THE STRUCTURE OF METALS 203 



metal undergoes any appreciable spontaneous change of the 

 kind at atmospheric temperatures, whether assisted by vibration 

 or not. There is some little evidence that vibration favours the 

 return of an unstable alloy to the stable state at the ordinary 

 temperatures but so far this case has not received much attention 

 from the practical point of view. 



The evidence for the popular opinion as to the influence of 

 fatigue on metals rests entirely on the appearance of the frac- 

 tured surface. The appearance of fractures is constantly used 

 in practice as a means of judging of the coarseness of grain of a 

 metal and very useful results are obtained in skilled hands from 

 the comparison of specimens broken under precisely similar con- 

 ditions, although the accuracy of the method is naturally less than 

 that of microscopical examination. On the other hand, a single 

 piece of metal may give two entirely different types of fracture 

 if broken in two different ways, as by slow tension and by 

 sudden shock. A metal which breaks with a so-called " fibrous " 

 fracture in an ordinary testing machine may have a coarsely 

 crystalline fracture when broken by shock or by fatigue. 



The manner in which fracture actually occurs has been 

 studied in detail by methods involving the fatigue of the metal. 

 For instance, a rectangular rod of steel may be fixed at one end 

 to a revolving shaft, whilst the other end is loaded by a weight 

 suspended by means of a stirrup passing over a polished sleeve. 

 The rod is thus subjected to a bending stress which varies 

 periodically in direction. By polishing and etching one surface 

 of the bar and interrupting the test at intervals, the course of 

 destruction of the specimen may be followed with the micro- 

 scope. 1 The development of slip-bands begins in a few crystals 

 and gradually spreads to others, whilst at the same time new 

 systems of lines appear in the grains which were first affected. 

 As the alternations of stress are continued, the lines broaden, 

 indicating the formation of a layer of amorphous material of 

 appreciable thickness along the rubbing surfaces ; after a time 

 actual cracks become perceptible. The cracks always pass 

 through the amorphous films, not between the crystals (that 

 is, in such materials as soft steel, from which brittle inter- 

 crystalline eutectics are absent). A crack once started tends to 

 spread by localisation of stress at its ends but only a few of 

 the cracks which appear reach any great development, the 



1 J. A. Ewing and J. C. W. Humfrey, Phil. Trans. 1902, 200 A, 241. 



