618 pke8ide.\t's address — section h. 



Spangeuberg, Bauschinger, Baker, and others. Osborne Reynolds 

 and J. H. Smith have, further, shown that the rate at which the 

 stresses are repeated has an important influence^ Working with 

 frequencies between 1,000 and 2,000 a minute, they showed that, 

 under a given range of stress, the number of reversals necessary be- 

 fore rupture takes place diminishes as the frequency increases, and 

 that the range of stress for a definite number of reversals before 

 rupture diminishes rapidly as the rate of the reversals is increased. 

 In a recent paper by Messrs. Stanton and Bairstow,* giving the 

 results of an investigation made at the National Physical Labora- 

 tory, they state, however, that these conclusions do not hold good for 

 reversals up to 800 per minute, and that there is no marked falling 

 off in the resistance of iron and steel due to increasing the rate of 

 reversals from sixty to 800 per minute. 



One of the strang-e features connected with fatigue is that 

 material which has been subjected to many reversals of stress does 

 not show any sig-n of weakness when tested in the ordinaiy way in 

 the testing machine. It is just as ductile and just as strong as ever 

 it was, and yet we know that somehow there is a deterioration, be- 

 cause if we continue to subject it to reversals of stress it fails in the 

 end. But an ordinary test in the testing machine will not show this 

 weakness. In accordance with this, tests of iron and steel taken 

 from structures where the material has been subjected to rough 

 usage and continual repetition of strain for many years almost in- 

 variably show that the strength of the material is unimpaired. 

 Another curious fact is that when fracture ultimately takes place in 

 a bar which has been acted upon by a great number of reversals of 

 stress, there is usually no trace of elongation or drawing out, 

 although the material may be of a veiy ductile charactei- wlien tested 

 in the ordinary way. It looks as though the effect of repetitions of 

 stress was to pick out weak spots in the bar, and that the deteriora- 

 tion that goes on is practically confined to the neighbourliood of 

 such spots, the rest of the bar lieing unaltered. 



Wohler's researches showed that iron, whose tensile strength 

 under a steady load is about '20 tons to the square inch, will break 

 down if it is exposed to perhaps a few millions of reversals of 

 stress of 8 or 9 tons per square inch, alternating between compression 

 and tension. Increase the range to 10 or 12 tons per square inch, 

 and the bar breaks with a smaller number of reversals. On the 

 other hand, we have numbers of examples, as in the balance spring 

 of a watch, the piston rod of a steam engine, or a railway axle, which 

 show that provided the range of the alternating stress be kept 

 sufficiently small many million repetitions may be made without 

 apparent harm. A mild steel axle of a railway carriage is exposed 

 with every revolution to continual repetition of reversed stresses 

 which in some cases reach as high a limit as 5 tons to tlie square 

 inch, seemingly without injury. Of course such axles do sometimes 

 fail, although fortunately failure is comparatively rare, and it seems 

 probable in such cases an explanation is to be found in tlie gradual 

 spreading of a crack from an origin supplied by an air l)ubble or 

 other small flaw in the material. Mr. Thos. Andrews, in a series of 



*Iiist. of C.K., Vol. Kid. 



