266 AN ACCOUNT OF THE 
cope of preat magnifying power, and vernicrs reading to 5 it does appear, 
but a fair suppesition, that angles could be observed to that degree of 
accuracy. ‘The divisions are however on brass, which renders them difficult 
to read with certainty. However judging from the extreme error In the 
sum ofthe three angles, and supposing it to be the same way on every angle 
of the three, we shall get 7 as the extreme possible error on each angle, 
“Now if we take an equilateral triangle, (not too favorable a supposition), 
we shall find that this error.on each of the two angles used in concluding 
the new side, and supposing them to be in ihe most unfaverabie sense, 
would only affect the result by =;;, part ofthe whole. But even in a few 
triangles, this error ought in a great measure to correct itself, so as to pres 
yentthe error increasing in the ratio of the number of the triangles, 
4. Now the closisg station is brought in at the 14th triangle, and if all 
those that only answer as checks be rejecied, will be but the 10th in 
order, ‘This would appear te be a sufficient warrant against any great ac- 
cumulation of error, but [ have as a check chosen io follow out the result by | 
other series, ‘Fhe 35th figure, furnishes the 3d value of the distance of 
the two principal stations Surkexda and Chandpur; the mean of the three’ 
values, is taken for the foundation of the large triangulation, Those after 
the 35th, are meant from some of the preceding results, to deduce the dis- 
iances of the intermediate stations of the great triangulation, and in one 
case, by means of a concluded angle. But this result is checked again by 
one of the great triangles, “4 
