Instrumental Error of his Horizon-Sector. 1 05 



indications of Fortin's level, to be of equal height. It was 

 discovered that the process of adjustment had, however, dis- 

 turbed the parallelism of the vertical wire to the plane of the 

 arcs ; the latter being now a little out of perpendicular, yet 

 by a quantity too minute to affect the I'esult of the measure- 

 ments, unless the figure of the cylindrical rings proved enor- 

 mously bad. 



The horizontal wire of the sector being scrupulously ad- 

 justed, the vertical one served to regulate the pearl slips of 

 the pi'oof-telescopes ; effected by making their edges parallel 

 to it. With this precaution, cross levels for the proof-tele- 

 scopes, otherwise indispensable, were quite superfluous. 



To insure the constancy of the adjustment of the zero-line 

 of each index to that of its corresponding arc, Mr. Lealand 

 had recently fitted to each index a powerful clamping appa- 

 ratus. And it may here be remarked, that the clamping of 

 the indices, the fixing of the levels, &c., were completed many 

 days before the measurements were undertaken, and had 

 passed in the interim tln-ough a range of temperature and 

 moisture of great extent. 



The trial of the reversing points of the great levels of the 

 sector took place, as soon as possible after every set of mea- 

 surements, with the instrument standing on the plank, to 

 which it was generally glued, in lieu of placing it, in a dif- 

 ferent temperature, on the stone pillar. To guard against 

 the slight unsteadiness of the plank, the square telescope was 

 stationed near the sector, the bubble of its highly sensible 

 level being kept stationary between two certain marks of its 

 scale some moments before the reading of the sector levels 

 commenced. The trials, which were very numerous, were 

 not so accordant with each other as might have been antici- 

 pated from the extreme care devoted to them. Some of the 

 discordancies wei'e wholly inexplicable ; one level continuing 

 constant during the repetitions, and the other oscillating 2" 

 or upwards about what proved to be the mean. * 



Process of Measurement. — The sector* (of which the line of 

 coUimation will represent the ray of light from the distant 

 object of our theory,) being placed at the northern end, and 

 Dollond's telescope at the southern end of the plank, the mid- 

 dle line of the pearl slip of the latter was got to be in a line 

 ■with, or of the height of the horizontal wire of the sector, the 

 vertical wire of the latter apparently bisecting the slip. The 

 level of the sector, then uppermost, and that of Dollond's tele- 

 scope being both noted, the other telescope was removed, and 



* Any other telescope wouhl have equally answered the purpose. 

 Tliird Series. Vol. 1. No. 2. Aii<r. 18.'J2. V 



