104 J. P. IDDINGS 



esses of differentiation according to generally accepted theories. — A 

 study of the crystallization of rocks has shown x that different 

 minerals and combinations of minerals may form from magmas 

 chemically alike, according to physical conditions influencing 

 the crystallization, proving that the chemical constituents in the 

 molten magma do not exist, wholly at least, in definite or fixed 

 combinations or molecules corresponding to distinct mineral 

 molecules. Moreover, if the molecular character of a molten 

 magma shortly before it solidifies is not fixed, but is flexible, its 

 molecular character at an earlier period, when differentiation 

 may take place is undoubtedly flexible or unstable, and prob- 

 ably to a greater degree. This is indicated by the variability in 

 the composition of minerals of the same group, such as pyrox- 

 enes, feldspars, etc., in rocks genetically related, and assumed to 

 have been derived by differentiation from a common parent 

 magma. Thus in genetically related rocks the feldspars may 

 range from those high in potash to those high in sodium and to 

 others low in alkali metals and high in calcium ; pyroxenes may 

 range from those rich in calcium and magnesium, with little iron 

 and aluminium, to others richer in the latter elements and to 

 those rich in sodium and iron. Several members of either of 

 these mineral series or groups may occur together in one rock, 

 and may occur in varying proportions in different rocks. 



The facts known regarding the mineral composition of genet- 

 ically related rocks, to which the theory of differentiation has 

 been made to correspond, are that the proportions of the com- 

 ponent minerals, as well as their chemical compositions vary not 

 only with different bodies of rocks, but not infrequently with one 

 rock-body. Furthermore, this variation is in many cases grad- 

 ual, with known transitions, while in other instances differences 

 of composition are marked, and transitional forms have not yet 

 been found. Consequently from the theory of differentiation 



1 Roth, J., Gesteinsanalysen in tabellarischer (jbersicht, etc., Berlin, 1861, p. 21. 



Iddings, J, P., Bull. Phil. Soc. Washington, 1892, Vol. II, p. 217. Also 

 1 2th Ann. Rep. U. S. Geol. Survey, Washington, 1892, p. 659. 



Lindgren, W., A Sodalite-Syenite and other Rocks from Montana, Am. Jour. 

 Sci., Vol. XLV, 1893, p. 297. 



