2l6 HOLLAND: SIVAMALAI SERIES. 



province, which, so far as we' know, is of Archaean age. The pro- 

 Differentiation of the ducts of differentiation recognised within this 

 petrographical province, pedograph ical province maybe chemically 

 divided as follows :— 



(a) a highly alkaline division comparatively rich in ferro- 

 magnesian silicates, giving rise to the various forms of 

 elaeolite- syenites ; 



(b) a division supersaturated with alumina and poor in the 



ferromagnesian protoxides, forming the corundiferous 

 felspar-rocks with chrysoberyl ; 



(c) a division approximately intermediate between (a) and [6) 



including the laurvikites, and 

 {d) the siliceous end-products, probably forming the associ- 

 ated acid pegmatites carrying aqua-marine (beryl) and 

 quartz-veins. 

 The elaeolite-bearing members of this province are in many ways 

 peculiar ; but new rock names do not at present appear to be 

 necessary. If new exposures are found elsewhere in the same 

 region, and show a probability of relationship to those near 

 Sivamalai, the question of correlating isolated exposures may be 

 simplified by using distinctive names for the different types. For 

 the present, however, the whole series of types may be known as 

 the Sivamalai series, so named from the most prominent mass 

 amongst the exposures. 



The corundum occurring as a constituent of the felspar-rock is 

 regarded as a primary constituent, being amongst the earliest of the 

 Origin of the corun- minerals to separate from a magma which was 

 dum * supersaturated with alumina and very poor in 



the ferromagnesian protoxides. The rock agrees very closely with 

 that which Morozewicz has described under the name corundum- 

 syenite from the Urals, and which, he points out, conforms to the 

 laws established experimentally by him as to the conditions essential 

 for the solution of alumina in alumino-silicate magmas, and the sub- 



( 48 ) 



