the Mean Density of the Earth, 157 



than before ; the single observed times of vibration varied from 

 770-3 to 813*2 seconds, and hence the force of torsion in the 

 ratio of 1:1-114. 



This variability in the force of torsion of a bifilar or unifilar 

 wire, which exhibits itself when the weight which imparts tension 

 to the wire remains the same, diminishes in time, is indepen- 

 dent of temperature, and which, it may be remarked, exhibited 

 itself in Baily's experiments in a still greater degree, is well 

 deserving of attention; for future observations it would be of 

 interest to ascertain its cause, and whether it is not possible 

 completely to annul it. 



The bifilar wire gave for the mean density of the earth, — 



Mean . . . 5-5910 

 Probable error 0-0169 ^ 



With the bifilar wire, therefore, we have not succeeded in 

 obtaining more coincident results. 



From the three series we obtain as principal mean — 

 5-5832, 

 with a probable error of 0*0149. 



A doubt very naturally suggests itself here, whether feeble mag- 

 netic or diamagnetic actions between the mass and the sphere 

 might not interfere with the correctness of the results, without, 

 however, subscribing to the view developed by Mr. G. Whitehead 

 Hearn (Philosophical Transactions, 1847, part 2, p. 122), 



To obtain certainty upon this point, I made a number of 

 experiments with a diamagnetic sphere of bismuth and a mag- 

 netic sphere of iron. 



The bismuth sphere gave — 



5-6728 5-3939 5-7153 



5-6853 5-7876 5-5063 



5-1436 5-5032 5-3551 



5-5033 



Mean 5-5266 

 with a probable error of 0'0402. ' 



