464 N. L. BOWEN 
the universal progressive “‘acidification’”’ of the residual base of 
volcanic rocks which the early crystallization of ferromagnesian 
minerals would necessitate. Finally it may be stated that from 
the theoretical side the early beginning of crystallization of quartz 
and alkali feldspar is to be expected in a rock whose chief charac- 
teristic is richness In quartz and alkali feldspar. 
It is possible, even probable, that Fig. 5 suggests too great a 
difference in time of cessation of crystallization of the various 
minerals. Certain it is that in most granites evidence of relative 
idiomorphism of the minerals does not leap to the eye. It must, 
as a rule, be carefully sought. 
It may also be objected that the med of formation of many 
rhyolites is of that nature which most favors supercooling with 
the possible result that the normal order of crystallization may suffer 
considerable change.t When, however, the quartz and feldspar 
phenocrysts of a rhyolite have attained considerable size they must 
have grown throughout a fair period of time in contact with fluid 
magma and therefore presumably in equilibrium with it. In 
short, they represent the normal early crystals. 
Inductive reasoning leads one to the same conclusion. In at 
least some rhyolites crystallization was instituted in depth and 
proceeded slowly for some time before extravasation (quenching) 
according to the order for a granite since conditions were identical. 
Some quenched rocks of granitic composition should, then, show 
the characteristics already pointed out as consequent upon the 
commonly stated ‘‘order of crystallization” if this order really 
holds and if the observed early crystallization of quartz and 
feldspar is the result of supercooling or any such complication 
ensuant upon difference of conditions under which granites and — 
other rhyolites crystallize. No quenched rocks of granitic com- 
position which show these characteristics are, however, known to 
exist. 
The resorption phenomena so commonly shown by the quartzes 
of a rhyolite might be taken as evidence that quartz may crys- 
tallize first in granitic magma, but is later resorbed and appears 
again only in the later stages. The gradation by insensible steps, 
often found in hypabyssal rocks, from vitrophyre through quartz- 
tA. Harker, op. cit., p. 213. 
