f 3 6 HOLLAND : GEOLOGY OF THE NEIGHBOURHOOD OF SALEM. 



that the magnesite is due to subterranean not to subaerial agencies 

 removes our South Indian peridotites from the class of exceptions 

 to this rule. 



As the result of a detailed survey made in 1895 for the pur- 

 pose of determining the distribution of the magnesite and chromite, 

 Mr, Middlemiss, whilst confirming my previous discovery of the 

 peridotites in the Chalk hills, grouped the minerals chromite and 

 magnesite together as the results of secondary changes in the dunite. 

 He says, 1 " there can be no doubt that originally nearly the whole 

 of the Chalk hills area was composed of this extreme form of perido- 

 tite known as dunite. But mineral changes rapidly set in, and 

 the two first of these changes that must be noticed are (a) the alter- 

 ation, partially or wholly, of the olivine into serpentine, and (6) the 



segregation of the chromite into nodules and veins 



Subsequent changes .... brought about the conversion of 

 much of the serpentine into magnesite [and] effectually destroyed 

 any ornamental qualities it might have possessed by giving it a dull 

 earthy appearance/' 



As to the chromite, it seems safer, in the absence of unusual 

 features, to follow the ordinary practice of regarding it as a mineral 

 of early consolidation segregated into nodules and bands during the 

 processes of primary crystallization. 



The magnesite and serpentine are, according to my view, formed 

 by secondary processes wholly distinct and independent of one 

 another, and the order of these secondary changes is possibly the 

 reverse of that indicated by Mr. Middlemiss. The magnesite, which 

 is by far the most abundant alteration product, was formed, as stated 

 by King and Foote, by the subterranean action of carbonic acid at 

 high temperatures, attacking, nevertheless, true eruptive, not 

 metamorphic, rocks. The serpentine was formed by simple hydration 

 of parts of the remaining ferromagnesian silicate, to some extent 



1 Rec.y Geol. Surv., Ind., Vol. XXIX, pp. 33 and 34 ; also p. 32. 



( 34 ) 



