Prof. Cams-Wilson. 



, we increase the stress over part of the section, and it is the 

 innximum stress that we have to reckon with, for the bar will begin 

 breaking there and tear across. 



There is no donbt that the added material increases the resistance 

 to shearing, and I am, therefore, led to the conclusion that it is this 

 increased resistance to shearing that causes the increase in strength ; 

 in otlier words, by adding material above and below the section, the 

 shearing stress for elements lying in the section is certainly reduced, 

 and, at the same time, the strength is certainly increased, the conclu- 

 sion drawn being that the true measure of the tendency to break is the 

 greatest shearing stress. 



The fact that longitudinal tension is equivalent to a uniform dilating 

 tension and a shearing stress, and that by the means above described 

 we can diminish the latter, without altering the former, and thereby 

 strengthen the bar, are strong reasons for supposing that Profess^ 

 lhir\vin's statement is correct, and that it is the shearing stress 

 duced by longitudinal stress that causes rupture. 



If it be true that the rupture of a steel bar under tension is 

 determined by the greatest shearing stress, we should expect to find 

 that a definite relation existed between the ultimate resistance to 

 direct shearing and the same to direct tension. 



Much has been written about the relation of these two resistances, 

 but the conclusions drawn are very misleading, since the tenoile 

 strength considered has been that calculated in the conventional 

 manner, which has, as has been shown, no real significance and is no 

 real stress. 



By a well known theorem, the greatest shearing stress is equal 

 one-half the longitudinal stress ; we should then expect to find th 

 one-half the true tensile stress at rupture was equal to the stress 

 rupture in a shearing experiment on a piece of the same steel. 



If the steel be soft, it will contract locally before breaking ; hen< 

 the greatest shearing stress will be less than half the longitudii 

 stress in the ratio of the sections of a cylindrical and contracted b 

 cut by planes parallel to dc and ab (fig. 5) respectively, the two 

 having the same cross section at coe ; in other words, in the ratio 

 v/2 (area across coe) to (area across 006), where boe = 45. If 

 ratio be called 0, and the true tensile stress at rupture be p, 

 shearing stress at rupture in a shearing experiment should be eqr 



Table III gives the resuls of some experiments made to investigat 

 this question. The tensile and shearing experiments were mac 

 i vspectively on pieces of steel cut from the same bar ; the former wei 

 iii;iile on circular specimens screwed at each end, and resting on nut 

 bearing on spherical seal ings; the shearing specimens were screw* 



