H. J. Loice — Rock Differentiation. 349 



to crystallize out is augite. This would take place earliest in the 

 upper part of the magma, which would presumably be the region first 

 reduced to crystallizing condition. The augite crystals thus generated, 

 having a specific gravity higher than their liquid matrix, would sink 

 therein until the consistency of the matrix prevented further gravi- 

 tational progress. Such movement taking place would render a zone 

 in the upper part of the magma free of augite, and cause some slight 

 upward flow of the more liquid matrix. This augite-free area would 

 contain eutectic material differing somewhat in constitution to that 

 from which the augite arose, and it may be inferred from the circum- 

 stances that the differentiation was such as to superinduce the 

 formation of hornblende during the final stages of consolidation. Such 

 possibilities being granted, it is obvious that less favourable conditions 

 for the formation of hornblende would prevail deeper in the mass, and 

 its mode of occurrence there as well as absence in the lowest parts is 

 consistent with the above hypothesis. 



In seeking for accounts of rocks similar to those under consideration, 

 in order to make comparisons therewith, I have been particularly 

 interested in a description by Messrs. R. Campbell, M.A., B.Sc, and 

 Andrew G. Stenhouse, F.Gr.S., of the "Geology of Inchcolm." ^ On this 

 island there is a sill composed of picrite and teschenites, the latter 

 forming the upper and lower portions of the sill. There, as at 

 Lurcombe, the teschenite varies considerably in respect to the 

 relative quantities of hornblende and augite appearing together, and 

 the varieties are distinguished as augite or hornblende types as either 

 mineral may predominate. The hornblende type preponderates on 

 the upper side of the sill, and the augite type, the hornblende occurring 

 with it, on the lower side. Their mode of occurrence in relation to 

 the picrite is that of ' merging ' as well as that of ' banding.' There 

 is " a well-developed 'rolling over' of hornblende teschenite and 

 augite teschenite along with the picrite — which, in a remarkable 

 way, is suggestive of flow structure," as explained by Mr. Harker 

 for the Skye peridotites, a "result of fluxion in a magma which was 

 heterogeneous at the time of intrusion." - A similar suggestion has- 

 been made to account for the remarkable acid and basic banding in 

 the ' granulitic ' of the Lizard area.^ But there can be no question of 

 the small Lurcombe mass being other than a single flow of a homo- 

 geneous magma differentiated in situ. The teschenites of Inchcolm 

 merge into each other in respect to the relative amounts of augite and 

 hornblende, but there do not appear to be any portions of that rock 

 quite free from either augite or hornblende. 



The analcite of the Inchcolm teschenites is a prominent feature, 

 occurring apparently in larger quantity relative to mass and crystal 

 development than is found in the rock described above. It is con- 

 sidered by the authors to be an original mineral. I have regarded the 

 analcite of the Lurcombe rock, which is quite subordinate in quantity, 

 as a secondary mineral. This has been suggested by its non-appearance 



1 Transactions of Edinburgh Geol. Soc, vol. ix, pt. 2, 1908. 

 ^ Tertiary Igneous Eocks of Skye : Mem. Geol. Survey, p. 75, p. 123, 1904, 

 and op. cit. 



^ Transactions of Eoyal Geol. Soc. of Cornwall, vol. xii, pt. 6, p. 440, 1900. 



