1921] 



on Elasticity 



391 



at the temperature indicated by the abscissa, the time at 100° C. 

 being taken as unity. 



Although on the whole the metals become less viscous at low 

 temperatures, there seems to be an actual increase, or slower rate of 

 decrease, of viscosity somewhere between 0° and - 100' C, which is 

 particularly noticeable in the case of copper, gold and silver. 



This is a matter which requires further investigation. The same 

 metal often gave different results in the annealed and hard conditions. 



There is no very apparent connection between the variations of 

 viscosity and the melting points, although the easily fusible metals 

 show the larger changes in this respect. 



A general conclusion which I think mav be drawn from these 



Y 



x - MeLUfi^ point ctjiei^roA 



„ _ RiyidUij <xt ALsoLuCi zvro 

 <r — -Rigidity at zzro- C€i*(7<jrae(< 

 on tl* assumption Che-f t/tr Vajiafmu of 



TJk* S/sety shtM* W» lalTot 



100O i5oo iOOo 



A \r solute. Te-m jxrcvturc 



ifoo 



~JSoo 



Fig. 7. 



and the former experiments is that the rigidity of any metal and its 

 variation with temperature is not much affected by the present state 

 of the test-piece or its previous history, being nearly the same for 

 hard, soft, cast, forged or annealed specimens ; but that the viscosity. 

 both as regards its absolute value as well as its temperature varia- 

 tions, is influenced by all these conditions, though not so much as 

 are the elastic limits. 



I have here an experiment which will show how small is the 

 difference between the elasticities of hard and soft steel, and between 

 these and wrought iron. 



A bar of hard steel is laid horizontally on supports at either end. 

 A weight can be hung from the centre and will cause a certain 

 deflection, indicated on an enlarged scale by the pointer. On sub- 



