Composition of Igneous Rocks. 411 



cause, we would expect to find the more acid facies of the 

 eruptive in contact with the more acid country rock and the 

 more basic facies in contact with the more basic country rock. 

 These conditions are not always fulfilled — often quite the 

 reverse. We are therefore justified in concluding that the 

 absorption of inclusions of country rock plays only a very 

 subordinate part in producing the variations observed in erup- 

 tive rocks. 



Quite recently Johnston-La vis- has drawn attention to the 

 part played by osmotic force in the production of variations in 

 igneous rocks. Suppose a laccolitic eruptive to be quite homo- 

 geneous at the time of its intrusion, then, according to this 

 writer, there would be a gradual interchange between the 

 magma and the country rock — material passing by osmotic 

 force from the magma to the country rock and vice versa. If 

 the magma were acid and the country rock basic the former 

 would become basic toward the contact and the country rock 

 would become more acid. The laccolite would be basic on the 

 borders and would become gradually more acid toward the 

 center, where it would possess the composition of the original 

 magma. True, this would seem to account for the variation 

 within the eruptive, but what must we conclnde with regard to 

 the heterogeneity caused in the country ruck \ We would 

 expect the metamorphosed country rock to resemble (chemi- 

 cally) the eruptive near the contact, while the less metamor- 

 phosed rocks would be less changed chemically. This, however, 

 is at variance with the conclusions based on analyses of series 

 of rocks from contact areas. Several series of analyses of con- 

 tact products agree in demonstrating that no important chem- 

 ical change accompanies such contact metamorphism. 



Homogeneous salt solutions become heterogeneous if slight 

 difference of temperature be maintained in different parts of 

 the solution. A solution placed in a long horizontal tube will 

 become slightly concentrated in the cooler end of the tube. 

 This principle of Soret's was first used to explain differentia- 

 tion in rock magmas by Teall.f Supposing the magma to be 

 an homogeneous mass at the time of intrusion and regarding it 

 as a solution, we would expect that the " dissolved" would con- 

 centrate in the cooler parts adjacent to the country rock. 

 Since the basic borders are characterized by the bases CaO, 

 MgO and FeO, they or their silicates are regarded as the " dis- 

 solved." More recently Harker^: has criticised this theory of 

 differentiation and concludes that it can account for only slight 



* Causes of variation in the composition of igneous rocks, Natural Science, 

 1894. 

 f British Petrography, 1888, p. 404. 

 % On the Gabbro of Carrock Fell, Quart. Journal Geol. Soc, 1894. 



