THE STARS AND NEBULAE 307 



ness. To quote him : "The redness of variable stars is, in general, 

 a function of the lengths of their periods of light variation. The 

 redder the tint, the longer the period." Whether the conditions 

 which produce redness are also the cause of long periods is an 

 unsettled question. These long-period red variables do not con- 

 form to definite time schedules. Their maxima may precede or 

 may follow prediction by a fortnight or a full month, but the 

 average length of twenty-five consecutive periods will differ al- 

 most not at all from the average of the twenty-five preceding or 

 following periods. 



Of the reasonably bright short-period variables there are 

 about 100 which pass from maximum to minimum and back 

 to maximum within less than thirty days; nearly all of these, 

 within ten days, and some of them within a few hours. In all 

 these cases maxima and minima arrive on time, and the periods 

 of most of them are known within a second. One cycle of change 

 is almost exactly a duplicate of the preceding and following 

 cycles. 



The key to all changes and states of brightness in 

 stars obviously lies in two things, namely; (1) in a close 

 analysis and comparison of the pertinent spectra, and 

 (2) in a systematic determination and tabulation of the 

 luminosities of the various minerals (metals, more 

 especially) at different temperatures. I note that some 

 work along this latter line has already been attempted, 

 but nearly all of it remains to be done. 



Entering into the matter is this feature also : Inas- 

 much as, according to my version, the light of a star 

 comes to us, not from its level surface but from its as- 

 cending geysers, it follows that the swifter the velocity 

 with which these rise toward us the higher radial veloc- 

 ity should the star's spectrum indicate. Furthermore, 

 where the star is variable it should appear at its maxi- 

 mum brightness when its deeper-seated, more refractory, 

 substances are ebulliated; which is to say, when the gey- 

 sers are thrown up highest and fastest. This reasonable 

 deduction is curiously confirmed by the action of the 

 classes of variables known as Cepheids and Geminids; 

 speaking of which Doctor Campbell says (p. 306) : 



Interpreted, we have the astonishing result that every star 

 investigated has its maximum brilliancy at or very near the time 

 of greatest velocity of approach toward the solar system, and the 



