1893.] Effects of Mechanical Stress on Electrical Resistance. 17 



crease of 2 per cent. ; and, when the drawing was continued, the 

 density began to diminish till, after drawing from diameter 2 mm. 

 to 1*3 mm., it showed an increase on its original value of ^ per 

 cent. Several other interesting results on alteration of density were 

 obtained. 



The most important part of the investigation, however, relates to 

 the alteration of specific resistance of copper, iron, and steel wire 

 due to stretching ; and, in connexion with this, the authors wish par- 

 ticularly to emphasise the advantages to be gained from using the 

 unit of specific resistance introduced by Weber, who always defined 

 it in weight measure, that is, as the resistance of a length of the 

 metal numerically equal to its density and section unity. 



Taking the expression for the resistance R of a wire, of section iv 7 

 length Z, and volume specific resistance a Vi R == a v (ljw), we have R = 

 a v p(ljwp) (where p is the density of the wire) = <r v p(l 2 jlwp) = o v p(l\w)l 

 = <r w (ljiv)l, w being the weight of the length Z, and a w the weight 

 specific resistance of the wire. We have thus obviated the necessity 

 of making the troublesome and uncertain measurement of the section 

 of the wire, and only require to measure the length and the length 

 per unit of weight, both of which can be done with great accuracv. 

 Moreover, the weight specific resistance is found to be more nearly 

 constant than the volume unit. Also, when only a comparison of 

 specific resistance is required, it is not even necessary to weigh the 

 wires, but only to measure lengths, thus making it possible to detect 

 very small changes with great ease and accuracy. 



The method used was a modification of what is known as Thomson's 

 (Lord Kelvin's) Double Bridge Method ("New Electrodynamic 

 Balance for Resistance of Short Bars or Wires," 1 Phil. Mag.,' 4th 

 series, vol. 24, 1862). The method being a zero one, the galvano- 

 meter could be made so sensitive as to be almost unstable, so that a 

 change of 1 in 10,000 could be easily detected. 



The tests showed that the maximum permanent stretching pro- 

 duced a permanent alteration in the weight specific resistance of 

 copper of 1 per cent. After the maximum stretching had been pro- 

 duced, it was found that there was no permanent alteration of 

 weight specific resistance due to renewed application of stress. 



In the tests of steel wire no permanent stretching was obtained, 

 but the effect of applying weight was to cause a very small per- 

 manent decrease of specific resistance, 0*06 per cent, at first, and 

 when additional weight was applied there was found to be an in- 

 crease of 0*06 per cent. These values are very small however, com- 

 pared with the temporary alteration of 1*6 per cent. 



In the tests on soft iron wire the permanent alteration due to per- 

 manent stretching was found to be -§- per cent. After the maximum 

 stretching had been obtained stress was again applied, and it was 



