402 COLOURS OF THE STARS. SECT. XXXVI. 



the earth's. There are no defective rays in the white light of 

 Sirius, Procyon, and others ; but Sir David Brewster found in 

 the spectrum of the orange- coloured light of Herculis a defective 

 band in the red space, and two or more in the blue ; consequently, 

 the orange colour of the star is owing to a want of blue rays ; for 

 flames in which certain rays are wanting take the colour of the 

 predominating rays. If the black rays in the solar spectrum 

 were owing to the absorption of the sun's atmosphere, the light 

 from the margin of his disc, having to pass through a greater 

 thickness of it, would exhibit deeper lines than that which comes 

 from his centre ; but, as no difference is perceptible, it may be 

 inferred that the analogous bands in the light of the coloured 

 stars are not due to the absorption of their atmospheres, but that 

 they arise from the different kinds of combustion by which these 

 bodies are lighted up. 



All the ordinary methods fail for finding the parallax when the 

 distances of the stars are very great. An angle even of one or 

 two seconds, viewed in the focus of our largest telescopes, does 

 not equal the thickness of a spider's thread, which makes it 

 impossible to measure such minute quantities with any degree of 

 accuracy. In some cases, however, the binary systems of stars 

 furnish a method of estimating an angle of even the tenth of a 

 second, which is thirty times more accurate than by any other 

 means. From them the actual distances of some of the more 

 remote stars will ultimately be known. 



Suppose that one star revolves round another in an orbit which 

 is so obliquely seen from the earth as to look like an ellipse in a 

 horizontal position, then it is clear that one-half of the orbit will 

 be nearer to us than the other half. Now, in consequence of the 

 time which light takes to travel, we always see the satellite star 

 in a place which it has already left. Hence, when that star sets 

 out from the point of its orbit which is nearest to us, its light 

 will take more and more time to come to us in proportion as the 

 star moves round to the most distant point in its orbit. On that 

 account the star will appear to us to take more time in moving 

 through that half of its orbit than it really does. Exactly the 

 contrary takes place on the other half ; for the light will take 

 less and less time to arrive at the earth in proportion as the star 

 approaches nearer to us ; and therefore it will seem to move 

 through this half of its orbit in less time than it really does. 



