MANSFIELD: POST-PLEISTOCENE DRAINAGE. 69 
and there are no well defined shoulders on its sides. Probably the 
greater depth may be accounted for by the fact that Bear Butte Creek 
has a more direct course to its junction with Belle Fourche River and 
delivers its water at a lower level than is the case with Whitewood 
Creek. 
The trunk stream probably had its original course directly outward 
through Spring Creek gap to the plains; but it was diverted north- 
ward at Boulder Park by a stream working in the soft strata of Crook 
valley. In this case, as in that of Bear Butte Creek, the capture 
must have taken place before conditions permitted incision; for 
the ancient stream, which would thus have become superposed on 
the anticline, did not have an opportunity to make any definite ex- 
cavation in it. At this time the lower terrace in Boulder valley was 
cut to a level corresponding with the altitude of the gravels in Crook 
valley. 
A second diversion northward occurred at the west end of Boulder 
valley, two miles below Deadwood. This time the stream was led 
on a somewhat steeper grade through the abandoned saddle, near 
locality 16, across or into the Red valley. Boulder valley was then 
permanently abandoned and its gravels formed a relatively broad 
and flat plain. 
Meanwhile a stream, that had been working along the strike of the 
rocks north of Whitewood Peak, succeeded in undercutting and 
diverting a third time the main stream, which was then flowing north- 
ward. The sharp bend eastward near locality 16 was thus produced 
and Whitewood Creek probably then assumed the course it holds 
today. The history of the creek below this bend is, however, not 
simple. The stream that made the capture was not a single sub- 
sequent stream that by its own unaided endeavor succeeded in working 
back sufficiently to undercut the other. Probably it was composed 
of sections of at least two, and possibly several streams, whose north- 
east-flowing parts coalesced through successive captures. It. seems 
certain that Spiegel’s Gap was cut by a stream that flowed from off 
Crook Mountain dome and was diverted by a subsequent stream, 
working in the soft strata at the base of the escarpment. Likewise 
Sandy Creek, in its earlier history, probably flowed outward into the 
Red valley and was later diverted in a similar manner. There are 
also two or, three streams from Whitewood Peak, whose subsequent 
branches may have taken part in the formation of the stream that made 
the capture, but their relations have not-yet been worked out. 
All of the captures described above appear to have taken place when 
