DEPARTMENT 0? MERIDIAN ASTROMETRY. 167 



was assumed that the zero azimuth was constant for each week assigned to 

 one of the given fundamental observers, R. H, Tucker, A. J. Roy, and W. B. 

 Varnum. It was found that the mire reading remained sensibly constant 

 during each week, but that it varied periodically during each 24 hours. This 

 periodic variation was assumed to change with the azimuth of the instru- 

 ment, and that change is probably due to diurnal variation in the relation of 

 the instrumental piers of the transit-circle and its parts, so that the results 

 of successive transits of circumpolar stars could be made comparable by 

 differential corrections. Each double transit by the same observer, in con- 

 nection with the mire reading, furnished an individual value of the azimuth 

 of the instrument, and in connection with the differential corrections applied 

 to the transits of clock stars led to the independent right ascensions of the 

 circimipolar stars. Both azimuths and right ascensions proved to be quite 

 consistent with the assumption made. 



The results for individual azimuths were carefully plotted and cur\'^es 

 drawn to connect these together. These curves were read off at intervals of 

 an hour and exhibited a close similarity throughout, save that they showed 

 small seasonal changes in the weekly constants. 



The mean systematic difference between independent right ascensions for 

 the 17 fundamental circumpolars (Obsd.-P. G. C.) appears to be relatively 

 small. It is : 



-}-?025 sec S. 



When these right ascensions were subsequently expanded to include 50 

 circumpolar stars, all based on independent azimuths, this became 



(Obsd.-P. G. C.) = +?027 sec 8. 



For the 17 stars, the probable error of a single right ascension is 

 ±?o2o sec 8, while for the miscellaneous circumpolar stars making up 

 the 50 the corresponding probable error equals 



±?023 sec S. 



In all the preliminary operations clock stars were chosen between -{-30** 

 and — 30° in declination, preferably between -|-30° and — 20°. The clock- 

 rates were determined from groups between these limits of declination, 24 

 hours apart, corrected the second group to the first, for change in azimuth, 

 usually small, as well as for collimation, level, etc., in a manner analogous 

 to the treatment of circumpolar right ascensions. The right ascensions of 

 the 217 clock stars selected from the Preliminary General Catalogue show 

 small differences, depending on the declination, in the observed right ascen- 

 sions compared with those of the Preliminary General Catalogue between 

 the limits of +30° and — 20° of declination. 



From the azimuth curves previously mentioned clock corrections have 

 been computed. These clock corrections were then arranged in 12-hour 

 groups to free them from systematic error dependent upon right ascension. 



