Alexander Scott — Principle of Saturation. 161 



below the temperatures of crystallization of many igneous rocks, 

 since the melting-point of fayalite is given by Vogt 1 as 1065°, and 

 fayalite and forsterite form a continuous series of solid solutions 

 conforming to Roozeboom's Type I. 2 



There is a very considerable amount of evidence against Professor 

 Shand's view that the compound Fe Si 3 cannot exist under any 

 conditions realized in nature. Thus an amphibole, grunerite, 

 consisting of almost pure Fe Si O g , has been described from Collobrieres 

 in France. 3 Hypersthene consists of an isomorphous mixture of 

 Mg Si 3 and Fe Si 3 , as the two salts form a series similar to the 

 corresponding orthosilicates,* and hence both must exist in a free 

 state. The occurrence of Fe Si 3 in the double salt hedenbergite, 

 CaFeSi 2 6 , also points to its free existence, in the liquid condition at 

 least. The rarity of crystals of pure Fe Si 3 is probably to be 

 explained by the fact that rocks with Fe generally contain Mg 

 as well, and by the tendency of Fe Si 3 to form double salts and 

 to enter into solid solution not only with enstatite but also with 

 monoclinic pyroxenes. The occurrence of fayalite in the lavas of 

 Pantellaria 5 is best explained by the dissociation hypothesis, the 

 unstable form persisting as the result of the undercooling to which 

 these rocks would be liable owing to the rapid fall in temperature on 

 extrusion. 



With respect to the felspathoids, the remarks in my paper regarding 

 the reaction 



kaliophilite + orthoclase =^=^ leucite 



referred to crystallization from the liquid state, and no mention was 

 made of the transformation, in the solid state, of leucite to pseudo- 

 leucite. What was stated was that a rock, which under some 

 conditions might crystallize as a leucite rock, would under other 

 conditions form a nephelite-felspar rock. That this case is very 

 pertinent to the question at issue, is shown by the following example. 

 A leucitite or leucite-basalt, in which the predominant minerals are 

 leucite and olivine, would have to be classed as unsaturated. A rock 

 of the same chemical composition could also solidify as a nephelite- 

 felspar- olivine rock, which would be classed as part-saturated. The 

 two rocks would therefore be widely separated in classification, 

 although, petrologieally, they would have many affinities, and would 

 exhibit the same tendency to react with any sediments with which 

 they were in contact. The same anomalies appear when analcite is 

 considered in place of leucite. 



The chief point in objecting to the garnets being classed partly 

 as saturated and partly as unsaturated is their probable origin 

 under particular conditions of temperature and, more especially, 

 pressure. The fact that most garnets break up on fusion and 

 recrystallize as mixture of other minerals is in favour of this, as is 



1 Die Silikatschmelzlosungen, ii, p. 66, 1904. 



2 Ibid, i, p. 152, 1903. 



3 Comptes Rendus, xxiv, p. 784, 1847; Bull. Soc. Min. Fran?., ix, p. 40, 

 1886. 



4 Vogt, loc. cit., ii, p. 112. 



5 Journ. Geol., xxii, pp. 16-27, 1914. 



DECADE VI. — VOL. II. — NO. IV. 11 



