184 Mr. Wollaston's description of the 
from the uncertainty how far sensibility might be carried in 
them. In one instance my thread was so fine, and I had 
made so large a bulb, that every degree on Fahrenheit’s 
scale was equal to 10 inches, and by connecting different 
threads to different bulbs, I have had them of all varieties 
from that length to half an inch. 
The instrument, with which I have made the greatest num- 
ber of observations, has a scale of 3,98 inches to every de- 
gree ; the thread, which is 22 inches long, was proved before 
its attachment to the bulb, and being found not cylindrical* 
the proper allowance has been every where made for the 
variations in the different parts of it. The degree was ascer- 
tained by comparison with a good thermometer at low tem- 
perature before closing the tube. The degrees I divided into 
100 parts on the scale, and by a Vernier into 1000. This has 
been compared with a common barometer, the height being 
always corrected for temperature according to General Roy’s 
Table in Phil. Trans. Vol. 67, p. 687. With this its agreement 
has been very close, after I had detected by means of it and 
had corrected two inaccuracies in my barometer, which would 
otherwise have escaped me. One was in the total length 
from the basin to the scale, and was ascertained by a com- 
parison made by means of a thermometer between my own 
barometer and two excellent mountain barometers, by 
Troughton and Cary. The other appeared by a want of 
agreement at low barometrical heights, when they agreed 
well above. For this I was at a loss to account, till I conjec- 
tured that it might arise from the greater quantity of mer- 
cury then expelled from the tube of the barometer and rising 
into the upper part of the basin, where the wooden box from 
