388 Mr. Gilpin's Observations on the Variation , 
variation observed in the house was found to be greater than 
that observed in the yard by 5', 4. The mean of nine sets of 
observations taken in the morning giving for the error 5', 5. 
And the mean of eleven taken in the afternoon giving for the 
error 5^3. The variation in those tables have therefore been 
lessened by the above mentioned quantity 5^4, as the error 
for the effect of the iron work of the room on the needle in 
the house. 
I must not omit to mention that of these 20 sets of obser- 
vations mentioned above, nine only were made with the 
compass in the same situation, and eleven in that of a diffe- 
rent one ; for, after nine sets had been taken, a pile of boards 
was put up between the compass and the mark to which it 
had been adjusted, which made it necessary to remove the 
post on which the compass had been placed, a few feet to 
the westward of its former situation, to clear it from the said 
pile of boards ; and eleven sets of observations were made 
from this new station, with the compass adjusted to the same 
mark it had been adjusted to before, and the angles that this 
mark made with the true meridian from each of these stations, 
were ascertained by placing a transit-instrument precisely 
where the compass had been placed, and observing transits of 
the sun and stars, in the same manner as has been described 
in finding the angle of the mark that the compass was ad- 
justed to in the house. And it is conceived that this accidental 
circumstance adds some weight to the accuracy with which 
these operations were performed, as the error from the two 
results of nine, and eleven, does not differ so much as o' 5 
from each other. 
