for determining the Mean Density of the Earth. 121 



merits, prosecuted in such a variety of ways and with such 

 different materials, that the several mean results, obtained 

 from the individual classifications, can be of equal weight. 

 In fact, the author himself has, in his investigations of the 

 subject, clearly shown that some of them are entitled to more 

 confidence than others ; and moreover that, in a few instances, 

 there may be a fair cause for rejection. On these points how- 

 ever there is no room for explanation in this place: and it 

 may be sufficient here to state, that, assuming every experi- 

 ment to be of equal weight, the mean result of the whole 

 2004- experiments is 5'67. Nor is there much probability that 

 the result of this immense number of experiments will be ma- 

 terially altered, even if those few experiments, which may 

 appear to be affected with some source of error or discord- 

 ance, should be wholly omitted. 



The author remarks that it cannot escape observation that 

 the general mean result, obtained from these experiments, is 

 much greater (equal to ^jth part) than that deduced either by 

 Cavendish or Reich, who both agreed in the very same quan- 

 tity, namely, 5*44 : but he does not assign any probable cause 

 for this discordance. It is evident, however, from the detail 

 which he has given of his own experiments, that perceptible 

 differences not only arose according to the mode in which the 

 torsion-rod was suspended, but also depended on the materials 

 of which the suspension-lines were formed : but it is somewhat 

 singular that none of the mean results, in any of these classi- 

 fications, are so low as that obtained by the two experiment- 

 alists above mentioned. 



In these remarks, no notice has yet been taken of the re- 

 maining 149 experiments that have been made with the brass 

 torsion-rod; a class of experiments that were undertaken for 

 the express purpose of ascertaining the effect of such a mea- 

 sure on the general result. This torsion-rod was nearly of the 

 same weight as the two 2-inch lead balls, and about half the 

 weight of the two 2|-inch lead balls. The experiments were 

 made not only with each of these balls successively attached to 

 the rod,but also with the rod alone, without anything attached 

 thereto. The results show that the attraction of the masses 

 on the rod should be diminished about ^th part, in order to 

 render these three several results consistent with each other, 

 and also accordant with the same balls and the same mode of 

 suspension, attached to the lighter wooden torsion-rods. 



