162 Alexander Scott — Principle of Saturation. 



also their comparative restriction to metamorphic rocks. Shepherd 

 and Rankin 1 found that grossularite broke up on melting to give 

 a mixture of anorthite, pseudo-wollastonite, and gehlenite (?), and 

 that it did not appear on the fusion surface of the ternary system 

 Ca — Al 3 3 — Si 2 Of the three garnets which commonly occur 

 in igneous rocks, pyrope is always a product of early separation from 

 the magma, and its probable instability at lower temperatures and 

 pressures is shown by the frequent occurrence of celyphite-borders. 

 Thus Rosenbusch says 2 : " Er [Mhra] nimmt an, dass der in grosserer 

 Tiefe gebildete Pyrop bei Abnahme des Druckes wahrend die 

 Aufsteigens des Magmas mit dem Olivin in Reaktion trat." Its 

 occurrence in a lava does not militate against its crystallization 

 under plutonic conditions, as such rocks are generally considered to 

 have two periods of crystallization, one ' intratelluric ' and the other 

 ' effusive ', the formation of pyrope belonging to the former. 



With regard to melanite, it seems that the chemical factor which 

 determines its occurrence is not so much the amount of silica as the 

 titanium-content, since rock-forming melanites are generally rich in 

 the latter element. Thus, the garnet of the nepbelite dolerite of 

 Oberweisenthal contains 10*84 per cent of TiC^. 3 Again, in the 

 Borolan complex, melanite seldom occurs along with primary sphene, 

 which, however, tends to increase as the melanite fails. The latter 

 also alters readily to secondary sphene. 4 This fact points to the 

 instability of melanite, and indicates that under other conditions 

 a mixture of sphene and, probably, pyroxene, would have taken its 

 place. Further, Weinschenk states that titanium free-lime garnet 

 occurs exclusively as a secondary product of volcanic activity. 5 In 

 some of the occurrences of spessartite it has been shown that the 

 mineral is of pneumutolytic origin, 6 while in practically the only 

 synthesis in which the composition was verified by analysis a mixture 

 of hydrogen and steam was used as a catalyst. 7 



Similar objections can be raised to many other minerals being 

 classed as saturated or unsaturated merely on the strength of their 

 field occurrence. If the variability of composition in the amphibole 

 group is an insuperable barrier to the correlation of mineralogical 

 and chemical composition, it is likewise a barrier to the saturation 

 classification. Numerous amphiboles and pyroxenes have been 

 described containing an amount of silica insufficient to ' saturate ' 

 the bases present, and this has led to the postulation of such mole- 

 cules as R'Al2 Si 6 , which can obviously take up more silica. 

 Whether these hypothetical molecules actually exist or not, the 

 fact remains that some of the minerals included in the pyroxene and 

 amphibole groups are deficient in silica. 



1 Journ. Ind. Eng. Chem., iii, No. 4, p. 10, 1911. 



2 Mikroskopische Physiographie, 4th ed., i, ii, p. 25, 1905. 



:! Eosenbusch, Elemente der Gesteinlehre, 3rd ed., 1910, p. 458. 



4 Shand, Trans. Edin. Geol. Soc, ix, pt. iii, p. 205, 1909; pt. v, p. 386, 

 1910. 



5 Zeit. Krijst. Min., xxv, p. 375, 1895. 



6 Penfield & Foote, Amer. Journ. Sci. (4), iv, p. 105, 1897. 



7 Gorgeu, Ann. chim. phys. (4), iv, pp. 515-56, 1883 ; Compt. Bend., xcvi, 

 p. 1303, 1883. 



