part 3] MINOR iN^TRrsio:t^s in the island of mull. 253 



points to a maximum temperature in the neighbourhood of 1400° C. 

 .as that at which the formation of the xenohthic minerals was likely 

 to have taken place. We must remember, however, that all the 

 work that has been done on the various ternary systems has been 

 carried out at atmospheric pressure and under anhydrous conditions. 

 In the case of the xenoliths we are, in the first place, not dealing 

 with pure ternary mixtures, but with melts containing a far greater 

 number of components the presence of which will undoubtedly 

 -depress the freezing-points of the various solid phases. Operating 

 in the same direction will be the large amount of water held in 

 solution at the high pressure obtaining within the magma-basin. 

 It is fairly certain, therefore, that the probable temperatures of 

 rseparation of the various crystalline phases that occur in the xeno- 

 liths, were lower than those judged to have existed in analogous 

 but not identical anhj^drous melts. How much lower the present 

 •state of our knowledge leaves an open question. 



So soon as crystallization commenced in the h^^-brid zone (c) we 

 must assume that transfusion of the components entering into the 

 •composition of the separating phases began, and was continued as 

 fast, and for as long a period as the viscosity of the melt and 

 magma would allow. That this diffusion accompanying crystal- 

 lization was a factor of considerable importance will be seen at 

 once on comparing the analj^ses of tholeiite (II), hybrid zone (III), 

 and buchite (IV), given on p. 236. From these it is obvious 

 that the ultimate composition of the hj^brid zone could not be 

 reached b}- simple mixture of fused aluminous sediment with 

 tholeiite-magma, but only by the selective diffusion of their respec- 

 tive component oxides. While Analysis III (p. 236) repiesents a 

 particular phase of interaction between tholeiite and fused 

 aluminous sediment, the reader is warned against regarding it 

 as expressing the composition of a simj^le mixture. The normal 

 crystallized zone consists of anorthite, corundum, and spinel, and 

 clearly indicates diffusion of alumina from the sediment and of 

 lime, and to a less extent of ferrous oxide and magnesia, from 

 the tholeiite. There has also been a relative concentration of 

 alkalies. 



Anorthite being an early and dominant phase to separate from 

 the melt represented by the hybrid zone, it follows that the magma 

 in the immediate neighbourhood would be impoverished as regards 

 lime and relatively enriched as regards magnesia, iron, and alkalies. 

 The slower transfusion of magnesia makes itself evident in the 

 precipitation of spinel, more particularly in that portion of the 

 hybrid zone which lies nearest to the magma. The extent of such 

 an impoverished zone would depend on the viscosity of the melt, 

 and thereby the extent to which diffusion could operate. We have 

 every reason to believe that this zone was not wide, and that, for 

 all practical purposes, we may regard the bulk of the magma within 

 the basin as unmodified by loss of bases or by the assimilation of 

 •sedimentary material. 



