268 Trans. Acad. Set. of St. Louis. 



These values of the latitude have not been corrected for 

 periodic variation, but the length of time over which the 

 observations extend, makes probable its almost complete 

 elimination from the mean result. This latitude depends on 

 B. J. stars only and the observations were made by seven or 

 eisfht different observers. The close agreement between this 

 value of the latitude and that deduced from the observations 

 of Groups I, II, III and IV, together with the small probable 

 error of the latter (± 0".022), seem to indicate for these 

 determinations a hi3:h degree of accuracv. 



Now the latitude being known, and neglecting for the 

 present the errors of the B. J. declinations of the stars, the 

 errors by which these observations in declination are affected 

 may be easily determined by taking the difference between 

 the observed latitudes and the corresponding true values of 

 the latitude. If it be assumed that all errors except flexure 

 have been allowed for then this process will give the flexure 

 for all zenith distances at which stars have been observed. 

 The means of the residuals found in this way were taken for 

 each ten degrees of zenith distance and the results with 

 probable errors are given in the table below : * — 



The horizontal flexure of the telescope had been determined 

 as about +0".ll Circle W. and + 0".38 Circle E., but no 

 correction for horizontal flexure was applied to the observa- 

 tions. The division errors of the lines under the micro- 

 scopes during observation of the nadir point had been care- 



* See Washburn Obs^y Publications, Vol. V., p. 12.* Appendix. 



