﻿478 DR. WHITMAN CROSS ON THE NATURAL [Aug. I9IO, 



made by Becke the type of the ' Atlantische Sippe/ for granites, 

 gneisses, gabbros, minettes. kersantites, and other igneous rocks of 

 the adjacent Erzgebirge are certainly of the contrasting kindred. 

 Systematic classification is, however, not for Tertiary 

 rocks alone. 



Petrographic provinces. — The interesting and significant 

 ' blood-relationship ' shown by the rocks of certain cycles of eruption 

 in many petrographic provinces, or comagmatic regions, is clearly 

 not adapted to systematic purposes. Many t}'pes are common to 

 various provinces, and the peculiar characters of a province are 

 seldom, if ever, so fundamental as the properties uniting the rocks 

 of different provinces. The type of consanguinity may be marked 

 only for the products of an epoch of eruption, rather than for all 

 rocks of a province, as seems to be the case of the East African 

 region above mentioned. 



III. Factors op Magmatic Differentiation. 



General considerations. — I suppose that all present-day petro- 

 graphers recognize magmatic differentiation as a general process by 

 which many of the observed chemical differences in igneous rocks 

 must be explained. In numberless volcanic centres there is evidence 

 of a sequence of lavas, by no means uniform, but resulting in final 

 complementary products of rhyolite and basalt. In certain masses, 

 like the Shonkin Sag laccolith, made classic by Pirsson, the differ- 

 entiation has occurred after intrusion, and we may see the results ; 

 in many localities a large central mass is attended by numerous 

 smaller bodies, the clanship of which and their derivation from the 

 main magma are demonstrable. 



The recognition of magmatic differentiation as a natural phe- 

 nomenon involves no acceptance of theories of explanation. It is 

 aiready clear that effects comparable in kind are produced by the 

 operation of various processes. When the complexity of the 

 magmatic solution and the multiplicity of conditions to which it may 

 be subjected in the eruptive history of a single centre are borne in 

 mind, it appears certain that we can as yet have but a faint con- 

 ception of the true character of what we include under the term 

 magmatic differentiation. 



The recognition of this vague but undeniable natural phenomenon 

 has furnished to some petrographers an all too alluring opportunity 

 for speculation and assumption, and on the other hand it has 

 suggested a most difficult and extensive field of physico-chemical 

 research. The genesis of igneous rocks is manifestly destined to be 

 an important department of petrology and a fundamental element 

 of the general science of geology. We are here concerned, however, 

 with the special question whether the systematic classification of 

 igneous rocks can be made natural, in one of the primary steps, by 

 the use of a genetic factor connected with differentiation. 



The now general beliefs that magmas are essentially mixed 

 solutions, and that most igneous rocks represent differentiation- 

 products, render the conception logical that all igneous rocks may 



