Instrumental Error of his Horizon-Sector. 315 



supported a deal plank 4^ feet long, 5 inches broad, and 3 

 inches thick, secured to them at a distance of 4 inches from 

 the wall*. 



To the upper surface of the plank, planed as horizontal as 

 possible, was glued towards its northern end the stand of the 

 sector. The cylinder contained, as before, its own and an 

 additional object-glass, but the latter was more effectually 

 fastened within it at the slit used for the eye-tube by two broad- 

 headed screws driven in at some distance from each other. 

 South of the sector stood the telescope, moveable, with the two 

 blocks of wood to which it was glued, along the surface of the 

 plank. Into its upper surface were screwed parallel to its sides 

 llie brass Ys of the proof-level, carrying Fortin's level-tube. 



As the plank did not prove perfectly steady, a mahogany 

 bar, mounted with two brass Ys supporting a level-tube, was 

 glued to its surface parallel and nearly close to the western 

 side of the sector-stand. The tube, 14 inches in length and 

 one in diameter, had a scale of y^th inch to 2". With the aid 

 of this reference-level the minute inclination given to the 

 plank by moving the telescope along it could be detected, 

 and readily corrected by placing one or more light weights 

 upon a proper part of the plank. 



As the axis of the reversed cylinder could not be made 

 parallel to the line of coUimation of the additional object-glass 

 without disturbing the cross wires of the cylinder, it was of 

 the utmost importance that one of the latter, when the bubble 

 of either transverse level stood at its marks, should be truly 

 horizontal. After this had been effected as correctly as was 

 practicable by the former plan, the telescope was pointed first 

 at one and afterwards at the other extremity of the wire; 

 Fortin's level being noted in both instances. The inclination 

 of the line being proved by the slight difference observed in 

 the two readings, it was gradually altered until it would endure 

 this severe test. The cylinder being inverted, the other trans- 

 verse level underv^ent a similar scrutiny. 



As the pearl slip was of unquestionable steadiness, one of 

 its divisions was substituted for the cross threads of the tele- 

 scope, and could be brought with rigid accuracy to be in a 

 line or level with the horizontal wire of the cylinder, by gra- 

 dually withdrawing from beneath one of the blocks fixed to 

 the telescope a sheet of thin paper previously introduced be- 

 tween the plank and block. At the same time the vertical 

 wire of the cylinder, as dimly seen through the diaphanous 

 pearl, divided the slip into two equal vertical sections. 



• Suggested by Captain Kater in his description of the floating colli- 

 mator. 



N. S. Vol. 10. No. r>0. Nov. 1831. 2 Y Mrf/iod 



