instrumental Error of his Horizon-Sector. 8 9 



First Method.— Ey finding the horizontal inclination of the 

 upper surface of the telescopic tube, resting within its Ys, with 

 the bubble of its level at the reversing point. 



Support for the Sector. -A block of compact micaceous sand- 

 stoneV (estimated to weigh two tons and a half,) hewn, with 

 triflino- waste of the material, into a pillar nine teet long, and 

 of the°average diameter of two feet, was set up, at the bottom 

 of a pit fou°feet and a half deep, with its thicker end based 

 on bedded rock. The excavation being gradually hlied up 

 with earth well rammed down, the pillar became so tirmly 

 fixed, that no change of situation of the observer, nor appli- 

 cation of his utmost force against its sides, affected its inclma- 

 tion more than a fraction of a second. After a apse ot some 

 days (extremely variable in the temperature), the top ot tlie 

 pillar was worked as even and as horizontal as practicable. 



Three cylinders of lead, each three inches in diameter and 

 nearly an inch thick, introduced within the lower part ot the 

 frame of the sector, served effectually to keep it steadily m its 



place upon the pillar. , , , i * 



Description of the Proof-Le-jcL-The level tube, made to 

 order by the celebrated Fortin of Paris, is twelve inches long, 

 one inch in external diameter, but considerably less withm, 

 the sides (to prevent flexure) being so thick as to require some 

 care in reading off the divisions without parallax bach end 

 of the tube is cemented into a brass socket or hollow cylinder 

 one inch and a half in length. The scale contains 230 divi- 

 sions, of one millimetre each, etched on the glass and stained 

 red, each division being equal to little more than halt a se- 

 cond, equivalent to one sixteenth of an inch for one second. 



The frame of the level was a well-squared bar of mahogany 

 full 17 inches long, IS broad, and 2 inches deep (or high). 

 At equal distances from the ends two horizontal brass plates 

 4 of an inch long, ^ of an inch thick, and nearly of the breadth 

 of the bar were fastened to its upper surface by two strong 

 screws. For the support to Fortin's level-tube each plate 

 carried a Y (formed out of the same piece of brass as the plate} 

 placed at a distance of 10 inches asunder, with their angular 

 points in a line parallel to the (vertical) sides of the irame. 



A brass plate | of an inch long, ^ of an inch thick, and ot the 

 breadth of the frame was secured to each end ot its under suriace 

 by four long screws driven in nearly at the corners ot, and 

 level with* the plates. These plates (previously attached to the 

 frame) were ground to a true plane upon twof contiguous 

 • In the measurements no part of the scrcw-heiuU would come in con- 

 tact with the cylindrical rings of the sector. 



t The diameter of neithc^r plate being equal to the length of the har^ 



N.S. Vol. 10. No. r,G. Aug. 1831. N non 



