ABSTRACTION OF OXYGEN BY IRON 321 



to get results that might represent the average composition of the 

 sedimentary rocks, and there can be no doubt that the figures give a 

 much closer approximation to the truth than would be afforded by 

 a larger number of analyses taken at random. There is the further 

 advantage that the analytical work is thoroughly reliable — a matter 

 of first importance, and particularly so when it is a question of ferrous 

 and ferric iron. 



As in the analyses of limestones, the iron is all given as ferric oxide, 

 the figures are not available for the present inquiry, but the iron 

 content of these rocks is too small to be of great moment. An effort 

 has been made to get from other sources evidence as to whether 

 the correction for limestone would be plus or minus, but as most 

 analyses combine the ferrous and ferric iron, and often the alum- 

 ina as well, no definite conclusion was reached. 



Not only does the small amount of iron in limestones affect the 

 magnitude of this correction, but it is further reduced by the fact 

 that the limestones are of minor moment in the mass of sediments. 

 Gilbert 1 estimates the limestones as making up one-fifth of the total 

 thickness of the sedimentary rocks. Reade's 2 estimate is one-tenth. 

 Therefore, in view of the small percentage of iron in the limestones 

 and their limited amount, they may be neglected without seriously 

 affecting the results. 



A more important question is that of the relative masses of shales 

 and sandstones, and the consequent values to be attached to the figures 

 for these rocks. No entirely satisfactory data are at hand to deter- 

 mine this point, but Gilbert's 3 estimate of equal parts of shale and 

 •sandstone is here used as being the best available. 



As pointed out below, a change of this ratio would affect the magni- 

 tude, but not the sign, of the results. In the following estimates, 

 therefore, the figures for shale and sandstone are given equal weight. 



According to Clarke's estimate, the older crust of the earth con- 

 tains 2.63 per cent. Fe 2 3 and 3.52 per cent. FeO. The analyses of 

 Gilbert's samples, by Stokes, give for shales 4 4.03 per cent. Fe 2 3 

 and 2.46 per cent. FeO. For sandstones 5 the figures are 1.24 per 



1 American Geologist, March, 1894, p. 214. 



2 Chemical Denudation in Relation to Geological Time, 1879, p. 53. 



2 Loc. cit. 4 Op. cit., p. 17, column C. 5 Ibid., column F. 



