22 BULLETIN 184, UlSIITED STATES NATIONIAL MTJSEfUM 



because of the fact that artificial iron carbide (cementite) is definitely 

 orthorhombic, having been so determined by many investigators. 

 Cementite and cohenite are identical, except that the latter (at least 

 in some cases) apparently contains a small percentage of nickel. 



Like schreibersite, cohenite is very hard (5.5 to 6), tin-white, 

 brittle, magnetic, and insoluble in dilute acids and therefore un- 

 changed by ordinary etching. Unlike schreibersite, it is infusible 

 before the blowpipe, is soluble in copper ammonium chloride, re- 

 duces copper from solutions of copper salts, and gives no precipitate 

 with ammonium molybdate. 



Occurrence oj cohenite. — Cohenite is found in comparatively few 

 irons. It is abundant in Magura, Bendego, Wichita County, Beacons- 

 field, and Canyon Diablo (Cohen, 1894, p. 115; 1903, p. 221) and also 

 in Cosby Creek, and it has been reported in a number of other irons. 



It seems to be confined to areas of carbon enrichment, as the anal- 

 yses of most of these irons show very small percentages of carbon. 

 The first four mentioned by Cohen show percentages of to 0.07. 

 Canyon Diablo in one sample analyzed yielded the high carbon con- 

 tent of 0.417 percent, in another none. In Canyon Diablo cohenite 

 sometimes is abundant over a considerable area, Cohen having iso- 

 lated 10 percent from one sample, while in other areas none what- 

 ever was found. 



In most of these irons the profusion of cohenite might seem incon- 

 sistent with the small carbon content shown by the analyses. 

 Cohenite, however, is virtually identical with the cementite of arti- 

 ficial irons; and, inasmuch as the atomic weight of iron is 56 and 

 that of carbon 12, one part by weight of carbon produces 15 parts 

 by weight of cementite. The calculation is: 



(3X56) + 12 

 12 



This striking ratio, together with the fact that there may be spots 

 of carbon enrichment, explains why cohenite is abundant in Magura 

 and Wichita County, although analyses of those irons have shown 

 respectively only 0.03 and 0.0 percent of carbon. 



Nickel in cohenite. — As nickel is not carbide-forming in artificial 

 irons, metallurgists might well doubt the presence of nickel in cohen- 

 ite. It has, however, been positively reported by a number of com.- 

 petent analysts who found the following percentages: Dafert in 

 Bendego 2.20 percent; Sjostrom in Beaconsfield 2.22 percent and in 

 Wichita County 2.37 percent; Weinschenk in Magura 3.08 percent. 

 In Canyon Diablo Fahrenhorst reported 1.77 percent nickel; Florence 

 0.13 percent and 2.21 percent; Tassin 2.47 percent, and in another 

 analysis 2.21 percent of nickel and cobalt. These analyses were 

 compiled by Farrington (1915). 



