274 HENRY S. WASHINGTON 
through O7, O7 and O°. Here also the influence of the introduction 
of anorthite is similar to that observed in the preceding cases, in the 
olivinic as in the pyroxenic salfemanes the lines LN and ON respec- 
tively falling so close together, in the otherwise homologous cases, as 
to be indistinguishable unless the diagram is drawn on a much 
larger scale than that used here. 
Finally, we may introduce molecules of magnetite (without either 
pyroxene or olivine) into peralkalic and domalkalic dosalanes and 
peralkalic salfemanes. The leucitic areas of these are shown respec- 
tively by green lines made up of dots, and of dots and short dashes, 
and by red lines made up of dots. Here the shift toward the right or 
less siliceous end of the diagram is very pronounced, owing to the 
entire absence of silica from the femic mineral.t But the triangles 
are of exactly the same sizes as in the homologous cases just studied, 
and the homologous lines are all parallel. Furthermore, the leucitic 
areas become smaller and smaller absolutely with increase in magne- 
tite, until they vanish at the point M, where SiO, =o, the locus of 
non-siliceous magmas, which lies to the right and outside of the 
diagram. 
In the preceding cases we have assumed magmas of definite com- 
positions, corresponding to certain center-points in the quantitative 
classification, as well as femic mineral molecules of simple and defi- 
nite compositions and introduced singly. ° But it is clear that greater 
or less admixtures of anorthite or of femic molecules, or the introduc- 
tion of several femic molecules simultaneously, will influence the 
positions of the various loci to a greater or less extent, but always in 
directions similar to those indicated by the examples given. Thus, 
if pyroxene and olivine, or pyroxene and magnetite, or cll three, were 
assumed to be present in different proportions, the positions of the 
corresponding leucitic areas would lie between those of purely pyroxenic 
and purely olivinic or magnetitic ones, corresponding to the s-lic con- 
tent, the exact position depending on the relative emounts end the 
compositions of the minerals present. 
But in all cases—and this point is of great importance in the sub- - 
sequent discussion when we come to deal with actual rocks—homolo- 
t Apatite, ilmenite, corundum, or any other non-siliceous mineral would have, 
of course, the same effect as magnetite. 
