Trigonometrical Survey. 449 
cisely over the points, by centering the holes in the cross- 
boards. 
To such a part of the staff as was judged to be the most con- 
venient, the lamp was buckled, and its direction obtained by 
bringing a mark in the middle of it to correspond with another 
on the staff, which was determined to be opposite the station, 
by directing a ruler to it from each side of the staff, and mark- 
ing the places touched. The distance between those marks 
was then bisected, which gave the situation for the middle of 
the lamp. 
In a very early stage of the business it was found, that the 
effects of heat and cold on the limb of the instrument were 
likely to produce the greatest errors ; for if the canvas parti- 
tions, forming the sides of the observatory, were open to wind- 
ward, streams of air passing unequally over the surface of it 
would cause such sudden effects, that little dependance could 
be placed on any observations made with the instrument in 
such a state. To avoid this ; it was the constant practice when 
the wind blew with any degree of violence, to prevent the ad- 
mission of it as much as possible, by keeping up the walls of 
the external tent, leaving only a sufficient opening for the dis- 
covery of the lamp or light ; and at other times when the wind 
blew moderately, and a greater difference appeared in the 
readings of the opposite microscopes, than an error in division 
might be supposed to produce, the walls of the external tent 
were entirely thrown down, and the instrument kept in an 
equal temperature by the admission of air on all sides. 
In taking the angles, it was a general rule for some person 
to keep his eye at one of the microscopes, and 'bisect the dot, 
as the observer moved the limb with the finger-screw of the 
