CHEMICAL CHANGES DURING DECOMPOSITION. 517 



erroneous supposition that the alumina is not dissolved at all. Since in 

 some of these rocks a part of the silica was present as quartz and the 

 rocks are not completely decomposed, it is certain that much more than 

 half of the silica of the silicates decomposed in the belt of weathering went 

 into solution and was carried elsewhere by the" underground solutions. 

 The number of instances given is large enough to make it probable that 

 they represent approximately the average fact as to the soluble nature of 

 the freed silica in the belt of weathering. It is therefore no inference, but 

 a fact proved by analyses, that the freed and dissolved silica of the belt 

 of weathering furnishes an enormous and adecmate supply of this compound 

 for the induration of the rocks of the belt of cementation, explained on 

 pages 617-640. 



Another notable thing shown by the table on page 515 is the small 

 range in solubility of the various elements. With some exceptions the 

 usual range for all the important elements except iron is not greater than a 

 ratio of 1 to 2. But in the case of iron the amount dissolved varies greatly 

 and irregularly. Disregarding the wider variations and considering only the 

 thoroughly decomposed rocks in which the variations in the amounts of the 

 other elements are not great, the amount of iron dissolved varies from 14.35 

 to 88.84 per cent, a ratio of more than 1 to 6. Doubtless this great varia- 

 bility in the amount of iron dissolved is dependent upon the fact that the 

 iron occurs in both the ferrous and the ferric forms. The former is much 

 more readily soluble. When the iron is mainly ferrous one would expect 

 that a large proportion would be dissolved; where ferric, a small proportion. 



In conclusion it may be said that, in order to comprehend the amounts 

 of the losses of the various elements in the belt of weathering, it is necessary 

 that the amount of loss of the supposed fixed substances should be known, 

 and the methods of calculation used, although a line of attack giving- approx- 

 imate results fails to convey a correct impression. Doubtless in many 

 cases the supposed fixed element is reduced to one-half or even to one-third 

 of the original amount. If this be true, it is evident that the percentages 

 of losses given fall far below the facts. It is, of course, impossible that more 

 than 100 per cent of any constituent shall be lost; but from this it does not 

 follow that the volume of the residual material may not be an exceedingly 

 small fraction of the original material. For instance, if in a given rock the 

 losses of all the various elements were exactly proportional, the conclusion 

 from the analyses would be that there had been no loss, and yet 0.90 or 0.99 



