178 SCHLESINGER— RADIAL VELOCITIES. [April 20, 



purpose fairly well. It would be still better to design an objective 

 with this specific application in view and to bring into focus the two 

 regions say from A 3,9CK) to A 4,500 and from A 5,900 to A 6,600, 

 without paying much attention to the intermediate region, from 

 A 4,500 to A 5,900. 



It would be quite possible to use a reflecting telescope for this 

 purpose. But as compared with a refracting telescope the reflector 

 seems to be less promising for two reasons : first, because of the 

 change in the focal length with change of temperature ; and, second, 

 because of the harmful effect of any change in the inclination of 

 the mirror. Thus, if one edge of a mirror whose aperture is one 

 tenth its focal length should be tilted with respect to the other edge 

 by only one micron, the lines in the spectrum would be shifted by 

 twenty times this amount, a quantity that is of the same order as the 

 shift due to velocity. In refracting telescopes the effect of a slight 

 tilt in the objective is of no consequence whatever. 



With the objective prism it is difficult to make long exposures 

 that will show sharp spectra. This is chiefly because the refractive 

 index of glass varies rapidly with the temperature and the amount 

 by which the rays of light are deviated is thus continually changing 

 if the prism is in the open air. Observers with slit spectrographs 

 have long been aware of the necessity of guarding their prisms 

 against changes of temperature, and have learned to surround them 

 with cases within which the temperature is artificially kept from 

 fluctuating. As I pointed out a few years ago,* there is no reason 

 why a similar device should not be used in connection with objective 

 prisms. In this case, the light from the star should first be admitted 

 through a window of optically plane parallel glass, whose aperture 

 is the same as that of the prism and the objective. This window 

 should form part of a temperature case enclosing the whole camera, 

 including the prism, the objective and the plate-holder. The case 

 can then be maintained at constant temperature in the usual way. 



Even with these precautions it would be well, in the experimental 

 stages at least, to calibrate the prism frequently by means of stars 

 whose velocities are known from observations with slit spectro- 



* Science, 30, 729, 1909. 



