THE ORIGIN OF IGNEOUS ROCKS. 
193 
was a consequence of the formation of minerals in a definite 
order. Diffusion, according to Soret’s principle, and the sub¬ 
sidence of crystals may assist in producing heterogeneity in 
the original magma. 
It now remains to be pointed out in what manner and to 
how great an extent the researches and observations of the 
present writer tend to modify and, it is hoped, to advance 
the theories just reviewed. 
The more detailed investigation of large areas of volcanic 
rocks fails to establish the individuality of the four princi¬ 
pal groups suggested by King, and proves that all of the 
volcanic rocks of one region belong to a single group. In 
such a group there are gradations in composition from inter¬ 
mediate to extreme varieties, and rhyolite and basalt repre¬ 
sent the extremes in many regions. This conclusion has 
also been reached by Hague in his work on the Eureka Dis¬ 
trict. 
The study of greatly eroded accumulations of volcanic 
rocks establishes the correctness of the conception definitely 
stated by Dana, that the crystalline character of all igneous 
rocks depends solely on the physical conditions under which 
molten magmas have consolidated, and therefore all of the 
igneous rocks which belong to the same period of volcanic 
activity, whether extrusive or intrusive, constitute a natural 
group. 
The consanguinity of the various rocks of such a natural 
group is shown by their mineralogical, chemcial, and geo¬ 
logical relationships; hence, instead of four separate sources 
for the volcanic rocks of a given district, we are led to the 
consideration of one common source for all of the igneous 
rocks belonging to an era of volcanic action in one region. 
The fact of a differentiation of some common magma, and 
not the compounding of two or more different magmas, is 
established by the geological occurrence and order of erup¬ 
tion of the various members of a natural group of igneous 
rocks, and also by their petrological and chemical character¬ 
istics ; but an investigation of the chemical variability of the 
28—Bull. Phil. Soc„ Wash., Vol. 12. 
