328 Mr. J. Nixon on the Measurement of the 



subtractive from elevations, according as the angle of depres- 

 sion exceeded or fell short of that of elevation. 



As the rays of light from the point of intersection of the 

 cross wires of a telescope issue from beyond its object-glass, 

 placed at the sidereal focus, parallel to each other, it is obvious 

 that we determine at once the error of a telescopic-level on 

 measuring by it the elevation of the line of collimation of a 

 similar telescope, and the depression of its reflected image. 



Description of the Apparatus, and Method of Observation. 



A cistern, formed of a block of wood 9 inches long, 5 inches 

 high, and nearly equally broad, excavated with a narrow mar- 

 gin to the depth of a quarter of an inch below its upper sur- 

 face, was glued lengthwise upon the horizontal plank (already 

 described) at about the middle of its length. North of the 

 cistern was placed the wooden stand, supporting within its Ys 

 the collimator telescope, which had been lately fitted up by 

 Mr. Dollond, with a stop (in lieu of the slip of pearl), consist- 

 ing of a plate of brass, having in the centre a minute circular 

 aperture, through which exclusively the light derived from the 

 pasteboard, or transit lamp* fixed beyond the eye-glass, passed 

 into the body of the telescope. South of the cistern, now 

 filled nearly to overflowing, stood the sector, placed with its 

 telescope in a line with the collimator, the middle of their ob- 

 ject-glasses being about the height of, and almost close to the 

 cistern f . On looking through the telescope of the sector, no 

 reflection of the circular aperture (or luminous disc) could be 

 discerned, as the collimator was gradually inclined from its 

 original horizontal position, until the depression exceeded 12'; 

 an angle quite beyond the limits of the levels. By making 

 the line of sight inclined by at least this quantity to the axis 

 of the cylinder, that axis, and with it the adjusted line of col- 

 limation (which lies in the same or parallel direction), would, 

 however, be truly horizontal whenever the deviated line of 

 sight, as the cylinder revolved within its Ys, should be found 

 to bisect alternately the minute disc and its reflected imagej. 

 In this case the difference between the observed position of 

 the bubble of either of the great levels and its reversing point 

 would give the constant error. 



* The tube being deprived of its lens, the aperture was closed by a disc 

 of gauze paper dipped in mastic varnish. 



+ Found by experiment to be better than when the cistern stood either 

 lower or higher. 



% Demonstration. — In fig. 1. let PB be the deviated line of sight bi- 

 secting a star, and PR its direction, and that of the reflected star subse- 

 quent to half a revolution of the cylinder. As the latter revolves without 



