Ivi Proceedings of the South African Philosophical Society. 



From 1900, with a new equatorial telescope presented to him by 

 Sir John Usher, and specially designed for this class of work, he 

 commenced an elaborate series of observations on what are known 

 as the Algol variables. These constitute a peculiarly interesting 

 class of objects. For many days together the star shines with 

 uniform light, suddenly at a particular moment the light of the star 

 begins to wane, diminishing until a certain minimum is reached and 

 again increasing in brightness till the normal magnitude is restored. 

 These periodic fluctuations recur with great regularity. The obvious 

 conclusion is that two stars revolve about each other nearly in a plane 

 directed tow^ards the sun, and consequently one star in the course of 

 its revolution obliterates the other. When the stars are not in the 

 same line with the sun we see as a single star their combined light, 

 when in a line we see the light of only one i)lus such part of the 

 light of the second as is not obscured by the first. There are thus 

 two kinds of minima, one when star No. 1 is in front of No. 2, and 

 vice versa. 



From the light curves expressing the amount of light at each 

 instant during the waning and waxing of the light Eoberts finds it 

 possible to determine the density and figure of the disc of the com- 

 ponents, and the elements of the binary system. 



Roberts found the accuracy attainable with his new instrument 

 was such as to warrant investigations of this nature, and these led 

 to conclusions bearing directly on the cosmical problem of Stellar 

 evolution. He found, for example, that the mean density of eight 

 southern Algol variables is ^ that of the sun. Further, in the case 

 of those double stars of which the components revolve in contact, he 

 found that the resulting oblateness in the figure of the component 

 stars agrees in a striking manner with that found by George Darwin 

 from purely theoretical considerations. 



Besides this he has undertaken an independent determination of 

 the magnitudes of all the, stars brighter than 9*2 magnitude which 

 are situated south of declination — 30°, and also the regular observa- 

 tion of about 120 variable stars. Eoberts has made in all about 

 250,000 independent estimations of stellar magnitude, and all this 

 as work entirely outside heavy duties in connection with the Love- 

 dale Institute, of which he has, in Dr. Stewart's absence, been the 

 responsible director. I know few instances of more successful 

 devotion of small means and limited opportunity to the attainment 

 of great scientific ends than the work of Dr. Roberts. 



In the same field a large amount of work and discovery has been 

 done by Mr. R. T. A. Innes, Secretary at the Royal Observatory. 

 He undertook the revision of the Cape Durchmusterung as a labour 



