70 Mr. P. W. Bridgman on Breaking '/'< sts under 



the test for one cylinder. Pressure was successively applied 

 and removed, each maximum being greater than the pre- 

 ceding maximum. The diagram shows distinctly the location 

 of the yield-point, and shows also tliat with every application 

 of pressure the new yield-point is not reached until the 

 previous maximum has been passed. This of course is only 

 a verification of common experience as to the possibility of 

 hardening by overstrain. But former experiments on 

 hardening by overstrain have usually been with tensile or 

 compressive or torsion tests, where the possible raising of 

 the yield-point is restricted in amount, a rise of 50 per cm nt. 

 being large. Here, in these collapsing tests, the possible 

 raising of the yield-point seems to have no limit except that 

 set by the complete closing up of the hole. In the diagram 

 shown above the yield-point has been raised about six-fold ; 

 in similar tests with copper, where the complete process 

 has been followed to complete closing of the hole, the yield- 

 point has been raised ten-fold, from 1000 to 10,000 atmos. 

 This was unexpected ; one might naturally expect a raising 

 of the yield-point to two or three times the original value, 

 but after this the metal might be expected to flow uniformly 

 towards the centre like any viscous fluid. 



The experiments with copper also showed the manner in 

 which the yield-point is connected with flow. The maximum 

 pressure reached in any cycle, plotted against the internal 

 diameter, gave a nearly straight line extending Erom the 

 original diameter at zero pressure to zero diameter at 10,000 

 atmos. That is, for equal increments of the yield-point, the 

 diameter decreases by equal amounts. One might expect 

 perhaps that the diameter would tend to decrease more 

 rapidly at the higher pressures. The material of the 

 cylinders alter these tests remained perfect h homogeneous, 

 without fault of any kind. Microscopic analysis has tailed 

 to reveal anything of interest except a slight elongation of 

 tho grain- in the direction of How. 



The diagram shows interestingly one other variation in 

 tho normal behaviour of metal under high pressure, which 

 is mentioned because of its intrinsic interest, although the 



bearing on theories of rnptnre is not BO immediate. This is 



the unusually large hysteresis which goes with the raising 



of the yield-point. The mere existence of hysteresis as an 



effeol apart from '* Elastische Nachwirkung "' baa even 

 been questioned by some. A diagram like the present 

 Bhows unmistakably the possibility o\' this effeel entirely 

 apart from an\ elastic-after-effeots. It also suggests that 



the hysteresis may iii some May be connected with an 



