BRIDGMAN, — ACTION OF MERCURY ON STEEL. 335 



on a steel surface or in the mass of the metal unless the amalgamation 

 has been started on the surface by some other means. This was 

 shown by subjecting rods of the three kinds of steel (hard and soft 

 nickel steel, hard and soft tool steel, soft and quenched bessemer) to a 

 pressure of about 6500 atmos. for twelve hours or more under mercury. 

 These rods were scoured bright with fine emery paper inmiediately 

 before being plunged into the mercury. The fracture after pressure 

 treatment showed not the slightest trace of amalgamation. Reference 

 may be made to some earlier experiments in which no gain of weight 

 could be detected in steel pieces subjected to hydrostatic pressure in 

 mercury. The possibility of amalgamation by pressure was also tried 

 in another form. A hollow cylinder of hardened nickel steel was sub- 

 merged in mercury, and pressure applied to the outside. The only 

 difference between this and the case of the solid rod is that in the 

 former the stress throughout the metal is not uniform, as it is in the 

 latter. It was thought conceivable that mercury might be forced 

 through metal in which the stress was not the same in every direction, 

 while it might show no tendency to work its way through a mass in 

 which the stress was already hydrostatic. The experiment, however, 

 showed no amalgamation in this case either. 



This test for amalgamation by examining the nature of the fracture 

 showed itself so easy to apply and so unmistakable in its indications 

 that it was now applied to the examination of the cylinders which had 

 formed the subject of the first tests. The possibility of the amalgama- 

 tion of the cylinders as an explanation had at first been discarded be- 

 cause the inner wall of the cylinder, where it was to be most expected, , 

 showed no indications of any amalgamation and because attempts to 

 detect the presence of mercury throughout the mass of the metal by 

 microscopic analysis of the polished and etched cross section had 

 given no result. Professor Sanveur had been kind enough to examine 

 four test pieces cut from different cylinders, three of which were rup- 

 tured with mercury and one with water. He found martensitic struc- 

 ture in the three former pieces, and only a very fine granular structure 

 in the other. There was no visible trace of mercury in the pores, nor 

 anything to suggest amalgamation. The differences of structure might 

 be due merely to the slight differences in heat treatment occasioned by 

 the separate hardening of the cylinders. More careful regulation of 

 the conditions would be necessary to settle this point. 



The fi-acture test was applied by cutting a scarf about \" deep all 

 around the cylinder with a thin emery wheel, and then breaking the 

 cylinder at this scarf with a hammer. All the cylinders broken with 

 the mercury showed the same characteristics. (See Figure 9.) The 



