part 3] MINOR INTETJSIONS IN THE ISLAND OF MULL. 245 



usually crowded with undissolved needles of this mineral (PI, IX, 

 figs. 1-3, 5, & 6) which impart to the crystals a striking flesh-pink 

 colour (p. 2M). Occasionally, they are coloured a sky-blue by 

 finely divided corundum (PI. YIII, fig. 6). In composition they 

 approximate closelj^ to anorthite, and are thus more basic than the 

 glomeroporphyritic felspars (bytownite) of the cognate xenoliths 

 previously described (p. 238). Their formation was often accom- 

 panied by the almost simultaneous separation of corundum and 

 spinel, in a manner that suggests that these three phases were 

 capable of existence in equilibrium one with the other and with the 

 melt at the time of their formation (p. 252). 



The early-formed anorthite has most often undergone consider- 

 able compositional modification as the conditions of equilibrium 

 between xenolith and magma were disturbed. There has frequently 

 been disruption of the coarsely crystalline zone and resorjDtion of 

 the anorthite and its contained sillimanite. This was followed by 

 an additional crystallization on the remaining anorthite of an in- 

 creasingly acid plagioclase, which ranges through labradorite to 

 oligoclase. Such later-formed felspar is free from included silli- 

 manite, and the excess of alumina furnished by its resorption has 

 separated either as corundum (sapphire) or as spinel (PI. VIII, 

 figs. 1, 2, & 5). Corundum, as might be expected, occurs more 

 frequently nearer the source of alumina (bucbite) ; while spinel 

 crystallizes nearer the source of magnesia andiron (magma). The 

 latest crystallization of felspar was of a rapid character, and ante- 

 dated but little the intrusion of the modified xenoliths into their 

 present position. It is usually of oligoclase composition, and may 

 occur either as an outer zone to pre-existing more basic felspars, or 

 as freshly grown individuals of skeletal form which have separated 

 from the residual melt (PI. X, figs. 4 & 5). With regard to the 

 formation and resorption of the earlier felspars, it is interesting 

 to note analogous processes going on in artificial melts. Mr. Gr. Y. 

 Wilson, writing on certain minerals formed in glass-furnaces, calls 

 attention to the fact that felspars entering into the composition of 

 an absorption-zone between the glass-melt and the furnace-brick 

 (aluminous) were full of sillimanite-needles, but that the sillimanite 

 of the melt, as of the buchite, was undergoing resorption in order 

 to supply alumina for the formation of felspar.^ 



Cordierite. — Cordierite appears to result in the xenoliths, either 

 from the complete solution of aluminous material by a tholeiite- 

 magma, whereby a cordierite-buchite ^ or cordierite-sillimanite- 

 buchite is formed ; or from the reaction of the magma with the 



^ ' Notes on the Formation of certain Eoek-forming Minerals in & about 

 Glass-Furnaces ' Trans. Soc. Glass Technology, vol. ii (1918) p. 177. 



2 For a description of cordierite-buchites, see J. S. Flett, ' Geology of 

 Colonsay, Oronsay, & part of the Eoss of Mull' Mem. Geol. Surv. 1911, 

 pp. 94, 95 ; Prof. A. Lacroix, op. jam cit. p. 21, comments on the remarkable 

 sharpness of the crystals and the triple twinning of the cordierite in rocks of 

 this nature ; see also J. J. II. Teall, ' The Natural History of Cordierite & 

 its Associates' Proc. Geol. Assoc, vol. xvi (1899) p. 61, with Bibliography on 

 p. 74. 



Q. J. a. S. No. 311. S 



