CLASSIFICATION OP THE ROCKS. 127 



production ; for in these rocks there is no sign of crushing. The 

 change which takes place when garnet is formed at the expense of 

 pyroxene is of the kind which Dana 1 termed metachemic, and 

 results in the elimination of silica with concomitant formation of 

 quartz. Pyroxene is stable at high temperatures, whilst it readily 

 changes to amphibole by dynamic action at low temperatures ; but 

 perhaps at some intermediate temperature, below the actual melting 

 point, the pyroxenic molecular compound breaks up into two bodies— 

 the basic garnet and the siliceous residue, quartz. The heat caused 

 by, or the heat accompanying, dynamo- metamorphism is probably 

 essential to this metachemic change ; but the change may occur when 

 the requisite temperature conditions are fulfilled without dynamic 

 action. 



Lacroix has called attention to the acicular inclusions some- 

 ...... . times found in the garnets of these rocks near 



Acicular inclusions in ° 



garnets. Salem ; these he regards, however, as rutile 



needles, 8 whereas the inclusions which I have found in, for instance, 

 the Nagaramalai garnets, show very wide extinction angles and are 

 probably monoclinic in crystallization. They are arranged with 

 remarkable regularity of crystallographic disposition within their 

 host the garnet, having their long axes parallel to the edge of the 

 octahedron, their orthopinacoidal faces parallel to the rhombic 

 dodecahedron and their clinopinacoidal faces parallel to the cube. 8 

 Diller found similar inclusions in a fragment of granulite near the 

 peridotite of Elliott County, Kentucky,* and Harker has mentioned 

 the wide extinction angles exhibited by the needles found in the 

 garnet of an eclogite from Port Tana in the north of Norway. 5 



l Amer. Journ. Sci., Vol. XXXII (1886), pp. 69-71. 



3 Lacroix, Rec, Geol. Surv., Ind., Vol. XXIV, p. 176. 

 a Holland, Rec., Geol. Surv., Ind., Vol. XXIX, p. 16. 



4 Diller, Bull. U. S. Geol. Surv., No. 38 (1887), p. 27. 



6 Geol. Mag., 3rd decade, Vol. VIII (1891), pp, 170, 171. Harker compares these 

 needles to kyanite in his original paper, but in his " Petrology for Students " (2nd edition 

 (1897), p. 327) refers them to rutile. The disposition of the needles parallel to the dode- 

 cahedral faces, and many details about the other constituents of the rock strongly recall 

 these occurrences near Salem. 



( 25 ) 



