Manchester Memoirs, Vol. xlviii. (1904), No. X n 



gives all the results in Unwin's book (Unwin seems to 

 support Wohler on this point) bearing on this matter. It 

 will be seen that in every case for a given range of stress 

 that with the highest mean stress is the weakest. This is 

 obviously of importance in engineering structures. 



The influence of the rapidity with which the repeti- 

 tions take place was discovered by Mr. Smith's experi- 

 ments. The reason for it is obvious from Fig. 2. The 

 more rapid the loading the greater is the lag, and this 

 effect will moreover be most pronounced at high speeds. 

 At very high speeds very few repetitions should suffice to 

 produce fracture, whereas at very low speeds an enormous 

 number should be necessary. This is quite borne out by 

 Mr. Smith's results from which Fig. 6 has been prepared. 



There are a number of other interesting points. 

 Wohler found for instance that sudden changes in the 

 cross section of a specimen were very weakening. This 

 is due to the more intense local stresses set up at the 

 places of change. It is noticeable that even in Mr. 

 Smith's specimens, where the changes of section were 

 gradual, the specimen nearly always broke at these 

 changes of section. This weakening is of course not so 

 great for a steady load. In this case if the yield point at 

 such a place is passed due to the local high stresses the 

 result is merely to put the whole specimen into a state of 

 relative ease. 



Another interesting point is that according to Mr. 

 Smith's experiments mild steel and cast tool steel (both 

 annealed) are approximately equally strong under repeti- 

 tions of stress (Wohler's results are not so conclusive). 

 This result may be due to the less time which the mild 

 steel would occupy in passing through the elastic exten- 

 sion being made up for by the greater time which it 

 would take to get through the permanent extension. 



