Survey of Northeast Greenland. 241 
As far as the Danmark-Ekspedition is concerned the matter 
stands nearly as follows: 
ad 1. The mean error of the astronomic determination of longitude 
is 36” or very nearly 256 m (see р. 129). 
ad 2. The two values of the azimuth Observatory —Cairn У 
deviate from each other by 2”. It will of course not do to base a 
mean error determination on this. On the other hand one may say 
that this discrepancy in regard to the division of the circle, reading 
methods, placing on the granite pillar etc., corresponds to what might 
a priori be expected, as well as that a difference of 10” would have 
been surprisingly great. (For further particulars see under Deter- 
minations of Azimuth pp. 101—103). If we put the actual error of 
the azimuth of Cairn V from the Observatory at 20”, we must be 
said to have reached the limit of what is practically possible. An 
error of twenty seconds in this place will, however, only entail an 
error of about 12m on the longitude of Haystack. 
ad 3. The mean error of the base measurement is 1 : 10000 of 
the base. If the base is presupposed to be erroneous to this extent, 
the longitude of Haystack will in this manner undergo a change of 
about 2m. 
ad 4. It is more difficult to form an estimate of the influence 
of the many small partial errors in the triangulation. The influence 
is twofold, in that it expresses itself as an error partly in the lengths 
of the sides of the triangles, partly in the azimuths of the latter. 
By comparing the computed lengths of the sides of the triangles, 
one will, however, get the impression that the error in the sides of 
triangles of 1906 cannot reasonably exceed 1:10000 of the lengths 
of the sides. In the case of the sides of the triangles at the stations 
XV, XXIII and XXIV, which are of any importance to the position 
of Haystack, the uncertainty, it is true, is somewhat greater, but 
may hardly, with any reason, be estimated at more than 1:5000 of 
the lengths of the sides. The lengths of the sides XXIII—Haystack 
and XXIV—Haystack are uncertain, but it must be borne in mind, 
partly that errors in the lengths of these sides have practically no 
influence on the longitude of Haystack, partly that these lengths of 
sides to a certain extent are controlled, because the latitudes com- 
puted for Haystack from north and south harmonize as well as one 
might a priori be entitled to expect. 
After this one may take it for granted that the uncertainty of 
the lengths of the triangle sides does not practically influence the 
longitude of Haystack. On the other hand, this hardly holds good 
of the uncertainty with which the azimuths of the sights at Haystack 
are encumbered. 
16* 
