59 



ducted entirely by Messrs. Dunkin and Ellis, Assistants of the Royal 

 Observatory. 



On discussing the results of the observations, there appears to be 

 reason for supposing that a change has taken place in one of the 

 pendulums after the Seventh Series. This appears from the circum- 

 stance that, though the Fifth and Seventh Series agree well, the Sixth 

 and Eighth are discordant ; and also from this circumstance, that the 

 abstract relation between the two pendulums given by the Fifth, 

 Sixth, and Seventh Series, agrees closely with that found at Harton ; 

 but if the Eighth Series is included, there is a considerable dis- 

 cordance. 



If the Eighth Series is rejected, it appears that Colonel Sabine's 

 coefficient ought to be increased by about -^th part ; and on intro- 

 ducing this correction into the computations of the Harton Experi- 

 ment, the result for the earth's mean density is 6-809. If the 

 Eighth Series is retained, the correction is reduced to less than one- 

 fourth of that just mentioned, and the earth's mean density is 

 6-623. 



The author subjoins an investigation with which he has been 

 favoured by Professor Stokes on the effect of the rotation and ellip- 

 ticity of the earth in modifying the numerical results of the Harton 

 Experiment. It appears that the numbers found in the paper ought 

 to be multiplied by 



, _ equatoreal centrifugal force 



gravity 

 e=ellipticity 

 1= latitude of place. 



On converting this formula into numbers, for Harton, the factor is 

 found to be 1*00012, which produces no sensible change in the 

 result. 



At the equator the factor would have been 1-00679. 



