448 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. 51. 



sometimes swelling into triangular or polygonal enlargements form- 

 ing a string of nearly disconnected beads. Shepard, however, did 

 not discriminate between the taenite and schreibersite, and the two 

 are often so closely associated and intergrown as to make this a matter 

 of difficulty. The most characteristic distinction is that the taenite 

 lies in very thin films parallel to the kamacite, while the schreibersite 

 is m knots, granules, and dendritic fonns, sometimes by itself but 

 often attached to or continuous with the taenite films. That these 

 forms are of the phosphide has been determined by separation and 

 microchemical tests. There is, further, a marked difference in the 

 manner in which the two irons etch, the Hidden iron etchmg quickly 

 and j^ielding a bright, lustrous surface, while that described by Shep- 

 ard, under precisely the same conditions, is acted upon much more 

 slowly and gives a dull surface, on which the figures show less dis- 

 tinctly. 



An analysis of the Shepard iron as given in the paper referred to 

 shows: 



Per cent. 



Iron (Fe) 94. 66 



Nickel (Xi) 4.80 



Cobalt (Co) 34 



99. 80 



There being reasons for doubting the accuracy of this analysis, it 

 was repeated at my request by J. E. Whitfield, with the following 

 results: 



Percent. 



Smcon(Si) 0.001 



Sulphur (S) 025 



Phosphorus (P) 095 



Manganese (Mn) None. 



Carbon (C) 004 



Nickel (Ni) 7.575 



Cobalt (Co) 550 



Copper (Cu) 016 



Platinum (Pt) Traces. 



Iridium (Ir) 002 



Iron oxide (FeA) 350 



Iron(Fe) 91.469 



100. 087 



A partial analysis of the iron described by Hidden shows a very 

 close resemblance, so far as the two essential constituents are con- 

 cerned, Nichols's results, as quoted by Farrington,^ giving: 



Per cent. 



Iron(Fe) 91.02 



Nickel (Ni) and cobalt (Co) 7. 38 



98.40 

 1 Mem. Nat. Acad. Sci., vol. 13, 1915, p. 155. 



