Survey of Northeast Greenland. 245 
- 
will then assume that the whole deviation of 5”.7 of the two values 
of the latitude of Muschelberg is due to an error in the base, which 
must then be supposed to have been measured too short by 1.8 metres 
or about 1/400 of its own length. The result of this would, however, 
only be that Haystack was shifted about 5” or about 40 metres to- 
wards the west. 
There still remains a valuation of the uncertainty with which 
the azimuths from Keferstein Berg and Kap Bremen at Haystack 
may be encumbered. As already mentioned by Koldewey, the former 
of these directions is given in whole minutes and may thus, if 
for that reason only, suffer from an error of !/s minute. If, how- 
ever, one maintains that the sight is possibly encumbered with errors 
of eccentricity, the error on the sight may be at most 3 to 4 minutes, 
in which manner the longitude of Haystack will undergo a change 
of about 100 metres. 
If one wishes to explain the discrepancy of the position of Hay- 
stack through an error of the sight Keferstein Berg—Haystack, this 
difference must be put at nearly one degree. In this manner the 
longitudes might be brought to harmonize, but at the same time 
discrepancies would arise between the latitudes of about 11/2 minutes, 
so that by this means nothing is gained. The same applies to the 
sight Kap Bremen—Haystack. In case the difference of longitude at 
Haystack has to be explained as a consequence of an error in the 
triangle Kap Bremen—Haystack—Keferstein Berg, one is forced to 
the supposition, as regards both of the azimuths Kap Bremen—Hay- 
stack and Keferstein Berg--Haystack, that they have an error of 
about 11/2 degrees towards east. A supposition of this kind is, how- 
ever, quite absurd. 
From the foregoing it will appear that the greatest uncertainty 
in the determination of the longitude of Haystack from Germania 
Harbour, also in this place, is due to the observation of longitude 
in the Observatory. It is true that, under certain rather improbable 
hypotheses the possibility of two other considerable sources of error 
has been demonstrated; but even when we have had recourse to 
these, we have not succeeded in finding a plausible explanation of 
the difference between the two determinations of the iongitude of 
Haystack, neither through the work of the Danmark-Ekspedition, nor 
through that of the Germaniaexpedition. One might suppose the 
observations to be encumbered by certain constant errors, which 
would cause the expressions of accuracy to become incorrect. As 
sources of such errors may be mentioned deviations of the plumb 
and errors in the refraction tables. Errors of this kind are pointed 
out by the Germaniaexpedition, as well as by the Danmark-Ekspe- 
