468 HENRY S. WASHINGTON 
in the rhyolite and keratophyre, and lowest of all in the 
tinguaites. 
It is evident, granting that these relations are not fortuitous, 
which the number of analyses seems to preclude, that this ratio 
does not vary with the SiO,, but seems to be dependent here on 
the general character of the magma from which the rocks 
solidified. Grouping these roughly into two classes according 
to their general characters we may say that the ratio is high in 
the granito-dioritic group and low in the foyaitic. 
It is, it must be confessed, somewhat surprising to find such 
a connection between the ratio of the two iron oxides and the 
petrographical character of the rocks. Is it indeed the fact 
that there is such a difference in behavior between the two 
oxides? Do they really differentiate with respect to each other, 
or is the relation only apparent and due to other causes, such as 
possible oxidation of the ferrous iron in the foyaitic rocks? 
In the flow rocks the ratio is low and it seems possible that 
their solidification at the surface may have induced oxidation. 
But all the other rocks are abyssal or hypabyssal, so that such 
an action would seem to be excluded, or at least equally effective 
in each. We have also seen that the granito-dioritic dikes show 
a uniformly higher ratio than their plutonic analogues. On the 
whole we seem driven, as far as the data at hand allow us to 
decide, to the conclusion that the two oxides differentiate with 
respect to each other and that their ratio is in some way con- 
nected with the composition of the magma. 
In connection with this ratio we may note the interesting 
case of Nos. XV and XIV, the tinguaite and sdlvsbergite, whose 
composition is very similar, the total iron oxides being about the 
same. Inthe tinguaite the high Fe,O, (ratio 0.52) has con- 
ditioned the formation of aegirite, while in the sélvsbergite the 
high FeO (ratio 3.31) has conditioned the formation of glau- 
cophane-riebeckite as the colored mineral. 
Na,O+K,0O 
SiO, 
Iddings* in the investigation of the relationships of rocks. Here 
This ratio has been recently employed by 
*IppINGs:: Jour. GEOL., Vol. III, 956, 1895, Vol. VI, 96, 189, Vol. VI, 219, 1898. 
