LATER STAGES OF EVOLUTION OF IGNEOUS ROCKS 51 



The quite exceptional degree of interaction between any two types 

 when one is injected into the other during these movements is 

 due to the fact, to be expected on the present supposition, that the 

 invaded rock, though soHd, was still hot, as Harker has supposed it 

 to be. The marked amount of solution of the peridotite by the 

 gabbro is simply the same reaction that would have happened 

 between the liquid and the individual olivine crystals had they not 

 collected into a body. The remarks made on p. 31 with reference 

 to the eruptibility of peridotite to form separate intrusions of that 

 rock would presumably apply to this natural example. 



The igneous complex of the Nischne-Tagilsk region in the 

 Urals is intrusive into Devonian sandstones, limestones, and basic 

 lavas. As mapped by Wyssotsky' it appears to be, like the Iss 

 complex, a concordant intrusion and suggests a sheethke mass, 

 dipping eastward, with the important platinum-bearing peridotite 

 and pyroxenite near its base and, passing eastward (upward), suc- 

 cessively gabbros, diorites, quartz diorites, and finally granites 

 extending beyond the region mapped. From the present point of 

 view the pyroxenite bodies occurring as borders about the dunites 

 (collected olivine crystals) are huge reaction-rims formed in the 

 same general manner as the borders of pyroxene seen about indi- 

 vidual olivine crystals in a great variety of rocks (see Fig. 0) . 



To the writer it seems that some of these examples, especially 

 the Skye complex, form an excellent transition type from those in 

 which movements have been practically absent and the relation 

 between the various rock-types is transitional, through those in 

 which the transitional relations are destroyed and even the simple 

 gravitative arrangement is on the verge of destruction as in the 

 Skye example, to those in which the evidence for both of these 

 has been obscured and for which the gravitative control of differ- 

 entiation is, therefore, not demonstrable. The Cortlandt series 

 of New York state which presents t3^pes more or less parallel to those 

 described by Harker though richer in varieties may, perhaps, 

 represent the last condition.^ 



^ Mem. Comiie Geol. de la Russie, Nouv. ser., livr. 62, 1913. 



^ G. H. Williams, Am. Jour. Sci. (3), XXXIII (1887), 33; G. S. Rogers, Ann. 

 N.Y. Acad. Sci., XXI (1911), 11-86. 



