ORIGINAL ARTICLES. 
Silica Percentages of Igneous Rocks. 
By V. A. Eytes, B.Sc., and J. B. Simpson, B.Sc. (Agri), B.Sc. 
INTRODUCTION. 
aes: I-VII give a graphic representation of the average silica 
_ percentage of a number of igneous rock types, so arranged as 
to show the relation between silica-percentage and_ rock- 
nomenclature. All the analyses used were obtained from the 1917 
edition of H.S. Washington’s Chemzcal Analyses of Igneous Rocks ”’ + 
This book contains some 8,000 rock analyses from many parts of the 
world. They are divided into two classes, superior and inferior. 
Only those of the former category (some 6,000 in all) have 
been utilized. 
We should at the outset like to make ourselves clear on two points : 
(1) The scope of our paper is limited by the acknowledged fact 
that the nomenclature of igneous rocks depends on many factors 
in addition to silica percentage. 
(2) Our graphs are intended to show what has oe been the 
practice in the past, rather than to supply a definite ruling for the 
future. 
We hope, however, that in view of the general habit of classifying 
igneous rocks broadly into three groups—acid, intermediate, and 
basic—our results may have general interest and suggestiveness. 
CONSTRUCTION OF CURVES. 
The method of construction of the curves may best be explained 
by reference to a typical example, such as that for basalt (Figs. Ia, 
Ib). Consider first Fig. Ia. One finds there a record of the fact 
that while 84 per cent of analysed basalts have 46 per cent of silica 
(reckoned to the nearest whcle number), 94 per cent of them have 
47 per cent of silica, and so on. It was found that the available 
evidence plotted in this fashion gives somewhat inconveniently 
uneven curves. Accordingly “smooth curves” have been derived 
by calculating mean points at the chief irregularities in the rough 
curves. Thus groups of points A, B, C, in Fig. Ia become points 
At, B}, Clin Fig. Ib. In addition to this, low-lying flat parts of curves, 
such as that which occurs under D in Fig. Ia, have been eliminated, 
since they obviously lie outside the extreme limits of the species 
considered. 
The following table explains exactly the data used in constructing 
each curve, stating, for instance, the number of analyses employed. 
Hypabyssal rocks, with the exception of dolerites, have not been 
treated, owing to the present relatively undeveloped state of their 
classification and nomenclature. 
1 Prof. Paper 99, U.S. Geol. Survey. 
a 
