252 DE. H, H. THOMAS ON XEISTOLITHIC [vol. IxXVlii, 



magma, and probably continuing up to the time of the magma's 

 intrusion into higher levels. 



The vitrification of the aluminous ' bank ' was followed almost at 

 once b}^ the separation of corundum and sillimanite from the 

 aluminous melt and, contemporaneously, a narrow diffusion -band 

 of hybrid character was formed between the normal magma and the 

 fused lining of the basin. Across the junction of sediment and 

 m.agma the following zones would appear to have been present : — 



(a) Unfused aluminous sediment ; (b) fused sediment from which silli- 

 manite and some corundum has separated ; (c) a zone of commingling of the 

 aluminous melt with the magma ; and (d) the magma in an unmodified 

 condition. 



It is the hybrid zone (c) that gives to the xenoliths their most 

 remarkable character. The liquid formed by the commingling of 

 aluminous melt and magma was of such a composition that 

 corundum, anorthite, and spinel were, relatively speaking, insoluble, 

 and were early phases of crystallization, and that the sillimanite of 

 zone (&) was unstable and tended to dissolve. 



Corundum and spinel have often crystallized together, and 

 there is evidence of the attainment of eutectic relations between 

 anorthite and spinel. Corundum was in most cases the primary 

 phase, and was followed rapidly by the separation of anorthite and 

 spinel, as the composition of the melt changed with continued 

 crystallization. 



According to the work of Rankin, the temperature at which 

 anorthite, corundum, and sillimanite can co-exist in equilibrium 

 with the melt is represented by a quintuple point on the ternary 

 diagram of the sj^stem CaO-Al203-Si02, and corresponds to a tem- 

 perature of 1512° C. In the xenoliths, however, this condition 

 does not appear to have been attained, for sillimanite was evidently 

 undergoing solution during the formation of the anorthite. The 

 actual conditions of temperature and concentration, therefore, are 

 more probably represented by some point on the boundary-curve 

 between the corundum- and anorthite-fields and away from the 

 quintuple point. The temperatures along this boundary-curve 

 range from 1500° to 1380° C., generally lower than that of the 

 quintuple point, and suggest that the separation of corundum and 

 anorthite took place at a less elevated temperature. 



Confirmation of a lower temperature is also furnished by the 

 mutual relations observed to exist between corundum and spinel. 

 Their obviously contemporaneous and early separation from the 

 melt would indicate some such conditions as those expressed by 

 the boundary-curve between the corundum- and spinel-fields in 

 the diagram for the ternary system anorthite-forsterite-silica, 

 and covering a range of temperatures from 1450° to 1325° C. 

 Again, we have the observed eutectic relation of anorthite to spinel. 

 From the work of Olaf' Andersen Ave learn that with relatively 

 small percentages of magnesia, anorthite and spinel can co-exist in 

 equilibrium with the melt at a temperature as low as 1440° C. 



All the evidence, therefore, so far as it can be co-ordinated, 



