I9I2] CLARKE— SOME GEOCHEMICAL STATISTICS. 227 



The higher of these figures is equivalent to a shell of rock com- 

 pletely enveloping the earth, 2,100 feet thick, or covering the land 

 area to the depth of 7,198 feet. This estimate is probably a maxi- 

 mum, but it gives fairly well the order of magnitude of the volume 

 to be determined. An exact estimate is of course unattainable, but 

 the total volume of the sedimentary rocks cannot much exceed 80,- 

 000,000 of cubic miles, or roughly_, one fourth the volume of the 

 ocean. If we apportion the volume actually found between the 

 different classes of the sedimentaries their volumes in cubic miles 

 become : 



Shales 62,662,400 



Sandstones 11,749,200 



Limestones 3,916,400 



78,328,000 



This leads to the surprising conclusion that the volume of the lime- 

 stones is less than that of the oceanic salts, or at least is a quantity 

 of the same order of magnitude. If the fundamental igneous rock 

 was more largely basaltic than the average analyses show, the pro- 

 portion of limestone would be increased, but probably not to any 

 very great extent. Some allowance should be made for sodium salts 

 enclosed within the rocks ; but it is easy to show that such a correc- 

 tion must be small. If all the sandstones had an average porosity 

 of 20 per cent., and if its pore space, once saturated with sea water, 

 retained all of its sodium, the total amount retained would be 

 no X 10^^ metric tons. This amount, which is evidently excessive, 

 is only 0.75 per cent, of the sodium in the ocean. Its inclusion in the 

 foregoing computations would raise the volume of rock decomposed 

 to 78,100,000 cubic miles, an increase smaller than the unavoidable 

 uncertainties of the computation. 



The foregoing estimate of the volume of the sediments obviously 

 includes those which cover a great part of the ocean floor, as well as 

 those which are now on land. Some of the latter, indeed, were once 

 oceanic deposits, and are now, by erosion, being partially returned 

 to the sea, either mechanically as salt and sand, or dissolved in the 

 water of rivers. An exact knowledge of the chemical work of river 

 water is therefore of great statistical importance from several points 



