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



element of danger in this work. On one occasion a specimen 

 5/1G in. diameter and 3 in. long penetrated 5 inches of wood 

 driven by a pressure of only G000 atmos. Tins effect might 

 very probably produce serious consequences for one designing 

 apparatus for the highest pressures without previous ex- 

 perience at lower pressures. The effect is insidious because 

 of the unexpected ways in which it may appear. 



The second type of test is very similar to the first in the 

 manner in which stress is applied, but diametrically opposite 

 in its effects. The material for these tests is in the shape of 

 a hollow cylinder, closed at the ends, and subjected to 

 hydrostatic pressure over the entire external surface, on the 

 ends as well as on the curved part of the surface. The 

 tendency of the stress, as is well known, is to collapse the 

 cylinder if the walls are comparatively thin. Such tests 

 are familiar to engineers ; the tube folds in on itself in two, 

 three, four, or more creases, depending on the dimensions of 

 the tube originally and on the very slight departure from 

 perfect geometrical symmetry. That there is collapse at all 

 must evidently be due to some slight geometric imperfection. 

 If the tube is made heavier, however, so that geometric 

 irregularities have less effect, the tube does not show collapse 

 by folding under pressure, but shows behaviour of a different 

 sort, depending on the material. Tests of this sort do not 

 seem to have been made hitherto, or at least are not well 

 known, probably because the pressure required to produce 

 the effect is fairly high. 



If the material of the cylinder is a ductile metal like mild 

 steel or copper, the effect of the pressure is to close up the 

 hole uniformly, the cylinder retaining its geometric figure. 

 Rupture is never produced in a test of this kind, the hole 

 eventually closing up perfectly tight if the pressure is pushed 

 Ear enough. This is perhaps as one would expect : the 

 interesting feature of this method of testing is the enormous 

 raising of the elastic limit that it is possible to produce, and 

 the unusual stress-strain relation below the yield point. At 

 one time tests were made simultaneously on seven Buch 

 hollow steel cylinders. The upper ends oi' these cylinders 

 were led out of the pressure chamber and connected to 

 graduated glass capillaries, The interior oi' the cylinders 

 was filled with mercury, so thai by observing the rise o^t 

 mercury in the capillary it was possible to follow the change 

 of internal volume with pressure. No change of length 

 accompanies the dosing of the hole, so that the rise of 

 mercury in the capillary gives directly the change of internal 

 volume with pressure. Fig. 3 shows B typical diagram of 



