[ 510 ] 
liars, differing confiderably in declination from one ano- 
ther acrofs the vertical wire of the feclor, and compar- 
ing; the obferved diiference of their tranfits with the 
known difference of their right afcenlions. If they agree^ 
it may be fafely concluded, that the inifrument is truly 
placed in the meridian. If not, by comparing the alte- 
ration that would be produced in the difference of the 
tranfits, by fuppoiing the inflrument out of the me- 
ridian, by any fmall quantity, as one degree or ten mi- 
nutes, vvith the obferved error, the deviation of the 
inflrument from the meridian may be inferred. In this 
manner I found, that the inllrument had been put 
very exadlly in the meridian by means of the meri- 
dian line; the difference by the two methods coming 
out only minutes of azimuth. As to the continu- 
ance of the inflrument in the plane of tb e meridian, 
I had a conflant proof of it by the fame means, and like- 
wife a farther fecurity, which I did not fail to attend to, 
by noting the degree and minute which an index de- 
pending on the vertical axis of the inflrument pointed 
out on a fixed azimuth circle. Being apprehenfive of 
error in an inflrument fupported on a wooden frame, 
I frequently examined the parallelifm of the fore arch to 
the back arch, by meafuring their perpendicular diflances 
at the two ends with a brafs fcale, whole vernier fliewed 
the five hundredth part of an inch, and found it liable 
to variations of a minute or two, owing probably to the 
force ufed in fetting the fedlor to different zenith dif- 
tances, and the weaknefs of fome fcrews at the top of the 
frame ; which fmall error I corrected, till I found it liable 
to 
