4 o SCIENCE PROGRESS 



With reference to the presence of water, the close resem- 

 blance between natural cordierite and the synthetic anhydrous 

 compound indicates that the water in the former is not an 

 essential part of the molecule, as was suggested by the synthesis 

 of anhydrous cordierite by Morozewicz (Tscher. Min. Mitt. 

 18, 28, 1898) and by the work of Weibull (Geol. Foren. For- 

 handl. 22, 33, 1900). It is very probable that the water in 

 the natural cordierite is analogous to that in such amphiboles 

 as tremolite, where it is supposed by E. T. Allen and J. K. 

 Clement (Amer. Jour. Sci. 26, 10 1, 1908) to be present in solid 

 solution. 



It is significant that the minerals which contain this " oc- 

 cluded " water are usually not those stable at the highest 

 magmatic temperatures, and hence it seems feasible that one 

 function of the water is to lower the temperature range at 

 which the rock solidifies, with the consequent production of 

 the forms stable at lower temperatures (e.g. amphibole instead 

 of pyroxene). The water, then, is occluded in the mineral and 

 plays no essential part in the molecule. It is possible that this 

 explanation also holds for the hydrated form of cordierite. 



Although most naturally occurring rock-forming metasili- 

 cates contain iron, most of the synthetic work which has been 

 done has been carried out with iron-free silicates, and hence 

 the effect of the presence of ferrous oxide, etc., on the equili- 

 brium of the various systems is little known. Several attempts 

 have been made to work out the system FeO — Si0 2 (cf. Hofman, 

 Trans. Amer. Inst. Min. Eng. 29, 701, 1899), but these have 

 not been very successful owing to the difficulties introduced 

 by the existence of the various iron-oxides and the fact that 

 the relations between these are not yet completely known (cf. 

 R. B. Sosman and J. C. Hostetter, Jour. Amer. Chem. Soc. 38, 

 807, 1 91 6). In a recent paper, A. McCance (Jour. Iron and Steel 

 Inst. 47, 239, 191 8) obtains results very different from those of 

 Hofman. In a number of melts obtained by fusing a mixture 

 of ferrous carbonate and silica, crystals form which are stated 

 to be griinerite (FeSi0 3 ). As, however, the composition is not 

 verified by analysis and the identifications are made by metallo- 

 graphic in place of the more usual petrographic methods, it is 

 very doubtful if the crystals are really ferrous metasilicate. 

 In any case it is highly improbable that griinerite, which is an 

 amphibole and seems generally to be of metamorphic origin 



