Il8 CALIFORNIA ACADEMY OF SCIENCES. [PROC. 3D SER. 



from such stars was found to fit in very well with those 

 deduced from the other stars. Furthermore, down to 85 

 zenith distance the observing was very good. In con- 

 sequence of these facts it was decided to take into account 

 all the stars observed. The zenith distances of the stars in 

 this list range from 21 21' to 89 12' (apparent). 



From 85 zenith distance down, the quality of the 

 " seeing" decreases quite rapidly. This can be seen from 

 the following table of average weights. These weights 

 were derived from the probable errors of the individual 

 determinations of dloga. 



Z. D. Av. Wt. 



The small weight for the small zenith distances is due to 

 the fact that in the expression for da the refraction occurs 

 in the denominator. The small weight for the stars at 

 zenith distances greater than 85 is, of course, due to 

 uncertainties in observing at such low altitudes. 



OBSERVATIONS. 



/. List. The following list of 31 stars was observed on 

 seventeen nights, from 1899 June 7 to 1899 July 6, 

 inclusive, and have been reduced according to the plan 

 outlined in the preceding section. Eleven other stars were 

 on the same observing list, but they have not been used 

 here. They were put on to obtain data for determining 

 bisection error, and for other purposes. 



The numbers of the stars are those of Newcomb's 

 " Catalogue of Fundamental Stars for 1875 and 1900, 

 reduced to an Absolute System." 



