J. D. Dana on Glacier movements along valleys. 237 
consequently the best evidence of the average direction of the 
movement in that region. 
n Vermont, where the course of the valley for the more 
southern half is S. 12° W., the scratches trend S. to S. 13° E. 
But north of White River Junction the course of the scratches 
varies between S. and S. 30° E., yet many scratches at the 
pe ay near the mouth of the Passumpsic, are 8. 5° E. 
In the part of the Connecticut valley south of Vermont the 
scratches conform closely in direction to the trend of the valley, 
and are the only scratches; while to the north there is a general 
southerly course in the scratches of the Connecticut river valley, 
yet at the same time about 15° less of westing than in the aver- 
age trend of this part of the river. Moreover, in this upper part 
of the valley there are often, besides the valley set of scratches, 
another set having the southeasterly course of the great glacier. 
The width of the region bearing the north-and-south scratches 
of the valley is generally twenty to thirty miles, but sometimes 
more. Going east or west of this there is a change more or less 
gradual to the course of the great glacier, and often also other 
scratches conforming to its course occur. In Massachusetts, in 
Heath, 15 miles west of the Connecticut, the course of th 
scratches given by Hitchcock is south with some westing, and 
the same on Mt. Pocomptuck in this town, 1888 feet high ; and 
i Rowe, 20 miles west of the river, the course is S. 2° W. In 
| southern Vermont, in Halifax, west of Vernon, and 10 miles 
| 
West of the Connecticut, the directions given by Hitchcock are 
mostly S. 12° W.; but also, in West Halifax, 16 miles from the 
Connecticut river, S. 53° E.; in Marlboro, north of Halifax, 
S. 20° E. on high land; and also, at another locality, two 
' Courses, S. 7° W. and S. 58° E., intersecting. __ 
3 The facts show beyond question that the abrading 2 i ve 
and also of the region on either side of it a little distant from 
€ river, 
