Geological Papers. 217 



the preglacial channel have some northern pebbles mingled with 

 the upper portion of the prevalent chert pebbles from the west. 

 Eventually the ice reached the Kansas valley and soon filled it 

 from near Lake View to Wamego. West of the latter point was 

 formed a large lake, first recognized by Robert Hay twenty 

 years ago, and named later by Mr. Smythe "Kaw lake." There 

 were probably several other smaller lakes in various tributaries of 

 the river, one of them being, as we have suggested, in the valley of 

 Rook creek. It should be remembered that the whole western 

 edge of the ice sheet from Montana south must have had its drain- 

 age turned first into the Kansas, and at the stage now considered, 

 diverted over into and along the present valley of the Wakarusa. 

 The result was that one after another the valleys became lakes 

 and overflowed the divides, at first by very winding, irregular and 

 shallow courses, scattering boulders, in some cases pushed along 

 by the current of the stream, but frequently by river ice or by 

 blocks from the glacier itself. To such conditions we are disposed 

 to refer the highest bouldery patches south of the Middle Waka- 

 rusa. Soon the water began flowing over the Oread limestone into 

 the easternmost valley and began to form the bouldery channel of 

 which the Three Sisters are the more notable remnant. For a 

 time, doubtless, quite a cataract existed a little west, as the Law- 

 rence shales were out out by the falling water, and the Oread was 

 cut through rapidly. One efiPect of such a situation would be that 

 the stream would frequently shift because the shale would cut 

 away more rapidly than the bouldery bed of the stream. The 

 course was cut back over the old channel south of Clinton, as al- 

 ready indicated. 



The front of the ice sheet was meanwhile looking over the divide 

 at several points, and streams past Stull, Pauline and Auburn were 

 at their prime and rapidly eroding, for the slope was steep for such 

 large streams. It might perhaps be supposed that most of the 

 peripheral stream would have come over at Dover, but judging from 

 the deposits, the Pauline branch was the largest and longest occu- 

 pied. Before the ice had receded much, the streams had cut down 

 to the level of this lower bouldery terrace. At that time the Waka- 

 rusa still flowed south of Clinton, and Deer creek did not join it 

 till several miles further east. The latter stream, not having to cut 

 down through the Oread so deeply — perhaps because it had taken 

 advantage of the previous erosion of the preglacial stream, or, more 

 likely, from some other advantage of direction of joints or softness 

 of rock, or possibly because of the northern dip of the rocks or the 



