DECLINATIONS OF CERTAIN NORTH TOLAR STARS I93 



giving the absolute declinations from the circle readings with- 

 out correction for flexure or division errors ; the second giving 

 the declinations of the list stars from comparison with the one 

 or more zero stars observed on the same night. 



In Table I the observed zenith distances are given, corrected 

 for runs, reduction to the meridian, and refraction. Bessel's 

 refraction tables were employed, as prepared by Professor 

 Eastman of the Naval Observator}-. The standard barometer 

 was repaired and tested a few years ago by the Weather Bureau 

 Office in Washincton. The thermometers also have been tested 

 by them and by the weather bureau official at Lansing. Dur- 

 ing observations the thermometer was hung near the object 

 glass of the telescope, and the readings were corrected w^hen 

 necessary, according to the table of corrections determined by 

 the Weather Bureau Office. 



The next column in Table I gives the observed zenith dis- 

 tances, reduced to Jan. o.o of the year of observation. These 

 reductions w^ere made with the "Independent Star Numbers" 

 G, H, etc., given for each day in the Berliner Jahrbuch. 

 They were checked by a sufficient number of identical reduc- 

 tions made with the Besselian Star Numbers, A, B, C, D, E, 

 from the Berliner Jahrbuch and the star constants from Dr. 

 Auwers's list in the Astronomische Nachrichten. 



The zenith distances are then reduced to Jan. o.o, 1900 and 

 the means taken of the different observations of each position. 

 In obtaining these means, a system of weights depending on 

 the number of settings in each case was adopted as follow^s : 

 probable error of one setting 0^.33 ; probable error of nadir 

 determination o".25 ; probable error in refraction tables 0^.30, 

 giving as weights : 



No. of Settings. Weight. 



1 .72 



2 .90 



3 ^-oo 



4 ^-04 



5 i-oS 



6 1. 10 



7 I.I I 



