UNIFICATION OF METHODS IN TESTING MATERIALS. XLV. 
which the application of the stress was continued. But experi- 
ments shewed that in addition to this, there was a secondary 
deformation depending upon the time during which the stress was 
maintained, but which appeared to approach a limit. Consequently 
the trace in the stress-strain diagram was not exactly straight 
anywhere ; and its curvature would slightly vary with the time 
of stress-application. A proposal therefore to produce the straight 
part of the curve, and adopting an arbitrary increase of the tangent 
or arc, to draw a second line, the intersection of which, with the 
trace, was to be taken as the arbitrary elastic limit, could not be 
regarded as having any distinct advantage over Professor Johnson’s 
proposal, viz., to draw a tangent to the curve with an arbitrary 
increase on the ratio of the deformation to stress. And his method 
had very little advantage over the older practice of marking 
the point where the trace commences to sensibly curve. The 
difficulty of ascertaining the point where the molecular constraints 
resisting deformation were so relaxed that the rate of plastic flow 
becomes very large, relatively to the elastic extension, was a 
difficulty inherent in the behaviour of the material and one which 
could not be avoided. Probably a higher order of accuracy of 
measurement now possible by such instruments as Prof. Martens’ 
extensometer, would, when thermal changes due to the stress were 
measured and discussed, indicate the solution of this question. It 
might be of interest to mention that in a test made to-day by Mr. 
Barraclough and himself with the instrument referred to, disposed 
as recommended in the paper recently read by the latter to the 
Royal Society, the evidence of plastic flow was noticed even in 20 
seconds and in 30 seconds. The loads in thousands of pounds, and 
the differences of the deflections in ten-thousandths of millimetres, 
were :— 
Load Sy 8. 18 16. 2S 
Diff. Extension 116 126 123 123 124 123* 138 1817 145 310 
* Flow in 20 seconds 4. + Flow in 30 seconds 4. 
It would be seen that while the rate extension was apparently 
uniform, plastic flow appeared to have occurred. In conclusion, 
