388 Intelligence and Miscellaneous Articles. 



appear to us reddish, changing at the perigee into white, changing 

 again into bluish as the star departs, and again into white at the 

 apogee. There are, however, variable stars whose colours undergo 

 exactly those changes, viz. passing from one colour to its comple- 

 mentary, and back again, with periods of white light intervening. 

 The binary stars, whose colours are frequently complementary to 

 each other, should, under the proper circumstances, exhibit the same 

 circuit of change ; but I am not prepared to say that they do so in any 

 case : it is in fact asserted that the larger star is usually red and the 

 smaller one blue. 



Supposing this train of thought to be sound, and that by extremely 

 careful observation a difference could be detected in the position of a 

 variably coloured star when it appears red, from its position when 

 blue, we should then be upon the track to calculate its distance. 

 We should need to know the differing rates of light-impulses re- 

 quired to produce the observed colours, the time occupied by the 

 star in passing from one extremity of its course to the other, and the 

 angular distance between those extreme positions. Let us assume 

 that such a variably coloured star has been found, which has a mea- 

 surable change of position in one line — in other words, the plane of 

 whose orbit coincides with our line of vision; and let us assume that 

 its extreme colours indicate a difference in the rate of arrival of light- 

 impulses (or in other words, a difference in the velocity of the light 

 arriving from that star at the two periods) equal to 2v : then the 

 actual speed of the star in its orbit =v ; and as the orbital period of 



v x t 

 the star has been found =t, it follows that - — — — is the real length 



3-1416 & 



of that diameter which is the measured angular distance between 



the two extreme positions of the star. Knowing the angle and the 



length of the base which subtends it, we have the distance of the 



star. 



If a pair of binary stars could be found whose colours alternate, 

 and which alternately eclipse each other, the matter would be sim- 

 plified by so much as accurate measurement of the angle of parallax 

 would be facilitated. 



Should it be objected that no such binary stars have been observed, 

 and that no change of position has been noticed in any single stars 

 of variable colour, I can only reply that possibly it may be worth 

 while to direct attention to those points. 



Should it be said that this hypothesis of the cause of variable 

 colour in stars would oblige us to believe that rays of different 

 colours are propagated at different velocities, and that this whole 

 suggestion is valueless until those velocities have been determined 

 by direct experiment, I answer that very probably rays of different 

 colours have different velocities, and that to determine them would 

 be a most important achievement. 



To conclude, I offer the analogy of the changing tone of a loco- 

 motive whistle as you rush past it on another train. Here, as you 

 approach the sound, its impulses reach you more frequently than if 

 its source and you were at rest. At the instant of passing you re- 



