136 
CLARKE. 
impurities in the silica and alumina, rendering the latter a 
trifle too high. 
For several of the less frequently determined elements the 
laboratory records of the U. S. Geological Survey furnish 
data. In 211 analyses of volcanic and crystalline rocks 
there recorded, titanium, manganese, and phosphorus were 
determined in a great majority of cases, and other elements 
appear frequently enough to prove something as to their 
relative abundance. Taking these 211 analyses all together, 
they show the following mean percentages for the constituents 
in question: 
Ti0 2 - 
_0-55 
P 2 0 5 - 
_ -22 
MnO . 
_ -10 
C0 2 - 
i 
i 
i 
i 
i 
cb 
—i 
S— _ 
_ *034 
Cr 2 0 3 - 
■ 
r 
i 
i 
i 
6 
to 
BaO_ 
— . -033 
SrO_ 
_ -009 
Cl _ __ 
_ *012 
Li 2 0 - 
_ -Oil 
All of these figures, obviously, are mder-estimates, for the de¬ 
terminations were not made in all cases. Furthermore, the 
rocks analyzed were varied enough in origin, locality, and 
character to avoid any cumulative error due to the peculiar¬ 
ities of any one formation or area. The value for titanic 
oxide includes whatever zirconia may have been present in 
the various rock samples, but, although the latter base is 
widely diffused, its proportion cannot be very high. Tita¬ 
nium, therefore, must be regarded as more abundant than 
phosphorus, manganese, or sulphur—a result hardly to have 
been expected. This conclusion, however, is borne out by 
evidence from other sources. Titanium is rarely absent from 
the older rocks; it is almost universally present in soils and 
clays, and it is often concentrated in great quantities in beds 
of iron ore. Having no very striking characteristics and 
but little commercial importance, it is easily overlooked, and 
so it has a popular reputation for scarcity which it does not 
