OF ARTS AND SCIENCES. 491 



kamacite consists of a purer iron. As stated above, Dr. Lawrence 

 Smith found in plates obtained from the Sevier County iron, which 

 is closely allied, if not identical, with the Cosby Creek, 27 per cent 

 of nickel. This supposition would harmonize with our idea of the 

 maimer in which crystallization takes place. For though a substance 

 in crystallizing may include foreign substances, still crystallization is 

 a purifying process. Hence, as the molten metal cooled, there would 

 be a tendency for the pure iron to crystallize first, thus forcing back, 

 as it were, a less pure material, which would solidify subsequently ; 

 and the natural alternation of such stages, during a very slow process 

 of crystallization, would result in a succession of plates of compara- 

 tively pure metal interlaminated with a richer nickel-iron alloy. 



This theory is strongly supported by the structure of the Pallas- 

 ites, where the iron occurs surrounding masses of olivine. Here, as 

 Reichenbach has so beautifully shown,* the silicate grains are first 

 surrounded by a deposit of kamacite, and the trias does not appear till 

 this layer of the purer iron has been deposited. 



Moreover, Tschermak has shown,-]- in regard to artificial irons, that 

 pure iron tends towards a cubic crystallization with markings similar 

 to the Hauptmansdorf meteorite, while impure iron, like cast-iron, 

 frequently shows imperfect octahedrons and a scaly structure, not 

 unlike that of many meteorites, and this becomes very striking in the 

 so-called Spiegeleisen. 



"We know as yet very little of the causes which determine the 

 crystallization of substances with an isometric structure in one or 

 another of the three fundamental forms, but we do know that con- 

 stantly the presence of some foreign material in the crystallizing men- 

 struum produces a marked influence on the result. Thus, common 

 salt, which usually cuystallizes in cubes, crystallizes in octahedrons 

 from solutions containing urea. Then again, galena may crystallize 

 under uid^nown conditions in all three of the fundamental forms. 

 Indeed, the crystallization of galena presents a striking analogy to that 

 of iron as it appears in meteorites. It was shown many years ago, by 

 Professor Cooke. $ that galena, like the iron, may be cleaved parallel 

 to faces of all three fundamental forms. Usually the cubic cleavage 

 is the most pronounced ; but in the remarkable variety from Lebanon 

 County, Pennsylvania, the octahedral cleavage is the more eminent, 

 although both the cubic and dodecahedral can also be easily obtained. 



* PoggendorfTs Annalen, Bd. cxiv. p. 90. 



t Akademie dcr Wissenschaften Wien, Bd. Ixx. Abtli. i. p. 447. 



I American Journal of Science, 2d series, vol. xxxv. p. 127. 



