[ 93 ] 



alteration of the ftem, will be fo little as to be quite imperceptible ; 

 but were that change ever fo great, fince the adjuftment is made 

 by what is afterward aftually the plumb-line, it can produce [no 

 error, provided the weight which is appended remains the fanie. 



By this means the level may be adjufted to a certain degree 

 of accuracy, but never to any thing near precifion ; this can be 

 eifeded by no means that I know, fo well as by the Hadley's 

 quadrant itfelf, to whofe ufe it is fubfervient, and from which 

 in return it moft readily receives its perfed adjuftment; which 

 is done by obferving,, when the image of a very remote objed, 

 as the fun or moon or a bright ftar in twilight, when near the 

 meridian and their altitude not great ; or the top of a very dif- 

 tant mountain feen in the mirror coincident or in contad with 

 the image of the fame bbjed refleded by the quadrant, at the 

 middle or edge of the tranfparent part of the horizon-glafs, will 

 preferve (in the fame place of that glafs) its coincidence unva- 

 ried, while the level is made to revolve half a turn round its- 

 axis : if the image in the level rifes or falls from that in the 

 quadrant, it muft be adjyfted by the fcrews as before. If the 

 quadrant is furnifhed 'with a telefcope that inverts, (by the help 

 of which the adjuftment may be rendered far more accurate than 

 it could be done by the naked eye) then the afcent of the image 

 in the mirror, from the fixed image in the quadrant, muft be 

 regarded as its defcent, and vice verfa. Any near fixed objed, of 



fucli 



