Alexander Scott — Saturation of Minerals. 323 



in a narrow temperature range. Numerous observers' have reported 

 that garnet, on fusion, breaks up and finally recrystallizes as 

 a mixture of otlier minerals such as anorthite, monticellite, melilite, 

 etc. Attempts to synthesize garnet have not met with much success, 

 though the formation of the mineral has been reported in a few 

 cases.* Hence we may assume that if a rock containing garnet 

 had crystallized by a normal cooling process, under ordinary pressures, 

 garnet would be entirely replaced by other minerals. In some cases 

 it is probable that some of the minerals did not form under such 

 high pressure as the garnet, and hence the latter would be liable 

 to undergo resorption by the residual liquid magma. The well- 

 known celyphite border which occurs round pyrope ciystals in 

 peridotites may arise in this way by the action of a liquid residual 

 magma on the pyrope, under pressures much lower than those at 

 which the latter form and therefore under conditions which would 

 render pyrope an unstable phase. 



Thus it will be seen that there are grave objections to the garnets 

 being considered as saturated or unsaturated in the way felspar or 

 nepheline might be. If spessartite be considered as saturated merely 

 on account of its co-existence with quartz in certain rocks, and the 

 other garnets as unsaturated on account of the absence of this 

 property, the result would be, not only to divide closely related 

 garnet rocks into two widely different groups but also probably to 

 class each with rocks from which they might differ considerably 

 both in chemical and mineralogical composition. The conclusions of 

 Boeke ^ render it very improbable that spessartite differs essentially 

 from the otlier garnets, as he finds that almandine forms a continuous 

 series of solid solutions, with both spessartite and pyrope, and that 

 grossularite and andradite form a similar series. The two latter, 

 however, form with each of the others systems which conform to 

 Roozeboom's fifth type, as does also the system pyrope-spessartite, 

 two limited series of solid solution being formed. Any differences 

 which do exist, are likely to occur between those minerals which 

 give systems of the latter type, but such differences, if any, would 

 probably be physical. Similar objections exist with regard to the 

 micas, whicli also seem to form generally under particular conditions 

 of pressure and temperature. 



Although the hornblendes and pyroxenes are found along with 

 quartz, it is probable that the members of these groups which 

 contain the sesquioxides — alumina and feri'ic oxide — can react with 

 free silica under certain conditions. The most generally accepted 

 explanation of the presence of these oxides is that they occur in 

 molecules of the type R"Al2SiOg (Tschermak's silicate), where R is 

 ferrous iron or magnesium, which molecules are obviously unsaturated 

 with respect to silica. Hai'ker * explains certain phenomena of 

 differentiation as due to a reaction taking place between diopside and 



' Bourgeois, Annales Chim. Phys., ser. V, xxix, p. 458, 1883 ; Vogt, 

 Mineralbilduiig in Schmchmasscn, 1892, pp. 186-8, etc. 



^ Goigeu, Annales Chim. Phys., ser. VI, iv, pp. 536-53, 1885. 



•■' Boeke, Zeit. Kryst. Min., liii, pp. 149-57, 1913. 



■* Geol. Small Isles (Mem. Geol. Surv. Scotland), 1907, p. 91. 



