M. G. Wiedemann on Torsion. 3 



sions in the same direction was first more fully investigated. 

 The apparatus employed was the same that had served for the 

 previous experiments. PL I. fig. 1 represents it with some 

 more recent modifications. The point which, in the former 

 experiments, held fast in the axis during the torsion the lower 

 extremity of the wire and the circle, bow, and metal rod sus- 

 pended from it, instead of working as before in an agate 

 bearing, passes now through a small well-polished steel ring 

 fixed on the axially pierced screw A x . Further, to the frame 

 with roller r and weighing-basin s (by the load of which the 

 wire is twisted) first a shorter brass wire wis soldered, on 

 which, by means of two silk threads, a longer, very thin and 

 light, smooth horn rod x x is suspended beneath the basin, 

 and, on the thread twisting, strikes against the vertical wires 

 w w as soon as, the string A A being let run down, the roller 

 is no longer supported by the latter, but by the thread o, 

 wound round the circle on the wire. This arrangement is 

 more suitable than the former one, in which the wire vv 

 struck direct against the wires w iv, since thereby the shakings 

 that might take place during the slow lowering of the roller, in 

 consequence of the elastic afteraction in the temporarily twisted 

 wire, and in the thereby occasioned sliding of the wire vv on 

 w w, are avoided. Finally, the string A A is fastened at t, 

 and carried over a pulley at u to the observer, where its end 

 is put round an axle Z of 5 millims. thickness, which can be 

 turned by means of a winch; Thus we can still better, when 

 the scale-pan is loaded with the weights, let down and raise 

 the roller very slowly, and twist and untwist the wire without 

 a shock. The wheel p turned so readily on its axis that when 

 a string loaded at each end with 200 grams was suspended on 

 it, an excess of less than 1 gram caused it to rotate. 



The mirror-reading was now constantly employed for read- 

 ing off the temporary and the permanent torsion within 

 wider limits. A strip of tin of 1 decim. height, bent to a 

 semicircle of 1 metre radius, and kept in this position by cross 

 bands, was placed concentrically upon the wire and lined 

 inside with a millimetre-scale, which, as usual, was observed 

 in the mirror d attached to the clamp c, by means of a tele- 

 scope, A displacement of the image of the scale 34*9 divi- 

 sions corresponds to 1° of rotation of the mirror. 



To the rod g a small bow with its edge horizontal was fas- 

 tened beneath, into which a broad hook was inserted carrying 

 on an iron rod 1 centim. thick, and 30 centims. long, a hori- 

 zontal circular iron plate of 20 centims. diameter. Upon the 

 latter were laid, very cautiously, in order not to cause any 

 torsion, flat circular iron weights slit radially on one side, 



B2 



