Points of Recalescence in Steel and Iron. 433 



But the two expressions <$>(p)—bt and f(v)—bt can only be 



identical at all temperatures iff(v) and cj> (p) are both equal 



to a constant c. 



Therefore m 7 . 



l = c— at, 



where c and b are constant. 



It appears then that the surface-tension of a liquid is 

 independent of the pressure and depends only on the 

 temperature, unless the film is very thin. 



The critical temperature is c/b, and can be found by deter- 

 mining the surface-tension at two very different temperatures. 



LIII. On some New Methods of Investigating the Points of 

 Recalescence in Steel and Iron. By Frederick J. Smith, 

 Trinity College, Oxford*. 



IT was discovered by Mr. Gore, in 1869, that when an iron 

 wire was heated to redness, and then allowed to cool, 

 a momentary elongation of the wire took place imme- 

 diately after it began to cool. The phenomenon was made 

 evident by means of a suitably arranged pointer (Proc. R. S. 

 1869). The phenomenon was next investigated by Prof. 

 Barrett, and it was shown by him to be due to a reheating of 

 the wire : the phenomenon has been called by him Recalescence 

 (Phil. Mag. 1873). Recently M. Osmond has observed that 

 there appear to be two points at which on cooling the pheno- 

 mena occur (Compt. Rendus, 1890). 



I have not mentioned other work which has been done on 

 recalescence, as it would have no immediate bearing on the 

 subject of this communication. I have been led to make some 

 experiments on recalescence, with a view to discover what 

 part it may play in the destruction of certain engineering 

 structures, such as heavy coiled guns, and those parts of 

 boilers which are exposed to the effects of fire. 



Up to the present time, observations have been made on 

 the changes of temperature apart from the lengthening and 

 shortening of the steel and iron under examination. I wished, 

 if possible, to discover the time-connexion which exists be- 

 tween the change of form and the change of temperature. 

 Several methods of experimentation were tried, and finally 

 the following ones were found to give satisfactory results. 



The upper end of the steel wire to be tested was fixed 

 vertically ; the lower end was attached to a long light lever of 

 aluminium, so arranged that a small change of length of 



* Communicated by the Author. 

 Phil. Mag. S. 5. Vol. 31. No. 192. May 1891. 2 I 



