THE METALLOGIIAPHY OF METEORIC IRON 39 



The eutectoid theory seems so consistent, and fits so well the 

 known structures of meteoric irons, that one abandons it with some 

 reluctance. It seems to the author, however, to be untenable. 



Apart from the original erroneous assumption that the magnetic 

 transformation line in the diagram marks a phase change, the exist- 

 ence of a defmite intermetallic compound was assumed. In pearlite, 

 which is a true eutectoid, the two components — ferrite and cement- 

 ite — ^have definite and constant compositions; and if an iron-nickel 

 eutectoid, the counterpart of pearlite, existed in meteoric iron its two 

 components, kamacite and taenite, likewise should have definite 

 compositions. Various writers have believed that taenite is such a 

 compound, its hypothetical composition being placed oftenest at 

 Fe2Ni. As will presently be explained, however, taenite has no 

 definite composition, and the prevailing view is that there is no inter- 

 metallic compound of iron and nickel. 



The composition of taenite. — As noted earlier (Chapter III) the 

 nickel content of taenite has been fixed by various analysts at per- 

 centages from 13 to 48, and recently Smith and Young (1939) by 

 thermomagnetic analysis found taenite in Butler to contain about 

 50 percent nickel, and m the narrowest lamellae as much as 55 percent. 



The impossibility of isolating "pure" taenite was a bar to the 

 efforts of the earlier analysts; but San Cristobal presented an excep- 

 tional case that seemed to obviate that difficulty. Inasmuch as 

 the taenite in that iron is apparently a clear segregate and easily 

 isolated, Pfann considered it an example of substantially pure taenite. 

 He found its composition to be 26.25 to 26.45 percent nickel, and he 

 therefore placed the saturation point of kamacite with respect to 

 taenite at 27 percent. The author, however, found the taenite in 

 San Cristobal to contain a profusion of kamacite particles, as well 

 as areas of a gamma-alpha aggregate and sorbitic patterns (pi. 28), 

 which indicate incomplete separation of the alpha from the gamma 

 phase; therefore the nickel content of the taenite itself must be 

 somewhat higher than the above percentages. It should be regarded 

 as further increased by the presence of cobalt, of which San Cris- 

 tobal contains at least 1 percent and which in this connection may 

 be considered identical with nickel. 



Melil and Derge (1937) found that only in artificial alloys with 

 less than 28.5 percent nickel could a pure gamma structure be re- 

 tained by quenching, and that at 27 to 28 percent the transformation 

 was only partial. This would be consistent with the structure of 

 San Cristobal. At 34 percent they found no transformation possible, 

 which is consistent with the structure of Santa Catharina (Ni-Co 

 35.47 percent). Smith and Young (1939) and Bradley (1939) hold 



