154 M. F. ReicVs New Experiments on 



the first and third with the mass in one position^ and the second 

 and fourth with the mass in the other position, I compared the 

 mean of the former two with the mean of the latter ; whereas 

 Baily drew his result from the first and third compared with the 

 second, and from the second and fourth compared with the third. 

 If, when the mass is in a certain position, the arm of the torsion- 

 balance always came to rest at the same place, both modes of 

 calculation would lead to the same results ; but the position of 

 equilibrium perpetually changes, and, as a rule, towards the 

 same side during a series of experiments ; hence Baily^s mode of 

 observation is the most correct ; and if the change in the position 

 of equilibrium were proportional to the time, which is not exactly 

 the case, it would furnish results totally free from error. In the 

 memoir, therefore, my former observations have been recalcu- 

 lated, and a final result of 5*49 obtained. 



Another circumstance appears, on the other hand, to render 

 it probable that the final result of Baily is somewhat too 

 high. In Table V. of his work the single series of observations 

 are arranged according to the weight of the spheres made use of. 

 Taking the arithmetic mean for each sphere, a more exact calcu- 

 lation with reference to the values of the single series being un- 

 necessary for our purpose, the mean density of the earth is found 

 to be — 



With the 2J^-inch lead sphere .... 5*595 



With the 2-inch lead sphere .... 5-634 



With the 2^ -inch hollow copper sphere . 5*658 



With the 1^-inch platinum sphere . . 5*627 



With the 2-inch zinc sphere .... 5*666 



With the 2-inch glass sphere . . , . 5*768 



With the 2-inch ivory sphere .... 5*775 



With the copper arm alone .... 6*024 



We observe here that the result increases as the weight of the 



sphere diminishes, the sphere of platinum forming only a trifling 



exception to the general rule. It is therefore probable that 



some correction has been inaccurately determined by Baily, which 



operates most sensibly in the case of the lighter spheres, leaving, 



however, a small error, on the same side, attached to the heaviest 



also. .^.^ 



The apparatus made use of by me was placed in a closed 

 cellar; to preserve the wood- work from decay, it was removed to 

 a room on the second floor, originally with the intention of pre- 

 serving, but not of making use of it in experiments, as I found 

 myself unable to take such anxious precautions against changes 

 of temperature and other disturbances as Baily has taken. But 

 by availing myself of an artifice described by the latter, and first 



