50 ME. LEONARD BAIESTOW ON THE ELASTIC LIMITS OF 



reality statements that ordinary extensometers are not sufficiently sensitive to detect 

 the first signs of want of elasticity, and that fatigue increases these signs to 

 recognisable magnitude. 



The Swedish Charcoal Iron and Bessemer Steel. These materials were dealt with 

 first, and the observations are in some respects less complete, as clear ideas had not 

 then been reached. The difficulties of experiment become greater as the steel gets 

 harder, the duration of the test being longer and the rate of recovery due to time 

 less rapid. 



For the Swedish iron the results were obtained with a fresh specimen for each 

 range of stress. The hysteresis loop did not become so constant in width, but other- 

 wise the experiments are in complete accord with those for the axle steel. One of 

 the specimens showed the whole of the extension up to fracture. It seems, therefore, 

 unlikely that anything- O f importance has been missed in the history of a fatigue test. 

 The Marten's extensometer is too delicate an instrument for use when fracture is 

 anticipated, and a less accurate extensometer was substituted in such cases. 



One series of experiments was made with the maximum stress very near to the 

 breaking stress, and again a case of actual fracture, observed nearly to the end, 

 showed the phenomena which may be expected to occur, and gives an indication as to 

 when observations have proceeded far enough. 



Recovery due to stoppage of the machine or other causes was somewhat rapid for 

 this material. 



The results for the Bessemer steel are very similar to the axle steel in all respects, 

 the ranges chosen being in closely corresponding positions. Some slight differences 

 might, perhaps, be referred to. In the case of equal stresses, the initial breakdown 

 of the unstable condition was not succeeded by a nearly horizontal portion ; in other 

 words, the width of the hysteresis loop continually increased. This is not an impor- 

 tant distinction, but is more easy to understand than is the behaviour of the axle 

 steel under the same conditions. 



The earliest case of unequal stresses was treated somewhat differently from the axle 

 steel. The first range of stress applied produced a gradual extension of the specimen, 

 and, after the rate of extension had reached its maximum, recovery was so very slow 

 that finality was not nearly reached after 17,000 repetitions. The range was then 

 decreased. From this point the observations agree with those on axle steel. The 

 rate of extension for the initial range, when this was returned to, was considerably 

 less than that previously observed. The question of recovery is obviously an 

 important factor in the behaviour of materials which are being fatigued. 



The remainder of the experiments do not call for special mention as no points of 

 difference from the axle steel occurred. 



The Elastic Ranges of the Materials. -The whole of the results are given in fig. 7, 

 together with similar figures for WOHLER'S experiments on fatigue to destruction. 



For two of these materials tests to fracture were made at moderately high speeds. 



