OF ARTS AND SCIENCES. 493 



tion. It has been thought that between certain limits in the per 

 cent of nickel present octahedral structure might result, while with 

 a different amount of nickel the crystallization would be cubic. With 

 an idea of determining this point if possible, the analyses of eighty 

 meteoric irons, as given in various scientific journals, were tabulated 

 according to their per cent of nickel, and at the same time the size 

 and character of the figures produced by etching were noted. Un- 

 fortunately, these results appeared so discordant that no trustworthy 

 conclusion could be drawn from them, for it was frequently the case 

 that the composition of the same meteorite as given by equally com- 

 petent analysts differed by several per cent. Thus the per cent of 

 nickel in the Babb's Mill meteorite has been variously stated us 14.7, 

 17.1, 4.7, and 12.4, and yet this appears to be a perfectly homogeneous 

 iron. 



Dr. Flight of the British Museum has made the remark, that 

 WidmanstJittian figures seldom appear in irons containing more than 

 nine per cent of nickel, and it seems to be true that most of the 

 irons giving well-marked Widmanstattian figures contain from five to 

 nine per cent of nickel, though they vary all the way from 3.12 to 

 17.37 per cent. On the other hand, the irons giving well-marked 

 Neumann lines contain generally only five or six per cent of nickel, 

 but the data at present are insufficient to afford any basis for gener- 

 alization. 



Moreover, it is by no means as yet established that the amount of 

 nickel present is the cause which determines the difference of feature 

 in the crystallization of meteoric irons, nor should we be led to infer 

 that such would be the case from the behavior of alloys of metals so 

 closely allied as nickel and iron. It is certainly quite as probable 

 that the effect may be due to some other impurity, — for example, to 

 the presence of phosphorus, which is known to produce such marked 

 effects on the physical properties of iron. According to the analysis 

 of Reichenbach already cited, the taenite plates of the Cosby Creek 

 meteorite contain 0.295 per cent of phosphorus, while the mass only 

 contains 0.089 per cent, and it is well known that Schrcibersite, which 

 is a phosphuret of iron containing fourteen per cent of phosphorus, 

 constantly appears on sections of meteoric iron, in lines along the 

 bands of karaacite. But our knowledge of the amount of phosphorus 

 in the different meteorites is even less complete than our knowledge 

 of the per cent of nickel. We do not know, in regard to any of the 

 impurities, either the average amount in any considerable poi'tion of 

 the meteoric mass, or to what extent the amount varies in different 



