ON THE ELLIPTICITT OF THE EAETH. 457 



all differences are softened down by the extent of the 

 comprehended arc. 



Now the ellipticities derived from the four combina- 

 tions are as follows: 1. Polar and equatorial stations, 

 __L._; (Sabine, Pend. Ex. 1825, p. 334.) 2. Equa- 

 torial and middle latitudes, -^-g ; (Kosmos, B. iv. S. 28 ; 

 Eng. tr. vol. iv. p. 28.) 3. Polar and middle latitudes, 

 -gi-g-; (Kosmos, idem.) 4. Polar and equatorial with 

 middle latitude stations, -fa ; (Kosmos, idem.) Com- 

 paring the differences thus shown with the statement 

 that has preceded them, we may perceive that they 

 correspond precisely to what would be occasioned by an 

 error in some part of the process, by which a strict in- 

 tercomparability has been supposed to be established, 

 between the results obtained by myself with the invari- 

 able pendulum in the equatorial and polar latitudes, and 

 those obtained in the middle latitudes with Borda's 

 apparatus. 



From a conviction of the great difficulty of establish- 

 ing an unexceptionable intercomparability between the 

 results of experiments in which the objects, methods, 

 and apparatus were dissimilar, I did not hesitate, when 

 investigating in 1825 the same question of uniformity 

 or otherwise of the ellipticity in the two portions of the 

 northern quadrant, to. prefer the combination of Captain 

 Kater's results at the principal stations of the British arc 

 of the meridian with my own in the equatorial and high 

 northern latitudes, to a combination of the experiments 

 on the French arc with my own. In the first combination, 

 that of Captain Kater's experiments and mine, the appara- 

 tus was the same; the method of experimenting identical ; 

 the connection between our separate series such as could 



