36 Mr. W. (ialbniilli on the Experiments 



places than theory supposes. On the same principles it is 

 less trom Figeac to Clifton ; and again, greater from Leith 

 to INIelviile Island. The irregularity of the densities of the 

 materials constituting the crust of the globe, combined with 

 the small errors of observation, produce a considerable num- 

 ber of discrepancies, as may be remarked from consulting 

 column A ,: but these appear to be insufficient to account for 

 such marked and regular differences, especially when regard 

 is had to their signs, as are shown in column A,. 



If it is not demonstrated by these observations, it is at least 

 rendered extremely probable, that the form of the earth differs 

 somewhat considerably from a regular spheroid. 



Since from the measurement of arcs near the equator, com- 

 pared with others in considerably high northern latitudes, the 



ellipticity is in general about — , while that derived from the 



same kind of measures at mean latitudes near the parallel of 



45° N. is about -—, nearly double of the former, — it is only 



necessary to give the arc of the meridian at 45° N. a greater 

 curvature than the elliptical hypothesis supposes, to satisfy 

 these conditions. 



Again : since the length of the pendulum is shorter at 45° 

 N. than theory requires ; if a protuberance or swelling out of 

 the earth be supposed to occur there, so as to place the ex- 

 perimental pendidum at a greater distance from the centre of 

 the earth, thereby rendering the gravitating force less, while 

 the equatorial and polar regions contract a little ; then the 

 theory would not disagree with the observed phasnomena. 



Thus, in the marginal figure, let PE/^Q 

 he a meridian of the earth formed by 

 a section passing through its centre, in 

 which P/» is the polar axis, EQ the equa- 

 torial diameter ; then suppose a zone or 

 belt passing round the globe parallel to 

 the equator, and nearly bisected by tlie 

 parallel of 45°, and of which zone, diminishing hi thickness 

 towards its boundaries, b a, c d, 8:c. are sections or opposite 

 meridians, the exterior curve P 6 6' E^j Q would be the form of 

 the meridian, to satisfy these conditions ; while the interior 

 curve is the regular ellipse. 



This conclusion must be looked upon merely as a very pro- 

 bable hypothesis, till by the accumulatioji of more numerous 

 observations it may be either completely verified or set aside. 



It is useful in the mean time to endeavour to connect ob- 

 servations, and reconcile these anomalies by the most pro- 

 bable 



