ROCK GORGES OF NORTHWESTERN ILLINOIS. 495 



in depth. But gorge No. 8, on Carroll Creek, has precipitous bluffs which 



are often perpendicular and in places overhanging, so that the bottoni width 



is nearly as great as the top except where the gorge is cut largely in drift; 



it there presents gradual slopes. Hershey's measurements of the gorge at 



Cedarville (No. 4 of the table) indicate that it may be about as abrupt as 



the gorge on Carroll Creek. He considers the Cedarville gorge the best 



illustration found in Stephenson County, and has presented the following 



description of it. 1 



The Cedarville gorge (No. 4 of the table) may be taken as the type, aud a descrip- 

 tion of it will apply, with some modifications, to all the others. Cedar Creek, situated 

 in the central part of Stephenson County, flows in a general southwest direction into 

 Richland Creek. After traversing a broad valley, with gently sloping sides covered 

 with drift aud loess, it suddenly enters a deep, narrow gorge with steep, rock-bound 

 walls. The contraction of the valley from 3,000 to 160 feet is conspicuous and readily 

 attracts attention. The old valley can be traced around by the south, but is partially 

 filled with sand ridges. The sides of the gorge are in some places perpendicular. 

 They generally slope at an angle of about 30 degrees. The bottom is flat and con- 

 sists of abed of dark-brown alluvium, through which the stream meanders, sometimes 

 touching one side and sometimes the other, undermining the walls aud widening the 

 valley. After about three-fifths of a mile the stream enters a small preglacial valley 

 and the gorge widens, but the same canyon-like character prevails to its end. A small 

 tributary occupies a portion of the old valley, and when Cedar Creek again enters 

 this valley the significant fact is learned that a stream one-tenth as large as the main 

 creek flows in a valley ten times as large. 



Hershey has determined that a large part of the erosion of these gorges, 

 estimated to be at least four-fifths, occurred prior to the deposition of the 

 loess associated with the Iowan drift sheet. The principal line of evidence 

 is found in the occurrence of loess within the gorges near their bottoms. 

 In one case near Freeport a gorge was abandoned because of the large 

 amount of loess filling, and the postloessial stream took a new course. The 

 character of the evidence is such that the conclusions drawn by Hershey 

 seem fully sustained. Within the limits of the Iowan drift in northwestern 

 Illinois there are several instances of the initiation of a new course for the 

 stream at the close of the Iowan stage of glaeiation. The channels cut by 

 streams having this date are so much smaller than the rock gorges included 

 in the above table that it is not difficult to separate the two classes of val- 

 leys. In the gorges under discussion the streams not only flow in valleys 

 having bottoms several times as wide as the stream beds, but so far as known 

 they are entirely free from falls or rock rapids. But in the streams which 



' Loc. cit., p. 316. 



