Pkksident's Addhess — Sect. A. 25 



■(letoriuine tlie ''candle power" of eucli star. And thus stellar plioto- 

 meters were introduced, and the study of variable stars became foi-th- 

 with an exact science. 



1 may not tairy here to discuss the various photometric means 

 adopted to ari-ive at absolute determinations of magnitude. Such a 

 discussion would have its proper place in a consideration of the science 

 of photometry. 



.Sutficient it is to say that various instruments have been de\ ised 

 — the wedge photometer used at Oxford Observatory ; the meridian 

 equalising photoineter used at Potsdam and Harvard, by which any 

 star can be compared with one fixed and invariable standard, and 

 the magnitude of that star thus determined to within a tenth of a 

 magnitude. 



.\t the Cape Obser\atory and at Harvard extensive use is made 

 of photography as a means of determining magnitudes. 



It is evident, therefore, that we have thus at our disposal two 

 m(^ns of determining the range and the mte of any star that has been 

 detected to vary in brightness. We may compare it as it passes 

 through its ever varying phases with stars — called for short "com- 

 parison stars " — whose magnitude is constant and photometrically 

 determined. This is Argelander's method, and that of the great 

 majority of variable star observers. 



We may, however, determine the brightness of the variable 

 directly by photometry, without any reference to comparison stars. 

 This is the method mainly adopted at Harvard and Potsdam. 



It may not be out of place for me to mention that the method 

 of observation emploj^ed at Lovedale is a modification of tliat of 

 Argelander. The variable star is compared at each observation with 

 all the comparison stars in the field, and its brightness at any instant 

 is determined from this intercomparison with often as many as eight 

 or ten stars. • This method greatly increases the toil of observation, 

 but I think there is a commensurate gain in accuracy. 



After a long series of observations, in the obtaining of which 

 every device making for accuiucy has been utilised and every pitfall 

 making for error avoided or accounted for, the observations are re- 

 duced by the ordinary processes of mathematics and then the mean 

 variation of the star deduced. 



AVith many stars it is easy to determine this mean curve, for 

 each cycle of variation is exactly like all the other cycles, and thus 

 with stars of this markedly regular type, as, for example, Algol, whoi 

 we have secured a mass of observations, by difi'erent observers, by 

 different methods, extending over a lapse of years, we are able 

 to deduce the average curve with considerable accuracy. With 

 some stars, however, it is not possible to determine this mean 

 curve, for no two C3'cles of variation are similar. Certain stars, 

 indeed, seem to follow no law of variation other than the rough one 

 of period. 



We shall now suppose that by careful and long-continued observa- 

 tions we have determined the light curves of all the variable stars 



