OF THE COSMOS. PERIODICALLY VARIABLE STARS. 159 



after having been at first rapid, comes after some time to a 

 stand, or at least to a very inconsiderable degree of diminution ; 

 after which the decrease suddenly resumes a most rapid rate. 

 It is as if the attainment of a second maximum was inter- 

 fered with. 



The question of whether there is, on the whole, more re- 

 gularity iu variable stars of very long than in those of very 

 short periods, is one difficult to answer. The deviations 

 from a uniform period can only be taken relatively, i. e. in 

 parts of the period itself. In order to begin with long 

 periods, x Cygni, Mira Ceti, and 30 Hydrse, must be first 

 considered. In x Cygni, the deviations from the most pro- 

 bable period, on the assumption of a uniform variability 

 (406-0634 days), is as great as 39'4 days. Even though a 

 part of this may be ascribed to errors of observation, yet 

 there will still certainly remain from 29 to 30 days, or -^th 

 of the whole period. In Mira Ceti, ( 28 ) in a period of 

 331*340 days, the deviations extend to 55'5 days, even if we 

 leave out of the account the observations of David Fabricius. 

 If, on account of errors of observation, we reduce the esti- 

 mation to 40 days, we obtain a quotient of -|-th, or, as com- 

 pared with x Cygni, a deviation almost twice as great. In 

 30 Hydrse, which has a period of 495 days, the deviation is 

 certainly still greater, perhaps amounting to th. It is only 

 within a few years, since 1840 and still later, that the 

 variable stars with very short periods have been observed 

 perseveringly and with due precision ; so that, in regard to 

 them, the question we are speaking of is still more difficult 

 of solution. As far, however, as experience hitherto can en- 

 lighten us, the deviations would appear to be less considerable. 

 In 17 Aquilse (Period 7d.4h.) they are only -^ or -p^-th of the 

 whole period j in Lyra (Period 12d. 21h.) only -^ or 



