146 I. P. Косн. 
It should be remarked that the .sights, both towards east and 
west, traverse the land. 
The total impression gained from the researches of the Danmark- 
Ekspedition as regards the astronomical refraction may hardly be 
said to support the surmise put forth by BORGEN and COPELAND. 
The observations of stars, it is true, seem to indicate that the Bess- 
lian refraction was too great; but the deviation from the Besslian 
value is at any rate much smaller than supposed by BORGEN and 
COPELAND. As regards the observations of the sun we did not suc- 
ceed in getting a fairly reliable indication of a definite tendency; 
these observations in reality only show the well-known general un- 
certainty which is inseparable from the great zenith distances. How- 
ever, the eight determinations of latitude by the lower culmination 
of the sun and with zenith distances of from 83° to 86° give a very 
good value of the latitude, and in the same manner the twelve 
determinations of time by means of the sun — six towards east 
and six towards west — with zenith distances from 82° to 86° give 
upon the whole rather good results. In the case of the lower 
culmination as well as towards east and west the sights traverse 
the land. 
On the other hand the eight determinations of the latitude by 
the sun at the upper culmination and with zenith distances between 
71° and 84° show quite a definite tendency, which, in all probability, 
has a certain bearing upon the fact that the rays of the sun pass 
very flat through strata in the air in which mirages occur i. e. strata 
of the air the index of refraction of which varies greatly. Even 
though the single strata must in the main be supposed to extend 
horizontally, there are, purely locally, undoubtedly still considerable 
deviations from this, and in particular deviations from the horizontal 
extension of the boundary surfaces of the strata might be expected 
near the coast. But so the base of the theoretically computed ге- 
fraction fails altogether. 
If the surmise here set forth holds good, we must suppose that 
the astronomical refraction at great zenith distances is subject to 
purely local anomalies which change with the seasons. The observa- 
tions of the Danmark-Ekspedition only contain a very faint indication 
of this (the difference between the latitude determined by means of 
observations of stars and the sun in the upper culmination and by 
zenith distances up to 70°). The observations in Germania Harbour 
where the observatory was situated on a small peninsula projecting 
from a comparatively mountainous coast point, in a higher degree, 
towards such a change according to the season, in that the determina- 
