part 3] Mi^^on intrusions in the island of mull. 251 



The relations borne by the corundum, spinel, and anorthite one 

 to the other indicate clearly that a melt was produced at the 

 surface of the aluminous material from which these three minerals 

 crystallized with more or less perfect equilibrium, but in which 

 sillimanite was underaroino: slow re-solution. The condition in- 

 dicated is one of slow cooling. 



We are here dealing with the crystallization of a quaternary 

 system subject to constant influx of material by diffusion ; but, 

 from the conditions controlling the separation of solid phases in 

 the ternary systems anorthite-forsterite-silica and lime-alumina- 

 silica, we are able to gather reasonable ideas as to the temperature- 

 concentrations that operated in tlie case of the xenoliths. 



Let us consider the probable condition of the magma within its 

 basin. We know from the cognate xenoliths present within the 

 tholeiite, that the maoma had commenced to crystallize before 

 its upward intrusion, and that crystals of bytownite and pyroxene 

 (both rhombic and monoclinic) were separating and being segre- 

 gated under the influence of gravity. The separation of pyroxene 

 and bytownite is what might be expected from a magma of the 

 composition indicated by Analysis II, p. 236. The gabbroic or 

 noritic matter so separated may be considered to have crystallized 

 under conditions of more or less perfect equilibrium, without 

 considerable under-cooling, and thus a temperature of initial crys- 

 tallization between 1200° and 1300° C. is indicated.^ 



We have a useful temperature-indicator in tridymite, the only 

 phase of silica that existed as a metamorphic product, either in the 

 xenoliths or in the Carsaig Sandstone in contact with the intruded 

 sill. The absence of cristobalite and presence of tridymite in both 

 instances would suggest that the temperature of the magma 

 within its basin and at the time of its intrusion was lower than 

 1470° C. 



Except for the presence of tridymite in the siliceous xenoliths, 

 we have little to guide us in an estimation of the initial tempera- 

 ture of the magma within its basin ; but from the strong evidence 

 that fusion of the basin-walls proceeded more rapidly than the 

 CO- solution of the aluminous melt and magma, we may infer that 

 the magmatic temperature was at any rate above that necessary to 

 vitrify aluminous sediments. Such a temperature has been proved 

 by experiment,^ in the case of drj' melts at atmospheric pressure, 

 to be in the neighbourhood of 1500° C. The retention of much 

 water, however, as well as the presence of small quantities of 

 fluxes, would undoubtedly lower the temperature of vitrification 

 considerably, probably below 1400° C. We are, therefore, fairly 

 safe in regarding the metamorphism of the xenolithic material as 

 comniencing at a temperature above, but not far above, that which 

 marked the separation of the cognate xenolithic material from the 



^ G. A. Rankin, 'The Ternary System CaO-iil^O.^-SiO, ' Amer. Jour. 

 Sci. ser. 4, vol. xxxix (1915) p. 1 ; 0. Andersen, ' The System Anorthite- 

 Forsterite-Silica ' ibid. p. 407. 



2 A. H. Cox, Geol. Mag. 1918, p. 59. 



