166 Mr. W. Galbraith's Remarks on the Experiments ■ 



The sign + denotes that the value by computation is 

 greater than that by experiment, — less. 



Here we find, by inspecting the column of differences, that 

 the greatest errors fall on the lengths of the pendulum at 

 Clifton and Arbury Hill. At the latter place the length by 

 experiment being greater than that by computation, shows 

 that the gravitating force is increased by some irregularity in 

 the structure of the earth's surface, and thus speaking the 

 same language that the zenith sector of Mudge did about 

 twenty years ago. The contrary takes place at Clifton. At 

 Shanklin Beacon the error is almost nothing ; it is very small 

 at Leith ; and at Unst, London, and Portsoy, it is but incon- 

 siderable, differing only about one or two ten thousandth parts 

 of an inch. 



Returning now to equation (9), since we have found the 

 values of y and z we have 



20919576 y 



e = % X 



or, 



20919576+1856062635 x 3'2505779 

 e = 0-0086386— -- =0-0086386— 0-0055435= 0*0030951, 



z 



or — —nearly. This value, being derived by the principles 



of minimum squares from the totality of the observations 

 made by Capt. Kater between Unst and Dunnose, will, as truly 

 as the observations can be depended upon, represent the ellip- 

 ticity of the meridian passing nearly through these points. 

 If we compare Unst and Dunnose together, which are the 

 extremities of the series differing by 10° 8' 4" of latitude, we 

 should get /=39-007336 +0-215565 sin 2 A, and e=0-003113 = 



— -, a compression nearly the same as that derived from the 



mean of the whole, though rather greater. 



We shall now treat the experiments made by M. Biot in 

 the same manner. 



Following 



