0. C. Farrington — Analysis of "Iron Shale." 307 



The andradite here reported was not observed, but its pres- 

 ence was indicated by the percentages of lime, iron and silica 

 remaining after the deduction of the other constituents. That 

 it could readily become mixed with the shale from the sur- 

 rounding sands seems probable. The other constituents indi- 

 cated are readily referable to the accessory constituents of the 

 meteorite and the results of the analysis are of interest as 

 showing what changes take place in them. The Canyon 

 Diablo meteorites, as is well known, contain as accessory con- 

 stituents, troilite, graphite, cohenite and schreibersite. Of 

 sulphur, representing troilite, only a trace was found. It 

 appears, therefore, that sulphates were formed and leached 

 out. No phosphorus was found to be present as phosphates 

 and only 0*64 per cent as phosphides. It thus appears that 

 while some schreibersite remained, all which had oxidized to 

 phosphate had been removed. The percentage of carbon 

 found, 0*15 per cent, indicated some cohenite or graphite yet 

 remaining. The percentage of chlorine found indicates the 

 presence of lawrencite, to which, as shown later, may be 

 ascribed probably to some extent the oxidation of the mete- 

 orite. The percentage of nickel oxide found is somewhat 

 higher than usually reported for the Canyon Diablo meteorites 

 and suggests that a concentration of nickel oxide has taken 

 place through a greater solubility of the iron. It is well 

 known that taenite, which is the more highly nickeliferous 

 constituent of nickel-iron, resists oxidation longer than kama- 

 cite, and this might tend to increase the nickel content of the 

 oxidized product. Whether the nickel oxide when formed 

 would be removed more or less rapidly than the iron oxide 

 does not seem to be known as yet however. Test was made 

 for metallic iron and nickel in the shale by treating the 

 powder with iodine, but practically no indications of the pres- 

 ence of these constituents were thus obtained. The shale also 

 when immersed in a solution of copper sulphate showed no 

 deposition of copper such as would occur if the above metals 

 were present. Undoubtedly pieces of the shale in which 

 oxidation had been less complete would show such a content. 



Comparison of the analysis of the shale with those of rust 

 crusts of other meteorites made by other authors shows dif- 

 ferences chiefly in the percentage of water found. Thus 

 Pugh, in the crust of Toluca,* Haushofer in that of Cran- 

 bournef and Cohen in that of Beaconsfield^: found an amount 

 of water approximating closely to 13 per cent. This indicates 



* Inaug. Diss. Gottingen 1856, pp. 5-14. 

 t Jour. pr. Chem. cvii, pp. 330-331 ,1869. 

 % Meteoritenkunde, Heft ii, p. 263. 



