496 G. W. Tyrrell—Petrographic Nomenclature. 
and if the term bandaite, for example, is as well defined and 
understood as granite, the compound orthoclase-olivine-bandaite 
becomes as self-explanatory as hornblende-biotite-granite. 
It is impossible to discuss the reform of petrographical 
nomenclature, in igneous rocks at any rate, apart from the con- 
sideration of the need for quantitative classification, preferably upon 
a mineralogical basis. Only when this is established will it be 
possible to make a serious attempt to systematize nomenclature. 
It may then be worth while to consider new prefixes and suffixes to 
be added to the roots at present im use, or whether it would be 
preferable to make a clean sweep, and begin de novo with a brand- 
new scheme of nomenclature. Only in relation to an exact scheme 
of classification will it be possible to discuss what amount of difference 
from already described and named types would justify a new 
name being given, In the absence of an accepted framework of 
classification, petrographers are justified in coming new names 
for rocks which, in their judgment, differ sufficiently from already 
recognized types for a new name to be of value; and when they 
consider the ascription of a divergent rock to a known class as a 
restriction of the usefulness of the latter by unduly widening its 
meaning. 
A perusal of Dr. Holmes’ glossary suggests many interesting 
points of nomenclature of which it will only be possible to discuss 
afew in this paper. It may be remarked in passing that Dr. Holmes’ 
definitions are admirably concise and well expressed. He has 
rightly chosen to record the customary and current meanings of 
the terms, and has abstained from critical comment, which could 
only have been satisfactorily conducted from the standpoint of 
classification, and would have overstepped the intention of the book. 
The abundance of andesites and basalts, and their importance in 
petrogenetic and distributional studies, renders their exact definition 
and difference a question of considerable magnitude. Dr. Holmes 
states (p. 32) that the modern primary distinction between andesite 
and basalt does not depend on the absence or presence of olivine, 
or on the relative proportions of felsic to mafic minerals (though 
each of these criteria have been applied in the past), but on the 
composition of the plagioclase. As bearing on this statement it is 
noteworthy that in the last comprehensive work on igneous rock 
classification!, the main distinction between andesite and basalt is 
based on the proportions of felsic to mafic minerals, the true basalts 
(p. 197) having equal or nearly equal amounts of each, the andesites 
being preponderantly felsic. It is recognized that the felspar in 
andesites is usually andesine or oligoclase, and that in basalts 
labradorite or even more calcic varieties ; nevertheless, labradorite- 
andesites are described as well as oligoclase-basalts. When con- 
taining olivine labradorite-andesites have often been described as 
1 Tddings, Igneous Rocks, vol. ii, 1913. 
