Prof. S. J. Shand — Saturation in PetrogyxqyJty. 493 



not a magnesian olivine. He adds that "the olivine found elsewhere 

 in highly siliceous rocks, as granite and rhyolite, is always fayalite, 

 not coniinon olivine or the magnesian forsterite ". 



Having disposed of olivine, I pass to Mr. Scott's other criticisms. 

 The possibility of the production of unstable phases, which undergo 

 transformation subsequent to the consolidation of the rock, is urged 

 us likely to render the saturation criterion inapplicable. The 

 alteration of leucite to nepheline and orthoclase is advanced in 

 illustration of this. But the recognitions of such transformations 

 has never presented any special difficulty. Pseudo-leucite is easily 

 recognized as such in most cases, and uralite, serpentine, iddingsite, 

 liebenerite, bastite, pinite, leucoxene, kelyphite, sprenstein, ' clilorite,' 

 saussurite, nearly always betray their parentage. The difficulties 

 which are introduced by such changes affect all mineralogical 

 classifications, and will equally affect the ideal classification of the 

 future. Uut it is not shown how such a transformation can change 

 a saturated rock into an unsaturated one, or vice versa; lience tlie 

 argument has no relation to the question at issue. 



With regard to the garnets, why does it seem impossible that 

 spessartite and alraandite can be stable in presence of excess of silicic 

 acid, while pyrope and melanite are unstable ? We have just seen 

 that this very difference exists, within a wide range of temperature, 

 between forsterite and fayalite ; and a parallel case can be found 

 in the different beliaviour of calcium carbonate and the isomorphous 

 carbonates of magnesium, iron, zinc, and lead towards an excess 

 of carbonic acid. The garnets themselves differ greatly in stability 

 towards mineral acids, some being easily decomposed while others are 

 very resistant. It is probably not too much to say that tlie end 

 members of most isomorphous series of compounds differ in stability. 

 As regards the effect of pressure on the production of garnets, 

 spessartite is known both in granites and in rhyolites; melanite both 

 in syenites and in phonolites ; pyrope both in deep-seated eclogites 

 and peridotites, and in serpentines which may have been lavas. 

 Here is no evidence that the production of garnet requires special 

 pressure conditions. The "observed facts of distribution", in the 

 case of the garnets, are not to be refuted by deduction from such 

 inadequate premises. If I am mistaken in my opinion of the relation 

 of the garnets to silica, then the proof of it exists in the field and 

 should be sought there. At the same time I freely admit that it 

 would be desirable to have experimental confirmation — or refutation — 

 of the unsaturated character of pyrope and melanite. 



In conclusion, I see no reason to modify ray views regarding the 

 saturation criterion in petrography. Its introduction into the basis of 

 the 'natural' classification of rocks, in conjunction with mineralogical 

 constitution (properly used) and texture, will do something towards 

 giving that classitication a meaning and a precision which hitherto 

 have been conspicuous by their absence. And its use will in no way 

 prejudice the reform — or it may be the abolition — of that classification 

 as knowledge advances. 



