XIV GEELMUYDEN. ASTRONOMICAL OBSERVATIONS. — [NORW. POL. EXP. 
In summer time only the Sun was available for observation. These 
altitudes were mostly taken with the sextant, but some also with the altazi- 
muth. The results of these observations, especially the clock error, are 
generally subject to greater uncertainty than those of the winter observations, 
by reason of the interval of several hours between the determinations of 
latitude and time. While in many cases the time could be safely computed 
by means of the latitude deduced from the nearest meridian altitude, it was 
in other cases necessary to allow for the drift of the ice, and the interpolation 
necessary for this purpose is of course always somewhat uncertain. When 
the latitude was determined by extra-meridian altitudes the same remark 
applies, so that sometimes repeated corrections were necessary. Only in a 
few cases was the clock error determined by equal altitudes of the Sun, 
which were, for the same reason, generally treated as absolute altitudes. In 
some cases when a series of circum-meridian altitudes had been taken, the 
moment of apparent noon could be deduced with sufficient approximation 
from the differences, and thus the time be determined as well as the latitude. 
In some few cases, when two or more altitudes of the Sun had been taken 
near the prime vertical, one of the computers, Mr. ALExanpER, has_ with 
advantage employed the differences for determination of latitude. 
Some observations taken with the altazimuth during the last summer 
(1896) have been treated in the same manner as the star observations, thus 
neglecting the drift in the interval. 
Occasionally altitudes of very low stars or a low Sun were measured 
in connection with the ordinary determination of time and latitude, especially 
during severe cold, in order to determine the refraction. 
For the reduction of altitudes measured with the sextant from the natural 
horizon it was deemed most correct to form a table for the dip, adapted to 
the peculiar circumstances, though the difference from the values ordinarily 
used are not of importance. The expression for the dip of the horizon may 
be written 
D= syFa—wy 
where S = 206265”, H is the height of the eye, @ the average radius of 
curvature for the part of the earth under consideration, and k the constant 
of terrestrial refraction. The theoretical expression for this constant contains 
