the Mean Density of t/w Earth. 155 



proposed by Professor Forbes,, I have been able in this new 

 locahty to make observations quite as regularly as before. The 

 artifice consists m covering the case surrounding the torsion- 

 balance within and without with a metallic coating, in my case 

 with tin-foil. The advantage of such a metallic covering in 

 observations with very mobile apparatus has been proved in other 

 cases also (by MM. Regnault and Kohlrausch) ; it may perhaps 

 be due to the lessening of the sensitiveness of the apparatus with 

 regard to differences of temperature ; or, which seems to me most 

 probable, to the removal of every trace of electrical development 

 from the apparatus. 



Another modification of the apparatus resulted from the fol- 

 lowing out of my original idea to make all corrections as small 

 as possible, or rather to do away with them altogether ; so that, 

 as far as practicable, we should have to deal with the simple 

 attraction between the mass and the sphere alone. For this 

 reason the thick wire which originally carried the mass, and the 

 attraction between which and the sphere and arm was by no 

 means inconsiderable, was omitted; instead of it the mass was 

 laid upon a kind of rotating disc which surrounded the sphere 

 and its case, so that when the position of the mass was altered, 

 the attraction of no other body for the sphere suffered any change. 

 In this case, however, experiments could no longer be made by 

 observing the position of equilibrium of the arm at various 

 distances between the mass and sphere ; but the same procedure 

 as that followed by Cavendish and Baily had to be adopted, the 

 mass being suffered to attract the sphere, first from one side and 

 then from the other. 



With the same spheres as were formerly made use of, which 

 were composed of tin with 10 per cent, of bismuth and a little 

 lead, and w^eighed 484*19 grammes, and also with the same masses 

 of lead which weighed 45031 grammes, three series of experiments 

 were executed. I will here state the single results, and refer to the 

 memoir itself for the more special indications of the experiments. 



The first series was executed with a copper wire 0*5 millimetre 

 thick and 2270 millimetres long, and gave — 



