ORIGIN OF THE CORUNDUM. 207 



the corundum is most abundant in the rock which contains little or 

 no elaeolite. 1 Both occurrences, therefore, point to the conclusion 

 that the conditions favouring the formation of elaeolite and corundum 

 are in some sense complementary, and what appears to be the 

 true explanation is suggested by the work of Morozewicz.* 



The formation of corundum [under experimental conditions], 

 according to Morozewicz, is not dependent on the basicity of the 

 magma, but on the ratio of the alumina to the sum of the other 

 bases, and one can consequently predict the saturation point in 

 pure alumino-silicate magmas. For instance, in pure anorthite 

 magmas the minimum limit is about 36*5 per cent. A1 A 3 , in 

 nepheline about 32 — 33 per cent., in labradorite 27—30 per cent., 

 and in an albite-magma about 195 per cent. 



Now, I find on analysis of the felspar-rock in which the corundum 

 occurs near Sivamalai that it (the felspar) corresponds very nearly to 

 the general formula of albite or orthoclase (R 2 0. A1 2 3 , 6Si0 2 ), being 

 in fact a mixture of albite and orthoclase. As this felspar-mixture 

 makes up the principal mass of the rock, and the accessories contain 

 still larger proportions of alumina, it satisfies as a magma the 

 conditions laid down by Morozewicz, and would become saturated by 

 a smaller quantity of alumina than the magma containing elaeolite. 

 From this consideration alone one would expect to find less 

 corundum in the elaeolite-syenite than in the felspar-rock if both had 

 similar opportunities of saturation with alumina. 



But there is still another feature in connection with the elaeolite- 

 syenite which militates against the separation of corundum. 

 Morozewicz points out that in supersaturated alumino-silicate mag- 

 mas, with the general formula RO. mAl 2 3 . «Si0 3 (R = K g , Na 2 or 

 Ca ; « = 2 — 13), the whole excess of alumina (?n — 1) separates in the 



5 Dr. Coleman also says the largest crystals of corundum occur in the ordinary 

 syenites. 



2 J. Morozewicz : Exprimentelle Untersuchungen fiber die Bildung der 



Minerale im Magma, Tschermak's min. undpetr. Mittheil., XVI I*, 1898, 1 — 240. 



( 39 ) 



