18 CORUNDUM IN THE UNITED STATES. r bull. 180. 
soon be worked out, while others might be explored for a hundred or 
more feet without any apparent change in their width. 
The most important observations leading to the adoption of this 
theory are: The occurrence of the corundum surrounded by granular 
dunite and also by serpentine; its occurrence with and surrounded 
by spinel; the sharp contact between the gneiss and the altera- 
tion products of the contact vein; the development, on both sides of a 
corundum vein penetrating dunite, of the same sequence of alteration 
products, which are almost identical with those on the dunite side of 
a contact vein; the usual narrowing and pinching of the dunite veins, 
the trend of which is toward the center of the mass of peridotite, while 
the contact vein seems to extend down indefinitely. 
At a number of the corundum veins in these peridotite rocks, feld- 
spar, which is undoubtedly one of the original minerals of the rock 
and not a secondary product, is found associated with the corundum. 
There is a marked difference in the associated minerals when the feld- 
spar is present; the chlorite is not so thoroughly developed, and 
enstatite is not so common, while margarite is rather abundant and 
zoisite is not uncommon. These last two minerals are rarely met with 
free from feldspar in the corundum veins. 
The separation of alumina from these peridotite magmas lias given 
rise to some interesting problems, which I have treated in a paper l 
published a short time ago. A summary of the conclusions reached 
is given here, as they bear directly upon the occurrence of the asso- 
ciated minerals, spinel, chromite, and feldspar, in a corundum vein. 
From what has been observed in nature and from the experiments 
that have been made in the laboratory, it seems that the separation of 
alumina as corundum from molten magmas is dependent upon the com- 
position of the chemical compounds thai are the basis of the magma, 
upon the oxides that are dissolved with the alumina in the magma, and 
upon the amount of alumina itself. 
1. When the magma is composed of a magnesium silicate without 
excess of magnesia, all the alumina held by such a magma will sepa- 
rate as corundum. 
2. Where there is an excess of magnesia in the magma just described, 
it will unite with a portion of the alumina to form spinel, and the rest 
of the alumina will separate as corundum. 
3. Where there is chromic oxide present in a magma composed 
essentially of magnesium silicate (as the peridotite rocks), and only a 
very little alumina and magnesia are present, these, uniting, separate 
with chromic oxide to form the mineral chromite, and no corundum 
or spinel is formed. 
4. When peridotite magmas contain, besides the alumina, oxides of 
the alkalies and alkali earths, as soda, potash, and lime, a portion of 
'Acq. Jour. Sci., 4th series. Vol. VIII, 1899, p. :X7. 
