282 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1909. 
yet found.” The researches of Sandberger, Stelzner, and Dieulafait 
also point to the same conclusion. 
Considering the quantitative distribution of the major constituents, 
silica is almost invariably present in igneous rocks, and almost 
always in greatest amount. In general, the percentage may vary 
from nearly 80, as in granites and rhyolites, to a minimum of about 
24 per cent. Indeed, it may even form 100 per cent, if some dikes 
consisting wholly of quartz are really of igneous origin, as is believed 
to be the case; while in some ores derived from igneous magmas the 
amount of silica may drop almost to zero. Alumina is usually the 
next most abundant constituent, the percentage varying from a maxi- 
mum of about 60, as in some corundum-syenites of Siberia and 
Canada, to a minimum of zero in certain rocks composed wholly 
of olivine. The two iron oxides each show maxima of about 15 per 
cent, except in the rare iron ores due to magmatic differentiation, 
where they constitute together almost all the rock. Magnesia attains 
‘maxima of from 45 to 50 per cent in dunites of New Zealand and 
North Carolina; while lime reaches a maximum of about 20 per cent 
in the anorthosites of Canada. Jron, magnesia, and lime may be 
practically absent in highly siliceous rocks, like granites and rhyo- 
lites, and in some’syenites. Soda may be present up to 17 per cent 
in some rare rocks in which nephelite is abundant; while potash 
attains a maximum of only about 12 per cent in some unusual rocks 
rich in leucite, which are found in Italy and in Wyoming. Both 
the alkalies may be wholly wanting in rocks composed essentially 
of pyroxene or olivine. 
The amount of water present in wholly crystalline rocks seldom 
rises above 2 per cent, if the rock is unaltered, though weathering 
very materially increases this quantity, and high figures for water are 
usually to be attributed to this cause. But some perfectly fresh, 
glassy lavas may carry up to 12 per cent of water, which was unable 
to escape from the magma owing to the rapidity of solidification. 
Titanium dioxide may rarely reach figures of about 6 per cent, as 
in some basalts of the western Mediterranean which I am now investi- 
gating, but is usually present in much less quantity, though it is 
seldom or never entirely absent. In some titaniferous iron ores of 
igneous origin, as those of the Adirondacks and Norway, it may even 
reach 18 per cent. The amount of phosphoric pentoxide is rarely 
over 3 per cent, and that of manganous oxide is scarcely ever above 
1 per cent, the higher figures sometimes reported for this latter con- 
stituent being almost invariably due to errors of analysis. It is but 
seldom that either of these three is entirely absent. 
The maximum amounts of the other minor constituents may be 
briefly stated, as attention will thus be called to their relatively great 
