GENESIS OF THE ALKALINE ROCKS 125 
Differentiation a reversible process.—Tolerance toward the idea 
of assimilation ought to be specially easy for petrologists who have 
studied the proofs and processes of differentiation, for differentia- 
tion is reversible. ‘This is shown in a variety of ways: First, the 
original formation of the unstable melt means the former existence 
of conditions compelling the thorough, mutual solution of all its 
components. ‘The local re-establishment of those conditions would 
lead to the re-solution of phases once separated from the original 
magma. Secondly, the resorption of phenocrysts is a very common 
proof of re-solution and that on a large scale. The crystals forming 
early in artificial melts are often resorbed before equilibrium is 
reached. Salt and cold water saturated with salt make a stronger 
brine at higher temperature. Can hot, primary basalt fail to 
dissolve some of the heterogeneous substances of the earth’s crust 
with which it makes contact during and after eruption? Thirdly, 
because the order of eruption is roughly parallel to that of crystalli- 
zation, and for other reasons, the temperatures of differentiation 
are to be considered as relatively low. Higher temperature should, 
then, presumably tend to restore homogeneity. Finally, the solu- 
tion of older gabbro in the red-rock magmas, shown at Pigeon Point, 
Duluth, near Nordingra, and elsewhere, proves the ability of even 
a late differentiate to make a new mutual solution with solid rock 
of essentially the same kind as that now representing the other pole 
of its own differentiation.* 
If, then, magmatic differentiation is reversible, assimilation of 
foreign, solid rocks is seen to be all the more probable. The case 
of the red rock and gabbro shows that very high temperature may 
not be necessary for solution; concentration of water is another of 
the important factors. 
Space cannot here be taken for a statement of the many facts 
(e.g., replacement of invaded formations by batholithic rocks) 
directly supporting the assimilation theory. They are amply 
competent to forbid belief that fractional crystallization is the only 
important cause for the diversity of igneous rocks. Some revision 
of the subject is called for when Bowen (p. 90) remarks that the 
™R. A. Daly, Amer. Jour. Sci., XLIII (1917), 442 (footnote); A. G. Hégbom, , 
Geol. Foren. Stockholm Forhand., XX XI (1909), 368. 
