THE METALLOGRAPHY OF METEORIC IRON 9 



be sections of rounded particles, and the same is true of the minute 

 rounded dots of schreibersite that abound in some irons. 



The nature of all such apparent structural features is often made 

 clear by repolishing a surface, the slight reduction of the surface 

 causing marked changes in the appearance of the structures. 



Very fine black particles may not readily be identified; for, what- 

 ever their composition may be, the strong attack of any solvent 

 etching reagent (not a staining reagent) along the interface causes 

 such a particle to appear as a black dot. Thus, even if they are 

 taenite, which is not affected by ordinary light etching, they may 

 appear black. 



Unless an iron is known to be phosphorus-free, however, such 

 black dots are most likely to be schreibersite; and that assumption 

 is practically confirmed if they are sparse or absent in the vicinity 

 of masses or large needles of that component (pis. 5, 54, 56, 59). 



In irons showing a gamma-alpha transformation structure (Chapter 

 XI) manj^ minute black particles may appear with ordinary etching 

 which are not phosphide. Their identification is discussed in 

 Chapter XVII. 



The plates afford many illustrations of the various structural 

 appearances referred to and of the effects of different etchants. 



Analyses. — The published analyses of meteoric irons cannot always 

 be accepted as dependable. Even in apparently homogeneous 

 hexahedrites and ataxites the composition may vary somewhat in 

 different places, while in octahedrites the structure is so diversified 

 that two samples taken from the same specimen may show marked 

 differences. 



It is often impossible to know whether the material used was a 

 piece cut off or a sample of borings or sawings, which would give a 

 fairer average of composition. Even borings or sawings may include 

 part of a large hidden inclusion, which would cause the analysis to 

 reflect an imdue proportion of sulphur, phosphorus, or other ingre- 

 dient. The difficulty in isolating kamacite, taenite, and plessite, 

 and the impossibility of separating the nickel-poor taenite from the 

 nickel-rich kind, have caused wide variations in the analyses of 

 these components. 



The methods used in determining nickel and separating it from 

 cobalt were formerly less satisfactory than now; in fact, the older 

 analysts often did not try to separate them, giving a combined per- 

 centage for the two. Recent analyses often show more cobalt and 

 less nickel than older ones of the same irons. This, however, is of 

 minor importance from the metallographic standpoint, as nickel 

 and cobalt have virtually the same eff'ects upon the transformation 

 of meteoric nickel-iron. 



507394 — 44 2 



