ON THE ENDTTEANCE OF METALS. 



427 



of a diflPerent but similar wire after having been for some time previously 

 in a state of rest or of less active oscillation. 



In the continuation of these experimental researches (after the publi- 

 cation of the paper, it would seem) the effects of the kind of fatigue and 

 rest here referred to manifested themselves very remarkably in the 

 oscillation of wires kept almost constantly in activity during most days 

 of the week, but getting rest usually from Friday evening till Monday 

 morning. The successive oscillations diminished in their amplitude, by 

 internal resistance or some condition like viscosity in their elasticity, 

 much less on the Monday mornings, after their Sunday rest, than at 

 other times, succeeding closely to previous activity. 



The experiments in connection with the subject carried out by A. 

 Wohler at Berlin, the results of which were published in 1870, are of 

 great importance. (See ' Engineering,' 1871.) 



These experiments proved that, in the case of wrought-iron and steel 

 of various qualities, rupture of the material took place after a certain 

 number of applications of a stress less than the statical breaking stress ; 

 that when the stress was alternately tensile and compressive, the range 

 of stress required to produce rupture, treating tension and compression 

 as of opposite sign, was but little greater than the maximum stress applied 

 a similar number of times in one direction only — i.e., simply tensile or 

 simply compressive. And again, when the stress varied from a certain 

 maximum compression to a certain minimum compression, or from a 

 certain maximum tension to a certain minimum tension, the range of 

 stress producing rupture after a similar number of applications differed 

 but little from that in the case where the stresses were in opposite direc- 

 tions. 



The following table — No. II. — gives the result of test-bars cut from a 

 steel axle and subjected to torsion. The bars, numbered from 1 to 5, were 

 twisted in one direction only ; those numbered from 6 to 9 were twisted 

 in opposite directions alternately, the range of stress being therefore 

 double the maximum stress. 



Table II. — Eesult of Test Bars cut from a steel axle and suhjeded 

 to Torsion. 



Not hroken hy this number of applications. 



