Hydrostatic Pressure and Conditions of Rupture. 67 



break glass rod or heavy glass tubing. The fracture is 

 beautifully clean, exactly at right angles to the axis. 



In the cases of the ductile materials, the test is complicated 

 somewhat by the fact that after the necking down has once 

 begun there is a tensile stress tending to pull the bar apart. 

 But this tensile stress cannot account for the beginning of 

 the necking down, so that the tensile stress present during 

 the actual rupture is only an incident due to the particular 

 form of: experiment, and is not at all the true cause of the 

 rupture. Evidently there is no such complication in the 

 case of the brittle materials which break with no necking 

 down. 



This first type of test disposes of the maximum stress 

 criterion, therefore, as applied to either ductile or brittle 

 materials. It shows a fortiori, therefore, that this criterion 

 cannot be applied to brittle materials in contradistinction to 

 ductile materials, as proposed by Scoble *. 



Furthermore, the yield or set point and the rupture point 

 practically coincide. No cases have ever been observed of a 

 bar receiving set under this type of stress without rupture. 

 The maximum stress criterion is applicable, then, neither to 

 rupture nor to set. 



Incidentally this test disposes also of the maximum stress 

 difference hypothesis, although more direct evidence is 

 afforded by another type of test. The principal stresses for 

 this first type of test consist of a compression equal to the 

 hydrostatic pressure on all planes including the axis, and a 

 small compressive stress due to the friction on the plane 

 normal to the axis. The maximum stress difference is equal 

 to the hydrostatic pressure decreased slightly by the amount 

 of the friction. In the similar case of a bar ruptured by 

 tension, the maximum stress difference is equal to the tension. 

 If the maximum stress difference theory holds, therefore, 

 the " pinching-off-efFect " should be produced by a hydro- 

 static pressure equal to the tensile strength in pounds 

 per square inch. As a matter of fact, the stress to produce 

 rupture always exceeded this by 25 or 50 per cent., except 

 for the glass, when the condition was more nearly fulfilled. 



Rupture of the first type has been encountered repeatedly 

 in all this high-pressure work. It is the greatest hindrance 

 to making connexions of any sort from the outside with the 

 interior of a high-pressure cylinder, particularly when con- 

 necting one cylinder to another by tubing, or leading 

 electrically insulated electrodes into the interior of the 

 cylinder. Possible rupture of this sort has been the greatest 

 * Scoble, Phil. Mag. xix. pp. 908-916 (1910). 



F2 



