370 ME. H. B. MILXER OX THE [vol. Ixxviil, 



it is not difficult to understand the similarity of composition or 

 their constituent minerals, their predominantly local derivation, 

 the ph^^sical features of the grains (particularly quartz) as con- 

 noting the ahrasiye action of sea-Avater, and the indication of the 

 ■deeper- \vater character of the St. Erth facies when compared Avith 

 that of the other localities. 



Lastly, the sources of origin of both the stanrolite and the 

 Ivyanite as ' foreign ' minerals, and their bearing on the wider 

 problem here discussed, haye to be considered. In the case of 

 the staurolite, the most probable source of derivation would seem 

 to be from the south, from a land-mass, though now submerged, 

 having geological and tectonic connexion with Britann}^ and the 

 North-AVest of France, where we know this mineral to occur in 

 quantity. The researches of Dr. H. H. Thomas^ have shoym that 

 the staurolite in the Bunter deposits of Devon and Somerset 

 probably had this origin, in view of the fact that the Armorican 

 massif is generallj^ considered to have extended much farther north- 

 westward at the close of Upper Palaeozoic times ; although it was 

 afterwards largely submerged, particularly in the late Cretaceous 

 Period, renewed folding and uplift in Miocene times, involving the 

 South of England and the North- West of France, determined 

 the reappearance of this Armorican continent in the area noAV 

 occupied by the English Channel, and the concomitant westward 

 recession of the sea. The effect of this uplift would be to promote 

 drainage both from Southern Eno-land and from the Armorican 

 continent into a broad central river flowing westAvards along the 

 basin of the English Channel, and probably emptying itself into 

 the sea at a point some miles south of the present Cornish coast. 

 The action of marine currents bordering the western shore-line of 

 this uplifted area Avould tend to the lateral distribution of the 

 detritus brought down, although with a strong inclination to a 

 northerly direction, owing to the probable south-westerly on-shore 

 wind prevailing from the open ocean, as is the case at the present 

 time. The comparative rarity of staurolite in the dejDOsits, how- 

 ever, suggests a gradual alteration in jDotency and direction of these 

 marine currents, such as Avould be occasioned by subsidence of 

 the land-areas and the advance of the sea eastwards up the 

 Channel once more ; such factors reflect the initial conditions 

 caused b}'' the early Pliocene submergence to which reference has 

 ah'eady been made. 



In the case of kyanite, the evidence is much more obscure, 

 though derivation from the north-east is suggested on petrogra- 

 phical grounds. Attention has already been called to the marked 

 similarity" between the Pliocene kyanites and those commonly 

 occurring in the Lower Greensand ; in order to test this point 

 further, it became advisable to examine the non-magnetic residues 

 from the LoAver G]'eensand in the vicinity of Seend (Wiltshire), 

 .some 4 miles Avest of Devizes. Accordingly, such an investigatir n 



1 Q. J. G. S. vol. Iviii (1902) p. 630. 



