THE TERRACES OF THE CONNECTICUT. 731 



wliole length of the State. We have, beginning back at the mountain, tlie 

 broad stretch of the liigli hike fiats (1 f), sinlving into a more limited area of 

 lake bottom (1 b t), and cut into this is a series of hitcr terraces, four in 

 number (t*t^), much broader than the con-esponding ones north of the 

 Westtield River, and combining with the terraces of this latter stream to form 

 a most beautiful succession of broad meadows, bounded back from tlie river 

 by sharp slopes, which swing in great curves — representing forn:ier curves 

 of the stream — up which one mounts to reach higher terrace flats as well 

 characterized as those below. 



THE INCOMPLETE TERRACES AS lEEUSTRATIONS OF THE STAGES IN 

 THE GROWTH OF TERRACES. 



All up and down the river broad sand flats may be seen extending out 

 into the stream at a level but little above low water and on the concave side 

 of bends, as north of the knee of the Hadley bend and at the first concavity 

 beloAv the Northampton bridge. 



Generally only one bank of the river is wooded at a given section of 

 the stream. Going up or down stream, one comes to a stretch where the 

 growth ceases and is replaced by a caving bank, beyond which the bushes 

 begin again. From the bushy banks the shallows extend far out, and the 

 conditions are favorable for the formation of islands. Against the cavino- 

 bank lies the thread of the stream. Each set of these sand flats and shallows 

 is connected diagonally across the stream with a corresponding set on the 

 other side, and at low water a series of disconnected deep water-pockets lies 

 in the line of the thread of the current, alternating against the right and left 

 banks of the stream, and so much of the water seeps through the sands of 

 the shallows between the pockets that the bed is not scoured out at all 

 between these long, curved deep-water stretches. 



It was a i-emarkable and interesting discovery of Gen. Theo. G. Ellis, 

 of Hartford,^ that at high water a large portion or the Avhole of this system 

 of bars is scoured out, and on the recession of the flood is re])laced exactly 

 in its old place and with its old dimensions, as a curtain held up by the 

 wind sinks to its old place as the wind falls. This is true, of course, in so 

 far as the banks of the stream above and below are unchanged, for these 

 bars are the mechanical solution of a complex equation in hydraulics and 

 change with any change of the factors. 



'Survey of Connecticut River: Ex. Doc. 101, Forty-fifth Congress, second session. 



