300 



Trans. Acad. Sci. of St. Louis. 



and Pg respectively. The resulting values of these probable 

 errors are as follows: — 



These results show a greater accuracy in the settings of the 

 position-circle which give the relative directions of the com- 

 ponent stars than in the settings of the micrometer which 

 determine their angular distance apart. This experience is 

 so general among double star observers that there seems little 

 room to doubt the superior accuracy of the measures of 

 position-angle. 



PERSONAL EQUATION. 



The personal errors which occur in measures of close 

 double stars bear so large a proportion to the quantities 

 measured that the need of investigation of the relative 

 personal errors of observers would seem to be more necessary 

 in this class of observations than in any other. As a basis for 

 such a determination of relative equations by systematic ob- 

 servers of double stars, and the consequent deduction of some 

 sort of systematic corrections which shall render the observa- 

 tions more comparable, the observation from time to time of 

 a selected list of double stars of small relative motion is very 

 desirable. Such lists have been proposed by O. Struwe 

 (Viertel jahrschrift Bd. XI, 227) and by Stone (A. N. 2201). 

 The first of these lists has been well observed by a number of 

 observers and I am at present comparing these observations 

 to determine whether they show systematic differences 

 between the observers. 



The results of the comparison as far as I have been able 

 as yet to carry it seem to indicate the following : — 



1. Systematic differences between observers do not seem to 

 depend on the aperture of the telescope used. 



