210 PROCEEDINGS OF SECTION A. 
is thought that alteration should be deferred until more definite 
steps, either by improvement of methods or more reliable observa- 
tion along the whole chain, materially add to the weight of 
evidence before us. 
The accompanying map (marked “K”) shows the positions of 
stations, the longitude of which with respect to the Sydney 
Observatory has been determined. Some of these have not been 
connected with by triangulation, and their places were fixed 
mainly for use in the construction of a map of the colony to take 
the place of one which has been in use for the last twenty years 
or more, and in which errors of as much as 8 or 10 miles have 
been revealed by the trigonometrical survey. The method of 
observing difference of longitude generally adopted in the United 
States and other countries has been somewhat varied in this 
Colony, and now it consists practically in causing the clock at 
(say) Walgett to record on the chronograph at Sydney, whilst the 
Sydney clock is recording on the same chronograph, and then the 
Sydney clock records on the chronograph at Walgett, whilst the 
clock there is recording its seconds. In this way it is possible 
to get the difference between the two clocks within a hundredth 
of a . second, and the problem is reduced to obtaining the actual 
clock errors at the time of comparison, and thus the difference of 
the time at the two places at the same absolute instant, and 
herein lies the whole difficulty. 
A programme is prepared of stars in groups of three or four, 
at or near the zenith, and one or two circumpolar stars, one if 
possible, being swb-polo. For this purpose those stars only are 
used whose right ascensions are well determined. It is necessary 
that the intervals between the stars should be as short as possible, 
so as to eliminate unknown variations in the chronometer rate. 
When possible the level is read, reversed, and read again just 
before and after each star. The times of transit of the stars over 
the wires are recorded on the barrel chronograph by the observer 
pressing the signal-key, every individual second being automa- 
tically “recorded by the chronometer. The instr ument is now 
rotated through 180 degrees, and another group of stars observed 
as before ; any error in collimation, and also inequality of pivots, 
is thus eliminated, and a mean of the two corrections deduced 
from these two groups is the clock correction at or about the 
middle time of all the observations. This process is repeated 
with two other groups of stars, and then, if practicable, clock 
signals are exchanged with Sydney. The observations then pro- 
ceed in same manner; that is, a group of stars, instrument 
reversed, another group, instrument reversed, and so on, until, if 
possible, a second exchange of signals is made, when two or more 
groups will close the evening’s w ork, A single night’s work (says 
Chauyenet), however, is not to be regarded as conclusive, although 
