Alexander Scott — The Grawfordjohn Essexite. 515 



the mineral of the essexites, and shows a greater difference between 

 the extinction angles of the inner and outer zones. This points to the 

 former mineral crystallizing during a rapid fall of temperature, 

 since the tendency of mix-crystals to show zonal structure increases 

 with the rate of cooling during crystallization, owing to the shorter 

 time available for readjustment of equilibrium. Hence it is 

 feasible to assume that the 'chilling' set in before the pyroxene 

 began to crystallize. There is, however, very little difference 

 between the olivine of the two rocks, which suggests that this 

 mineral commenced to crystallize before the inception of rapid 

 cooling. While the relatively greater amount of olivine in the 

 chilled rock is doubtless partly due to the rapid cooling ensuing 

 before the olivine had finished crystallizing, and. hence preventing, 

 to some extent, resorption and subsequent reprecipitation as 

 pyroxene, 1 the main factor seems to be the difference in chemical 

 composition, i.e. the greater amounts of magnesia and ferrous oxide 

 and the smaller silica content. 



This difference in composition may be explained in two ways. 

 Firstly, it may be due to the migration of orthosilicate molecules to 

 the cooling margin during the crystallization of the olivine. This 

 migration may have taken place by diffusion, as suggested by Harker, 2 

 or by means of convection -currents, as advocated by Becker 3 and 

 Pirsson. 4 "Washington's view 5 that it is to be attributed to "the 

 force of crystallization ", which is supposed to be capable of acting 

 at an appreciable distance, can be explained in terms of the diffusion 

 hypothesis in the following way. The concentration of a solution 

 in the immediate neighbourhood of a growing crystal will be lowered 

 by the tendency of the molecules of the solute to attach themselves 

 to the crystal. Hence an osmotic-pressure gradient will be set up, 

 and in order to restore equilibrium there will be a transference of 

 molecules by diffusion from the remainder of the solvent. The 

 second explanation of differentiation of this type is due to Bowen, 6 

 and is based on the gravity-separation of the crystals. While the 

 minerals of early formation (in this case olivine) are crystallizing, the 

 action of gravity tends to make them sink. In the chilled margin, 

 however, the rate of cooling is sufficiently rapid to prevent this 

 sinking taking place to the same extent, as it is hindered not only by 

 increasing viscosity but also by the crystallization of the remaining 

 minerals. The second explanation has the advantage over the first in 

 being based on experimental work and being less dependent on 

 hypothesis. In the present case, however, the depth of exposure is 

 too small for the detection of any gravity sinking in the essexite, 

 though it is possible that a more basic layer may exist at a greater 

 depth. It is noteworthy that while the essexite has nearly the same 

 chemical composition as the Brandberget rock, the monchiquite 



1 N. L. Bowen, Amer. Journ. Sci. [4] , xxxviii, pp. 256-8, 1914. 



2 A. Harker, Natural History of Igneous Rocks, 1908, pp. 317-20. 



3 G. F. Becker, Amer. Journ. Sci. [4], iii, pp. 21-8, 1897. 



4 L. V. Pirsson, Bull. U.S. Geol. Surv., No. 237, pp. 187-90, 1905. 



5 H. S. Washington, Bull. Geol. Soc. Amer., xi, pp. 409-10, 1900. 



6 N. L. Bowen, Amer. Journ. Sci. [4] , xxxix, pp. 175-90, 1915. 



