EARTH'S CRUST — WASHINGTON. 279 



A few words may be said of boron, glucinum (beryllium), and 

 scandium, as these enter into a later phase of the subject. The an- 

 alytical difficulties involved in their determination, for the extremely 

 small amounts that are present, are so great that the percentage of 

 these is seldom recorded for any rock. Yet they are all known to be 

 rather widely distributed among the igneous rocks, boron in tourma- 

 line, glucinum in ber}'l, and both in some other rarer minerals, 

 while the widespread occurrence of scandium among igneous rocks, 

 though in very small amounts, has been shown spectroscopically. 8 



THE AVERAGE IGNEOUS ROCK. 



We come now to the consideration of the average chemical compo- 

 sition of the earth's crust — that is, of all igneous rocks. Apparently 

 Dr. F. W. Clarke was the first to undertake this estimation, 9 basing 

 his conclusions largely on the numerous analyses that had been made 

 by the chemists of the United States Geological Survey. Since then 

 he and others, Harker, Mennell, Knopf, Mead, and the writer, have 

 published other estimates, which, it may be said here, do not differ 

 greatly the one from the other. The latest discussion of this subject 

 is to be found in the last edition of Clarke's Data of Geochemistry, 10 

 where numerous references to the literature are given. 



The true estimation of the average chemical composition of the 

 igneous rocks is by no means such a simple matter as it may appear 

 to be at first thought, and, before we deal with it, it will be as well to 

 state very briefly some of the disturbing factors that are involved. 

 The matter will be treated in greater detail in a forthcoming paper 

 by Dr. Clarke and the writer. 



In the first place, we know but little of the exact chemical charac- 

 ters of the igneous rocks of many districts of the earth. This is true 

 of the great continents of Asia and South America, as well as of 

 Africa and Australia, in all of which we have for the most part a 

 knowledge only of the rocks more or less near the coasts and know 

 only in a general and very imperfect way the rocks that constitute 

 the vast expanses of the interior portions. The same ignorance, 

 either total or partial, holds true for many countries, such as China, 

 Arabia, and even Brazil, India, Egypt, and Spain, in which the 

 number of analyses is quite disproportionate to the number and 

 masses of igneous rocks that are known to occur. A most striking 

 example is furnished by the West Indies, where, of the igneous rocks 

 of the otherwise well-known and readily accessible larger islands 



8 Eberhard, C, Sitzb. kg. preuss. Akad. Wiss., 1908, p. 851. 



9 Clarke, F. W., Bull. Phil. Soc. Wash., xi, p. 131, 1889; also U. S. Geol. Survey, Bull. 78, 

 p. 34, 1891. 



"Clarke, F. W., U. S. Geol. Survey, Bull. 695, pp. 24 ff„ 1920. 



