196 Mr. A. E. Dunstan on the Coefficient oj 



Discussion of Results. 



Generally speaking very regular straight-line curves obtain 

 for each wire. The extension increases quite steadily with 

 time over a long range, so long as the load is not excessive. 

 It is very clearly shown that increase in concentration of tin 

 lowers the viscosity, increasing the stretch per minute. Pure 

 tin is markedly mobile under a load of 500 gnus., pure lead 

 is proportionally scarcely affected by such a load. 



For a constant load the curves, after a long period of stress, 

 tend to fall towards the horizontal, a state of fatigue sets in; 

 but after an extra load is suddenly put on the mobility 

 markedly increases, a sudden loosening of molecular attrac- 

 tion occurs, which soon passes away however, when the 

 necessary rearrangement of groupings comes into play. From 

 this point a normal state of affairs obtains, but with a noticeable 

 increase on the previous rate of extension. 



Whether similar curves would be obtained with virgin 

 wires remains to be settled. In all probability the gradual 

 change of curvature would disappear and a discontinuous 

 series of straight lines would be obtained. Undoubtedly, 

 much depends on the previous history of the wires. Too 

 much stress should not be laid, however, on the curves after 

 a certain limiting line, because the progressive thinning out 

 of the wires causes an apparent diminution of viscosity, and 

 ultimately the rupture of the material. In fig. 1 values 

 for the coefficient of viscous traction are plotted against per- 

 centage composition. The inner curve is for an approximately 

 uniform load of 500 grams, and the nature of the curve is 

 undoubtedly hyperbolic, asymptotic to the axes. The effect 

 of adding tin to lead is very great, 10 per cent, tin lowers the 

 arbitrary viscosity coefficient from 160 to 37; beyond this 

 further addition of tin is almost without effect ; similarly, 

 addition of lead to pure tin produces no noticeable change, 

 till 80 per cent, has been alloyed with it. 



The greatest divergence of this curve from the straight 

 line joining the extreme points of the curve is at 70-75 per 

 cent. lead. 



Now it has been found in previous work that such a point 

 of maximum divergence always lies at some simple molecular 

 concentration. The nearest molecular concentration in this 

 case is 2 Pb 1 Sn. 



The addition of lead to tin does not markedly affect the 

 viscosity of the lead ; it does, however, throw the curve well 

 out of the normal. One must assume, therefore, that there 



