“74 H. G. Fourcade.—On the [April 26, 
graduation, that is starting from the left hand station at near zero. 
Care was also taken not to disturb the focus as the line of 
collimation did not coincide with the axis of the draw tube. 
Observations were made only in favourable weather (except at 
Rondevallei), during the period of steadiness of the images ; generally 
between sunrise and nine in the morning and from 3°30 to sunset in 
the afternoon. ‘Two end readings were taken and the mean of the 
results from O and 2, and from 1 and 2+1, repetitions, assumed to 
give the mean value. Thus with ten repetitions one-tenth was taken 
of the mean of the angles between 0 and 10 and between | and 11 
repetitions. 
These precautions, gradually introduced during the course of the 
earlier portion of the triangulation, were followed by the noteworthy 
increase in accuracy which is indicated below. It should however 
be also mentioned that the symmetry of the large stone beacons 
used in the first few triangies left much to be desired, an important 
source of error with sides of a few miles, whereas in the later portion 
of the work, good wooden signals were put up, chiefly to test the 
accuracy of the method. 
First 2 triangles mean close 64 
Next 5 a as 3°2 
Last 6 ae - 13 
‘The first two triangles, and two of the next five, were beyond 
‘the chain connecting the stations Hoogeberg and Belvidere of Capt. 
‘Bailey’s survey which were used as a base, so that their lesser 
-accuracy did not affect that of the connection. 
In a repeated angle, the a prior probable error cannot be deduced 
from the agreement of the observations with their means, as 
in the case of an angle obtained by reiteration ; but if the obser- 
-vations at Roode between Spijoenkop and Hoogeberg, which have 
already been given, be taken singly, and also those of another set 
taken at the same station between Spijoenkop and Rondevallei, we get 
4he results contained in the annexed table : 
