BRIDGMAN. — ACTION OF MERCURY ON STEEL. 



331 



of the elastic limit. The two cylinders above were tested at pressures 

 beyond this limit. To test whether a soft cylinder would be broken by 

 a pressure under the elastic limit if the pressure were suHiciently pro- 

 longed, a cylinder of soft nickel steel filled with mercury was exposed 

 for three weeks to 4000 atmos. without rupture. At the end of three 

 weeks, pressure was increased and rupture took place at 12000 atmos. 

 The break was remarkable, as the cylinder showed no preliminary 

 stretch as all soft steels do, but snapped like 

 a piece of hardened tool steel. This may 

 have been due to fatigue from the prolonged 

 application of pressure, or it may be that 

 weakening by the action of mercury had 

 started at the lower pressure and that the 

 cylinder would have finally broken at 4000, 

 had the application been sufficiently pro- 

 longed. 



An attempt was made to make a series 

 similar to that on the nickel steel on cylin- 

 ders cut from a bar of tool steel containing 

 1.25 per cent carbon. Several of these cyl- 

 inders cracked in hardening so that a com- 

 plete set of observations could not be made. 

 The few successful tests made with the hard- 

 ened tool steel cylinders, however, confirm 

 the conclusions reached with the hardened 

 nickel steel. Thus one cylinder with no 

 hardening flaw burst at 4000 atmos. when filled with mercury, and a 

 similar cylinder, which had a hardening flaw in it, broke at 6500 on the 

 second application of pressure when filled with water. 



Another series of tests was now tried which gave the clue to the final 

 explanation. Pressure was applied to cylinders of the form shown 

 (see Figure 3), in which the end opposite the piston is left solid, and 

 may be of varying thickness. In the first tests this bottom was left 

 only I" thick. The cylinders were made of the same nickel steel as 

 above, hardened in oil. One, filled with mercury, ruptured by blowing 

 out of the bottom at a pressure of 2400 atmos. The bottom blew out 

 of a second in which the transmitting fluid was watei', at a pressure of 

 12700 atmos. The manner of rupture was very diff"erent in the two 

 cases. "When water was used, the bottom was blown out in the form 

 of a clean punching, slightly less in diameter than the hole and slightly 

 bulged as one would expect. (See Figure 4.) On the other hand, the 

 piece blown out by the mercury was in the form of a conical cap twice 



Figure 3. Another form 

 of test cylinder for water 

 or mercury, made of hard- 

 ened nickel steel. 



