228 ROYAL SOCIETY OF CANADA 



tions causing changes in the radiation from the regions affected. Now 

 although these changes may not be great enough to seriously modify the 

 total radiation of the sun, nevertheless in the case of variable stars we 

 may assume that similar changes on a larger scale account for the lighi> 

 variations. 



§4. By thus attributing light-variations to changes in convection 

 currents an explanation is forthcoming for the fact that nearly all con- 

 tinuously variable stars take a longer time in passing from maximum to 

 minimum brightness than from minimum to maximum brightness. 

 This is illustrated in the case of the solar spots/ faculse and promin- 

 ences which show a more rapid rise to maximum than fall to minimum. 

 In other words, the accentuated convections in the solar atmosphere rise 

 to a maximum more quickly than they sink to a minimum; and it is 

 natural to suppose that a, disturbance in an atmosphere will be slow to 

 die out completely. If the radial and transversal currents be accelerated 

 in any way they will induce other disturbances. If the ejections of 

 matter be increased the subsequent returns of cooled matter will be 

 greater and there will follow outbursts and further ejections due to the 

 sudden expansions of the cooled materials, and possibly too there will be 

 changes in the phases of matter — solid and liquid to gas — with accom- 

 panying electrical effects. Or, the observed effect will be made up of a 

 primary effect and an induced effect due to the primary effect, i.e., 



Observed Effect = Primary Effect + Induced Effect, 

 or, . E = P + J. 



'J'he nature of the jirimary and induced effects will, of course, de- 

 pend on tlie disturbing action, the length of period and the character of 

 the atmosphere affected. In the case of the sun it appears that the in- 

 duced effoet, I, is a function of the primary effect, P, involving F to a 

 power higher than the fir^t, since the greater the maximum elfect the 

 greater is the fraction of the period spent in the fall away to minimum 

 (foot-note heloiv). It is to be expected that the effects produced in gases 

 of low density, such as hydrogen and helium, will differ from those pro- 

 duced in heavy and easily condensed vapors such as iron and calcium, 



1 Joseph Baxendall in Monthly Notices, 21, 141-142, 1861, discussed Wolf's 

 sun-spot cun'es and calculated the average value of the ratio, (Interval from min. 

 to max.) : {Interval from max. to min.) or R., to be 0.76 and concluded that 

 " R is least in long periods and periods of greatest activity, and greatest in short 

 periods and periods of diminished activity," and he pointed out the resemblance 

 of the sun-spot curve to the light-curves of variable stars, and supposed that they 

 had a common explanation. 



