300 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1920. 



instead of along a narrow line, is to be attributed to the complica- 

 tions that may be introduced in such correlations by the presence of 

 silica, lime, and possibly aluminum or titanium. 



It may be mentioned here, en passant, that, curiously enough, the 

 same correlation between these two pairs of elements, soda and iron, 

 and potassium and magnesium, seems to hold good in the organic 

 world. 27 This is apparently shown by the following facts : In auto- 

 trophic plant metabolism potash is an essential element, as is also 

 magnesium, in that chlorophyll (which in the leaves acts as the car- 

 bon-transferring substance) is a magnesium salt of a complex or- 

 ganic acid, while sodium and iron are generally toxic toward (at least 

 the higher, gymnospermous and angiospermous) plants. On the 

 other hand, sodium, rather than potassium, is the alkali metal essen- 

 tial to the higher animals, salt being a very necessary article of diet 

 (in part because of its chlorine, and in part because of its sodium, 

 content), and sodium chloride is present in the blood plasma; and at 

 the same time, hemoglobin and its derivatives (which act as oxygen 

 carriers, and are analogous to chlorophyll in plants) are iron salts of 

 organic acids closely related to that of chlorophyll ; while, similarly, 

 potassium and magnesium are more toxic toward the higher animals 

 than are the other pair. 



Let us now pass briefly in review some of the correlations that are 

 shown in igneous rocks by the rarer, and generally petrogenic, ele- 

 ments with the most abundant ones. In the first place, the rocks that 

 are dominantly sodic seem to show the greatest tendency toward 

 the segregation of many of the rarer elements. Thus, lithium, zir- 

 conium, cerium (and some of the other rare earth metals), chlorine 

 and fluorine, and probably glucinum and tin, are found most often, 

 both as components of minerals and in rocks, that are high in soda. 

 Barium seems to be most abundant in those that are high in potash ; 

 titanium, 28 manganese, vanadium, nickel, and cobalt, in those that are 

 specially high in iron ; and chromium and platinum in those that are 

 high in magnesium. Of the proclivities of the more truly metallo- 

 genic elements, as gold, silver, mercury, lead, and zinc, we know little 

 as yet, but further study may indicate such relations, if they exist. 



It is needless to enlarge here on the bearing of such observations 

 on the practical search for ores and metals, especially those of the 

 rarer kinds, some of which are now coming into prominence, such as 

 tungsten and tantalum for electric lights, and zirconium for refrac- 

 tories. It will be self-evident that a knowledge of the rocks of a 

 region can thus give us a clue as to what elements, or their ores, may 

 be most likely met with, so that, for instance, we would not search for 



27 Washington. H. S., Proc. Natl. Acad. Sci., ii, p. 623, 1918. 



* Titanium also evinces preference for sodium, like its congener zirconium. 



