366 Prof. Stromeyer's Analyses 



According to an average of three analyses of this mineral, 

 which all agreed very well with each other, it was found to be 

 composed, in 100 parts, of 



Silica 38-48 



Magnesia 48*42 



Protoxide of iron 11 "19 



Oxide of manganese 0*34 



Alumina 0'18 



98-61 

 Its composition is therefore wholly the same as that of olivine 

 and chrysolite. But the quantity of protoxide of iron is some- 

 what greater per cent than in these two minerals ; and, what 

 is very remarkable, the oxide of nickel is wholly absent. 

 Howard, indeed, says he found some in it : but as this chemist 

 does not state that he separated the olivine from the mass of 

 iron with all possible care, it is easy to conceive that from his 

 experiment nothing can with certainty be concluded ; and be- 

 sides, his datum of a quantity of oxide of nickel is no refu- 

 tation of the above experiments. 



As, from these researches, it is now clear that olivine, chry- 

 solite, and the mineral resembling olivine from the Pallas iron, 

 have the very same components, it is no longer a subject of any 

 doubt that these three minerals belong to one and the same 

 species, and can be distinguished from one another only in re- 

 gard to their appearance. 



As, moreover, the proportion of the silica to the magnesia 

 is the same in the three minerals, and they exactly correspond 

 in the proportions of their equivalents, which compose y^ths of 

 the whole, while the iron alone varies ; so it is, consequently, 

 very probable that the silica in them is combined only with the 

 magnesia, and that this silicate of magnesia is also their only 

 essential constituent. On the other hand, the protoxide of 

 iron with the oxide of nickel, the oxide of manganese, and the 

 aluminous earth, appear to be merely diffused in this silicate. 

 That no oxide of nickel should be contained in the Pallas 

 mass, although it occurs in an iron which so clearly contains 

 nickel, is certainly at the first view extremely strange. If, 

 however, it is supposed that this mass of meteoric iron existed 

 in a fused state, and at the same time the ready reducibility 

 of the oxide of nickel and its slight congeniality for siliceous 

 combination is kept in view, it is not improbable that these 

 circumstances have prevented the admission of this metallic 

 oxide into the Pallas iron ; just as is the case with the smalt, 

 where, notwithstanding the use of niccoliferous cobalt ore, 

 the oxide of cobalt alone, with a mixture of iron and arsenic, 

 combines with the vitreous flux, while the nickel becomes se- 

 parated 



