.308 



THE SAYEE OBSEEVATOEY. 



A series of measurements was also made for the determination of the periodic 

 errors, but the resulting probable error was of the same order of magnitude as the 

 correction ; moreover this will be pretty effectually eliminated from the mean of the 

 observations ; for these reasons no correction of this kind was applied. 



The Reticle. 



This was provided with three vertical threads and one horizontal, the latter for 

 marking the middle of the field. As there were no means of making a close adjustment 

 in collimation, the single bisection was made as formerly at the instant when the star 

 crossed the meridian, as shown by the clock, which did not necessarily coincide with the 

 time of crossing the thread. Except, however, in case of the few stars included whose 

 declinations were large, the difference was not important. 



The micrometer was fitted with five horizontal threads, the intervals being approxi- 

 mately equal to ten revolutions of the screw. These are numbered consecutively 

 I, II, III, IV and V, a small piece of brass near one end showing the number of each 

 thread for the purpose of avoiding mistakes in identification. 



When the difference of zenith distance of the two stars forming a pair was not more 

 than twenty revolutions, both stars were commonly bisected with number III. For 

 greater distances II and IV were employed. I and V were not used. 



It was therefore necessary to determine accurately the distance between II and IV, 

 and as this was found to be variable to some extent, the operation must be frequently 

 repeated. Writing II + x = IV, the following values of x, in terms of revolutions of 

 the screw, were employed, each being the mean of ten or more determinations : 



1889 December 3 — 1890 February 1 a; = .0904 J? 



1890 February 1 — Juno 1 .0872 

 June 1 — August 1 .0812 

 August 1 — November G .0831 

 November — End .0875 



The Star List. 



This comprises 111 pairs, 16 of which were found in the old list. Much care was 

 given to the selection of the stars employed, the effort being to include all favorable pairs 

 from h to 24 h which could be observed without overlapping. As ideal pairs were not 

 always to be found, a number were included which were objectionable in one way or 

 another. It was somewhat doubtful whether these should be included in the final 

 discussion, but with the exception of two or three cases no important difference could be 

 discovered between the results from these and the more ideally perfect ones. 



