174 The American Geologist. March, is>j7 
of quartz whicli are such a characteristic, as well as the latest, 
feature of the pegmatite. We hold, with Lehinann, that the 
most satisfactory explanation of this blending of pegmatite 
with quartz vcnns on the one hand, as with normal granite on 
the other, is to be found in acoi-responding blending of aqueo- 
igneous fusion with igneo-aqueous solution ; and this dynamic 
gradation, it appears to us, can only result from the gradual 
hydration of the residual magma during the slow centripetal 
solidification of a body of magma and a consequent elimina- 
tion of water. 
In any case, the solidification and crystallization of a body 
of magma must be conceived as centripetal ; and this consid- 
eration makes it impossible for us to agree with Brogger that 
the water of any considerable volume of magma would pass 
outward through the constantly thickening shell of granite 
into the wall-rock. And before the solidification begins the 
ditfusion of magmatic water through the wall-rock appears 
equally improbable, since the magma is to be regarded as a 
chemical union of water and fused rock, a true hydrate, which 
is dissociated by crystallization, but while actively molten 
under pressure,absorbing water rather than giving it off. Fur- 
thermore, it appears to us probable, or at least possible, that 
magma injected into the wall-rock may become more liquid, in 
spite of cooling, through absorption of water; provided the 
wall-rock has a high temperature, or the magma is already 
sufficiently hydrated to have a low fusing point. Local ab- 
sorption of water from the wall-rock would best account for 
the pegmatitic character of the small dikelets or apophyses of 
granite, to be observed on the shores of Marblehead and 
Swampscott, Mass. The main body of granite is quite free 
from ])egmatitic characters; but the innumerable branches or 
apophyses of granite ramifying through the bordering diorite 
are very generally pegmatitic, at least in part, with frequent 
gradations, the most typical pegmatite passing by insensible 
steps into the normal granite on the one hand and ordinary 
vein quartz on the other. The dikelets vary in size from half 
an inch or less to several feet in width, even the smallest often 
exhibiting perfectly the gradation just noted; and the pegma- 
titic character is quite commonly accompanied by a more or 
less distinct banding or zonal structure — symmetrical bands 
