986 A TREATISE ON METAMORPHISM. 



It therefore appears to me that we must look somewhere else to explain 

 the greater part of the discrepancy. It is certain that when the shales (the 

 sedimentary rocks into which the aluminous minerals mainly pass) are meta- 

 morphosed in the zone of anamorphism, the secondary minerals show an excess 

 of alumina. All of the chief original aluminous minerals, viz, feldspar, 

 pyroxene, amphibole, and mica, maybe reproduced; but there also develop 

 other aluminous minerals which are not known in the original rocks. Of 

 these, garnets, staurolite, and the aluminum-silicate minerals— andalusite, 

 sillimanite, and cyanite — are by far of the greatest consequence. The 

 latter minerals are especially characteristic of rocks of the approximate 

 composition of the pelites metamorphosed in the zone of anamorphism. 

 They clearly mark an excess of aluminum and furnish almost conclusive 

 evidence of the segregation to a considerable extent of the aluminum in* 

 the pelites beyond the amount which is present in the original rocks. 



This argument has such weight that I believe more careful analyses of 

 shales with reference to their mass and chemical composition will show that 

 the alumina ought to be considerably higher than in the analyses given. It 

 has already been noted (p. 962) that correlative with the deficiency of 

 alumina there is apparent excess of silica in the shales, and the deficiencies 

 and excesses are remarkably accordant. It has just been seen that the 

 apparent deficiency in alumina is 3.68 per cent. A small fraction of a per 

 cent can be accounted for by the bauxite deposits. The< excess of silica as 

 shown by the calculations (p. 961) is 3.247 per cent. Consequently the 

 two nearly balance. I A 7 enture to predict that when more satisfactory 

 average analyses are available the shales will be found to contain several 

 per cent more alumina and several per cent less silica than is given in the 

 average analyses of shales on page 938. 



IRON. 



According to Clarke's estimate of 1891 iron composes 5.08 per cent of 

 the outer 10 miles (16.1 kilometers) of the crust of the earth, including 

 the lithosphei'e, hydrosphere, and atmosphere, and 5.46 per cent of the 

 original rocks alone. In his estimate of 1900 Clarke estimates that the iron 

 forms only 4.64 per cent of the original rocks." This gives iron fourth 

 place among the elements, it being' surpassed only by oxygen, silicon, and 

 aluminum. Reckoned as an oxide, according to Clarke's estimate of 1891, 



« Clarke, tit., Bull. 78, p. 39; Bull. 168, p. 15. 



