220 I. P. Kocx. 
was built and with a centred instrument. The station was situated 
on a rather small ledge, projecting from the wall of the rock and 
with a vertical fall towards the valley. After the observation, during 
the building of the cairn, it proved difficult to make the axis of the 
latter coincide with that of the instrument. Therefore it is very likely 
that there is a small centration error in the angles of Station X. 
Through the two recorded triangles a double computation of the 
f 4.26942 
side VIII—X was performed, which gave the result log VIII—X = \ 4.96940" 
The accuracy attained. 
Before I passed on to compute the sights at the stations VI, XIII 
and M, I endeavoured to find an expression for the accuracy attained. 
The single groups of triangles were, it is true, as will appear from 
the preceding, computed under a fully satisfactory control of obser- 
vation and computation, but this did not exclude the fact that the 
errors from group to group might have accumulated and so have 
become quite considerable. An expression of the accuracy might be 
gathered by computing the side X—XI, partly direct from the base 
net through triangle [IX—X—-XI and partly from the side УШ--Х | 
through the triangles УП-УШ-—Х and VII—X— XI. 
Triangle VII-— VIII—X The double computation 
Уи 116°31'27” + 00” of the side VII—X gave 
VIII 37 4953 +04” i f 4.10542 
X 253835 +01" a zn 
179°59'55” + 03” 
Triangle VII—X— XI The computation gave: 
X 34942'26” log X— XI — 3.87307 
XI 112 3250 Se un 
VII 320°44'44” 
Triangle IX—X—XI The computation gave: 
IX 96914702” + 03” log X—XI — 3.87297. 
x 282427 +05 Dee 
XI 55 2120 + 03 
179°59'49" -L 11” 
The length of the side X—XI corresponding with the two loga- 
rithms is 7465.7 m and 7464.0 m. The difference between these two 
values, 1.7 m, is nearly — of the side X—XI. As the control has 
now been looked for on a side, which is placed so unfavourably as 
it could well be, in that the one value is found by a computation 
through the whole of the triangulation net and by using one triangle 
of a less favourable shape, it seems probable to estimate the un- 
