ABUNDANCE OF THE CHEMICAL ELEMENTS. 137 
deserve. Among all the elements it probably ranks tenth 
or eleventh in point of absolute abundance and is rare only 
as regards obvious concentrations. 
For phosphoric acid and manganese the data given are 
probably not far out of the way, but in the case of carbon 
the estimation is more troublesome. The percentage of CO 2 
in volcanic and crystalline rocks, 0‘37, is doubtless untrust¬ 
worthy, for the reason that surface rocks are mainly repre¬ 
sented, in many of which alteration may have begun. The 
figure, however, may be used as an offset for the undeter¬ 
minable carbon existing in coal, shales, petroleum, etc. As 
regards the limestones, rough estimates of their quantity must 
suffice, and we may provisionally accept that of T. Mellard 
Reade as given in his essay on “ Chemical Denudation in 
Relation to Geological Time.” According to Reade, the ex¬ 
isting bodies of limestone are equivalent to a layer of the 
rock 528 feet thick and completely enveloping the globe. 
This is approximately one per cent, of the crust under con¬ 
sideration, and represents 0*44 per cent, of C0 2 . To this we 
may add the 037 per cent, given above, making 0.81 per 
cent, in all—an estimate which can hardly be too low. 
In the case of sulphur, the figure given, 0‘034 per cent., is 
surely much too low; for sulphur is abundant both in sul¬ 
phates and in sulphides, and iron pyrites especially is widely 
diffused. The absolute proportion of this element seems to 
be hardly determinable. It should be at least 0*05 and prob¬ 
ably not over 0T0 per cent. For chlorine, chromium, barium, 
and strontium the figures are minima, but cannot be very 
largely increased. The value for lithia is probably not far 
out of the way, for that oxide is almost universally present 
in minute traces in the older crystalline rocks, although it is 
rarely estimated in ordinary analyses. 
Now, taking the mean of 880 analyses as cited in column 
H of the table, we may insert in it the additional values so 
far determined, but with certain qualifications. In about 
half of the analyses Ti0 2 , P 2 0 5 , and Cr 2 0 3 were not esti¬ 
mated, but their amounts appear in the figures for silica and 
