340 PROCEEDINGS OF THE AMERICAN ACADEMY. 



a soft cylinder with the sides of the punching amalgamated, are to be 

 explained by the fact previously noted that a freshly fractured surface 

 is amalgamated by mere contact with mercury. Breaking across of the 

 punchings showed no evidence of amalgamation in the mass of the 

 punching, so that all the amalgamation must have taken place after 

 rupture. A similar experiment on the hardened nickel steel punching 

 mentioned above showed more or less complete honeycombing of the 

 metal with mercury. 



This is as far as the explanation has been carried. Enough has been 

 done to show that there is here a genuine effect, so that pressure can 

 not be transmitted directly by mercury in hardened steel cylinders, 

 and that the effect is due to amalgamation. One-sided pressure is 

 necessary to start this amalgamation, so that when steel is entirely 

 surrounded by mercury there is no danger of amalgamation or of 

 penetration of the mercury into the pores. This fact was made use of 

 in modifying the design of the apparatus spoken of in the first part of 

 the paper. 



Summary. 



The fact has been established that cylinders of hardened steel will 

 burst at very much lower than the natural bursting pressure when the 

 fluid exerting the pressure is mercury. Soft steel cylinders show the 

 effect hardly at all, the yield point being reached before the pressure 

 can be raised high enough to produce the effect. The fact that this 

 rupture is due to the amalgamation of the steel is established by the 

 examination of the fracture of such cylinders. The unexpectedly great 

 affinity between steel and mercury was established by the complete 

 amalgamation of surfaces broken under mercury, and the enormous 

 effect of the slightest contact with the air was shown. When this 

 amalgamation is once started, the rapidity with which it spreads 

 through the metal is greatly increased by the action of hydrostatic 

 pressure. The spread of mercury through the mass of the steel and 

 the subsequent destruction of the hollow cylinders is produced by two 

 causes, both of which must act together. One is the natural chemical 

 affinity between mercury and steel, shown by the ready amalgamation 

 of freshly broken surfaces. But the amalgamation is never started by 

 the action of pressure alone. In all those cases in which we have had 

 amalgamation, we have had in addition to the chemical affinity a strain 

 of such a nature as to distend the pores of the metal. This allows the 

 entrance of mercury into the pores so that amalgamation may begin, 

 and also facilitates its further growth, which is most rapid in the 



