180 SCHLESINGER— RADIAL VELOCITIES. [April 20, 



of the F type and beyond, the neodymium hne or band at A 4,272 

 becomes involved with lines that are proper to the star itself, and 

 accurate settings upon it become impossible. 



A number of other suggestions have been made for utilizing the 

 objective prism for determining velocities, but I shall stop to con- 

 sider only one of these. Like the two proposals that we have already 

 described, this method owes its origin to Pickering,^ who seems to 

 have studied this problem to good purpose long in advance of any 

 other astronomer. Let us suppose that we have secured by means 

 of the objective prism a photograph showing the spectra of a number 

 of stars in the same region, and that the prism has been so oriented 

 that the violet end of each spectrum is toward the north. Suppose 

 further that we have measured accurately the position of the K line 

 in each spectrum. From these measurements we might compute the 

 relative declinations of the stars, but these declinations would be in 

 error by a slight amount on account of the unknown radial velocities 

 of the stars. Thus a star that is approaching us would have its K 

 line shifted toward the violet end, or in this case, toward the north, 

 and the computed declination would be too great. Let us then se- 

 cnve another photograph of the same region with the prism reversed, 

 so that now the violet end of each spectrum is toward the south. 

 The approach of a star would as before shift its lines toward the 

 violet, which is, however, now toward the south. We should there- 

 fore derive a relative declination for the star that is as much too 

 small as it was too great in the first instance. It is obvious, then, 

 that the measurement of such a pair of plates would theoretically 

 give us the data from which the relative velocities of all the stars 

 on the plate might be inferred. Pickering further proposed that the 

 prism be reversed by simply reversing the telescope on its equatorial 

 mounting, and that the comparison of the two plates could be facili- 

 tated by taking one of them with the glass side outward. The plates 

 could then be put film to film and the measurements would become 

 differential. 



Although the theory of this method is simple enough, its practical 



® Harvard College Observatory Circulars, No. 13 ; see also Astronomische 

 Nachrichten, 171, 137, 1906. 



