The Rupture of Steel by Longitudinal Stress. 247 



than at any other point in the bar ; this would tell in its favour 

 against a plain bar which is free to break at the weakest spot. The 

 steel on which these experiments were conducted was very homo- 

 geneous, and the variation of strength within a short distance on the 

 same bar would not be more than 4 or 5 per cent., so this would 

 scarcely account for the phenomena. 



On the other hand, a grooved specimen is at a disadvantage when 

 compared with a plain specimen for four reasons : 



(i) The stress is much more unevenly distributed over the least 

 section in the grooved bar than in the plain. 



(ii) If the pull is not exactly parallel to the axis of the bar, bending 

 stresses are induced, which are very prejudicial in the grooved bar, 

 whereas their effect is largely neutralised in the plain bar by the 

 ready flow of the metal.* 



(iii) The load at rupture can be observed with accuracy in testing 

 the grooved bar, for it breaks off short, and the required load is also 

 the maximum load. In testing the plain bar, however, in con- 

 sequence of the very rapid contraction of area immediately before 

 rupture, the load has to be reduced, in order to keep the lever hori- 

 zontal ; sometimes the load cannot be run back quick enough, and the 

 bar may break while the lever is resting on the bottom stop, so that 

 too high a load may be observed as the load of rupture ; this would 

 tend to give a higher breaking .stress in the plain bar than was 

 actually the case.* 



(iv) The grooved bar has a crystalline fracture. The plain bar 

 has a silky fracture. Experiments will be quoted later on to show 

 that the ultimate resistance to rupture is less, the more crystalline is 

 the steel at the moment of rupture. 



' Careful measurements have been made of the test piece No. 831 

 (the others being very similar), to ascertain the least area of all 

 planes passing through any point in the narrowest section at 45 to 

 Jhe axis (i) in the grooved bar, (ii) in the plain bar ; the section of 

 these planes is shown at ab in fig. 4 ; the diameter at the narrowest 

 section was the same in both specimens. The ratio of the area of 

 this plane in the grooved bar to that in the plain was found to be 

 |83 : 100. 



If, now, rupture is an overcoming of a resistance to shearing, the 

 igrooved bar ought to be stronger than the plain in the ratio of 

 183 to 100, other things being the same in both bars. 



For the resistance to shearing, at rupture, will be the resistance of 

 all the planes similar to those shown in the figures, equally inclined 

 to the axis, and if the area of any one of these planes in the grooved 



<* This is only stated as a possible source of error ; no result was accepted if there 

 any suspicion of its being thus influenced. 

 OL. XLIX. S 



