12 EWING, The Structure of Metals. 



practical importance, namely, the fatigue which occurs in 

 metal when we expose it to repeated variations of strain. 

 To take an experimental laboratory example, suppose we 

 have a rod fixed in a thing like the mandrel of a lathe, 

 which is caused to revolve, and let the rod be loaded at 

 its extremity by a weight or by the pull of a spring 

 balance. This causes it to be bent, and the surface is 

 being compressed and extended alternately at every half 

 revolution of the rod. You may cause this to revolve 

 thousands or millions of times, and what you find is that 

 it will ultimately break under the action of a bending 

 force which would be perfectly safe if the rod were 

 standing still. In other words, if there is repeated alter- 

 nation in the strain, then the metal suffers what is 

 technically called fatigue, and this fatigue may lead to 

 rupture under many repetitions of the straining action. 



The nature of this fatigue manifests itself when 

 the piece is examined under the microscope. Watch an 

 individual crystal in the surface of the bar, and at first 

 you may not detect any effect. After a few thousand 

 reversals you may just begin to see the -appearance of 

 slip-lines ; with more reversals these increase in number 

 and in visibility, and finally, after some hundreds of 

 thousands of reversals, the piece breaks. The process 

 of fatigue is associated with the gradual development of 

 more and more prominent slip-lines in the piece which is 

 ■undergoing alternations. Mr. Humfrey and I showed 

 that what actually happens is that some of these slip lines 

 develop into cracks, and when a crack starts it rapidly 

 spreads from crystal to crystal, and at last there is a 

 sudden rupture.* 



I will now ask your attention for a short time to a 

 molecular theory which I have endeavoured to formulate 



* Ewing and Humfrey, Phil. Trans., ser. A, vol. 200, p. 241, 1902. 



