324 PROCEEDINGS OF THE AMERICAN ACADEMY 



not occur, but its inner surface always exhibits the flaws and irregu- 

 larities observed when a metallic mass contracts in cooling from a 

 fluid state. These balls were used for reducing saline bodies to powder 

 in revolving cylinders containing several hundreds of them, and the 

 conditions were such that the balls, impinging on each other at mere 

 points as it were, received light blows over every part of their sur- 

 faces. It would perhaps be inferred that the diameters would have been 

 reduced by the metal being forced into the void space as the effect of 

 percussion. Instead of this reduction, the balls first became elongated 

 pear-shaped, they then exhibited protuberances, and finally an elon- 

 gated mammillary form, in which the diameter was one half longer 

 than that of the original, while the whole bulk was increased from 

 one to one and twenty-four hundredths. 



" A careful examination of the surfaces showed that the uniformity 

 of the indentations from impinging was constant, and the conclusion 

 was, that the new forms assumed were in no wise affected by any in- 

 equality of this action. 



" On breaking the specimens, the internal structure of each ball 

 was nearly the same, exhibiting an effort to form prismatic crystals 

 radiating from a centre on one side of the void, while every particle 

 seemed to have changed its j^lo-ce and made a new aggregation. 

 Where before the texture was small-granular, broad and brilliant- 

 bladed crystals were found, with open interstices, while in the space 

 originally void the terminal points of many crystals made it a geode 

 in appearance. 



"• In offering an explanation of this extensive change among the 

 molecules, I think we may consider the polarized state of the outer 

 surface of the ball suddenly cooled as continuous in its action. The 

 attraction of the interior molecules for this part is seen in the forma- 

 tion of a void space ; and when the vibrations of impinging points in- 

 duced a movement, the molecules united their dissimilar poles in the 

 ordinary way of building up a crystalline aggregate. The natural 

 crystal of this alloy being prismatic, room for the radiations which this 

 form must exhibit would be found only by an enlargement of the 

 exterior crust, which, owing to the slight degree of malleability in this 

 case, occurred without fracture. Unequal aggregation of crystals 

 formed would produce the concretionary and mammillary masses into 

 which the balls were converted ; and it seems probable that the taking 

 on of this form was but one step in passing to one still more simple, in 



