Manchester Memoirs, Vol. xlviii. (1904), No. 7. 13 



cases be unidirectional and generally not sufficiently often 

 repeated to have brought the specimen very near the 

 breaking point. Hence cracks would probably be non- 

 existent or small in which case annealing would weld the 

 grains together at the slip surfaces. If however serious 

 cracks have been formed it seems unlikely that mere 

 annealing without forging or some means of compressing 

 the hot metal — a sudden cooling of the surface might do 

 a little good — would do much in the way of restoring 

 the metal, hence the discordant results of Mr. Smith's 

 experiments. 



Another very important point is that of the restoring 

 effect, if any, of a period of rest. If no permanent 

 extension has been reached then the lagging extension 

 being elastic should disappear with a rest and the speci- 

 men would be as good as ever. In confirmation of this it 

 is found that metal railway bridges subject to heavy 

 traffic during the week-days sag a little but recover 

 during Sunday's quiet. On the other hand if many slip 

 bands have appeared, these being permanent not much 

 can be expected from a period of rest except it be very 

 prolonged or accompanied by gentle heating, say at 

 100° C, which has been shown to have a kind of slow 

 annealing effect on metals. Even this would probably be 

 absent in the case of a specimen which had bad cracks in it 



The raising of the elastic limit in tension by repe- 

 titions of tensile stress is evidently due to the permanent 

 extension produced, and the consequent lowering of the 

 elastic limit in compression is probably due to the loose- 

 ness of the gearing between adjacent slip surfaces in the 

 grains allowing a considerable inelastic motion to take 

 place at reduced stresses (the forces of cohesion are 

 supposed to vary inversely as the fifth or sixth power of 

 the distance). 



