for determining the Mean Density of the Earth. 119 



a sensible and material error is likely to occur : but this sel- 

 dom happens if due precaution has been taken to screen the 

 torsion-box effectually. Yet the author is still of opinion that 

 discordances sometimes arise which cannot wholly be attri- 

 buted to change of temperature, but to some other occult in- 

 fluence with which we are at present unacquainted. The re- 

 gular march of the resting-point of the torsion-rod is one of 

 the most important objects of attention ; since any considerable 

 deviation therefrom is the source of great discordance, and 

 therefore requires to be watched with care. 



The torsion force comes next under consideration. Mr. 

 Baily justly remarks that the torsion force of a wire is that 

 elastic power in the body, by means of which it is enabled to 

 return to its original position, after being drawn aside by any 

 external impulse. It varies with the substance, magnitude, 

 and length of the wire ; but it is generally considered to be 

 constant for the same wire, whatever be the weight suspended 

 thereto. This, however, must be taken within certain limits, 

 since the time of vibration (which is one of the elements for 

 determining the force of torsion) will frequently differ very 

 considerably without any apparent or sensible alteration in 

 the component parts of the apparatus. For the author states 

 that we frequently have in the same hour very considerable 

 variations in the time of vibration, which evidently show that 

 the force of torsion has undergone some sensible change. But 

 this alteration in the torsion force does not appear to affect 

 the results of the experiments, since we find that, when the 

 time increases, the deviation is also increased in due propor- 

 tion. The magnitude, therefore, of the force of torsion is not 

 a necessary object of inquiry in these investigations. 



The only two objects requiring close attention, for the pur- 

 pose of obtaining results from any of the experiments, are the 

 determination of the mean resting-point of the torsion-rod, and 

 the time of its vibration. Now, it fortunately happens that 

 these two objects can, in all cases, be observed with the greatest 

 ease and accuracy, however anomalous they may be; and 

 they are never accompanied with any doubt or difficulty. 

 There is however another subject that is required also to be . 

 accurately ascertained in every experiment ; namely, the ex- 

 act distance of the centre of the masses from the centre of the 

 balls. This has been effected by means of plumb-lines, which 

 abut against the masses, and the distances between which 

 are measured, at every experiment, by means of a micro- 

 scopical apparatus, carefully adjusted. 



From the results of the several experiments that the author 

 has made, it would appear that single wires, of different dia- 



