TYLOR QTJATERNAEY GRAVELS. 71 



Jime, to show that the work of denudation performed by the main 

 stream flowing at the bottom of a deep valley becomes insignificant 

 compared with that of the side-streams flowing into the main river. 



The river itself only operates on a narrow slip of ground at the 

 bottom of the valley, although, by the momentum it possesses and 

 its constant action, it has a powerful denuding force. The side- 

 streams flowing into the river, on the contrary, although more inter- 

 mittent, act on the whole of the remaining surface of the valley, 

 either through themselves or their smaller tributaries. They thus 

 denude an area twenty times as great (or even more) as the river 

 itself can touch. The side-streams are less fixed in position, and 

 run with many times the speed of the river itself. Their power as 

 an agent of denudation is not in proportion to the average amount 

 of water which they contain, but to their speed in flood-times. 



The side-streams occupy channels of infinitely greater length than 

 that of the main river. The points where the greatest force of water 

 is employed now at the bottom of the Taff valley are where the water 

 in the side-streams and the main stream unite, as at the north-east 

 corner of the horseshoe bend at Quaker's Yard at H. 



At the south-eastern corner is a large quarry close to the road, in 

 which 60 feet of massive-bedded sandstone is seen. It is also seen 

 in the river-bed. 



As the rails near the viaduct are 96 feet above the level of the 

 river, and as rock is seen 120 feet above the rails, and 96 feet below 

 them, we have at least 216 feet of soKd sand-rock exposed in these 

 sections, without any intermixture of clay or shale. This is an im- 

 portant fact, as we may observe that a horseshoe bend of as per- 

 fect a form as in alluvial soil may be excavated in solid rock. Mr. 

 Fergusson has attempted to prove that the sizes of the bends and 

 curves in rivers depend upon the velocity of the river, and that a 

 curve in a river falling 10 feet in a mile will be smaller than one in 

 a river falling o feet in a mile ; but at Quaker's Yard we have a 

 well-formed large horseshoe bend in a river falling 50 feet in a 

 mile. The laws which regulate the flow of rivers may be different 

 in sand -rocks from what they are in alluvial soil ; but I stiU hold 

 the view as to the cause of these curves which I mentioned to the 

 Society in 1852, but which it is unnecessary to discuss at the present 

 moment *. 



At A, Plate lY. fig. 1, the gravels are exhibited very clearly in the 

 cuttings at and near Quaker's- Yard Station ; these are 200 yards long 

 and 40 feet deep, and are represented at the north end of the section. 

 The local sand-rock furnishes the largest, most numerous, and least- 

 rolled pieces, many of them 8 and 9 feet long. The boulders of Mill- 

 stone-grit are generally well rolled, and not more than 6 feet long ; 

 those of Old Red Sandstone are throughout well roUed, and not often 

 more than 2 feet. JS^ot more than one stone out of 500 is derived from 

 the Old Red Sandstone or Millstone-grit. 



The gravel is supposed to be from 80 to 40 feet thick at A, covering 

 the side of the hiU. to a height of 200 feet or more above the river. 

 * See Phil. Mag. 4th series, vol. v. p. 275, 1852. 



