THE CLASSIFICATION OF THE IGNEOUS ROCKS 3 
than zncorrect. It should be made possible for the geologist 
to determine correctly at least the family to which a rock 
belongs, leaving to the petrographer the determination of 
rock species as well as the solution of the purely petrological 
problems. 
To aid his eye the field geologist has only his pocket lens, 
and whatever rock species are fixed upon by petrologists they 
should be grouped into a comparatively small number of families, 
limited by simple and easily tested characteristics. In the case 
of the volcanic rocks it would be necessary to adopt terms broad 
enough to cover all rock types which it is found impossible to 
easily distinguish in the field. This reform would be made in 
the interest of the petrographer quite as much as of the geologist. 
If this be done the petrologist may multiply terms as he will to 
express any extension of his refined methods of study without 
in any way disturbing the composure or the effective work of the 
great body of field geologists. 
BEARING OF RECENT PETROGRAPHICAL STUDIES ON ROCK 
CLASSIFICATION 
From the point of view of the systematic petrologist the two 
most significant developments of petrology during the closing 
years of the nineteenth century have been, first, the numerous 
observations showing that the time honored families of igneous 
rocks, once supposed to be more or less sharply delimited, pass 
by insensible gradations into one another; and second, the return 
of chemical composition as a basis of rock classification to a 
position of prominence nearer to that which it formerly occupied. 
The attention of petrologists was first drawn to the marked 
facial differentiation of a rock magma when the late Professor 
George H. Williams showed that rocks as diverse as quartz-mica- 
diorite and peridotite occur in the same stock near Peekskill, 
N.Y... Since that time other investigators, but notably Iddings, 
Brégger, Ramsay, and Weed and Pirsson, have multiplied the 
1G, H. WitLiaMs, The Gabbros and Diorites of the “Cortlandt Series,” on 
the Hudson River near Peekskill, N. Y. Am. Jour. Sci. (3) XXXV, pp. 438-448, 
1888. 
