Survey of Northeast Greenland. 147 
tion of latitude, by means of the sun, in the upper culmination 
within the period 
sth to Sth August (midsummet) . .. 2... gave ф = 74°32'10" 
28th June to 1st July (the snow melting period) „ g = 74°32'20". 
Terrestrial Refraction. 
The frequently occurring mirages, both directly on the sea and 
in the higher air strata over land, were suggestive of peculiar re- 
fraction conditions in the atmosphere. The sending up of kites and 
balloons showed that reckoning from a few metres above sea level 
there was as a general rule an increase of temperature (inversion) 
for the first two hundred metres, whereupon the temperature fell, 
so that at a height of 500 metres there was nearly the same tempe- 
rature as at the place of ascension (about 5 metres above sea level). 
It was thus — for this reason only to be expected that the ter- 
restrial refraction at Danmarks Havn must be different, according 
as the lines of vision ran under about 200 metres or extended to 
higher altitudes. 
Already by the computation made during the winter of 1906—7 
of the triangulation net measured in the course of the autumn, it 
became evident that the co-efficient of refraction varied rather much 
(see the computation of the altitudes of the points of triangulation 
pp. 229—230), and this was the reason why after my return from the 
great sledge-trip in the spring of 1907 I made up my mind to carry on 
researches of ihe terrestrial refraction. These researches were intended 
to comprise the determination of the mean value of the co-efficient 
of refraction in the case of sights across the land as well as those 
across the sea. 
In the case of the latter it would be sufficient from the perma- 
nent Observatory to level at one or more points of the sea horizon. The 
co-efficient of refraction might then be computed from the formula: 
z=90°\? В 
Lek = (=== ) Si 
where z is the zenith distance of the sea horizon, R is the radius 
of curvature of the earth corresponding to the sight, while h is the 
altitude of the telescope, and k is the co-efficient of refraction. 
In the case of sights across land it proved the best method to 
use reciprocal, simultaneous measurements of zenith distances. When 
measuring from À the zenith distance z, of the telescope in B, and 
measuring at the same time from B the zenith distance z, of the 
telescope in À one gets with approximation 
R о 
Е = 1+ р @, +2, + 180°) 
и 
where D means the distance between A and В. 
