240 J. D. Dana on Glacier movements along valleys. 
seems to prove that the direction of movement thereby indicated 
characterized the ice of this part of the valley through the whole 
of the Glacial period. 
6. Again, if a local glacier occupied the valley having a thick- 
ness of say one, two, or three thousand feet, or such as wou 
lie below the level of the Green Mountain summits, the glacier 
would have had through its breadth a nearly southerly course 
corresponding to the trend of the valley, and in that case south- 
erly scratches should have existed over the whole surface, evel 
localities remote from the Connecticut river—where they are 
not found. 
In another place I have sypposed that the southeasterly 
course which occurs in the scratches to the west of the Connec- 
ticut river might have been a resultant between the tendency 
to a southerly movement down the valley, and that down the 
slope into the valley. But this was so only to a very sm 
egree. For the ice, after passing over the valley, resumed on 
the east its southeastward scratching. 
c. In the part of the Connecticut valley north of Massachu- 
setts, the course of the scratches is not that of the valley, but 
differs 10° to 15° from it to the eastward. This greater easting 
shows that the southerly movement of the ice induced by the 
valley was modified by some force pressing it eastward, and 
: t the valley ice of the Con- 
necticut had through its southern half (across Massachusetts and 
modifications in the valley movement just pointed out, and 
acier. If the southeasterly 
westerly scratches were the oldest, but admits that there ® 
much doubt with to it. 
This movement of the bottom of a glacier six or eight thou 
sand feet thick along a different course from its main ™ 
wherever it lies in + valleys, is an result of m& 
