Survey of Northeast Greenland. | 279 
From the result it will appear at once that the observation of 
7 Andromedæ is not good, and there is every justification for thinking 
that a considerable error must have crept into the 34 pair of observations. 
However, the observation journal contains no remark to this 
effect, so that from a purely theoretical point of view this pair of 
observations should scarcely be condemned. At observations of this 
kind, theoretical considerations are, however, of no great importance. 
In practice it will undoubtedty be most correct to condemn the 
whole observation of у Andromedæ and stick to € Herculis. 
In all the hitherto given examples, from determinations of latitude 
as well as from those of time, we have exclusively had to do with 
observations of zenith distances, the most primitive of all methods, 
and consequently the most suitable on sledge trips, where the condi- 
tions are also very primitive. When gradually realizing the consi- 
derable accuracy which might be attained by means of zenith distance 
observations with the small Hildebrand universal instrument, the 
question presented itself, whether we might not, by other methods, 
carry the accuracy still further. As a rule, a greater accuracy than 
the one attained through our zenith distance observations was not 
essential from a practical point of view, but in exceptional cases tasks 
might occur, where a higher degree of accuracy would be desirable, 
for instance, at the above-mentioned determination of longitude 
between Danmarks Havn and Pustervig in the autumn of 1907. 
With the large instrument permanently set up in the Observa- 
tory the determinations of time, in the course of the winter, were 
nearly always performed by transits on the vertical of a or 2 Ursæ 
Minoris. It was therefore, quite natural that we should try whether 
this method might be suitable for the small travelling instrument. 
For this purpose we used the travelling universal instrument No. 3521, 
which had a particular reticule, with three vertical threads. 
The method is explained on p. 116. 
The observation was performed in the open air, in the shelter of 
the house at Danmarks Havn. As usual, the instrument was placed 
on the transportable tripod. During the observation the horizontal 
axis was covered with rime, which had to be wiped off, together with 
that on the feet of the level of the axis. For this purpose a stick 
covered with chamois leather would have been very convenient, but 
as we did not possess it, we had to use a piece of ordinary chamois 
leather. In the same manner the readings of the level became 
rather uncertain, which is particularly awkward in high latitudes. 
The computation also seems to show that the reading of the level 
is encumbered with grave errors; particularly in the case of 7 Piscium, 
