26 Kkpout 8. a. a. Advanckment of Sciexck. 



known to astronomeis. This brings us naturally to the thiid portion, 

 of this address — 



Interpretation of the Results. 



When we examine the light curves of variable stars we find that 

 they group themselves into five very distinct classes. 



(1) There are those stars which vary regularly and symmetrically ;. 

 that is, they descend down to a minimum and ascend up again to a 

 maximum at precisely the same rate of change. Further, when they 

 reach their maximum they remain at this lirightness for a considerable 

 portion of their period. Their actual duration of variation is often 

 confined within a few hours. The prototype of this class is the famous 

 star Algol, the second variable discovered by astronomers. 



As we ha^e already stated, the explanation of the variation of 

 this star — and consequently of all stars of the Algol class — was ffrst 

 given moi'e than a hundred yeai's ago by Goodricke, and the passage 

 of time has only certified as to its efticiency and sufficiency as an 

 explanation. 



Like the \\inking eye of a lighthouse, Algol owes its brief en- 

 shadowment to eclipse. Round a central sun there circles a dark, 

 contiguous, almost equally massive body. Each circuit this dark 

 companion partially eclipses for a brief space the light of its primary. 

 This is the simple explanation of "algol" variation, as light changes of 

 this first class are designated. 



Now since the conditions under which the two component stars 

 revolve round one another are unlimited — they may be contiguous to 

 one another or far apart ; they may be equal in brightness and size ov 

 markedly unequal ; they may circle round one another in a circular 

 orbit, or pursue an elliptic path : they may eclipse one another 

 totally or only partially ; they may be deformed in shape or practically 

 spherical orbs — the character, limits and duration of eclipse will be 

 difi^erent for each algol system. Still the common type — the family 

 likeness — is there. It will require no enlargement of explanation on 

 my part to make this matter fuller and clearer. 



One of the most interesting problems in \ariable star work is to 

 determine the elements c)f orbital movement, size, form, density, from 

 the observations secured. 



Each mean light cur^'e is treated harmonically, and from this 

 analysis the conditions producing and determining the eclipse are 

 deduced. This is one line of work specially pursued at Lovedale, and 

 the instrumental eijuipment, mode of observing, and system of reducing 

 observations are all made subordinate to this aim and end. 



In the case of C)ne star alone no fewer than 2000 sets c)f obser\a- 

 tions, each set being the measurement of thirty-six stars, has been 

 secured in order to deduce one element -the slow secular change in the 

 periofl of \ai-iation of the star. 



Is such a mass of observations necessary it may be urged. When 



