STUDY OP IGNEOUS ROCKS. 269 



United States Geological Survey and in the Table of Analyses edited 

 by Justus Roth, and more recently by H. S. Washington. The 

 extremely variable nature of these data and their great abundance 

 present such an exceedingly complex set of numerical relations that 

 their statement, or discussion, requires the aid of diagrams by which 

 the problem may be greatly simplified. 



In addition to the chemical elements noted in ordinary rock 

 analyses there is a much greater number known to occur in rare 

 minerals that crystallise from rock magmas in special instances, or 

 that oftener appear in certain varieties of igneous rocks, such as 

 pegmatites. A consideration of all known pyrogenetic minerals with 

 respect to their chemical composition calls attention to the compounds 

 that are repeatedly formed in igneous magmas by the union of the 

 elements that existed in the magmas before their solidification. And 

 by arranging them in accordance with the oider of their constituent 

 elements in the Mendeleeft" series valuable information as to certain 

 chemical relationsliips among these compounds is at once furnished. 



The substances occurring in igneous rocks are in most cases 

 solids, less often liquids or gases. The solid compounds are always m 

 crystallised condition. Amorphous, glassy, solids that sometimes occur 

 in igneous rocks are seldom, if ever, definite chemical compounds, but 

 are mixtures. The crystallised substances (minerals) are rarely present 

 as uncombined elements, such as gold, graphite, metallic iron. A few 

 are simple compounds with invariable composition, as Si02 (quartz), 

 Ti02 (rutile), AI2O3 (corundum). Most of them are complex, and 

 variable in composition owing to the presence of isomorjjhous mix- 

 tures, as the feldspars, olivine, amphiboles. There are very few 

 examples of pleomoi-phism, such as quartz and tridymite. The 

 apparent difference in the ciystallisation of orthoclase and microcline 

 is probably due to submicroscopic multiple twinning in the apparently 

 more sjonmetrical form. Isomerism of some of the pjTogenetic com- 

 pounds, as (Mg, Ca) SiOs, which is known in laboratory products, is 

 not clearly developed in pyrogenetic minerals. 



Certain isomorphous mineral compounds are not developed in 

 igneous magmas with like frequency, or in certain cases not at all. 

 Such, for example, are the hexagonal compounds, NaAlSiO^ (sodium- 

 nephelite) ; KAlSi04 (kaliophilite) ; LiAlSiO^ (eucryptite). Compare 

 also the potash- lithia- and soda-micas. Other compounds that are 

 analogous chemically and might be expected to crystallise isomor- 

 P'hously in igneous magmas have quite different ciystal symmetry, as 

 li the case with KAl(Si03)3 (leucite) ; NaAl(Si03)2 (jadeite); LiAl 

 (8103)2 (spodumene). 



Various silicate compounds involving different silicic acids; 

 orthosilicic, metasilicic, polysilicic, and in rare instances disilicic, 

 besides uncombined silica, may ciystallise from the same rock magma. 

 And even base-forming elements, as iron and aluminium, may under 

 some conditions separate from rock magixias without combining with 

 silica, which may itself separate as SiOo. That is, hematite, magne- 

 tite, or corundum may ciystallise in the presence of quartz. On the 

 other hand, certain lower silicates do not form when there is sufficient 

 silica to form higher silicates with the same bases. Thus, KAl(Si03)2 

 (leucite) and NaAlSi04 (nephelite) do not occur pyrogenetically withi 

 Si02 (quartz). 



