NEIQHBOTJEHOOD OF SAEN, CAERNARVONSHIRE. 447 



matter of mere conjecture ; but the relation of the rock to subsequent 

 earth-movements, of probably pre -Llandovery times, renders it likely 

 that the granite belongs to the Bala age. It is impossible to say 

 whether this plutonic mass was connected with volcanic centres ; 

 acidic lavas and ashes are found, however, less than five miles to 

 the east, and are associated with Bala strata. 



III. The Gabbro, Diorite, and Gneissic Diorite. 



These rocks, grouped together for reasons which will appear, 

 have not been noticed before. They occur only in two small patches, 

 and are quarried in both places, viz. in the little hill of Craig-y-fael, 

 two miles south-west of Sarn, and on the banks of the stream at 

 Plas Llangwnadl. 



The Craig-y-fael rock is, to the eye, a medium- or rather coarse- 

 grained rock of granitic habit, apparently either a diorite or a 

 gabbro. A similar or identical type is exposed above Plas Llang- 

 wnadl ; but going north we find the rock assuming a more or less 

 gneissic and even almost a schistose appearance, becoming, as it 

 seems, a hornblendic gneiss. The microscope explains clearly the 

 relations of these various types. 



The typical rock of Craig-y-fael may be described as a partially 

 amphibolized gabbro. It consists mainly of an aggregate of some- 

 what altered felspars and grass-green grains of hornblende and 

 augite, none of the constituents ever showing idiomorphic contours. 

 The minerals seen in the thin sections are augite and diallage, horn- 

 blende and actinolite, felspar, and a black opaque mineral, pre- 

 sumably an iron-ore, with some pale green decomposition-products, 

 which mostly present the characters of chloritoid. 



The black mineral forms shapeless patches without any indication 

 of leucoxenic alteration, and is probably magnetite. The felspar is 

 a plagioclase, showing the usual albite-twinning, sometimes crossed 

 by twin lamellse on the pericline-law ; the extinction-angles are 

 those proper to labradorite. The crystalline plates are much strained 

 and bent, so that the lamellae are curved ; it is possible, though not 

 clear from the specimens examined, that the twinning may be in 

 part induced by the strain. 



The augite is in long plates of a very pale greenish tint. It 

 shows sometimes the ordinary augite-cleavages, but other plates 

 have a very marked diallagic structure. In spite of decomposition, 

 it can be seen that both augite and diallage are converted at the 

 margins into hornblende. The hornblende is of a dull green colour, 

 and the pleochroism is expressed by : — a, very pale to almost colour- 

 less ; /3, grass-green, rather pale ; 7, a rather deeper and bluish 

 green. The usual prismatic cleavage-traces are seen in the more 

 compact portions ; but a part of the amphibole is in granular masses 

 or in blade-like imperfect crystals with actinolitic structure. 



There appears to be no augite or diallage in the Llangwnadl 

 rocks, but their essential identity with the amphibolized gabbro of 

 Craig-y-fael can scarcely be doubted after an examination of the 



Q.J.G.S. No. 175. 2h 



