84 BULLETIN 184, UNITED STATES NAiTIOKlAL MUSEIUM 



The author believes, however, that the few examples of supposed 

 martensite or troostite cannot be accepted unreservedly. As was 

 mentioned in Chapter VII, acicular transformation figures consist- 

 ing of alpha and gamma needles suggest a martensitic structure, 

 and black unresolved alpha-gamma aggregates resemble troostite 

 or sorbite. Apparently the structures referred to by Vogel were 

 not verified positively by tests for carbide. 



Instances of Fe-C structures. — The author has found only two 

 examples of unquestionable iron-carbon structures in meteoric irons. 

 One of these was found in an inclusion of extraordinary character 

 in Locust Grove (pis. 65, 66). 



The inclusion consists mainly of a rather coarse dendritic mass 

 in which light gray dendrites appear, surrounded by white areas, all 

 in an almost black groundmass appearing like troostite. Etching 

 with alkaline sodium picrate leaves the gray dendrites unaffected, 

 showing that they consist of a gamma-alpha aggregate. The clear 

 white areas darken, indicating the presence of phosphide. 



The question remained whether the black areas represent merely 

 a phosphide-iron solution or whether they also contain carbon. 

 That was determined by examining an unetched surface on which 

 the component that appears white in plate 65, figure 2, and black 

 in figure 3 is visible, with a gray color similar to that of graphite 

 inclusions in artificial irons. The gray dendrites and the dark mot- 

 tled groundmass appear white and structureless, which is consistent 

 with the assumption that they are wholly metallic (gamma-alpha) 

 and therefore would show no structure without etching. It would 

 seem fairly certain, therefore, that the matrix of the dendrites con- 

 sists of the iron-phosphorus-carbon eutectic known as steadite in 

 artificial irons (Fe, FcaP, FcaC). A very small proportion of carbon 

 is sufficient to produce visible segregations of steadite in artificial 

 irons containing as much as 0.03 percent of phosphorus. 



This conclusion is substantiated by the demonstrable presence 

 of carbide in a narrow zone of acicular appearance surrounding the 

 dendritic area (pi. 66). When this is etched with nital it shows a 

 precipitation of a white substance along crystallographic planes, 

 and also along grain boundaries, suggesting in appearance the struc- 

 ture sometimes found in hypereutectoid cast steels in which free 

 cementite is rejected partly in the form of oriented needles on crys- 

 tallographic planes and partly to the grain boundaries. 



That such is also the structure in this case is proved by etching 

 with alkaline sodium picrate, a common test for cementite in arti- 

 ficial irons. The white component is now blackened, proving it to 

 be cohenite (cementite). The structure resembles that found in 

 a 3 to 5 percent nickel carburi7ed steel and seems beyond question 

 to be identical in nature. 



