18 
[oa 
BULLETIN OF THE 
deep trough which constitutes the eastern part of the fiord known as 
Narragansett Bay. At present there is about three hundred feet differ- 
ence in the altitude of the bottom of the channel and the top of its 
eastern wall. It is likely, however, that the débris which has accumu- 
lated since the Ice Period has diminished the depth of this trough to 
the amount of a hundred feet or more. The axis of the depression lies 
approximately in the path of the ice stream. It was, indeed, partly 
excavated by the same glacial movement which conveyed the fragments 
of the boulder trail, and therefore it naturally has the same general 
direction as that followed by the glacier. 
An important effect of the channel which forms the main or western 
part of Narragansett Bay has been to hide a great part of the surface of 
the train beneath the waters of the sea. As will be seen from the map, 
the train appears on the shore upon both sides of the Narragansett main 
channel, and upon the islands in the bay wherever they exhibit deposits 
of till. It is evident, however, from the distribution as shown on the 
map, that the ice stream, while pretty closely following the path of the 
depression, inclined to surmount the eastern border of the channel. 
This tendency becomes more marked as we approach the seaward end of 
the bay, for at that point we find the train inclining away from the 
mainland, or western shore of the embayment, and crossing Aquidneck 
Island to its eastern side. It will be observed that this eastward ten- 
dency is probably increased in the part of the train which lies below the 
level of the sea. This is indicated by the fact that the area occupied 
by the boulders touches the southwestern end of Martha’s Vineyard. A 
glance at the map will show that this requires a very sudden turn of the 
train to the eastward on the area beneath the sea between the mouth of 
the bay and the district occupied by Martha’s Vineyard. : 
At first sight, it seems improbable that such a sharp ‘flexure in the 
path of the ice as is above noted would have been likely to have 
occurred. There is evidently no barrier in the region beyond the mouth 
of Narragansett Bay which could have served to bring it about. While 
still in a state of doubt concerning this feature I had occasion to ob- 
serve a similar sharp eastward turn in the ice which escapes from the 
southern end of the fiords at Mt. Desert, Maine, where we have valleys 
comparable to that of Narragansett Bay as regards their direction, 
debouching into the open sea. In the Narragansett Basin we are not 
able to find very clear evidence as to the path of the ice after it escapes 
from the groove between Aquidneck and Conanicut Islands ; but on Mt. 
Desert we can see on the islands which lie to the southeast of Somes 
