I'RKSIDENT's Al)l)KEt<8 ^SECTIOX H. 6l7 



cleavage along one of these g-liding surfaces or the grains may 

 separate from one another at their boundaries. The foiiner usually 

 takes place in a pure and the latter in an impure metal, but. in 

 either case, the plastic yielding that precedes fracture is effected in 

 the same "nay, by the slipping on one another of the granular 

 lamella'. 



The phenomena we have described are curiously similar to many 

 long known to geologists on a much larger scale, by means of which 

 the most extraordinary distortions of boulders and quartz pebbles 

 are produced under great pressure. Indeed, it has been shown* that 

 the defonnation of marble under stress is of precisely the same 

 character as that of iron. The movements are caused by the con- 

 stituent eiystals changing shape by sliding over their gliding planes 

 or by the production of twin crystals, and the agreement between the 

 two is so close tliat the term flow is just as correctly applied to the 

 movement cif marble in compression as to metals. 



One of the most important and most puzzling phenomena with 

 which sti-uctural engineers have to deal is that of the " fatigue" of 

 metals. The results of the earliest enquiries into this subject appear 

 to be containefl in " The Report of the J^oyal Commission on the use 

 ot Iron in Railway .Structiu'es," issued in 1849. At that time cast 

 iron was a material of much greater structural importance than it is 

 now, and most of the experimental work refei'red to in the report was 

 made upon it. The experiments showed that cast iron bars or giitiers 

 which could stand any given load once before breaking could only 

 stand 'lialf that load wdthout fracture if the load were repeated an 

 indefinite number of times. The cast iron itself, on the other hand, 

 did not appear to be any the worse for tliese repeated loadings, for 

 filter a bai- had been broken down in this way the broken halves 

 were often retested, and never showed any signs of inferiority. At 

 the request of the Board of Trade, Fairbairn, in 1860, carried out a 

 series of experiments on a wrought iron girder, subjected to repeated 

 loads. About this time also Wohler, working from 1860 to 1870, 

 began his laborious and now classical researches on the fatigue of 

 iron and steel. In these experiments bars w^ere subjected to alter- 

 nations of stress at the rate of about seventy alternations per minute, 

 and the number of reversals of stress i^equired to break some of the 

 bars was such that some single experiments must have taken between 

 three and four years to complete. His work established a number of 

 fundamentally important residts. It was shown that wrought iron 

 and steel will ])reak with stresses considerably below the ordinaiy 

 breaking stress as applied gradually in a testing machine, provided 

 the stresses are repeated a sufficient number of times. It was shown 

 that the mnnbei- of icversals necessary to produce rupture depends 

 on the range of stress and not on the maximum stiess, and that, as 

 the range of stress is diminished, so the number of repetitions 

 necessary to produce rupture increases. Further-, within a certnin 

 limiting range the stress may be reapplied an indefinite number of 

 times without producing rupture. These results have since been 

 amply verified and considerably extended ])y the researches of 



* Adams and Nicliolson, Experimental Invpstigation into the Flow of Marble— 

 Prop. R.f^., 1901. 



