in order to afcertain the height of Mountains . 517 
mean of all was taken. The fame was done with each 
angle at a, b, and c, horizontal as well as vertical, viz, 
the elevation of c above a and b was taken ; and alfo the 
depreflion of a and b below c. The advantage of this 
method was, that the error of the line of collimation, the 
effedt of refraction, and of the curvature of the earth’s 
furface, all became equal and contrary ; by thefe means 
the little errors were diminifhed, and great errors abfo- 
lutely avoided^. I {hall, however, beg leave to fetdown 
the operation at length refpedting this one triangle, in 
order to fhew the precilion that may be expected from 
fuch a geometrical procefs; to remove the fcruples of 
thofe gentlemen who fufpedt that accuracy is only to be 
obtained by large quadrants ; and laftly, to do juftice and 
fatisfadtion to the celebrated artift who invented and 
made this valuable inftrument. 
(c) I muff acknowledge here, that the attra£lion of the mountain-creeps into 
the account uncorre&ed for, but only half of this quantity influences the mean 
refult, as at the top it was nothing, and at the bottom of the mountain it could 
not exceed 10" in the dire&ion ac, as I find from a rough computation, the half' 
of which zz 5" would give only four inches for the corre&iom 
