Survey of Northeast Greenland. 281 
where the correction of the level is greatest and very much deviating 
from the others, the result differs a good deal from the other three. 
As the thread intervals were not known, we have simply used 
the mean of the transits above the three threads. Collimation correction 
is not applied, because the collimation had not been determined. 
The result is not altogether bad, and the certainty of the readings 
of the level — the weak point of the method — might perhaps, as 
mentioned above, also be increased. Still the experiment clearly 
indicates — as might a priori also be expected — that with an 
instrument as primitive as the little travelling universal instrument 
one ought to restrict oneself to the primitive zenith distance obser- 
vation. 
Determinations of azimuth. 
Whereas in the case of determinations of latitude and time one 
will nearly always be protected against the influence of gross errors, 
because already before the observation one obtains a rather good 
idea of the latitude and time, this is not the case as far as the 
azimuth is concerned. It is true that one may attain some control 
by means of the compass; but the slight horizontal intensity in high 
latitudes makes even a very sensitive compass so slow-working that 
one can hardly trust it within a few degrees. Consequently a deeply 
rooted distrust is quickly felt respecting this instrument, and its use is 
avoided as much as possible. Fortunately it is easy enough to attain 
control without having recourse to the compass, when the observation 
is arranged in such a manner that the azimuth may be computed 
both by zenith distance and time. This demands a simultaneous 
pointing of the sun (the star) both on the horizontal thread and on 
the vertical thread (Fig. 39), which dexterity 
| | 
one quickly acquires by a little practice. Ту | 
Where the greatest accuracy has to be 
s Е 5 : 3 | 
achieved, a proceeding of this kind is | 
perhaps objectionable, because a simul- 
taneous pointing of the two limbs will, 
in reality, easily result in perceptible errors”); but, as a rule, the 
utmost accuracy is not demanded in the case of determinations of 
azimuth on sledge trips. An error of 10’ in the azimuth will in 
Fig. 39. 
1) If the utmost accuracy is to be attained, the azimuth must, I think, be determined 
at noon or midnight by pointing the easterly or westerly limb of the sun, but 
without measuring the zenith distance at the same time. The azimuth will 
then be determined by the hour angle only, and therefore requires good 
determinations of time on both sides of the observation of azimuth, besides, 
of course, a good determination of latitude. On the Danmark-Ekspedition this 
method was never used on the slege trips. 
