194 ON THE OBLIQUITY 



December, 1808, and June, 1809, and is to be accounted for from the 

 great expansion to which the intense heat of these climates, subjects every 

 kind of metal. It is true that this cannot be supposed to affect the ad- 

 justment within the limits of an observation, but I never observed twice 

 at the distance of an hour without having occasion to alter by some 

 seconds the mircrometers, when set to the horizontal mark. 



6. The microscopes which read on opposite points of the circle are of 

 the usual construction, and contain micrometers the heads of which are 

 divided to 2 of a degree, and by estimation read to one. The wires in- 

 side act as in all such instruments, one as a fixed, the other as a moveable- 

 wire. 



7. There is but one level which answers for all adjustments. It is: 

 suspended on two short arms, projecting from the horizontal axis of the 

 circle, and is exquisitely sensible ; the tube hangs freely so that the air 

 bubble remains upwards, whatever be the position of the circle. This 

 level is supplied with the proper adjusting screws and perfectly com- 

 petent for its purpose, subject to the checks to which I had recourse. 



8. The error of centring is very trifling, and is variable. I never 

 observed it at its maximum to exceed 3. The error of collimation has 

 frequently varied from accidents, or other causes ; but since I refered to 

 an horizontal mark this became of no sort of importance, because pre- 

 vious to every observation, the moveable wires were always set at zero 

 against the mark. The thickness of the wire in the focus of the telescope,, 

 (which subtends nearly 6 ) is considered in the altitudes,* by observa- 

 tions of opposite limbs.. 



* In place of wires, a large spider's web, found in numbers in the Coorg country, is made 

 use of. It is exquisitely delicate : it is perhaps extraordinary that it, should have been found 

 on a great number of observations to subtend so considerable a quantity.. 



