466 SCIENCE PROGRESS 



It will be well, at the beginning, to describe in detail the 

 behaviour of solid bodies when subjected to stress — that is to 

 say, applied force. Let us suppose, for the sake of exactness, 

 a bod}' - in the form of a wire held fixed at one end, and to which 

 a longitudinal stress may be applied by loading the other end 

 with weights. When the load is attached, there results an 

 immediate deformation in the form of an extension of the wire, 

 which, if the load is sufficiently small, does not increase with 

 time, even though the weight is allowed to remain on the wire. 

 The removal of the stress effects a contraction of the wire, which 

 at once regains its original length. A repetition of this procedure 

 produces identical results. Now let the wire be loaded with 

 a greater weight. The increase in length produced will again 

 be immediate and afterwards constant, provided that this weight 

 also does not exceed a certain limit. The extension will, how- 

 ever, be greater than in the first case in the direct proportion of 

 the weights, but the contraction subsequent to the release from 

 stress will again be instantaneous and equal to the extension. 

 Results similar in all respects to these are obtained as greater 

 and greater loads are experimented with, until, when the stress 

 has reached and exceeds a certain limiting value depending 

 on the material and dimensions of the wire, other effects appear, 

 the investigation of which forms the subject of this article. The 

 stress on the wire is now said to have reached the " elastic 

 limit." The wire no longer behaves as a perfectly elastic body, 

 i.e. it does not now immediately regain its original length upon 

 the removal of the load. The deformation or " strain," as it is 

 called, resulting from the applied stress, does not remain equal 

 to its initial value, but increases with the progress of time. It 

 is no longer a function of the weight only, but also depends 

 on the time during which the latter has been applied. If the 

 load is allowed to remain on the wire, the gradual increase 

 of length continues for many hours, and, when the load is 

 at length removed, the immediate contraction is followed by 

 a slow but long-continued recovery towards the original length. 



An exactly similar series of observations may be made on 

 a wire clamped at one end and twisted at the other. Provided 

 the torsion-couple does not exceed the elastic limit for the 

 particular case, it produces torsion of constant value, and the 

 recovery is complete on release. But so soon as the couple 

 exceeds that limit, the initial twist is followed by an increase 



