A TYPE OF IGNEOUS DIFFERENTIATION 657 
said before, some were selected with the expressed intention of 
showing what the intermediate rocks are like. 
In several large igneous bodies one of the differentiates has 
assumed intrusive relations to the others. In such cases it might 
be assumed that the disturbance is responsible for the abrupt 
change from one extreme to another; possibly the large amount of 
intermediate rock required by theory has been eroded, or is con- 
cealed in depth. Such arguments may apply to batholiths, but 
not to an intrusion with a floor, especially if it is as well exposed 
as at Duluth. Granted that there is a lack of intermediate rock 
at Duluth, the corresponding lack elsewhere indicates that inter- 
mediate rocks were never formed in any large quantity. 
An argument for immiscible separation.—Ii this apparent break 
_ in the series is as normal and characteristic as it appears from the 
table, it gives a rather different impression from the simple sequence 
that has been suggested for the results of differentiation. Some 
modification or restatement of the outline of crystallization differ- 
entiation may bring it into closer accord with known series, but the 
clear impression from the series here tabulated is that two processes 
have been at work either successively or simultaneously. 
If crystallization differentiation produced the banded gabbro, 
with differentiates ranging from peridotite and iron ore to anortho- 
site, what other processes can be conceived which might yield the 
red-rock differentiate? It cannot be that one is a regional and the 
other a local variation, for it is all in the same chamber. It is very 
likely to be a contrast between wet and dry. The first type of 
differentiation occurs in the presence of a very small amount of 
water (though biotite occurs even in the gabbro); and if the crys- 
tallization results in a concentration of water in the mother liquor, 
a point may be reached where the water has an effect not known 
in the earlier stages. Even in this case it is not clear why the water 
concentrating gradually would not yield a series of intermediate 
rocks in considerable volume. 
The abruptness of the gradation leads almost conclusively to 
the idea of a separation of an immiscible liquid. This would be 
favored by the increased amount of water. Water was certainly 
abundant, as is indicated by numerous miarolitic cavities in the 
