220 



CLARKE— SOME GEOCHEMICAL STATISTICS. [April 20, 



more truly the general magma. They would be balanced, or more 

 than balanced, by other areas of granitoid or dioritic character. 



The tables so far given serve only to show the probable uni- 

 formity of the accessible magmas. In order to complete them, a 

 number of minor constitutents of igneous rocks must be considered ; 

 which, however, have only been determined with adequate fre- 

 quency in the laboratory of the United States Geological Survey. 

 For example, in round numbers, 1,100 nominally complete analyses 

 have here been made, and in 736 of them barium oxide was deter- 

 mined or proved to be absent. The mean of these determinations, 

 counting absences as zero, is 0.100 per cent., which may be a maxi- 

 mum. If the remaining 364 rocks were all free from barium, and 

 so regarded, the mean percentage of BaO would be 0.067, a mini- 

 mum. Between these two figures the most probable value would lie, 

 but nearer the maximum than the minimum, the mean being 0.084. 

 Upon this basis of computation the following table of percentages 

 has been constructed. 



BaO. 



SrO. . 



Li=0. 



NiO. 



Cr203 



V2O3. 



S 



F. . . 

 CI... 

 Zr02. 

 CO2 . 



Minimum. 



0.067 

 0.022 

 0.005 

 0.007 

 0.012 

 0.002 

 0.076 

 O.OIO 

 0.015 

 0.007 

 0.330 



Mean. 



0.084 

 0.031 

 0.008 

 0.017 

 0.031 

 0.013 

 0.094 

 0.055 

 0.038 

 0.016 

 0.410 



It is evident that these figures are not of equal significance. 

 Some of them rest upon too few determinations, especially those 

 for fluorine and vanadium. The experience of the Survey labora- 

 tory, however, leads me to believe that vanadium is very widely 

 diffused in the igneous rocks, and that the mean assigned to it may 

 possibly be exceeded. Chromium, nickel and zirconium are also 

 more abundant than they were formerly thought to be. As for 

 carbon dioxide, the figure given is probably much too high, for its 

 presence in igneous rocks is mainly but not entirely ascribable to 



