262 HENRY S. WASHINGTON 
rather than nephelite, if there is silica enough, the magnesia and fer- 
rous oxide remaining in olivine. Leucite should thus be formed only 
when there is insufficient silica for all of the potash to form ortho- 
clase, even after all of the soda, magnesia, and ferrous oxide have 
been reduced to their lowest silicated conditions.t These relations 
are somewhat complicated by the presence of the diopside, magne- 
tite, ilmenite, and other molecules; but, as they have been fully dis- 
cussed elsewhere, further details are uncalled for. 
The above-described behavior of the various oxides is the norma- 
tive one; that is, the ideal result of the application of the assumed 
orders of affinity of the bases for silica and for alumina. It remains 
to be seen whether it is the normal? one or not; that is, whether the 
actual facts are in accordance with these assumptions, as regards the 
occurrence of leucite. 
In this theoretical discussion we shall first consider the norms 
only, not the actual rocks, and we shall assume that there is neither 
an excess of alumina over the alkalies and lime (resulting in normative 
corundum), nor of alkalies over alumina (resulting in normative 
acmite or alkali metasilicates), so that alumina may be eliminated 
from consideration, as well as corundum, acmite, and alkali metasili- 
cates. The possibility of the formation of kaliophilite will be dis- 
regarded, as well as the presence of such usually minor constituents 
as Cl, SO,, TiO,, P,O,, etc. The numerical relations of the various 
oxides will be expressed in the following pages as molecular ratios, so 
as to make them comparable chemically. 
« It may be noted that, were the direction of affinity reversed, so that silica had an 
affinity for the bases, rather than the bases for silica, in the order named, the presump- 
tive order of formation of the minerals would be the reverse of that described. On 
such a basis we would expect leucite to be formed in preference to orthoclase, then 
nephelite rather than albite, and hypersthene rather than olivine. On this assump- 
tion, therefore, leucite, nephelite, and hypersthene rocks should be more common 
than olivine ones, which is not the case, and leucitic and nephelitic rocks much more 
abundant relatively to feldspathic ones than these are known to be. 
2 On page 2 of the paper cited above Michel Lévy apparently considers that the 
term ‘‘normative” is synonymous with “normal,” and he objects to it on the ground 
that the normative minerals are seldom normal ones. He has evidently overlooked 
our express statement of the distinction between the two words: “normative” being 
employed by us in the accepted English sense of “establishing or setting up a norm,”’ 
while ‘‘normal’’ is recognized as having the meaning of ‘‘usual” or ‘‘common,” and 
is therefore not used in this connection. (Cf. Quantitative Classification [1903], p. 147.) 
