EFFECT OF TENSION ON CERTAIN ABNORMAL METALS. 55 



tions and removals of 250 gm. The effect was found to be normal in 

 sign, the resistance increasing under tension. The effect is smaller 

 than for some of the pure metals, and accordingly was not so regular. 

 The maximum departure of any of the observed points from a smooth 

 curve was 4.5% of the maximum change. The relation between 

 tension and change of resistance is very nearly linear, but there are 

 nevertheless consistent small departures from linearity greater than 

 the errors of measurement; the change becomes proportionally 

 greater at the greater tensions. The average tension coefficient over 

 the entire range of tension is not more than one or two per cent greater 

 than the initial coefficient. 



The three runs on the hard drawn sample gave the following values 

 for the initial coefficient respectively: +5.78, 5.72, and 5.63 X 10"^ 

 for a tension of 1 kg/cm^. The average, +5.71 X 10-^, is taken as 

 the most probably correct coefficient. 



A length of wire contiguous to the hard drawn specimen was 

 annealed by heating to redness. Measurements were made on the 

 first application of tension. 250 gm. was distinctly beyond the elastic 

 limit. The specimen was seasoned by several applications of 250 gm., 

 and then by a number of applications of 200 gm., until finally the 

 changes of resistance had become steady. The initial coefficient of 

 the annealed wire was +5.88 X 10-^. After accommodation to the 

 range of 200 gm. the coefficient had increased to 6.75 X lO"^. This 

 latter is the average coefficient over the range of tension. After 

 accommodation the relation between resistance and tension is not 

 linear or single valued, but the relation involves hysteresis of the 

 usual character, the maximum width of the hysteresis loop being about 

 10% of the maximum effect. 



Young's modulus of this wire was determined directly from the 

 increase of length under a given load. A piece about 3.5 m. in length 

 was hung in an elevator shaft, and observations made on fiducial 

 marks at the top and bottom with microscopes. The supports of the 

 microscope were entirely independent of the supports of the wire, so 

 they could not be affected by the load applied to the wire. The hard 

 drawn wire was used for this determination.. It was seasoned by a 

 number of applications of 500 gm. ; the maximum load for the determi- 

 nation of the modulus was 400 gm. There was a slight amount of 

 hysteresis, but there were no perceptible time effects. The maximum 

 width of the hysteresis loop was 5% of the maximum extension. The 

 mean of points with increasing and decreasing tension lie on a straight 

 line within the limits of error, which were not more than 0.3%. 



