ABSAROKITE-SHOSHONITE-BANAKITE SERIES. 959 
must naturally be considered as a series variable in two principal 
directions chemically: in the ratio of alkalies to silica, and also 
in the silica percentages. The variations of the other chemical 
constituents are to some extent functions of these variables. 
In a region where highly differentiated igneous rocks occur as dis- 
connected geological bodies in great numbers, it might always be 
possible to select some particular series of isolated rocks as a gen- 
etic series, from which a law of differentiation might be derived. 
But it would seem that the several instances of “‘laccolithic differen- 
tiation” described by Pirsson must be the surest basis upon which 
definite laws may be established. And on the other hand, any- 
thing like the frequent occurrence of differentiated lavas in a 
region must involve a multiplicity of lines of differentiation 
which cannot be disentangled, and which must of necessity be 
treated collectively. As Professor Brégger observes in the work 
already quoted the various members of the grorudite-tinguaite 
series in the Christiania region are mostly disconnected rock 
masses, and are probably related genetically with rocks differing 
slightly from them in chemical composition, so that members 
of a parallel series, differing in its general character from the 
one described, may be genetically more closely related to certain 
members of the first series, than to some members of the chemi- 
cal series in which they may have been placed. 
The question how far the genetic relationships between 
igneous rocks can be made to serve as a basis of classification is 
an open one. They certainly constitute the foundation of a 
genealogy of igneous rocks, but it may be doubted whether 
they can properly mark the lines along which a systematic 
classification can be established. 
JoseprH P. IppINGcs. 
