[dELURY] 



CONVECTION AND STELLAR VARIATION 



283 



irregular intensity, while in the case of the short-period variables there 

 is one close satellite whose attraction changes very rapidly and the 

 changes in convections fall in with a precise oscillation. For the Cepheid 

 or day- variables there is a pronounced tendency to a secondary maximum 

 which is not always of the same magnitude : this may be regarded as a 

 marked induced effect following the maximum primary effect, or it may 

 be explained in the way mentioned at the end of §6. 



(3) Many stars which varij in luminosity sîiow simultaneous 

 changes in the character and position of the lines in their spectra. If 



(a,)/l?onfh t^na6/e- HraarO Ce//'. 



■iij ^tviMs 



SJi L. 





(c)Moorliihablè-A^99 M Me C/u^rO) Cenfauri 



■il . fl âj il 71 f, i 



\Pavi 



/lgr2 



the variation in light is attributed to changes in conviections, for reasons 

 already given (§5) it is natural then to assume that there will be sim- 

 ultaneous variations in the character and positions of the spectral lines, 

 and that possibly some new lines will appear, bright lines for instance, 

 caused by the upheaval of glowing matter above most of the filtering 

 action of the cooler gases of the outer atmospheres or by changes in the 



