Hydrostatic Pressure and Conditions of Rupture. 75 



cylinder, PL II. fig. 4, broke by separation of the fibres along 

 an axial plane. Break occurred suddenly with explosive 

 report, so that it was not possible to ascertain whether 

 rupture really started from the outside or not. But with 

 the smaller cylinder, figs. 5 and 6, break started more 

 gradually, and it was possible to watch the whole proceeding. 

 In fact, two strokes of the piston were necessary to enlarge 

 the crack to the condition shown in the figure after the first 

 beginnings of the fissure appeared at the outer surface. 

 The fissure appeared first as a small longitudinal crack, 

 which extended itself axially, the central portion gaping 

 wider and wider as the metal on one side protruded by 

 slipping out along a shear plane. 



These two experiments showed that the rupture, which 

 apparently begins at the outside, is not produced there by 

 flaws in the steel. There might still be some question as to 

 whether the rupture really did begin at the outside, since of 

 course it is conceivable that the slip had begun at the inside 

 and travelled to the outside, there first becoming noticeable. 

 That the rupture actually does begin at the outside was shown 

 by other experiments on nickel-steel and on copper cylinders. 

 The nickel-steel cylinders do not stretch so much before 

 rupture as the Bessemer cylinders, so that it was possible to 

 produce rupture of these cylinders with a true liquid trans- 

 mitting pressure. That the rupture really begins at the 

 outside was shown by the appearance of the crack on the 

 outside and its gradual growth, just as above, without any 

 liquid leaking through from the inside, as of course it would 

 have done under the high pressure, 30,000 atmos, if there 

 had been the slightest crack reaching from the centre out. 

 The copper cylinder showed the same thing. Here it was 

 possible by a special arrangement of shrunk-on steel rings to 

 make the necessary connexions and produce rupture with a 

 fluid. Rupture again appeared at the outside first, and 

 spread toward the centre along a shear plane, leak not 

 finally occurring until the crack had opened to the dimensions 

 shown in the photograph, fig. 7 (PL II.). 



One feature common to all these tests, whether the rupture 

 through the mass of the metal takes the form of a tearing of 

 the fibres or a shear, is that at the inner surface the break is 

 always along a shear plane for a short distance. If the 

 rupture from the outside is by tear, at the inside there are 

 almost invariably two shear planes, running down into the 

 tear. The result is that a sliver of metal in the form of a 

 triangular prism is expelled through the crack. This sliver 

 has been caught on several occasions in a block of lead. It 



