106 Mr. J. Nixon o?» the Measurement of the 



the square one occupied its place. Keeping the sector-level 

 exactly at its mark, a certain division of the pearl slip of the 

 square telescope was made coincident with the horizontal wire 

 of the sector. Fortin's level was then read off. 



Substituting for the sector Dollond's telescope, with its level 

 stationary at the previous mark, the division of the narrow 

 slip of the square telescope was got in a line with the middle 

 division of the broader slip of Dollond's telescope, (both slips 

 being longitudinally bisected by the same (imaginary) vertical 

 line,) and the position of Fortin's level subsequently noted. 



It is now evident that when (the centre of) the bubble of 

 Fortin's level stood at a point of its scale between, and equi- 

 distant from, its last and previous positions, the line of colli- 

 mation of the square telescope must have been truly horizon- 

 tal. This level point, as it may be termed, came out, 



Feb. 9th (1832), at ]24°-5 of Fortin's scale. 

 10th 124 -7 ditto. 



11th 124 -5 ditto. 



As one degree of the scale answers to 0"'6, the deviation from 

 the mean was limited to a range of 0"'l. Hence the method, 

 from its great simplicity and accuracy, might be successfully 

 applied to astronomical instruments. 



Lastly, the square telescope, being still in its place with its 

 bubble at the level point, the sector was placed to the north 

 of it (in lieu of Dollond's telescope), and its horizontal wire 

 made coincident with the division of the slip before as well as 

 after the cylinder of the sector had been inverted within its 

 Ys; the reading of the great level uppermost succeeding each 

 coincidence. 



With the difference between the reversing point and re- 

 gistered position of each level we compute the observed de- 

 pression of the (level) line of collimation of the square tele- 

 scope, which will be equal to the error of the sector. 



During the experiments a thermometer, constantly on the 

 plank, ranged between 55° and 57° Fahr. ; but a more certain 

 criterion of the uniform temperature of the levels was afforded 

 in the lengths of their bubbles, which occupied constantly the 

 same number of degrees of their scales. 



The results of the measurements were, 

 1832. Feb. 9th. Temp. 57° F. Error = IS"-!) 



loth. 57 = 20 -1 VMean 19"-1 



11th. 57 =]9-lJ 



Repetition of the Fourth Method, "with the Ohject-Glass fixed 

 lioithin the Eye-end of the Cylinder. 

 Being unable to explain why the average of the measure- 



