﻿139 



The results of the three observations recorded subsequently to January 27, 1903, are inconsistent with 

 each other and individually with previous observations, and cannot therefore be considered as affording 

 values as reliable as the previous observations, of which there appears no reason to doubt the reliability. 



The results of all the observations are given in the following Table IV : — 



Table IV. 



In the recorded times of transit of the sun's limbs, there is an inconsistency in the observations of dates 

 January 27 and October 20, 1903, amounting in the former case to 7 seconds and in the latter to 

 5 seconds. These may possibly have been caused Isy a change of atmospheric conditions causing refraction, 

 or by an error in noting the times. In either case the result of the ob.servation is not affected to any 

 appreciable extent. 



No explanation can be found for the apparent inconsistency of the results of the observations dated 

 December 9 and 18, 1902, but they appear to be quite outstanding, and it is considered advisable to 

 discard them. 



Omitting the results which appear to be doubtful, viz., those of dates December 9 and 18, 1902, 

 October 20 and 21, 1903, and January 11, 1904, there remain the following : — 



This mean value has been used in reducing the observations for absolute Declination. 



The omission of the doubtful observations confines the accepted results to a short period of two months, 

 December, 1902, and January, 1903, and it has therefore been necessary to assume that the azimuth of the 

 mark did not alter during the two years in which the absolute observations were made. 



In this connection it is of interest to note that the mean of the results of the observations of dates 

 October 20 and 21, 1903, gives the value 121° 23' 20", which differs by only 5' 6" from the 

 accepted value. 



The mean of the results of all observations gives the value 121° 28' 26". 



In reducing the absolute Declination observations, no correction for tor.sion has been made, the necessary 

 data not having been recorded. 



T 2 



