EiAMBAUT — Determining Didance of a Double Star. 667 



for the value of V. In the Greenwich results, velocities as small as 

 this are sometimes recorded, which shows that they can be detected 

 by the eye ; and though, in comparing the spectrum of a star with that 

 of some terrestrial substance in a vacuum tube, this small displacement 

 may not be worthy of much confidence, as it might be altogether due 

 to some want of adjustment of the apparatus : yet in the case before 

 us, where the two spectra are presented in the telescope under exactly 

 the same conditions, any displacement at all visible must be due to 

 some relative velocity of the components. The photographic process, 

 also, may detect small displacements to which the eye is insensible. 

 Professor Pickering, speaking of the Draper photographs of stellar 

 spectra, says: "The present results encourage the expectation that 

 the movements of stars in the line of sight may be better determined 

 by the photographic method than by direct observation." But, even if 

 the limits I have chosen should be thought too small, the only change 

 would be to increase what may be termed the critical value of k. 



ISFow it will be interesting, I think, to find for what stars h is 

 greater than unity ; and although in deducing a value of the parallax 

 from the observed value of V, it will be necessary to compute it from 

 the complete formula ; yet in merely searching for stars, for which h 

 may at any time exceed this value, a rougher method will be suffi- 

 ciently precise. There is a very simple graphical method which will 



hr 



give us what we require with considerable accuracy. Since p = -jj, 



equation (5), may evidently be written 



7 I 



n V= — — sin (^ - A,) sin y ; 



or, since pp' = h"^, p' being the perpendicular from the other focus, it 

 may also be written 



Z«jK»' sin (<^ - X) sin y 



nF= 



F.b 



Now ^' sin ((^ - X) is the orthogonal projection of ^ on the line of 



