March, 1907. Iron Meteorites — Farrington. 109 



nickel.* While few meteorites probably vary to this extent, such de- 

 terminations show the need of as thorough sampling as possible if a 

 mass analysis is to be made. Occasionally a marked variation in 

 the analyses of a meteorite seems explicable only on the assumption 

 that the material analyzed did not belong to that meteorite. Such, 

 for instance, seems the most reasonable explanation for the percentage 

 of nickel, 12.67 pei" cent, reported by Hayes for Limestone Creek, as 

 compared with the percentages, 25-30 per cent, obtained by other 

 analysts. Errors of this sort are obviously difficult to detect, and can 

 only be surmised in extreme cases. Another and more serious cause 

 of discrepancies in analyses is the imperfect separation by the analyst 

 of nickel and cobalt from the iron. The methods for this separation 

 are not altogether satisfactory, even at the present day, and in earlier 

 years they were much less so. Consequently the results of the 

 earlier analysts were for the most part too low in these ingredients. 

 The determinations of specific gravity shown in the tables appear in 

 some cases to have been equally open to sources of error with the 

 analyses. It can easily be calculated that the specific gravity of an 

 iron meteorite is likely to be between 7.6 and 7.9, since the specific 

 gravity of pure iron, 7.85, will be increased by that of nickel, 8.8, 

 according to the proportion of the latter. It will be decreased by 

 accessory minerals, such as troilite, which has a specific gravity of 4.7, 

 schreibersite, 6.5, graphite, 2.2, and oxidized ingredients. Any poros- 

 ity of the meteorite will also lessen its specific gravity. It is obvious, 

 therefore, that determinations of specific gravity made on small frag- 

 ments can hardly represent that of the mass as a whole, since they 

 may contain a disproportionate quantity of accessory ingredients or 

 may be more oxidized than the main mass. It is hardly credible that 

 porosity or accessory ingredients of a meteorite would in any case 

 reduce its specific gravity below 7. Determinations below this figure, 

 therefore, probably indicate that oxidized material was used. From 

 the showing in the tables that large numbers of meteorites have prac- 

 tically similar composition, it is evident that similarity of composition 

 cannot be used, as has often been done hitherto, to prove identity of 

 origin of meteorites found at different places. This method at one 

 time obtained considerable vogue. Dissimilarity of composition, on 

 the other hand, as a rule indicates separate falls. The only marked 

 exception to this rule seems to be furnished by the two masses of 

 Babb's Mill, one of which shows about 11 per cent, the other about 

 17 per cent, of nickel. The only alternative supposition possible here 



* C. R., 1893, cxvi., 290. 



