1882.] Stress and Strain on the Action of Physical Forces. 277 



1 inch thick between them, stood upright on a stout table. The ends 

 of the wire to be tested, and of the other wire, which will be called 

 the comparison-wire, were clamped into three brass blocks which 

 rested upon a support of hard wood placed on the top of the chamber. 

 One of the blocks was twice the length of each of the other two, and 

 into this was clamped one end of each of the wires ; the other ends 

 were clamped into the other two blocks. The blocks were provided 

 with terminal screws, and a " Wheatstone's bridge " was formed, 

 having four branches consisting of the wire under examination, the 

 comparison-wire, and two sets of resistance-coils each of about 

 100 ohms, but capable of variation by such small amounts as '1 ohm 

 at a time. These resistance-coils were connected by caoutchouc- 

 covered copper wire, several feet in length but of small resistance, to 

 two of the brass blocks, and were also united to each other by a 

 platino-iridium wire having a resistance of *1 ohm, which was stretched 

 along a scale divided into millimetres, and was traversed by a sliding- 

 piece, which, by means of a suitable spring and catch, could be readily 

 clamped to any part of the wire. By means of the resistance- coils 

 and the platino- iridium wire an alteration of less than one in a million 

 could be measured. As the change of resistance was in general very 

 small, it was necessary to take every precaution to avoid sudden 

 changes of temperature. It was necessary also to keep the galvano- 

 meter-circuit always closed in order to avoid errors which would 

 otherwise have arisen from thermo-electric currents. A single Le- 

 clanche cell was employed for the current-motor, and with this it was 

 possible in a large majority of cases to measure with the aid of a 

 delicate reflecting galvanometer an alteration of resistance not exceed- 

 ing one in a million. The wire to be strained was provided with a 

 moveable pulley 2 inches in diameter, to which was attached by means 

 of a stout wire passing through a small aperture in the table a scale- 

 pan, and both wires were, before suspension in the air-cbamber, 

 surrounded with caoutchouc tubing, silk, or other insulating material. 



The electrical resistances of all the substances which were examined, 

 were, with the exception of nickel, increased by temporary longitudinal 

 stress. With nickel, however, of which metal a wire nearly chemically 

 pure was at length with difficulty procured,* the resistance was found 

 to diminish under longitudinal stress not carried beyond a certain 

 point ; bat after this point had been attained further stress began to 

 increase the resistance. The effect on nickel appears still more 

 remarkable when we reflect that the change of dimensions produced 

 by the stress, namely, increase of length and diminution of section, 

 would increase the resistance. 



The specific resistances of all the substances, except nickel and 



* Through the kindness of Messrs. Johnson, Matthey, and Co. 



