50 BUTJliETlN 184, UNITED STATES NATIONAL MUSEfUM 



ular, the grains ranging from very minute to macroscopic. Such 

 grains may be a primary structure, or a secondary structure caused 

 by reheating. 



As already stated, the large grams first formed from the melt in 

 artificial iron change to smaller grains in the upper gamma range 

 and disappear, although their original outlines may remain visible 

 because of the presence of impurities along the former larger grain 

 boundaries. 



In meteoric irons then" preservation in that manner is common. 

 In some of the very coarsely granular (inaptly termed brecciated) 

 hexahedrites a polished surface shows grains with dimensions up to 

 2 or 3 inches. Such grains usually have diverse systems of Neumann 

 lines and are bounded by a dark substance. On weathering such irons 

 are likely to break up into rounded nodules. 



Smaller grains, but still of macrosize, are observable in numerous 

 hexahedrites, coarse octahedrites, and nickel-poor ataxites. The 

 boundaries of grains, preserved from the high gamma range, are 

 sometimes marked by copious precipitation of phosphide, which 

 could have taken place only at a high temperature. In Soper (pi. 11) 

 this is especially prominent. 



Within large grains a finer granulation may be observed, caused 

 by the normal breaking up of the original large gamma grains whose 

 former outlines remain, or by the recrystallization accompanying 

 the gamma-alpha transformation. A uniform structure of macro- 

 scopic grains is found in some ataxites (e.g., Mejillones) their diverse 

 orientation giving them the appearance of a mass of glistening flakes 

 with macroetching. 



Granulation is usually coarsest and most conspicuous in the low- 

 nickel irons, where the gamma-alpha range was narrow at a high 

 temperature, and therefore the Widmanstatten structure was trans- 

 itory and relatively slight, there being little difference between the 

 higher and lower nickel phases (see fig. 5). In octahedrites a stronger 

 Widmanstatten pattern, the production of which continued through 

 a longer and lower temperature range, often obscures the grains; 

 although they are well marked in many coarse and medium octa- 

 hedrites. Numerous examples are illustrated in the plates. 



In nickel-rich ataxites the grains usually are obliterated, although 

 the outlines of large gamma grains may sometimes be traced in a 

 dense paraeutectoid groundmass (e.g., Illinois Gulch, pi. 21). 



In hexahedrites the complete transformation of the gamma iron 

 into a unigrain mass of kamacite is shown where Neumann lines 

 pass without change of dhection through the outlines of the original 

 large gamma grains. 



The large grain boundaries do not penetrate plessite fields, though 



