ON THE ELLIPTICITY OF THE EAKTH. 463 



Sierra Leone and Spitzbergen should be diminished on 

 account of the more accurate correction for the differ- 

 ence in the temperature of the pendulums, as it was 

 previously shown that it ought to be increased by 

 O s '34 as a more accurate correction for the reduction to 

 a vacuum ; both the more recent values of the correc- 

 tions being consequences of Bessel's discovery. Sierra 

 Leone and Spitzbergen are stations of nearly extreme 

 difference, both in the temperature of the pendulums 

 and in the temperature and density of the air: the 

 effect of Bessel's discovery on the acceleration as ori- 

 ginally computed, and, consequently, on the ellipticity 

 originally deduced, is seen, therefore, to be wholly 

 insignificant. 



Thus it has happened that, from my having adopted 

 what appeared to me a more practical mode of ex- 

 amining the influence of differences of temperature on 

 the time of vibration of the pendulums, viz., by caus- 

 ing them to vibrate at one and the same station in 

 temperatures differing so widely as to include the whole 

 range of temperature experienced elsewhere, (instead 

 of deriving the correction from the expansion of brass 

 measured by pyrometric experiments, as previously 

 practised), the acceleration originally computed from 

 my experiments in different latitudes has stood in need 

 of little (and that little quite unimportant) correction as 

 a consequence of Bessel's discovery in 1828. The reduc- 

 tions to a uniform temperature are indeed everywhere too 

 small, if regarded as representing only the effect of the 

 expansion of the metal on the vibrations of the pendu- 

 lum ; but they are experimentally all but correct when 



