Septembek 17, 1935] 



SCIENCE 



387 



articles of Harrington and Kraus and Golds- 

 berry. (For analyses with more than two or 

 three per cent, of gangue the percentages have 

 been recalculated.) The small triangle fur- 

 nishes a key to the larger diagram which rep- 

 resents one sixth the area enlarged tenfold. 

 The diagonal line crossing the diagram is the 

 locus of analyses of minerals in Kraus and 

 Goldsberry's series, Cu^Fe2S|+3. This ranges 

 from CuFeSj on the left to Cu^S on the right. 

 Most of the analyses are ranged along this 

 line. Those much above the line are probably 



that very few of the massive bornites are en- 

 tirely free from other minerals, but chalco- 

 pyrite and chalcocite, the two most common 

 impurities in bornite, tend to neutralize the 

 eflFect of each other for 



CuFeS, 4- Cu,S = Cu^FeS, 

 and 



GuFeS^ + 2Cu,S = Cu.FeS,. 



Because of this, and because the impurities 

 are often trifling in amount, the analyses may 

 be used with caution. 



due to errors in the analyses or to the presence 

 of oxidation products. 



The only clustering of points in the diagram 

 is around Cu^FeS^. Some may interpret this 

 as evidence that bornite has the formula 

 CUjFeSj, but on the solid solution hypothesis 

 advanced later on it may simply represent the 

 average solubility. Most of these analyses were 

 made upon massive material and as the study 

 of polished sections proves, apparently pure, 

 massive bornite usually contains small amounts 

 of chalcopyrite, chalcocite or covellite, and 

 occasionally other minerals. It is safe to say 



The lower limit of bornite seems to be rep- 

 resented by Cu3FeS3 with iron content of 16.36 

 per cent.; the Cornish crystals approach this 

 formula and an artificial bornite made by 

 Booking* is very close to the theoretical for 

 Cu3FeS3. Only seven analyses out of the fifty- 

 nine show more than 16.3 per cent. iron. A 

 slight admixture of chalcopyrite (Fe = 30.5 

 per cent.) will easily account for the excess 

 iron. The lowest iron content on record is 6.4 

 per cent. The low iron content of some of the 



* Hintze, ' ' Handbuch der Mineralogie, ' ' Vol. I., 

 p. 914, 1901. 



