trigonometrical measurements at Spitzbergen. 293 
the northern and two on the southern meridian, the extreme 
differences in the latitude deduced from the results did not 
amount to 7 seconds ; and as each zenith distance in the pre- 
sent determination is a mean of several repetitions, they may 
be presumed to be free from any error which could affect the 
conclusion. It was not considered necessary to go through 
the process of repetition in the horizontal angles, especially 
as there were four verniers on the circle ; each angle, how- 
ever, is the mean of separate observations by Mr. Foster 
and myself, both of which are inserted. It will be seen, that 
in the triangle A. B. C. the three observed angles fell short by 
57 seconds of 180°. It was not however deemed necessary to 
repeat the observations, for the purpose of detecting an error 
which does not make a difference of one inch in the length 
AC, nor, consequently, in the altitude of the hill ; a third of 
57 seconds, or 19 seconds, has however been added to each 
of the observed angles to complete their sum to 180°. 
A corroboration of the measurements from A and C was 
obtained by including the Observatory, which was visible 
from the three stations, and from whence the zenith distance 
of the cone was also observed ; the angles were taken, as 
nearly as could be judged, to the middle of the door of the 
Observatory, which faced the south ; as however it did not 
present so definite an object as a station pole, and as more- 
over the distance of the cone from the Observatory was much 
greater than from A and C, and the angle of elevation con- 
sequently much less, it may be preferable as a fixed deter- 
mination to take a mean of the results at A and C, and to 
consider the one obtained at the Observatory simply as 
corroborative. 
