286 TROF. T. G. BOXN-Er AND luss c. A. RAisix ON [May 1899, 



a deep reddish-brown spiriollid, probably chroniite, in the portions 

 corresponding to the sphortdes. 



In structure, also, the two parts generally show a contrast. 

 The spherules consist largely of short tibros, often in tufts or with a 

 matted arrangement (PI. XXI EI, tigs. 2 & 4) The matrix is formed 

 sometimes of an uniform greenish substance : this in places is inter- 

 rupted by irregular st-rings or clots of iron-oxide, and separated 

 into darker and paler areas. Sometimes this matrix is patchy and 

 streaky : in other examples it shows a network of ferruginous 

 strings, with grannies of iron-oxide, deposited towards the centre of 

 the mesh (PL XXIII, fig. 1). This structure has probably existed in 

 other cases, though it is not now recognizable owing to subsequent 

 disturbance. With crossed uicols the greener areas always give 

 the u>sual reactions of serpentine, varying from parts almost inert 

 to strings with the wonted fairly bright polarization-tints. In 

 ordinary light the fibres of the spherules are colourless, usually 

 acicular and finely pointed ; with crossed nicols they give clear blue- 

 grey tints and extinguish at a small angle. Thus they are almost 

 certainly actinolite.^ 



Only rarely is a concentric structure well marked, but in two or 

 three slices (from east of Pwll-pillo) it appears as broad zones 

 alternately clear and dark, producing a strong resemblance to 

 ordinary spherulites. 



A radial structure is apparent in the spherules with concentric 

 bands, as well as in some others which do not show them ; but in 

 most cases only an external rim, often rather narrow, is made up 

 of radial fibres. In these the outline is circular, but much more 

 commonly it is subrotund, even subangular, not infrequently rhom- 

 boidal with roui^.ded angles. A passage may be traced from the 

 more regular shapes to irregular patches or mere clots. Within 

 the crust or rim, and in the less circular forms, the matted fibies 

 run sometimes mainly along two directions in the plane of the 

 slide, crossing at a fairly definite angle, which (though it is not 

 easy to obtain an exact measurement) often seems to be between 

 50° and 60°. 



A dusty-looking finely granular material, which may be partly a 

 carbonate, sometimes obscures the interior of the spherule. Often 

 this tnclot«es distinct crystals, sometimes rhomboidal in form, with 

 irequent cleavages and giving bright colours with crossed nicols : 

 these are probably dolomite, or possibly ankerite. Ihe boundaries 

 of the rhombohedra are often imperfect, and the acicular crystals 

 are intercrystallized with, and appear to penetrate into, the 

 carbonate. 



A quarry north of Graig-fawr (described on p. 283) produces speci- 

 mens which seem to carry the change a step farther. They vary in 

 character, one being gteyish-green. slightly speckled with a purple 

 tint and streaked with paler, more compact-looking bands ; others 

 are of a brighter, more apple-coloured green, distinctly mottled with 

 purplish- grey, and in these an occasional hint of a spherulitic 



I They were tested mariy times for straight extinction, which might iodioate 

 serpentine or anthophyllite, but it was r.ot obtained with certainty. 



