniform period. The differences 

 between calculation and observation amount to 50 days ; 

 tind these differences are found to be nearly the same and 

 on the same side for several successive years. This cir- 

 cumstance clearly indicates that the luminous phenomena 

 are affected by a perturbation of long period. More exact 

 calculation has proved, however, that one perturbation does 

 not suffice, and 'that we must assume several, wLi-eh may 

 indeed proceed from the same cause, one returning after 

 1 1, a second after 88, a third after 176, and a fourth 

 after 264 single periods. According to these assumptions 

 we may derive a formula of sines ( 278 ) with which the se- 

 veral maxima now shew a very near accordance, although 

 there still remain deviations which cannot be explained 

 by errors of observation. 



2. ft Persei, Algol; E, A, 44 36'; DecL + 40 22'. 

 Although Geminiano Montanari first remarked the varia- 

 bility of this star in 1667, and although it was also ob- 

 served by Maraldi, yet it was Goodricke who, in 1782, 

 first made out the regularity of the variations. The 

 reason of this may very probably be, that this star does 

 not increase and decrease gradually, as do most of the 

 other variable stars, but for 2 days and 13 hours shines 

 constantly with the same brightness (2'8 m,), and only 



