238 " S*. W. SARDESON THE REDSTONE QUARTZtTE 



bottom a mile wide. This is the channel of the glacial Miimesota or 

 "river Warren/' where that river eroded into the glacial drift. At New 

 Ulm the valley widens over the top and a system of terraces appears. The 

 river Warren was also once divided here by an island which is now the 

 site of the city of New Ulm (6, volume 1, page 582). At this place the 

 valley is cut not only in the glacial drift, but also into the Big Cottonwood 

 formation. At Eedstone, 4 miles below New Ulm, the river Warren en- 

 countered the mass of Courtland quartzite, and this obstruction caused 

 the river to widen its valley and build terraces even several miles above 

 the obstruction. The terraces at New Ulm correlate largely with the 

 stages of the valley's making at Eedstone. 



From the direction of the valley as it now lies at Eedstone, it is readily 

 seen that the river Warren flowed first directly across the present area of' 

 Courtland quartzite. At that early stage the river had only glacial drift 

 to encounter, but after cutting its valley to a small depth the quartzite 

 was met. This rock was highest on the left, and, being very resistant, it 

 swerved the river gradually to its right side, and finally entirely beyond 

 the west end of the quartzite formation. As the valley was gradually cut 

 down, Eedstone hill was thus uncovered and left extending half a mile out 

 into the valley. 



The making of terraces by the river might be attributed to three fac- 

 tors, namely, variation in flow of the stream, inequality in the quartzite 

 barrier, and inequality of the other formations across which the river was 

 diverted. Of these the first may be set aside, since elsewhere, especially 

 above New Ulm, the valley indicates rather a constant stream. The rela- 

 tion of the stream and quartzite appears also to have been rather a uni- 

 form one. On the highest and earliest iincovered part of the quartzite 

 its surface bears prior glacial striae (6, volume 2, page 165; 10, page 21), 

 thus proving that the river scarcely cut the quartzite there ; also, at lower 

 levels, not much of the quartzite was removed. The surface there is in 

 fact grooved, furrowed, and worn into irregular loiobs and depressions, in 

 the direction and manner of rapid stream erosion, by the river Warren; 

 but yet the cause of this appearance of great erosion is that the leached 

 and softened portions of the quartzite, as before described, was alone 

 worn away. The firm quartzite appears thoroughly polished and gener- 

 ally rounded, but was in fact so little cut away that much, if not the 

 greater part, of the softened or superficially leached rock was protected by 

 it. While Eedstone hill appears at first sight to have been worn into 

 a rugged remnant hj the river Warren, there is still better reason for 

 thinking that the hill was practically an impregnable barrier to the 

 stream. The course of the valley around the base of the hill is further 

 evidence to this effect. The present form of the hill is therefore nearly 



