94 Mr. Nixon on the Measureinent of the 



against the other arc*. The cross-level of the sector being 

 exactly, and that of the proof-level nearly at their respective 

 marks, the latter was rendered perfectly so (a little at the ex- 

 pense of the sector cross-level) by moving the arcs a second 

 or so out of a vertical plane. The proof-level and that of 

 the sector being read off, the former was replaced revei'sed 

 in direction, and read off, and finallj*, the cylinder being re- 

 versed within its Ys, the sector-level uppermost was read off, 

 but not before the proof-level had been replaced, as its weight 

 was found to alter the sector-level about 1". — The process 

 was then repeated, on the opposite surface of the cylinder, 

 with the other level of the sector. In these observations it was 

 always noted whether the telescope pointed north or south, 

 and the readings for all the levels were registered in two co- 

 lumns, one for the north ends, and the other for the south 

 ends of their bubbles. 



The conical figure of the object-end ring of the sector has 

 already been pointed out. With the reduced plates the proof- 

 level could, however, be pushed close to the shoulder of either 

 ring without causing a deviation of more than half-a-second 

 in Fortin's level. In reversing the sector the effect of the 

 conical ring would be even more insignificant; the play within 

 the Ys being scarcely perceptible. 



The measurements with the reduced plates were made, 

 April 13th, 1830; Fortin's level tube being secured within 

 the Ys of the frame by broad ribbon passed round the sockets 

 of the tube and the frame, and made perfectly tight by stitch- 

 ing together the ends of the ribbon with strong thread, — On 

 placing the eye in the plane of the horizontal plates of the 

 proof-level, their contact with the surface of the cylinder was 

 so uninterrupted, that scarcely a glimmering of light could be 

 seen anywhere between. The temperature being 49°, the 

 horizontal inclination appeared to be 31"' I by the right-hand 

 level, and 29"'2 by the other; mean 30"'15. The discrepancy 

 may arise from that surface of the eye-end ring, which is 

 uppermost when the right-hand level is made use of being 

 slightly corroded; or partly because the scales of the two 

 levels have not had their angular equivalents determined with • 

 equal accuracy. 



Second Method. — The cylinder being placed within its Ys 

 with the bubble of its level at the reversing point, the incli- 

 nation is found by placing the proof-level upon it direct and 

 reversed; half the run of the bubble giving the inclination re- 

 quired. 



In this method, especially when the cylindrical error is con- 



* The arcs are not exactly parallel to each other. 



siderablc, 



