4 68 SCIENCE PROGRESS 



of the stretching weight. This will cause a small shortening 

 of the wire, which, even though in a state of over-strain, still 

 possesses certain elastic properties. It will be seen that, pro- 

 vided this diminution in stress does not reduce the total stress 

 to a value below the elastic limit, this decrease in length will 

 only be temporary, and that soon the wire will again be 

 stretching. As soon as it reaches the same length as before, 

 remove another small amount of the weight, and wait until the 

 wire is again of the length decided upon above. Similar re- 

 movals of weight will produce similar results, until at last the 

 remaining weight is equal to the elastic limit, when no further 

 tendency to increase in length will be observable in the wire. 

 By this procedure the wire has been kept approximately at 

 constant length. It is obvious that by removing the stress 

 continuously at a suitable rate, instead of in finite portions after 

 certain intervals, an absolute constancy of length could be main- 

 tained, and the necessary rate of removal of stress recorded. 



The two chief methods of investigation are therefore : 



(i) To observe the changes in length and torsion in wires 

 subjected to longitudinal and torsional stresses respectively, 

 and the recovery following the removal of the stress. 



(2) To observe the diminution of stress in wires kept at 

 constant length or constant torsion. 



The aim of these experiments has been, in the first case, to 

 discover the law connecting strain with time, both while the 

 stress is applied and after it has been removed. In the second 

 case it has been to find what function of the time the stress 

 is when the strain is unaltered. But before recording the 

 results of these attempts, attention may be called to certain 

 general features of the elastic after-effects. Referring again to 

 fig. 1, it will be noticed that the immediate extension repre- 

 sented by oa is shown as being equal to the immediate 

 recovery indicated by c d. This is actually the case even when 

 the stress is applied for a very long time. That is to say, 

 wherever the point b is chosen, cd remains equal to oa. 

 The same statement is, however, not true of the after-effect 

 represented by de. The shape of the curve de is not inde- 

 pendent of the time of duration of the stress. N. A. Hesehus 1 

 has shown that if the stress is applied only momentarily, the 

 after-effect is not perceptible, and that it increases for stresses 

 1 N. A. Hesehus, Journal de la Soc. Phys. Chem. russe (1882). 



