SAND-WAVES IN TIDAL CURRENTS 309 



sandbanks in that part of the Severn below the 

 bridge where the gradient of the channel is steep, 

 all of which had dried out with a smooth surface. 

 The Dun Sands, however, was all in large sand- 

 waves. A series of fifteen had an average length 

 of 37 feet 8| inches, and an average height of 

 I foot 11-22 inches, the length being, therefore, 

 19-49 times as great as the height. The mean 

 difference between the lengths of successive waves 

 was 26-4 per cent, of the average length, and the 

 niean difference between successive heights was 

 37-7 per cent, of the average height. These ridges 

 had somewhat undulating crests and sinuous fronts, 

 and in rriany places pools of water are left in the 

 troughs, where, my boatmen told me, salmon are 

 sometimes impounded. This state of things is no 

 doubt due to a current which sets across this part 

 of the shoal after the higher south-west part un- 

 covers during the ebb. In some places there were 

 two sets of ridges of almost equal size crossing 

 one another. The fact that the variation of height 

 is much greater than the variation in wave-length 

 is probably due to the effect of the cross current. 

 Its action, I think, would be more to scoop out holes 

 in the troughs than to raise peaks on the crests. 



I was on the Dun Sands at noon. At 2 p.m. 

 the water began to rise. I left for the western 



