Section III., 1909. [ 209 ] Trans. R. S. C. 



Xll.—Slit Width and Errors of Measurement in Radial Velocity 



Deicrniinalions. 



By J. S. Plaskktt, B.A. 



Presented by Dr. W. F. King. 



The following investigation is a continuation and conclusion of a 

 similar one carried out about a year ago and recently published.i It 

 will be necessary here to give a short summary of the purposes, methods 

 and results of the original investigation in order to render the present 

 work intelligible and complete. 



In the detei-mination of the radial velocities of stars by means of 

 the spectrograph, one of the chief diiKculties to be contended with is the 

 very long exposure required to produce measurable intensity of spectra 

 in all except the briglitest of the stars. For example, under ordinary con- 

 ditions with our three-prism spectrograph attached to the 15-inch equa- 

 torial, an exposure of about three hours would be required to photograph 

 the spectrum of a fifth magnitude star. This is about the practical limit 

 of exposure desirable owing to the large change in hour angle involved, 

 and therefore in the position of the spectrograph, with the consequent 

 possibility of systematic displacements of the spectrum lines due to tem- 

 perature changes, flexure, or other causes. 



The modern spectrograph is exceedingly wasteful of light so that 

 probably not more than five per cent of the star-light incident upon the 

 objective of the telescope is effective in producing the spectrum on the 

 photographic plate. This loss occurs by absorption and reflection in the 

 telescope objective and correcting lens, by occultation at the slit jaws, and 

 by the absorption and reflection of the collimator and camera objectives 

 and the prism train. We have practically no control of the loss of light 

 after passing through the slit, as in order to obtain the necessary dis- 

 persion we are obliged to use media which absorb and reflect the light, 

 Some recent experiments of mine have shown 2 that the loss of light 

 at the slit jaws when normally opened, say to 0.03 mm., is very consider- 

 able, amounting to about seventy per cent, and that for slit widths up 

 to about 0.13 mm. the exposure time required on star spectra is very 

 approximately inversely proportional to the slit width. The importance 



1 Astrophysical Journal, XXVIII, p. 259 ; also Report of Chief Astronomer, 



1907-1908. 



2 Astrophysical Journal, XXYII, p. 139. Also Pweport of Chief Astronomer, 



1907-08. 



