196 I. P. Kocu. 
Upper sea horizon 90°06’18” 
Middle sea horizon 90°07’06” 
Diff. 48” 
False sea horizon 90°07’55” Diff: 40” 
The difference between the boundaries of the mirages in the sea 
horizon was occasionally observed, as appears from the above-men- 
tioned and from the table p. 162 of levellings at the horizon. When 
going over the weather journal in Medd. om Grönland XLII one gets 
an impression of the general occurrence of the phenomenon; here 
we likewise find a great number of diagrams and a few photo- 
graphs of mirages. In chapter VI, “Features of the Geography of 
Northeast Greenland”, the mirages are likewise mentioned and illu- 
strated. A systematic investigation of mirages has not taken place 
on the expedition. 
Concluding remarks. 
As a practical geodetic problem the question of the terrestric re- 
fraction might be considered treshed out about the middle of the 
last century, in that it was recognized as impossible to surmount the 
practical difficulties, inseparable from trigonometrical levelling in con- 
sequence of the variation of the co-efficient of refraction, even though 
one might, by means of various measures, protect oneself against 
the worst drawbacks. The object of the treatment of the material 
of the Danmark-Ekspedition in this direction was thus a purely 
scientific one: By determining the co-efficient of refraction and the 
variation of the latter in various circumstances to attempt to deepen 
the understanding of certain peculiarities of the physics of the at- 
mosphere. 
In the course of the observations in Greenland and partly also 
by the treatment of the material, certain practical experiences have, 
however, been gained, the passing over of which in this place would 
presumably be wrong. 
The great difference between the co-efficient of refraction over 
land and over water, combined with the strong and sudden oscilla- 
tions of the co-efficient, encourages one to be still more distrustful 
of trigonometric levellings in Greenland — and in particular in North- 
east Greenland — than in more southerly zones. Considerable dis- 
agreements between the results of the levellings are unavoidable. 
In chapter Ш I have given examples of how — in the course of 
geographical survey — to form an estimate of the reliability of the 
measuring of altitudes performed. Here I shall only add that the 
observation journal ought to contain particulars of observations, 
which may point towards special conditions of refraction, as in 
