Torsion of Lead Wire beyond the Elastic’ Limit. 539 
is produced and the rate of change in the stress required for 
preserving it observed. We have not so tar made experi- 
ments on the rate of disappearance of the stress after the 
obliteration or forcible removal of the strain, the case 
corresponding to the recovery from strain on the removal of 
the stress referred to above. 
The investigation of the changes in the stress of an over- 
stretched or over-twisted wire when the strain is kept con- 
stant would appear to be fundamental. Here the movements 
which have to be considered can only be of the nature of 
internal rearrangements of the particles, the material on the 
whole never changing its position. 
The experiments which we have made may be divided into 
three parts. First,a determination of the rate of decrease of 
the restoring couple on a wire kept twisted through a con- 
stant angle. Second, a determination of the decrease of 
iongitudinal stress in a wire kept at constant length. Third, 
a determination of the rate of stretching under a constant 
force and the rate of recovery on removing the force. 
1. Rate of Decrease of Couple in a Wire kept twisted 
through a Constant Angle. 
A copper rod was soldered to each of the ends of a lead 
wire about 1 m. long and 1 mm. in diameter. It was then 
suspended, the top copper rod being fixed. The lower one 
was made T-shaped for the application of a couple. It was 
necessary to have some means of keeping the wire accurately 
twisted through a fixed angle. The angle of twist was for this 
purpose observed by aid of a lamp and’ scale and a mirror 
attached to the lower rod. The couple was applied by méans 
of the apparatus A A (fig. 1, p. 540), which simply consisted 
of weights hung from the points CC, which could be pulled 
aside by horizontal fibres fastened to the arms DD of the 
rod at the end of the wire under test. The couple is very 
approximately proportional to the deflexion of the points BB 
on the scale. As the restoring couple in the wire diminished 
further twisting began to take place, and the angle of twist 
was kept constant by moving the stands AA inwards, thus 
reducing the applied couple. The rod E was kept in a fixed 
position by means of the wire F passing through a small. 
circular hole (but not touching the sides), and by always 
moving the stands AA in such a manner as to keep the wire 
F always central in this hole. 
Owing to the necessity of fastening the fibres on toDD 
after twisting, it was impossible to read the value of the 
~ couple immediately after twisting, and the first reading was 
