ii8 Tlie American Geologist. August, i898 
as those of Iddings, Lang, Zirkel, Rosenbusch, Weed and 
Pierson, Washington, Lasaulx, Fouque, Rust, Becke, Du- 
parc, Ritter, Termier and Brogger, are represented by such 
triangles in a series of plates, and on the same plates are rep- 
resented the triangles of the leading minerals of these rocks. 
The ferromagnesian triangle is found to be divisible, first 
into four parts, or positions, according to the relative propor- 
tions of magnesia and lime included in feldspars, viz: mag- 
nesia zero; magnesia less than the lime; magnesia equal to the 
lime, and magnesia greater than the lime; and each of these 
is further separable into three subdivisions depending on the 
presence or excess of alumina, lime or of soda. 
The alkaline-earthy triangle is also divided into four parts, 
based on the amount of lime present in the feldspars, and each 
part is again divided into four more according to the potash it 
may contain. 
Each analysis, therefore, is represented by two triangles 
placed ill juxtaposition upon the chosen abscissas and ordin- 
ates, and being differently colored give a striking expression 
of the composition of the rock concerned. The scale of the 
diagrams is two millimetres for one per cent., drawn on or- 
diary engineer's paper lined in squares of one millimetre. 
In a general table all rocks, expressed by such triangles, 
are also grouped in vertical and horizontal columns, the verti- 
cal columns expressing the alkaline or alkaline-earthy constit- 
uents and the horizontal the ferromagnesian, with the same 
distinctions as employed in the construction of the triangles. 
This classification is intended as a fundamental framework 
for the comparative grouping of crystalline magmas. It ig- 
nores all such accidental elements as structure, age, texture, 
and even the mineralogical composition, and embraces only 
the quantitative chemical composition. It serves as a point of 
departure for the investigation of the question of the origin of 
the various igneous rocks and the cause of their variation and 
succession, a line of inquiry which the author does not fully 
enter upon in this paper— only outlining in a series of critical 
conclusions the methods necessary to follow in such an 
inquiry. 
As an illustration he applies this chemical classification 
to a series of igneous rocks recently studied by Weed and 
