518 J'EOF. T. a. BONNET ON THE SO-CALLED BIOEITE 



about 10-9. Thus there is 4-5 at least to spare, more than is 

 wanted for a mica. Hence in the one case we may have, for 

 instance, a rock composed of more than one variety of hornblende 

 (pargasite, arfvedsonite &c.), a potash-iron mica, olivine or enstatite, 

 and a little felspar ; in the other case, if non-aluminous and non- 

 alkaline hornblende were formed, we might have more than half of 

 the rock composed of felspar *. Thus although there is probably 

 some variation of chemical composition, the great difference in the 

 component minerals may be quite as much due in some cases to a 

 difference of circumstances. 



I have carefully revised all my collection of slides in the hope of 

 discovering characteristics which might help in the identification of 

 the boulders, but, as might be expected in the case of rocks which 

 exhibit so much variability (due no doubt in part to their coarseness), 

 I have not met with much success. 



I do not find any distinct traces of enstatite in slides cut from 

 boulders from Pen-y-Camisiog, Peny-y-cnwc, the gate-post, or the 

 road-side near Ty Croes, but in each, a colourless augite is more or 

 less distinct. This mineral is not present in the boulder on the 

 shore at Porthnobla, but it contains a mineral like enstatite. The 

 Pengorphwysfa rock exhibits the augite and some enstatite ; that 

 from Caemawr, augite and perhaps enstatite. In the Little-Knott 

 slides I find enstatite, but have not certainly identified augite ; here, 

 however, it is generally not difiicult to detect some felspar. This 

 accords with the result of the partial analyses quoted in my last 

 paper, so that it is a less typical picrite than the rock of the Welsh 

 boulders. Thus the lithological evidence rather favours the deriva- 

 tion of the Anglesey boulders from dykes in that island, and this 

 appears from the evidence supplied by Prof. Hughes to be most 

 probably correct. 



The specimens sent to me from Caemawr and Pengorphwysfa are, 

 indeed, not so coarsely crystalline as some of the boulders. This 

 predominance of coarse varieties in the boulders (^true also of the 

 Little-Knott rock) may be an instance of a survival of the fittest ; for 

 I noticed at Schriesheim (and I have seen it elsewhere in doleritic 

 rocks) that the most coarsely crystalline parts had a nodular habit, 

 and were often well preserved, when the more fine-grained parts 

 immediately around them were decomposed t. 



Hornhlende-Picrite Boulder near St. Davids. 



l^ot long after the reading of my last paper Dr. H. Hicks informed 

 me that he had recently discovered a boulder of hornblende-picrite in 



* If the alkaline percentage were 2'5 (and the joint amount is often quite 

 that), then, assuming a felspar of the composition of labradorite, 55*5 of the 

 rock would be felspar. 



t But the Henslow collection contains specimens, labelled " N. of Llanerchy- 

 medd," as coarse as any boulder which I have found. Whether these come 

 from rock in situ or not, they bring this variety very close to Caemawr. Sir. 

 Harker also has sent me specimens of boulders from near Llanerchymedd quite 

 typical of those which occur in the district south-west of Ty Croes. 



