iiistnimental Error of his Horizon-Sector. 91 



not be more than V or 2' out of plane; a quantity too small 

 to vitiate the measurements *. 



Measurement of the equivalent angle of the divisions of the 

 level scales. 

 1. Sector levels. — The sector was placed upon the pillar 

 in the shade, weighted, and the bubble of the cross level brought 

 to its mark. When the sector had acquired the temperature 

 of the air, (noted by a thermometer placed close to the pillar,) 

 the zero of the index carrying the level then uppermost was 

 made coincident with that of its arc. With the nut of the 

 stand the bubble was brought nearly to the beginning of the 

 scale, (numbered onwards from the eye-piece towards the ob- 

 ject-glass,) and by-and-by noted and registered for both ends 

 in one column A. With the nut of the index the latter was 

 moved until the bubble had run nearly to the object-end of 

 the scale, and its ends, perfectly stationary, registered in another 

 column B. This process forms one observation. Secondly, 

 with the nut of the stand the bubble was again brought towards 

 the eye-end, and by the nut of the index afterwards moved 

 towards the object-end of the scale. This constituted the se- 

 cond observation. When five observations were completed, 

 the total arc was read off without disturbing the instrument, 

 and the process repeated up to about ten observations, when 

 the final arc was read off. Half the difference between the 

 sum of the column A and that of the column B is the de- 

 grees of the scale answering to the arc last read off, and di- 

 viding the latter by the above difference we obtain the value 

 in seconds of one division of the scale. This value we can 

 confirm by the arc first read off and the difference of the two 

 columns up to the time of reading off The results are sub- 

 joined. 



Right-hand Index Level. 

 March 6th, 1830. Arc 6' 17" 1° = l"-74 Temp. 36°'5 

 7 1-76 36 



Left-hand Index Level. 



• Admitting tiie rings of tiic sector to form together the frnstimi of a 

 cone of which the axis (during the ineasiirenicnts) is inclined 1(5" and the 

 upper surface 2H", the error {mhiiis) introciuced by iin inaccuracy of adjust- 

 ment of the cross level amounting to 10', would be equal to (28— 16) = 

 1 2" X versed sine of 10' ; or to C'-OOOOr.. 



N2 III 



