1906] 



IX CHEMICAL GEOLOGY. 27 



It would be foolish to ascribe any high degree of accuracy to 

 these figures, for the data are confessedly of very unequal value. 

 They do show, however, clearly and conclusively, the order of mag- 

 nitudes with which we have to deal. We may claim to know, for 

 example, that oxygen forms about one half of all known terrestrial 

 matter and silicon about one fourth. Next in order of abundance 

 comes aluminum, then iron, and then calcium, followed by mag- 

 nesium, sodium and potassium in nearly equal proportions. So 

 much is established, and we are also able to say that certain other 

 elements appear in minor, but determinable, amounts. In its general 

 drift, the table is satisfactory; but its details are subject to revision. 

 The question now is, what legitimate uses can be made of it ? \Miat 

 problems can it help us to solve? 



My answers to these questions, I fear, can hardly be satisfac- 

 tory. The more closely I scrutinize the uses which have been made 

 of the averages, the more questionable I find them to be. For in- 

 stance, it is possible to compute from the average chemical or min- 

 eralogical composition of the igneous rocks their average physical 

 properties ; and as the component data in such a calculation are 

 usually quantities of similar magnitude, the results obtained will 

 probably be quite near the truth. Mr. W. H. Emmons,^ for ex- 

 ample, has determined, from one of my averages, the average min- 

 eralogical composition to which it corresponds, using for that pur- 

 pose the norms of the new quantitative classification. Then, from 

 the known coefficients of expansion of the minerals, he has calculated 

 the mean coefficient for the igneous crust of the earth. The value 

 found is 0.0000199, and its uncertainty cannot be very large. But, 

 after all, what can be done with the figure? If we use it to discuss 

 the swellings and shrinkings of the lithosphere, we are limited to 

 surface phenomena alone, in whith the disturbances due to cracks 

 and crevices in the rocks are exceedingly large. If we go below 

 the surface, to a point wdiere the rocks are presumably continuous, 

 this particular coefficient of expansion ceases to be applicable, for 

 it has been modified by changes in temperature and by increased 

 pressure. In other words, our constant is only a constant under 



^ In Chaml^rlin and Salisbury's "Geology," vol. i, p. 546. 



