3+2 Mr. Nixon o?< the Measurement of the 



level lube (having a scale of rf'oth of an inch for 1") placed 

 parallel to the sides of the plank ; and no measurements were 

 made b}' the great levels until the ends of the bubble of the 

 reference-level were brought, by an altei'ation of tlie situation 

 of the observer, to be stationary between two certain divisions 

 of its scale. 



{Nevo) Verification of the Adjustments of the Foci. — It has 

 already been stated that the cylinder and telescope were placed 

 on the plank in a line, and level with each other. In this po- 

 sition, before the stand of the sector was glued to the plank, 

 the line of collimation of the telescope was made parallel to 

 that oi \he proper object-glass of the cylinder (that is, the in- 

 tersection of the wires of the cylinder appeared to be in a line 

 with, or intercepted by the intersection of the threads of the 

 telescope) ; when first one and then, on inverting the cylinder, 

 tlie other great level were carefully read off". In the next place, 

 the stand of the sector was taken off" the plank, and replaced 

 with a board 0*85 inch thick, introduced between it and the 

 plank. The cylinder being now I'aised nearly an inch above 

 the level of the telescope, their lines of collimation were once 

 more made parallel, and the great levels again read off*. The 

 results were: 



Bubble. Bubble. 



Sector-stand placed on \ 42°; 113° Left do. 33°; 123° 



plank ; right hand level J 

 Ditto on removal of ioaz-rf... 40i; 130^ 31|; 122| 



Mean 411:1311 32^:123 



Sector-stand placed on Joorrf 41 ; 131 — — 33; 123 

 (The divisions of the scales are about 2" each.) 



Flexibility of the Plank. — As one end of the cylinder was 

 very probably a few ounces heavier than the other, the distri- 

 bution of gravity might be sensibly varied, in case the flexure 

 of the plank was considerable, on reversing the cylinder 

 within its Ys. As a two-pound weight could however be 

 placed on any part of the plank without affecting the paral- 

 lelism of the lines of collimation, its flexibility was undoubtedly 

 too slight to vitiate the measurements. But on fixing a se- 



* With Kater's horizontal floating collimator the parallelism of the rays 

 might be elegantly demonstrated. Point the wires of a telescope having 

 a very large object-glass exactly at those of the collimator, the telescope 

 being placed as much as possible above the level of the collimator without 

 losing sight of its wires. The tube of the collimator may now be raised 

 several inches in level by gradually augmenting the quantity of quicksilver 

 within the trough, yet without disturbing the apparent interception of the 

 wires of the collimator by those of the telescope. 



cond 



