338 Mr. Nixon on the Measuremetii of the 



the instrumental error. Then if we place a second or proof 

 telescope with its object-glass nearly in contact with that of 



the imperfect telescopic-level P, and make their lines of col- 

 limation parallel to each other, we shall find, on reversing the 

 latter, that the line of coUimation of an object-glass substi- 

 tuted for the eye-tube E will be inclined to that of the proof 

 telescope HR by twice the instrumental error. As the rays 

 of light from the intersecting point of the cross wires of the 

 telescopic-level pass out of the additional object-glass into 

 that of the proof telescope parallel to each other, make Q'R 

 parallel to QO, and Q'RE will be double the angle PEO = 

 QOE. Or, more simply, if we make E'O' parallel' to the 

 horizontal line EO, the line QR is elevated, and RE de- 

 pressed by an angle equal to the instrumental error. 



Description of the Apparatus. 

 On an oak table (A) 3*5 feet long and 2 feet broad, which 

 stood against the eastern wall of a long room having but one 





M 



m 



^ 



dfczsP. 



window (to the south), was placed a deal plank (B) 6 feet 

 long, 3-5 inches broad, and very nearly 6 inches thick, or 

 high ; equal lengths of the plank overhanging each end of 

 the table. To prevent the plank upsetting, a bar of wood 

 about a foot long and 3 inches square, was fixed to each side 

 of it, their bases being slightly elevated above that of the plank, 

 that the latter might not bend by the bars coming in contact 

 with the table. 



Two blocks of wood (CC) 3-5 inches high, 2-5 inches 

 broad, and of the length of the breadth of the plank, were 

 glued to its upper surface, one at a distance of 6 inches, and 

 the other at that of 17 inches from its south end. Firmly glued 

 to these blocks, stood an achromatic telescope set in a stout 



mahogany 



