Survey of Northeast Greenland. 139 
staff and monopolised the instruments. However, as often as the 
opportunity offered, we busied ourselves with the matter. 
From the following will appear what the Danmark-Ekspedition 
has done in the way of investigating astronomical refraction, and 
what may be deduced from these researches. 
As a link in the investigations of the astronomical refraction we 
undertook a series of zenith distance observations of stars according 
to У. STERNECK’s method (for particulars see pp. 130—135). In the 
course of the computation of the latitude of the Observatory we have 
already learnt that, in view of the refraction, it was necessary to 
distinguish between stars culminating to the north and south of the 
zenith. In a similar manner we find below the results of forty-nine 
observations of stars arranged according to the zenith distances (in- 
cluding the eighteen observations, which have already been used for 
the computation of the latitude). The reduction to the meridian 
and the declination are not given in the table. The temperature 
given under the pressure of the air is that of the barometer. 
The forty-nine stars are arranged in two groups, in that those 
with a zenith distance exceeding 70° are collected in a group by 
themselves. This because we were taught by experience that when 
the zenith distance exceeded 70°, the uncertainty as regards the 
refraction became very great, as will also appear from a cursory 
glance at the latitudes computed on the strength of the latter group. 
Consequently one gets a more correct idea of the state of affairs 
when leaving out — for the present — these ten stars, so that the 
research only comprises zenith distances up to 70°. 
The first mentioned group of thirty-nine stars is divided into 
two sub-divisions, of which the former consisting of eighteen stars 
with zenith distances up to about 51°15’ is the one on the strength 
of which the above-mentioned latitude was computed. A glance at 
the thirty-nine computed latitudes shows directly that the uncertainty 
is considerably greater in the last sub-division (58° < z << 70°). This 
appears more clearly when computing the latitude from the whole 
group, in the same manner as it was computed above from the 
eighteen stars. Result: 
of 16 culminations towards north on = 76°4613"”.21 + 1".0 
„ 93 R „ south g, = 76°46'16".14 + 0".6. 
If we strike the mean of these two values, whereby the error of 
the computed refraction may be supposed to be eliminated, we get 
p = 76°46'14”.68 0.6. 
The mean error of a single observation is about 3". 
