Instrumental Error of his Horizon-Sector. 341 



focus had been ascertained in the way described. On com- 

 mencing the verification, the latter was placed, purposely, out 

 of focus, but when altered so as to produce distinct vision of 

 the threads, its eye-tube was found, on every trial, to be at the 

 circle marking the sidereal focus. 



The object-glasses within the cylinder were placed at a di- 

 stance from the cross wires fixed between them equal to their 

 respective sidereal foci (the cylinder and telescope being si- 

 tuated as represented in the figure, p. 338) by drawing out or 

 pushing in the tube containing the proper object-glass of the 

 cylinder, until the wires, as viewed through the telescope, ap- 

 peared as distinct as possible. The cylinder being reversed, 

 the additional object-glass, now brought close to that of the 

 telescope, was similarly adjusted. 



(Netio) Method of setting a Wire vertical. — The upperjnost 

 (L) of the transverse levels of the cylinder having its bubble 

 at the mark, the slip of pearl had its divided edge rendered 

 parallel to the upright wire of the cylinder by unscrewing in 

 a very slight degree the eye-tube containing it*. The cylin- 

 der was then reversed within its Ys, and the bubble of the 

 transverse level brought to its mark by turning the cylinder 

 within its Ys. On looking through the telescope, if the slip 

 and wire were found parallel, both were considered as perpen- 

 dicular ; otherwise one half of the deviation was corrected in 

 the slip and the other in the wire, the bubble of the deranged 

 level being in the last place restored to its mark by its ad- 

 justing nuts. After the cylinder had been inverted and the 

 wire placed parallel to the (undisturbed) slip, the other trans- 

 verse level was similarly adjusted. In setting the lines of col- 

 limation parallel, or in reading off the great levels during the 

 subsequent observations, the bubble of the transvei'se level 

 uppermost was carefully kept to its marks. 



Rcfcrence-Lcvcl. — The horizontal piece of brass (K) in 

 which works that part of the screw situated between its head 

 and the threads was so thin as to be elastic, thus rendering the 

 measurements made with the micrometer discordant and un- 

 certain. Recourse was therefore had to the great levels of the 

 sector (omitted in the figure), not however without reluctance, 

 the unsteadiness of the floor now requiring the addition to the 

 apparatus of a reference-level. To the upper surface of the 

 westernmost (O) of the pieces of wood projecting from the 

 sides of the plank was fixed a bar of oak (P) with its eastern 

 side glued to the plank. The bar was mounted vviih a spirit- 



• The slip being only a few seconds distant from the intersection of the 

 threads, the eye could estimate the parailcliini with sufficient accuracy. 



level 



