Mr. J. Nixon on his Horizon-Sector. 327 



and at the Scilly Islands, indicate a cleavage, or tendency to 

 split in the same direction. 



I apprehend that granitic ranges occur in various parts of 

 Europe possessing identical characters, as far at least as the 

 joints are concerned, respecting which it seems desirable to 

 obtain specific information bearing on the points in question ; 

 for I think it must be evident, that wherever a conformity in 

 the cleavage, or indeed in the joints of different ranges of 

 granite can be established, especially where such ranges are 

 found in contact, there is good reason to consider them con- 

 temporaneous, however differently they appear to be circum- 

 stanced in other respects. r # -y^ p ox# 



LIV. Particulars of the Measurement, by various Methods, of the 

 Instrumental Error of the Horizon- Sector described in Phil. 

 Mag. vol. lix. By Mr. John Nixon. 



[Concluded from the Lond. & Edin.Phil. Ma<r. and Journ. vol. i. p. 108: 

 with Figures ; Plate IV.] 



By the Second Method. 



Theory. — A RAY of light falling, at any angle of depression, 

 f~ on the surface of a fluid at rest, is reflected from 

 it at the same angle of elevation. Hence parallel rays are 

 reflected from a level surface at their previous equal inclina- 

 tion to the horizon. The elevation of any direct ray (A. fig. 1.) 

 at any point (R) of a horizontal surface (HH') will be the same 

 as that of the reflected ray (C). Another ray B, parallel to 

 A, intersecting C at any point * of it P, will be seen from that 

 point at an elevation equal to that of A at R, and equal to the 

 depression (at P) of the reflected ray C in the direction of R. 

 PL being parallel to HH', the angle ARH will be equal to 

 BPL, and also to LPR, and the sum of the two latter angles 

 equal to BPR. The rays A, B, although sensibly parallel, 

 might have diverged from one point situated at a considerable 

 distance. Rays from a (fixed) star would be of this description. 

 At P, a star observed in the direction of B, would also be 

 seen by reflection from a level surface below, as HH', in the 

 direction of R. As the elevation of the star should equal the 

 depression of its reflected image, an instrument having a con- 

 stant error would not give the two angles equal. However, 

 as the one would be exactly as much in excess as the other 

 was in defect, half their sum would be the correct quantity, 

 and half their difference the constant error, additive to, or 



* The horizontal distance of P to R should not exceed a few feet, or the 

 direction of gravity at the two points will not be sensibly parallel. 



