1880.] Torsion- Grcwimeters of Broun and Babinet. 145 



tied up should retain its torsion indefinitely. The force thns per- 

 manently stored represents the constant force against which gravity 

 is to be measured. 



To nse it, I imagine the ends of the double or twisted cord clamped 

 in two frames of suitable make. The upper one grasps the junction, 

 the lower one the bight. The wires issue from the jaws of the clamps 

 at 08 inch from each other. The upper clamp is attached to a 

 support capable of sustaining, without shake, the intended weight of 

 200 lbs. The lower is attached to the weight so as to have as little 

 loose motion as may be r but so as also to be easily detached for trans- 

 port. Before detaching it the wires should be allowed to twist upon 

 each other. They would then be placed, with their clamps, in a box 

 specially prepared to guard them from all imaginable injury. 



There would be an arrangement above by which, when attached, 

 the upper clamp would be raised slowly, so as to take up the weight ; 

 which would be turned without rising as the wires untwisted, until 

 finally, it would hang ; the wires then being in a state of torsion, the 

 lower ends in a different vertical plane from the upper, according to 

 the weight. I assume that on thus commencing the weight is in 

 excess, so that the angle between the two planes (which will be the 

 6 of the theory) will be about 90°. 



In this condition we want means of relieving the suspension of 

 so much of the weight as will allow the mass to reach the position of 

 unstable equilibrium, and of determining exactly how much. This 

 I propose to effect as follows : — 



The weight is to consist of a cylindrical drum, capable of holding 

 3 cubic feet of water. Near the bottom it will have a small stop- 

 cock. When the apparatus is in repose, the cock is to be turned. 

 The flow of water will relieve the weight, the drum will slowly turn, 

 under the solicitation of the wires. If allowed to run freely it can be 

 shut off when the critical point has been passed. This is the first 

 approximation. A little water is put back, and now the cock is 

 turned so as to reduce the flow to drops only. The effect can be 

 watched with a microscope. The flow can at any time be stopped or 

 accelerated. Experience will enable this determination to be made 

 with a degree of precision which it is impossible to over-estimate. 



The drum should be as light as possible, and silvered inside or pla- 

 tinised. The object is to secure a minimum of friction. As the critical 

 point is approached the force is so small that it would take long to 

 move the whole mass if solid ; and if the whole mass were in motion 

 its momentum would be troublesome. For the same reason the obser- 

 vation cannot be made until the body of water is at rest. 



The same drum will serve for all weights, and makes it possible to 

 obtain a series of results with different pairs of wires — all prepared in 

 the same way, but of different torsions and widths apart. 



