THE RIVER THAMES AT LONDON BRIDGE AND THE SEA. 
19.3 
position at each observation, and making a complete revolution in its collars : 
my assistant read off each, and noted it down in a book. I then went to the 
staff and examined its position on the picket, and read off the last observation 
myself ; my assistant then read from his book the last observation ; the two, 
of course, when correct, would correspond. 
The telescope of the instrument was then reversed in its collars (which is a 
good check to the adjustment at each observation), and by the motion in 
azimuth directed to the next station staff, when the same mode was used in 
observing : the spot was marked, and the instrument moved to the next station, 
and the station staff turned half round in its collars, and gimbles ready for the 
next level. At the end of the day’s work an additional picket was driven in 
about twelve or fifteen feet from that j ust used, and compared with it ; this was 
to ensure the detection of any alteration in the pickets during the intervening 
time, either from mischief or accident, being compared before the commence- 
ment of the day’s work. After a few days the ground was gone over again, 
generally from the opposite end, and two observations taken at each picket, 
which were sometimes more and sometimes less in number than in the former 
levellings ; these were the proof-levels. The distances were then measured, 
and the necessary angles taken to lay down the work. In this manner proof 
levels were taken up to picket No. Ill; but finding in this method that I was 
liable to great inconvenience and loss of time from many circumstances, and 
amongst others by the pickets being mischievously drawn or moved ; and if a 
trivial mistake occurred in the levellings (of several days back perhaps), it 
was only detected at the summary of the levels. I determined to endeavour 
to adopt some method of proof, not liable to the inconveniences of the former. 
I accordingly made use of the following mode, which I have found, after 
repeated trial, to be a most correct proof. 
After having finished the four observations at one picket, I threw the instru- 
ment out of adjustment by the foot-screws ; and after adjusting it again most 
correctly, I took a pair of observations ; but instead of reading off the staff' 
from the picket, it was now read from above, the staff being imaginarily con- 
tinued to the length of ten feet. The mean of these observations, therefore, 
gave the complement to 10,000 of the true observation : by this mode much 
time was saved, and an error, if any, was immediately detected of a tenth of a 
2 c 
MDCCCXXXI. 
