OP ARTS AND SCIENCES. 



389 



TABLE XL — LiGHT-CuRVK op DM. 81°25. 



annular eclipse. ' In this case, if the motion was uniform, the dura- 

 tion of the minimum light would equal only one ninth of that of in- 

 crease or decrease. Tiie effect of the curvature, or ellipticity, of the 

 path would not greatly affect this conclusion. A very great ellipticity 

 is not admissible, or at the periastron the satellite would strike the 

 star. We are therefore obliged to admit that the eclipse is total (that 

 is, that the star is entirely covered by the satellite) , and that the light 

 during the minima is due to one of the two following causes : first, that 

 the satellite is self-luminous, and that the light at the time of the mini- 

 mum is that received from the satellite, the star itself being completely 

 obscured. In this case we should expect to find a corresponding dimi- 

 nution midway between the minima when the star was in front of the 

 satellite, and accordingly cut off a portion of its light. The loss of 

 light would, however, be small, and might easily escape detection. 

 The greatest effect would occur when the transit was central. In this 

 case, to produce the observed duration of the minimum, assuming the 

 motion to be uniform, the diameter of the satellite should be about 1.3, 

 that of the star being taken as unity. Since the light of the satellite 

 is supposed to be 0.110, that of the satellite and star together being 

 taken as unity, it follows that if the star passes in front of the satel- 



