428 



COMPARISON OF CATALOGUES OF STARS. 



Comparing the deviations between the different systems in a century, we have the 



following results : — 



It will be seen from the foregoing comparison that if ive may consider the system of 

 Auwers as an absolute standard, three classes of errors will be introduced into any system 

 of observations which depends upon the systems, either of Safford, Boss, or Newcomb. 

 These are : — 



(a) A constant error which is a direct function of the time. 



(b) Systematic errors which result from those errors in the proper motion, which 

 are functions either of the right ascension or of the declination, and whose magnitude 

 is proportional to the distance from the assumed common epoch. 



(c) Errors which will be introduced through the constant n, resulting from a 

 change in the relation between the places of the ecpuatorial, and of the polar stars in 

 the fundamental system chosen. 



The effect of these errors will be sensible in the following cases : — 



Class (a) Safford in right ascension. 



Class (b) Boss and Newcomb in declination. 



Class (c) Newcomb (Gould) in right ascension. 



But there is no decisive evidence that the system of Auwers is to be preferred to the 

 system of Newcomb in right ascension (from 8 = — 30° to 8 = +60") or to the system 

 of Boss in Declination. 



We must therefore conclude that the problem of the construction of a homogene- 

 ous system of stellar coordinates which will satisfy all past observations and which 

 future observations may be relied upon to verify, has not yet been completely solved. 



