332 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



r^v^ 



Figure 4. Form of the 



the diameter of the hole, as indicated in Figure 5. The structure of 

 the ruptured metal from A to B was very coarse and granular, remark- 

 ably like the structure of brass made rotten 

 by mercury. Between the grains of the steel 

 were minute drops of mercury, and the frac- 

 tured surface was partially wetted by mercury 

 when plunged beneath a mercury surface. 

 The only clean break was around the edge 

 of the cap (BC), where the rupture took the 

 form of a clean shear as for the piece broken 

 with water. 



, , . ,, ,. , . The corresponding tests with tool steel 

 break ol the cyhnder ot ,. , i i , i r 



Figure 3 when the fluid cylinders were hard to carry out because ot 

 exerting the pressure is loss of the cylinders on hardening. However, 

 w^ater. two successful 



tests were made. v\yv 

 The bottom of one cylinder, when tilled with >'^\^ ^-^''^'Ny^ 



mercury, blew off at 2700 atmos., while a 

 similar cylinder filled with water did not fail 

 until a'iOO atmos. The manner of failure 

 was different from that of the nickel steel, 

 the detached piece being in the form of a 

 frustum of a cone, as shown in Figure 6. 

 The form of this cap was the same for both 



water and mer- 

 cury, the only 

 difference being 

 that the one frac- 

 tured with mercury showed unmistakable 

 amalgamation over a limited region (AB). 

 This, together with the curious fracture of 

 the nickel steel piece, suggested the amal- 

 gamation of steel by pressure as a possible 

 Figure 6. Form of the explanation, 

 break in a cyUnder hke that Experiment was also made on two cylinders 

 of Figure 3 of hardened tool ^^,^i--, ,,, i j. J.^ j. 



steel when the fluid exerting of tool steel Similar to the above, except that 

 the pressure is mercury. the bottom was |" thick instead of ^". The 



cylinder filled with water stretched non- 

 elastically and leaked at a pressure of 8000 atmos., while that filled 

 with mercury failed at 4000 by the blowing out of mercury along a 

 crack. This break was more like that described by Amagat than any 

 other in the course of these experiments, but even here the break was 



A^ 



C 



Figure 5. Form of the 

 break of the cylinder of 

 Figure 3 when the fluid 

 exerting the pressure is 

 mercury. 



