76 Mr. P. W. Bridgman on Breaking Tests under 



is difficult to .sec how this shear along two planes could get 

 started it' the crack originally started from the inside. 

 Evidently the crack runs in from the outside until so close 

 to the centre that the prism slips into the crack, driven by 

 the high internal pressure. This universal manner of 

 rupture affords additional evidence, in those cases where the 

 rupture occurs so suddenly that it cannot be observed, that 

 the crack starts from the outside. 



The particular bearing of this third type of test on the 

 theories of the conditions of rupture is in showing that the 

 maximum extension criterion does not hold, although of 

 course all the other criteria are also ruled out, if the stress- 

 strain relation is calculated up to the rupture-point by the 

 ordinary theory, because every one of the supposed critical 

 quantities has its maximum value at the inner surface. But 

 the striking feature of all the tests is the enormous stretching 

 of the inner surface without rupture. The Bessemer-steel 

 cylinders of PI. II. figs. 4 and 5 show a circumferential 

 elongation at the inner surface of 175 and 125 per cent, 

 respectively. The tool steel cylinder of fig. 8 has an interior 

 elongation of 120 per cent. The copper cylinder of fig. 7 

 shows an interior elongation of 300 per cent, and the lead 

 cylinder of fig. 9 several thousand per cent. There does not 

 seem to be any connexion between the value of the elongation 

 at the exterior surface, where rupture actually does occur, 

 and the elongation under pure tensile tests at rupture. AVe 

 may have values either greater or less than the tensile 

 elongation. The nickel steel cylinders, those of specially 

 toughened steels, the copper cylinders, and the 3 in. Bessemer 

 cylinder all showed a circumferential elongation at rapture 

 very much less than the elongation at rupture under pure 

 tension. The 2 in. Bessemer cylinder showed an elongation 

 almost exactly equal to that snowed by the rings cut from 

 the same piece, while a drawn tube of annealed steel has 

 shown an elongation of 100 per cent, at the outside, and the 

 lead cylinder shows 300 per cent, 'flic tensile value Eor 

 ilic steel i- about 20 per cent., and lor the Lead 25 per cent. 



Tests on brittle materials like glass also show that the 

 maximum elongation criterion is not Eulfilled. Ot course, 



in the case of a brittle material like glass, rupture conic- BO 

 suddenly and is BO complete when it doc- conic that it is 



impossible to tell by any examination of the fragments 

 whether rupture began at the inside or the outside. But 

 the maximum value of the Btress which the glass is capable 

 of -lauding before rupture is considerably in excess of the 

 theoretical Limit. On no theory ought the glass to be able 



