ON THE ELLIPTICITY OF THE EARTH. 481 



results computed from different portions of the meridian 

 should be the same. By multiplying the number of 

 stations the effects of the variations of superficial density, 

 sometimes in excess and sometimes in defect, tend to 

 counterbalance each other, and thus to diminish the 

 general amount of error ; whilst the increased extent of 

 the included arc gives to such error a diminished in- 

 fluence on the ultimate deduction. 



The discordances which were formerly found in the 

 results of pendulum experiments, when the number of 

 stations were comparatively few, or when the arc which 

 they included was of small dimension, appear to have 

 produced in M. de Humboldt's mind an impression 

 (p. 29), that the pendulum is a less likely means of ob- 

 taining a well-assured conclusion respecting the figure 

 of our planet than the measurement of degrees. It is 

 not surprising, indeed, when we look back upon the 

 formidable array of distinguished geometricians and 

 astronomers who have taken part for more than a cen- 

 tury past in the measurement of degrees, that there 

 should be a bias in favour of the ultimate conclusions 

 from a method on which such a prodigious amount of 

 time, means, and skilled labour have been expended. 

 Still, there have not been wanting eminent persons who 

 have been of a different opinion ; and who, viewing the 

 discrepancies which have also presented themselves in 

 the ellipticity deduced from different measured arcs 

 discrepancies arising from causes similar to those which 

 affect pendulum results have anticipated that the pen- 

 dulum would eventually be found amongst the most 

 accurate means of determining the general configuration 

 of the globe. Happily, we have arrived at the time 



VOL. IV. I I 



