596 THEORY OF ISOMORPHISM. [Ch. XXVHL 



very rarely associated together in the same rock ; and that when this 

 happened, as in some lavas of modern date, the hornblende occurs in 

 the mass of the rock, where crystallization may have taken place more 

 slowly, while the augite merely lines cavities where the crystals may 

 have been produced rapidly. It was also remarked, that in the crys- 

 talline slags of furnaces, augitic forms were frequent, the hornblendic 

 entirely absent ; hence it was conjectured that hornblende might be 

 the result of slow, and augite of rapid cooling. This view was con- 

 firmed by the fact, that Mitscherlich and Berthier were able to make 

 augite artificially, but could never succeed in forming hornblende. 

 Lastly, Gustavus Rose fused a mass of hornblende in a porcelain fur- 

 nace, and found that it did not, on cooling, assume its previous shape, 

 but invariably took that of augite. The same mineralogist observed 

 certain crystals in rocks from Siberia which presented a hornblende 

 cleavage, while they had the external form of augite. 



If, from these data, it is inferred that the same substance may 

 assume the crystalline forms of hornblende or augite indifferently, 

 according to the more or less rapid cooling of the melted mass, it is 

 nevertheless certain that the variety commonly called augite, and 

 recognized by a peculiar crystalline form, has usually more lime in it, 

 and less alumina, than that called hornblende, although the quantities 

 of these elements do not seem to be always the same. Unquestionably 

 the facts and experiments above mentioned show the very near affinity 

 of hornblende and augite ; but even the convertibility of one into the 

 other, by melting and recrystallizing, does not perhaps demonstrate 

 their absolute identity. For there is often some portion of the mate- 

 rials in a crystal which are not in perfect chemical combination with 

 the rest. Carbonate of lime, for example, sometimes carries with it a 

 considerable quantity of silex into its own form of crystal, the silex 

 being mechanically mixed as sand, and yet not preventing the car- 

 bonate of lime from assuming the form proper to it. This is an ex- 

 treme case, but in many others some one or more of the ingredients 

 in a crystal may be excluded from perfect chemical union ; and after 

 fusion, when the mass recrystallizes, the same elements may combine 

 perfectly or in new proportions, and thus a new mineral may be pro- 

 duced. Or some one of the gaseous elements of the atmosphere, the 

 oxygen for example, may, when the melted matter reconsolidates, 

 combine with some one of the component elements. 



The different quantity of the impurities or refuse above alluded to, 

 which may occur in all but the most transparent and perfect crystals, 

 may partly explain the discordant results at which experienced chem- 

 ists have arrived in their analysis of the sanie mineral. For the reader 

 will find that crystals of a mineral determined to be the same by physi- 

 cal characters, crystalline form, and optical properties, have often been 

 declared by skilful analyzers to be composed of distinct elements. 

 (See the table at p. 608.) This disagreement seemed at first subver- 

 sive of the atomic theory, or the doctrine that there is a fixed and 



