I9I-'-] SCHLESINGER— RADIAL VELOCITIES. 183 



cation of this method are undoubtedly serious ; nevertheless, they do 

 not appear to be of the character that patience and perseverance on 

 the part of a skillful observer will not overcome. 



If the objective prism is ultimately to be used for determining 

 velocities, it would be a great advantage to be able to utilize the full 

 apertures of modern telescopes without necessitating objective prisms 

 of corresponding size. In the case of reflecting telescopes, this might 

 be done by replacing the flat secondary (of the Newtonian form) 

 or the hyperboloid (of the Cassegrainian form) by a convex para- 

 boloid, with its axis and focus coincident with the axis and focus 

 of the primary mirror. The beam of light reflected from such a 

 secondary would be a parallel one, contracted to perhaps one fourth 

 or one fifth the diameter of the original beam. This reduced beam 

 could then be made to pass through an objective prism of moderate 

 size. Similarly, in the case of refracting telescopes, a diverging lens 

 might be placed in the position that the correcting lens usually occu- 

 pies when the telescope is to be used with a slit spectrograph.^"^ This 

 diverging lens can be so designed as to make the emerging beam of 

 light parallel in any portion of the spectrum desired. For practical 

 reasons it would not be advisable to contract the original beam too 

 much ; or in other words, to put the converging lens too near the 

 focus of the visual or the photographic objective. In the case of 

 reflectors, too, though for not quite the same reason, the paraboloid 

 should not have too small an aperture and should not be placed too 

 near the primary focus. 



In conclusion, it seems to me that the prospect of obtaining radial 

 velocities by means of the objective prism is good enough to warrant 

 a trial of all three of the methods that have been reviewed above. 

 If I were asked which of the methods seemed to me the most prom- 

 ising, I should say that the one which makes use of neodymium chlo- 

 ride would probably lead to immediate results, if we are to remain 

 satisfied with a moderate degree of precision; but that the method 

 which is concerned with the length of the spectrum might ultimately 

 be developed to give considerably more accurate results. 



Allegheny Observatory, 



Allegheny, Pennsylvania. 



" Compare with the paper by Wadsworth, Astrophysical Journal, i6, 12. 

 1902. 



