112 HOLLAND: GKOLOGY OF THE NEIGHBOURHOOD OF SALEM. 



undulose extinctions and ragged granulated margins. The amphi* 

 bole is of a very pale-green colour, almost colourless, in thin section- 

 The double refraction is strong and the axes of optical elasticity dis- 

 tributed according to the usual arrangement in the hornblendes ; the 

 angle of extinction in clinopinacoidal sections (c:c) is i8°. The 

 crystals are generally elongated slightly along the direction of the 

 vertical axis, but show no definite crystallographic outline. Gliding 

 planes parallel to the base are common ; but definite twinning was 

 observed in one case only. The crystals are stained by ferruginous 

 depositions along cleavage and fracture cracks. 



The presence of hematite with magnetite in the iron-ore has 

 been ascertained simply by determining the 



Intergrowths of hematite , . c f r . -, . ". .,* 



and magnetite. ratio ot ferrous to ferric oxide by titration with 



permanganate of potash. In pure magnetite 

 there is twice as much iron (Fe) in the ferric (Fe 2 3 ) as there is 

 in the ferrous (FeO) condition,; but in specimens collected from the 

 Godamalai beds the iron in the form of ferric oxide was found to be 

 quite three times the amount occurring as protoxide, the actual 

 figures for the average of four tests being 3*41 Fe as Fe 2 3 and i'io 

 as FeO. This corresponds to 4*86 Fe 2 3 and 1*41 FeO, respec- 

 tively, or to 4*54 parts of magnetite with 1*73 parts of hematite. 

 This is the only specimen of Godamalai ore which I have tested, but 

 on examining the results of the analyses made at the Imperial In- 

 stitute it was found that practically all other Salem magnetic ores 

 showed by excess of the sesquioxide the presence of a considerable 

 proportion of hematite with the magnetite. The analyses published in 

 the Imperial Institute Journal for June 1896 (Vol. II, p. 223), gave 

 the amount of each oxide as well as the percentage of pure iron (Fe). 

 By multiplying the percentage of FeO by \ 9 (which represents the 

 molecular ratio of FeO to Fe 3 4 ), we obtain the amount of magnetite 

 present, which, deducted from the total oxides, gives the hematite, 

 This calculation, applied to the 16 analyses of Salem magnetic ores, 

 brings out the fact that in 7 of the specimens the hematite exceeds the 



c 10 ) 



