PitKsiDKNTs Adduks.s — Skct. A. 27 



I say tliat the period of this southern star to which T refer lengtliens 

 only hy one-thousandth of a second every revolution, the need for long- 

 continued and carefully made observations will he cleai*. Indeed, some 

 oi the deductions arrived at depend often on such minute differences of 

 brightness as the change in intensity between a candle at a distance of 

 100 yards and the same candle at 101 yards. 



(2) The second class of variable stars are those which vary 

 regularl}' and continuous!}' and uns3'mmetrically in a few days. 

 The outstanding character of all these stars is their rapid rise to 

 maximum. Slowly the star — take S Cephei as a type — sinks to 

 its minimum brightness. This reached, it swiftly rises to its maxi- 

 mum. But there is no tarrj'ing here. Inmiediately the star starts 

 again on its downwai'd track, only once more to leap up into fullest 

 lustre. 



Each C3'cle, as we hii\e said, is performed in a few days, sometimes 

 in a few hours. One southern star belonging to this class goes through 

 all its light changes in ten hours, taking only two hours to rise from its 

 faintest phase to its brightest. 



]Many explanations of this type of variation have been offered. 

 But no single explanation is satisfactory. Thus we know that 

 in some definite, but as yet undiscovered, way revolution and 

 variation are inseparably linked together. Such stars of this class 

 as are bright enough to come within spectroscopic examination 

 have been proved to be binary stars, and binary stars with the 

 same period of revolution as photometric observations show to be 

 also the period of variation. But further than this we cannot go with 

 safety. 



It is possible, nay, probable, that tidal deformation, consequent on 

 the contiguity of the components and the ellipticity of their orbit, ma}'' 

 be the main cause of that period variation. Or it may be physical 

 changes, making themselves manifest in a periodic variation of light 

 and heat intensity, due to the regular alteration in surface pressure as 

 the stars sweep through their orbit, that are the background of this 

 type c>f variation. For the present we must be satisfied simply to 

 gather material against the da}' when the veil will be drawn aside 

 and we shall know what conditions and causes underlie short period 

 variation. 



At Lovedale twenty southern short period variables, the majority 

 of which have been discovered at Lovedale, are under regular observa- 

 tion, and over 25,000 observations have been secured of these most 

 interesting stars. 



(3) Our ignorance unhappily as regards causes is not confined to 

 short period variation alone. There is a third class, and they embrace 

 three-fourths of the known variables in the sky, concerning which it 

 may be said that we know the type and laws of their variation, but 

 nothing more. 



This important class of stars, commonly called long period 

 variables, possesses an unmistakable type of variation. True, no two 

 sta-rs follow exactly the same type ; and often the same star does not 



