182 PROCEEDINGS OF SECTION A. 
directions measured after and before noon, and the general con- 
clusion was come to that steps were to be taken to reduce the 
effects of refraction by varying the conditions under which the 
observations were made rather than by seeking precision from 
extraordinary accuracy of the instruments used. In this view 
a 10-inch theodolite was substituted by Dr. Gill for the 18-inch 
instrument previously employed, and throughout the remainder 
of the work morning and afternoon observations were combined, 
the very satisfactory result being that with an instrument of 
less than one-fourth the weight the probable error of an observed 
angle was reduced from + 0-49" to + 0:33”. The question 
of weight of instruments is so large a factor in the cost of sur- 
vey, affecting as it does not only the expense of moving from 
station to station but also the length of time occupied by the 
work at each station, that anything which will serve to reduce 
the weight of apparatus appreciably without militating against 
the accuracy of the results demands fullest attention, and it is 
with the hope that the subject will receive the general attention 
it deserves that it is here introduced. 
MEASUREMENT OF ANGLES. 
The method of measuring angles, followed during the earlier 
years of the survey of New South Wales, was to observe the direc- 
tion of one point as a referring mark and then to sight to each 
station consecutively in order of azimuth, completing the surround 
by again observing the referring mark ; the instrument was then 
reversed and the same round of stations was observed in this 
position, similar rounds being taken after moving the graduated 
circle so that each station was observed upon different parts of the 
arc in order to eliminate graduation errors. About twenty 
observations on to each station were deemed sufficient. In the 
later angle measures, the method has been to reverse the telescope 
after each round and observe in reverse order of azimuth back 
to referring mark, twenty such double sets being altogether 
observed on different parts of the circle. 
COMPUTATION OF TRIANGULATION. 
Prior to 1890, when the survey was resumed after having been in 
abeyance some years, the azimuthal directions derived immediately 
from the observations were tabulated, without adjustment for 
correction of reading or bisection of referring mark, and Peirce’s 
criterion was applied in order to determine whether any should be 
rejected ; also, in the computation, no adjustment for “figure” con- 
ditions was made except in the case of one or two quadrilaterals. 
In the primary work done since the year named, the practice in 
