Vol. 58.] THE RIVER-SYSTEM OF SOUTH WALES. 219 
It is a reasonable assumption, therefore, that the river-system 
of South Wales was initiated upon a slope of Upper 
Cretaceous rocks, by which all features in the Paleozoic strata 
were blanketed over, and that in this may be sought the explanation 
of the complete disregard by the rivers of structures due to the 
Armorican and Charnian movements. The case is comparable to 
that of a part of Dorset, where folded Jurassic rocks form the sub- 
stratum to the Upper Cretaceous. ‘There also the rivers, initiated 
upon a Chalk-plain by post-Cretaceous movements, have maintained 
their courses after the removal of the Cretaceous rocks, and occupy 
valleys which have no connection with the structures in the Jurassic 
strata.’ 
V. ConNECTION WITH THE CoURSE oF THE SEVERN, AND OF THE 
THAMES AND FROME, 
The courses taken by the Severn and the eastward-flowing rivers 
of the South of England were discussed by Sir Andrew Ramsay in 
the paper already quoted. The southward direction assumed by the 
Lower Severn was accounted for, on the supposition that the country 
received a tilt after Miocene times, by which a slope to the north- 
west was produced. The Chalk having been denuded along the 
route which the Severn was thus induced to follow, the Chalk- 
escarpment came into existence. In order to account for the east- 
erly direction taken by the Thames and Frome, he found it necessary 
to suppose a later tilt of the country by which a slope to the east 
was produced. The argument, however, in favour of the tilt to the 
north-west was not regarded as convincing, and the course taken 
by the Severn remained unexplained.” 
The evidence on which the history of the eastward-flowing rivers 
of the South of England may be reconstructed is fairly complete. 
The initiation of the two main lines of drainage, the Kennet-Thames 
and the Frome-Solent, has been discussed elsewhere,’ and it will 
suffice now to recall the fact that they follow the London and 
Hampshire synclines respectively, and collect their tributaries from 
the intervening anticlines. Further, it was shown that these 
synclines and anticlines, while running east and west, are arranged 
in échelon along a line which is parallel to the Chalk-escarpment. 
The Chalk-escarpment now forms the main water-parting through 
miuch of its range across England, as shown on the map (Pl. V). 
There is reason to think that originally the coincidence between 
escarpment and water-parting was closer even than it is now. 
The escarpment, in that part of it which extends from Dorset to 
the borders of Hertfordshire, diverges from the water-parting three 
times, namely, in the Vales of Wardour and Pewsey, and in the 
valley of the Upper Thames. In all these cases, rivers rising in 
* Mem. Geol. Surv. ‘Geology of the Isle of Purbeck’ 1898, p. 233. 
* The theories put forward by Mr. 8, 8. Buckman in Proc. Cotteswold Nat. 
Field Club, vol. xiii (1900) p. 175, following the lead of Prof. W. M. Davis, 
appear to me to transgress the limits of legitimate speculation. 
* Mem. Geol. Surv. ‘Geology of the Isle of Wight’ 2nd ed. (1889) p. 248, 
& Proc. Geol. Assoc. vol, xiv (1896) p. 406. 
