794 WHITMAN CROSS 



The series is almost unique as to its manganese minerals, if they 

 are constituents of true igneous rocks. Garnet, manganiferous 

 pyroxenes, and graphite are, however, common constituents of 

 associated metamorphic rocks in India. The first and heavy 

 burden of proof is then to demonstrate that the kodurite rocks are 

 igneous. Known field relations do not do this, as Fermor freely 

 admits. The lack of definite evidence to the contrary scarcely 

 permits the acceptance of the igneous prigin in view of the known 

 occurrences of the pecuHar minerals. 



I am perfectly willing to admit the possibihty of manganese- 

 rich magmas as differentiation products, and even assume that 

 they will be found; so that while not regarding the evidence 

 adduced by Fermor as sufficient to estabHsh the Kodurite Series 

 as magmatic differentiates, I am quite ready to discuss the problems 

 of classification which would be presented by such rocks. 



termor's quantitative classification of kodurite and 

 garnet rock 



Assuming that rocks like kodurite may be of igneous origin, let 

 us review Fermor's discussion of the problems involved in the Quan- 

 titative Classification of the Kodurite Series. That discussion is 

 based on four hypothetical or "calculated analyses." The kodurite 

 proper is the original rock now much altered through replacement 

 of feldspar, believed to have been orthoclase, by opaline sihca. 

 Two of the freshest rocks were analyzed, one from Kotakarra in the 

 Vizagapatam district, and one from Boirani in the Ganjam district. 

 Assuming the replacement of orthoclase by opal, Fermor made an 

 estimate of the original mineral composition of these rocks. The 

 former is thought to have consisted of: apatite 3.36, garnet 

 (spandite) 55.04, orthoclase 41.29, TiO, 0.29, CuO 0.02; the 

 latter of apatite 2.62, orthoclase 57.80, albite 2.79, garnet (gran- 

 dite) 36 . 55, and TiOa o . 24. From these figures equivalent chemical 

 "analyses" have been calculated, which serve as the basis of classi- 

 fication of kodurite. 



The two nearly pure garnet rocks, one from each of the locahties 

 above mentioned, are assumed to have the composition of the 

 garnets in the two kodurites, as calculated from the rock analyses.^ 



■ Memoirs, etc., pp. 256-61. 



