Stress or Permanent Strain to a Bright-red Heat. 263 



weight. The wire was heated by a burner about three inches 

 from the clamp. When the temperature reached a bright red, 

 the wire began to bend permanently very rapidly by its own 

 weight ; and when the burner was removed and the wire 

 cooled, a sudden further bending took place at a temperature a 

 little below bright red ; and when cooled to the temperature 

 of the room, the wire remained permanently bent. Here the 

 attempt to unbend, which the wire no doubt made at the cri- 

 tical temperature when it was heated, was masked by the 

 permanent bending. Another piece was therefore taken and 

 heated at a point further away from the clamp, so that the 

 bending-stress should not be so great. With care it was 

 found possible to detect that the wire suddenly straightened 

 itself when the critical temperature was reached on heating. 



Similar trials were made with other specimens of iron and 

 steel not so well annealed ; and w T ith these, when cooling, the 

 sudden bending took place at a temperature lower than that 

 at w r hich it occurred with the very soft iron. With one spe- 

 cimen of wire, 1^ millim. thick, it occurred at a temperature 

 apparently below that of visible red* ; but, as Professor Bar- 

 rett justly remarks, the internal temperature of the thicker 

 wires is no doubt masked by the cooling of the surface, 

 whereas in thin wires the cooling throughout is extremely 

 rapid. I found also that a piece of this specimen of iron ap- 

 peared to lose its magnetic properties at a temperature below 

 visible red, whereas with the thinner wires the apparent tempe- 

 rature at which this took place was somewhere about a dull red. 

 The very small amount of heading- stress which is required 

 to bring out the phenomenon makes it a little difficult to de- 

 tect with certainty the opposite effect of hending- strain ; but 

 it may be managed after a few trials in the following manner: — 

 Experiment VII. — A piece of the very soft iron wire was 

 bent, as in fig. 2, w r ith the portions A C, A B 

 in a vertical position ; the end C was secured Fie;. 2, 

 to a clamp, and the bend A heated by a burner. 

 If A B has been so arranged that its centre of 

 mass is very nearly vertically above the part 

 heated, there will be little bending-stress. In 

 this experiment, on heating the bend A to a 

 bright red, the end B jerked suddenly towards 

 C, and when the burner was removed almost 

 immediately jerked back again. If AB is 

 arranged so that there is a little more bending- 

 stress, it is curious to notice the struggle wdnch V J 



sometimes ensues at a bright red heat as to A 



* Prof. Barrett also remarks that the phenomenon occurs on coolino- at 

 an apparently lower temperature with thick wires than thin ones. 



3 



