148 DHE JOURNAL VOR NGPOLOG V: 
occurrences not being represented, these factors are to a large 
extent eliminated, and the present occurrences offer a good 
opportunity for study of the relation of structure to composition. 
As regards this relation they are quite in accordance with 
Iddings’ conclusion that “the batic magmas exhibit a much 
greater tendency to crystallize than the highly siliceous or alka- 
ine OMNES.” 2 
But it seems to me that the difference between them is not 
entirely one of simply greater or less tendency towards crystal- 
lization, but that this tendency in the two classes differs in 
quality, and that it is probable that further research will show 
that the highly acid magmas have a tendency to crystallize about 
comparatively few centers, the result being an eminently porphy- 
ritic structure, with a vitreous, felsitic, microlitic or extremely 
fine-grained holocrystalline groundmass, due probably to the 
rapid cooling of the still liquid groundmass magma; while in 
the basic rocks the tendency is to crystallize about very numer- 
ous centers, the result being a fine-grained rock with few pheno- 
crysts and the groundmass containing many small crystals and 
microlites, but comparatively little glass base. It would follow 
from this that in these last rocks the transition between pheno- 
crystic and groundmass crystals will be much more gradual than 
in the acid rocks. 
Such a relation is clearly shown on comparing the basalts 
with the trachytes or rhyolites, where the generally porphyritic 
structure and vitreous or minutely crystalline groundmass of the 
latter is in striking contrast with the slightly porphyritic struc- 
ture and groundmass full of small crystals in the former. That 
the same law probably holds good for the plutonic rocks as well 
is indicated when we compare the granites and diorites; where, 
as a rule, the former are much larger grained than the latter, the 
larger grains being the natural result of fewer centers of crystal- 
lization in a magma which was finally completely crystallized. 
Such tendencies may be most closely connected with the 
‘IDDINGs, Crystallization of Igneous Rocks, Bull. Phil. Soc., Washington, 11, 
1889, p. 107. 
