Origin of Pegmatite. — Croshtj and F^dler. 171 
ing the formation of giant granite or pegmatite in the central 
portion. 
Cooling, crystallization and dehydration are processes in- 
volving contraction, and thus, as in a septarian concretion, 
there is a manifest tendency to form vacant spaces or cracks 
in the interior of the boss, in which the highly hydrated and 
liquid residuum of the magma collects and slowly crystallizes. 
If these cracks extend through the hard crust of the boss into 
the surrounding or overlying schists, or if fissures due to ex- 
traneous causes penetrate the boss, they will become the seats 
of sharply outlined and typical veins of pegmatite. 
The gradual character of every phase in the refrigeration 
and solidification of the boss makes it absolutely necessary to 
suppose that ther.^ must b3 somewhere a perfect gradation 
from the normal granite to the poarsest pegmatite. As previ- 
ously stated, this gradation may be observed to some extent 
in the natural ledges; but we may fairly suppose that, while 
the cooling of the boss would be partly lateral, it M'ould take 
place mainly downwards; and since distinct veins of pegma- 
tite represent an actual extravasation of the aqueo-igneous 
magma, it is apparent that the gradation in question should 
besought mainly in depth. It appears to us probable that 
some of the coarser plutonic rocks, like the syenite of Marble- 
head, Mass., in a portion of which the crystals of feldspar are 
from four to six inches in length, should be regarded as in- 
termediate between those of the more normal texture and peg- 
matite. 
It is, perhaps, not impossible that a portion of the water re- 
quired for the more perfect hydration and liquefaction of the 
residuum of the magma may be derived from extrmeous 
sources. The boss is a great body of magma which has been 
forced up to, or developed in, a position of the earth's crust 
far above its normal level. It brings the deep-seated temper- 
ature and other conditions up into the region of comparatively 
active aqueous circulation; and it may be that shrinkage 
cracks or faults, if not interstitial percolation in accordance 
with Daubree's law, would admit meteoric or relatively super- 
ficial waters to a participation in the differentiation of the 
magma. 
There appears to be an opportunity presented, also, to ajiply 
Soret's principle, upon which Iddings so largely relies, in his 
