370 THE SCIENTIFIC PAPERS OF 



as to be unsafe for this mild steel. Two large rivets had been put 

 in a forward position and other rivets to make up the holding 

 strength between the piece under test, and the shackles yielding 

 were put back. In dealing with a material like this mild steel, 

 the strain comes entirely upon the forward rivets, whereas the 

 backward rivets being, as it were, the elongation of a much larger 

 piece of metal, do not feel the strain at all. As it appears to me 

 these two forward rivets were near the edge of the metal, and they 

 set up a tearing action. Now, it was remarked that although the 

 sample was expected to elongate 25 per cent, before rupture took 

 place, it did not elongate at all, but broke in the manner already 

 alluded to. In further testing this metal it was found to be per- 

 fectly homogeneous and capable of an amount of elongation of 

 25 or 28 per cent, before breaking. 



If you take a strip of india-rubber you may elongate it to several 

 times its length, but if you make the least nick in the edge of the 

 india-rubber, and try to elongate it, it will tear readily. In the 

 same way, if you give this mild steel a chance of tearing it will tear. 

 In that case no elongation will take place, because in tearing the 

 strain is confined to an infinitesimal amount of area, and its 

 course in fracture will be the result not of the total amount of 

 resisting area, but of the direction the tearing action may choose 

 to take. With regard to riveting, Mr. Parker states, and perfectly 

 correctly I need hardly say, that in punching a plate such as is 

 used in the construction of large boilers, the metal in the act of 

 punching is reduced 30 per cent, in strength. Mr. Kirk has 

 already offered an explanation which is similar to the explanation 

 I should give ; but with your permission I will draw attention, 

 perhaps a little more, to the precise action that I conceive takes 

 place in punching a thick plate of this material. Suppose the 

 plate is being punched, the punch being above and the supporting 

 die below it. As the punch penetrates into the metal, the latter 

 does not immediately sever ; it will first yield by compression, 

 and it will, as it were, be forced laterally into the mass of metal 

 surrounding the punch, where it will create a zone of highly com- 

 pressed metal all round the hole that is being punched. The metal 

 beyond the zone of compression will be to a certain extent under 

 tension. If the piece punched out is afterwards examined it will be 

 found to be about 10 per cent, thinner than the original thickness of 



