rrilE METALLOGRAPHY OF METEO'EIC IRON 51 



the latter may show their own grain boundaries. They also do 

 not cross the taenite lamellae separating kamacite bands. 



In the neighborhood of schreibersite inclusions granulation is 

 usually finer (e.g., Mejillones, pi. 11). In artificial irons phosphorus 

 inhibits grain growth. 



Alpha granulation. — On the analogy of artificial ferrite, there 

 should be no separate alpha granulation, so it may be assumed that 

 all primary granulation in meteoric irons originated before the gamma- 

 alpha transformation. 



In ferrite there is often a faint reticidation within the ordinary 

 grains, which is called alpha veining and has been explained in vari- 

 ous ways. The author has observed no clear example of alpha vein- 

 ing in kamacite. In a few instances there is an appearance suggest- 

 ing it, but the appearance of lines in such cases is probably due to 

 minute particles of impurity along the outlines of obliterated gamma 

 grains. 



Smithland, a nickel-rich ataxite, shows a phantom grain pattern 

 due to the suppression of the paraeutectoid structure along former 

 grain boundaries (pi. 26). 



Secondary granulation, a result of re-reheating into the gamma- 

 alpha range or into the lower gamma range, is found in nickel-poor 

 ataxites and also in zones of alteration near the surface of octahedrites 

 and hexahedrites (see Chapter XIII). Such granulation is fine and 

 irregular. 



The twinning structure in gamma grains is discussed in Chapter XI. 



Composition of kamacite. — The composition of kamacite, as chemi- 

 cally determined, has already been referred to in Chapter III. As 

 the saturation ratio of nickel in iron is about 6 percent, this is the 

 maximum nickel content of kamacite. Thus the upper limit of its 

 nickel content is definite. Below that point, however, it is indefinite 

 because any iron with less than 6 percent nickel would necessarily 

 be in the alpha phase at room temperature and therefore would be 

 kamacite. 



The best analyses of hexahedrites, which consist of kamacite, 

 actually show on the average about 5.5 percent nickel, with a very 

 few slightly less than 5 percent. Fractional percentages of cobalt 

 also are always present, and the combined nickel-cobalt percentages 

 vary from about 6 to 6.5 percent. 



As nickel and cobalt are practically identical in their effects upon 

 the phase transformations, they may be considered as a unit metal- 

 lographically. Rinne and Boeke (1907) fixed the saturation point 

 of nickel and cobalt in iron as about 6 percent, which is fairly con- 

 sistent with the average nickel-cobalt percentages found in the 

 analyses of hexahedrites. This may be taken as a fair approximation 



