76 G. W. Tyrrell— Alkaline Igneous Rocks, West Scotland. 



poikilitically enclosing both olivine and aiigite. It is a red-brown 

 variety belonging to barkevicite. A very little ilmenite (3 per cent) is 

 enclosed in the tiornblende, but the rock is remarkable for the scarcity 

 of iron-ore. A few flakes of red biotite, of later crystallization than 

 the augite, are seen occasionally. There is no apatite, as that mineral 

 always seems to go with abundant analcite. This rock differs from the 

 Inchcolra type ' in the absence of felspar, the abundance of liornblende, 

 and the granular habit of the augite. It may be referred to as the 

 Lugar type. The rock is so rich in olivine that it should perhaps be 

 called a peridotite. Towards the top of the ultra-basic mass, however, 

 the proportion of olivine lessens, and a considerable quantity of plagio- 

 clase and analcite comes in. The percentage mineral composition 

 of this rock is shown in column VII of Table III (p. 77). There is 

 also a variety in which the augite totals over 50 per cent of the mass 

 (column VI, Table III). These rocks are true picrites, comparatively 

 rich in the bisilicate constituents, and containing some felspar. 



It is not intended here to discuss the cause of the remarkable 

 differentiation which has given rise to the varied assemblage of 

 rock-types within the Lugar complex. Briefly the writer believes 

 that it is due to a stratification of the magma under the influence 

 of gravit}' — a separation of heavier from the lighter and more volatile 

 constituents — which began before crystallization commenced and 

 continued for some time after. This stratification took place within 

 a shell of viscous or solid material slowly growing out from both 

 contacts, and was rendered possible by the abundance of a water-rich 

 mother-li(juor which subsequently crystallized as analcite. The 

 presence of this mother-liquor is believed to have maintained the 

 magma in a state of great liquidity for a long period of time, and thus 

 facilitated differentiation under the influence of gravity. Numerous 

 facts and observations, for the full statement of which a later detailed 

 paper must be consulted, give support to this hypothesis. 



Table III shows the percentage mineral composition of the rock- 

 types appearing in the complex, arranged as a series from the top 

 to the bottom of the sill. These have been obtained by the Rosiwal 

 method, to which these fresh, medium-grained rocks lend themselves 

 admirably. 



A. picrite-teschenite sill, very similar to that at Lugar, occurs on 

 the Ayrshire coast at Castle Craigs, Ardrossan. It has been described 

 by Falconer,^ who, whilst noting the abundance of analcite, has 

 identified the rocks as hornblende-dolerite and picrite. Close to the 

 upper contact the rock is a nepheline-tesch^nite of the Cathcart type, 

 containing abundant red hornblende, which is indeed a characteristic 

 mineral of the whole suite. Nepheline is also fairly abundant. This 

 rock passes downwards into a teschenite of the Glasgow type, but 

 with more hornblende than the normal variety. According to 

 Falconer the picrite, which belongs to the Inchcolm type, occupies 

 more than half the thickness of the sill. He describes the underlying 

 teschenite as more nearly allied to the picrite than the overlying rock. 



' Campbell & Stenbouse, Trans. Edinb. Gaol. Soc, vol. ix, pt. ii, pp. 121-34, 

 1907. 



2 Trans. Roy. Soc. Edinb., vol. xlv, pt. ill, pp. 601-11, 1907. 



