482 GEORGE I. FINLAY 
County, N. C. Such an association as that last mentioned, (3), was 
observed in but one locality. Here it is quite possible, in the light 
of the experiments made by Morozewicz, that the corundum is an 
original pyrogenic constituent, as suggested by Dr. Pratt. The other 
occurrences, (1) and (2), can hardly be regarded as anything but 
secondary alteration products in veins. They show at times a remark- 
able banded structure, and they are made up largely of the minerals 
of alteration. 
T. H. Holland? in the Geology of India records the occurrence of 
corundum in a variety of rocks. While he states that the occurrence 
of this mineral ‘‘as a constituent of normal lavas in the form of tabu- 
lar crystals characteristic of those obtained artificially by Moroze- 
wicz during the devitrification of a slag leaves no doubt as to the 
possibility of its free crystallization from an igneous magma, as is 
admitted for other simpler though commoner oxides,” he does not 
admit that the occurrences which have come under his observation 
are other than accidental, in the sense that the excess of Al,O, was 
contributed from some extraneous source, or influenced by some rock 
other than the matrix of the corundum. Thus at Karutakalaiyam, 
in Madras, corundum occurs ‘in a coarse feldspar rock,” but only 
near its contact with an eleolite-bearing rock rich in Al,O,. Corun- 
dum in India occurs in association with basic rocks, but near pegma- 
tite dikes; and, according to Mr. Holland, influenced by such intru- 
sions. The occurrence of corundum with sillimanite imbedded in 
orthoclase in an acid rock near Paparapatti, Madras, is referred to 
the contact effects of veins of pegmatite on pyroxene granulite. 
The pegmatite dike near Canyon City, Colo., does not show any 
structures in the field pointing to its having been sensibly changed by 
metamorphic processes. The arrangement of the constituent 
minerals is not such as we look for in a metamorphic rock like gneiss. 
The development of the mica plates is not that which we should 
expect in a schistose rock. Garnet was not observed. The minerals 
in the pegmatite are essentially those in the granite which is near at 
hand. In the dike they are not arranged in lines parallel to the walls, 
as they have been observed to be in pegmatitic veins or dikes filled 
« T. H. Holland, Geology of India, p. to. 
