GLACIATION IN THE CRAZY MOUNTAINS 565 



to have reat-bed a late mature stage, while the later one was somewhat less 

 advauced. The present deep canyons were probably exc-avated aud broadened 

 and the lateral spurs cut back during these periods of erosion. The evidence 

 of these two later cycles is less well marked within the mountains than in the 

 bench land outside. In Sweetgrass canyon, however, just inside the contact 

 zone, the rocks are very hard and the canyon boxes. The walls at this point 

 are divided into four minor slopes, with noticeable breaks in grade between 

 successive portions. The uppermost set of slopes indicates a relatively broad 

 and open valley with gently sloping concave sides. The two next lower sets 

 are smoothed and somewhat concaved. The lowest slopes descend steeply to 

 the present stream bed. The uppermost set of the slopes here indicated ap- 

 pears to represent an earlier, broadly opened valley of the ancient topography, 

 while the three lower sets represent respectively the two later erosion stages 

 aud the present cycle. The trough-like appearance of the head of Big Timber 

 canyou (plate 37, figure 1) is very likely due merely to glacial modification 

 of the valley of the .second of the later cycles, while the upper slopes represent 

 the first, the more ancient topography having been obliterated in that region. 



ALTERNATIVE HYPOTHESES 



The alternative hypothesis, that the benches and slopes mentioned do not 

 represent revival, but merely difl'erential erosion in rocks of varying hardness, 

 does not find support in the case of the Sweetgrass locality just cited, for the 

 slopes are cut in hard igneous rocks that have practically uniform vertical 

 distribution. The benches, too, are very distinct in many localities rather 

 widely separated, and diiferiug in geological horizon and structure, yet they 

 seem to maintain fairly constant relations to each other. 



The hypothesis that the benches are merely alluvial terraces is not sup- 

 ported by the facts. The benches are only thinly covered by gravels. Near 

 their tops exposures of bed rock occur in the later ravines. The forms in 

 question are true rock benches due to revival by uplift, which must have been 

 sufficiently extended to include the area observed. 



ADVENT OF ICE 



Morainic material lies on the uppermost benches, aud also at lower levels 

 within the lower canyons, but in the latter case it is not certain whether the 

 glacial deposits rest on rock benches or simply represent morainic fillings. In 

 the Sweetgrass gateway above described, however, the two intermediate sets of 

 slopes have been glaciated, while the lowest is not striateu and appears to be 

 post-Glacial. Hence it seems probable that occupation by ice succeeded the 

 second of the later cycles of erosion. 



DUEATION OF GLACIATION 



It has been shown that in the lower canyons the trough form includes nearly 

 or quite the entire valle.y, and that in the bench lands the ice at times even 

 overspread the valley walls. In the upper canyons, however, the troughs form 

 much less conspicuous parts of the valleys. Slopes of normal erosion extend 

 2,000 to 3,000 feet above the tops of the troughs. Though the divides are gen- 

 erally sharp, they are more frequently the product of normal erosion than of 



