APPARATUS FOR MEASURING STRAIN AND APPLYING STRESS. 291 



A case which bears considerable resemblance to the case of permanent set last 

 quoted is one by M'Farlane,"* who has shown that if a wire is twisted to nearly its 

 limit of torsional elasticity, an increase in pull will cause the torque to give the wire a 

 permanent set. This latter case can be easily explained in the same manner as the 

 one described above. 



XII. — Effect of Annealing. 



It has long been known that iron and mild steel stressed beyond the limit of elasti- 

 city regain their elastic properties when heated to a red heat and allowed to cool slowly. 

 The process may be repeated many times without apparently changing the elastic pro- 

 perties of the material. The yield-point, however, is found to alter in position as the 

 annealings proceed. In a particular case f a mild steel bar, which in an ordinary test 

 would give an extension of 25 per cent, upon a ten-inch length, and a yield-point of 

 about 18 tons per square inch, was stretched approximately \ inch, and annealed in 

 the ordinary manner after each operation. Throughout the experiment the bar appeared 

 to recover its elastic properties after each annealing, and finally broke with a total 

 extension of approximately 100 per cent. The yield-point remained fairly constant, 

 except at the end, when it experienced a rise. 



Copper treated in the same manner has been drawn out by the author to consider- 

 ably more than double its length in this way without causing fracture. Remarkable 

 advances in our knowledge of annealing have been recently obtained by Mum,| acting 

 upon a suggestion of Professor Ewing. 



Muir has shown that comparatively low temperatures, such as boiling water, will 

 restore a strained bar to its elastic condition. The yield-point, however, alters during 

 the process, and is always higher than in the original condition. 



After a few applications of stress, followed by heating in boiling water, or even water 

 at 50° C, the bar fractures, with a total extension not very different from a bar stressed 

 to breaking without special treatment. The annealing at low temperatures, therefore, 

 appears to be less complete than that at a high temperature. 



In order to discover what effect a temperature of 100° C. would exert on a bar over- 

 strained l>y a torque, the steel bar which had been used for experiments on the recovery 

 of elasticity with time (see Section VI., Table IV.) was selected, and after being boiled for 

 fifteen minutes was overstrained, giving results (col. X., Table XIV.) practically identical 

 with those of col. I., Table IV., for the first part of the curve. As in practice it is 

 troublesome to get exactly the same calibration value for each setting of the instrument, 

 this latter (stripped of the reading microscope and wire holder) remained on the bar 

 during the heating, and the labour of comparing readings whose unit values differed bv 



* Art. "Elasticity," Enc. Brit., Part 21. 



t (Joker, "Note on the Endurance of Steel Bars subjected to Repetitions of Tensional Stress." Proc. Inst. O.E., 1899. 



X "The Recovery of Iron from Overstrain." By James Muir. Phil. Trans., 1899. 



