292 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1909. 
more jron, magnesia, and lime than alumina, soda, and potash are of 
very exceptional occurrence. The rule mentioned above is so gen- 
erally true that it may be regarded as the normal one for igneous 
rocks in general, and is commonly accepted as such in petrology. 
At the same time, there is considerable evidence that certain sub- 
sidiary relations obtain among the constituents other than silica, 
which, while by no means universal, are at times very pronounced, 
and occasionally seem to supersede the more general law. ‘Thus, 
soda not uncommonly tends to vary with the iron oxides, while potash 
shows similar relations to magnesia, resulting in the presence of 
potassium minerals in highly magnesian rocks and the abundance 
of sodium minerals in those high in iron. Again, while neither iron 
nor magnesia shows any marked affinity toward or tendency to vary 
with alumina, this constituent and lime are occasionally found to 
occur together in great abundance and to the general exclusion of 
the others. These relations are also evident in certain facts of chemi- 
‘cal mineralogy, as the usual predominance of magnesium over iron 
in the potassic biotites and phlogopites, the abundance of soda and 
absence of potash among the highly ferriferous augites and horn- 
blendes, and the numerous silico-aluminates containing much calcium, 
while those with iron or magnesium are comparatively very rare. 
But this tendency to selective and correlated variation among the 
chemical constituents of igneous rocks is not confined to those which 
are present in greatest amount. It is equally well, and indeed in 
some respects more strikingly, shown among the rarer elements, both 
as compared with those which are most abundant and with each other. 
Furthermore, the distribution of some of these rare elements would 
seem to have important bearings on some of the problems of economic 
geology and the distribution of ore deposits. 
The general facts of this distribution and variation of the rare 
elements have been summarized in several recent publications,* but 
many of the details are still. uncoordinated and widely scattered 
through the vast mass of petrographic literature, and there are cer- 
tain aspects and recent developments which are either neglected, or 
only briefly alluded to, in the publications referred to. 
It is now commonly understood that certain elements are prone 
to occur most often and in largest amounts in rocks which are high 
in silica, the. so-called “ acid ” rocks; while others are met with simi- 
larly in those low in silica, the “basic” rocks. This is essentially 
the only set of relations recognized by Vogt, while De Launay in 
“Jj. H. L. Vogt, Zeitschrift fiir praktische Geologie, p. 326 (September, 1898) ; 
J. KF. Kemp, Ore Deposits, 3d edition, pp. 34-387 (1900); H. S. Washington, 
Manual of the Chemical Analysis of Rocks, p. 14 (1904); L. De Launay, La 
Science Géologique, p. 687 (1905); W. F. Hillebrand, Bulletin of the U. S. 
Geological Survey, No. 305, p. 21 (1907). 
