1894.] and Steel and the Effect of Straining and Annealing. 187 



Experiment IY. 



Mild Steel Bar, No. 1436. 



Original diameter = 0*750 in. Area = 0'4418 in. Clips 8 in. apart. 



Maximum load 11'56 tons = 26'17 tons per sq. in. 



Breaking load 9'84 =22-27 



Load at yield point 773 = 17'49 



Total elongation on 8 in 2 '28 28*5 per cent. 



Time of test np to yield point . . 3 minutes. 

 Time of whole test 8 



In this experiment a bar of the same steel as in the previous test 

 was broken in the first loading without any annealing, in order that 

 its autograph diagram might be compared with that of a bar broken 

 after eight annealings. The diagram is drawn dotted in fig. 8. 

 There is no marked difference in the two diagrams, if the decrease of 

 area due to successive tests and annealings is allowed for. 



It is clear from these tests that the change produced in the iron or 

 steel bar by straining beyond the yield point is completely reversed 

 by simple annealing, and this apparently after any number of repeti- 

 tions of the process. It appears hardly possible to understand either 

 the singular abruptness of the load strain curve at the yield point, or 

 the difference of the load strain curves for an unstrained and a 

 strained bar, or the reversal of the change of properties produced 

 during straining by simple annealing, except by supposing that at 

 the yield point there occurs a chemical or molecular or allotropic 

 change of the same kind as that which occurs in ordinary tempering 

 by sudden cooling. 



If mere stress can produce a change of this kind, it may be that 

 the stresses produced in sudden cooling also are factors in the 

 changes which occur in that case. If a red-hot bar is plunged into 

 water the surface must cool first. A condition of stress must arise 

 in which there is a ring tension and radial thrust in the section of 

 the bar. It is well known that the stresses so set up are often of 

 great intensity. Bars tear or crack in hardening. The material 

 may be strained beyond its yield point by the stresses due to cooling. 

 When bars are subjected to a Wohler test, that is, to a large number 

 of repetitions of straining action, they break ultimately with a stress 

 much less than that which would break them in a simple loading and 

 with a very characteristic fracture. The outside of the fracture is 

 fine grained, and resembles a crack extending gradually across the 

 section of fracture. However ductile the material, initially, there is 

 no sign of contraction at the fracture. The bar breaks like a brittle 

 material. It seems possible that here also the repetition of the 

 straining action may produce a molecular change in the bar in the 

 neighbourhood of the fracture. 



