164 The American Geologist. September, i896 
1. Identity of alignment indicates probable identity in age. 
In the Narragansett bay region this is the most reliable and 
l^robably the only criterion which can be used. 
2. Identity of elevation of the limiting level of delta con- 
struction in the same field without barriers indicates deposi- 
tion in the same water-body or at the same stand of the land 
with reference to the sea, and hence approximate synchro- 
nism. Allowance must be made for postglacial deformation. 
3. Identity as regards the number and kind of post-deposi- 
tional changes indicates approximate identity in date of 
origin. 
The sand-plains of Narragansett bay are so nearly of the 
same epoch that it is difficult to distinguish differences in 
geological history, and this method of differentiation has not 
been applied. 
Matched by alignment, the following sand-plains appear to 
be clearly along the same front: 
Went side. East side. 
Gaspee Point stage. Harrington stage. 
Occupasspatuxet stage. Nayatt Point stage. 
The sand-plains and cones to the south on the west side of 
the bay are not so clearly traceable into connection with the 
plains and terraces of which traces are recognizable on the 
eastern side of the bay. At present little can be said of this 
eastern region except that the retreat of the ice-front is 
marked by the existence of a kame-like terrace along the Fall 
Kiver shore and by a partially submerged plain at the north- 
ern end of Aquidneck island. That these are frontal deposits 
is indicated by the occurrence of till on the western shoi'e of 
the Taunton river. A similar distribution of drift is found 
on Watuppa pond east of Fall River. The eastern shore of 
the pond exhibits a remnant of stratified drift of extraglaeial 
deposition: the western side of the pond is till; between the 
two deposits probably lay the ice edge. These two stages, tlie 
Watuppa and the Fall River, on the Fall River atlas sheet, are 
probably to be correlated with sonue of the frontal stages be- 
tween Greenwich cove and Wickford. 
Westward the sand-plains of the bay are clearly to be cor- 
related with boulder belts, but the details of this correlation 
are not yet worked out. 
