£12 HOLLAND: SIVAMALAI SERIES. 



There is still a further circumstance to complete the parallel 

 between the occurrence in the Urals and that near Sivamalai : the 

 same eruptive zone in the Urals also includes elaeolite-syenite 

 (miascite), which Morozewicz regards as genetically related to the 

 corundiferous syenites. 



In these three instances, the corundum has apparently separated 

 out as a primary, original constituent from a magma supersaturated 

 with alumina, the alkaline nature (with a predominance of soda) 

 having permitted the previous solution of large quantities of 

 alumina. There are no circumstances connected with these occur- 

 rences to indicate formation by secondary or by contact action, 

 and no facts to suggest that the alumina has been locally picked up 

 at the time of intrusion. That magmas rich in alkalies and rich in 

 alumina have existed may be postulated without necessarily com- 

 pelling an explanation as to how such magmas were formed. It is 

 important, however, not to confuse crystallizations from a definite 

 aluminous magma with mere local concentrations of that compound 

 due to contact action and local absorption of aluminous rocks. The 

 corundiferous rocks near Sivamalai do not break through any 

 unusually aluminous rocks, the associated gneisses be?ing of the 

 ordinary intermediate to acid kind with probably never more than 

 15 per cent, of alumina. The same thing appears to be true of 

 the Canadian occurrence, and as the elaeolite-syenite and corundi- 

 ferous rocks of Ontario occupy a belt some 75 miles long and 

 about 5 miles wide, the peculiar character of the rocks could hardly 

 be the result of an accidental and local contact. 



( 44 ) 



