269 



evidently, be measured from the rocky base on which it stands. 

 This is a partial explanation of the fact that so many inland 

 drumlins have elevations incommensurate with their limited 

 areas and overlooking the large drumlins near the shore. But 

 it must be borne in mind, also, that some of the inland 

 drinnlins are much larger than they appear, only their summits 

 protruding from the sandplnins. 



As the map shows, there is a noticeable difference in trend of 

 the drumlins in the northern and southern parts of the town, 

 the contrast being almost as marked here as in Nantasket and 

 Cohasset. North of the railroad the pi'cvailing trend is south- 

 east or between that and east-southeast, while south of this 

 line it is between southeast and south-southeast. The Turkey 

 Hills belong in the northern division. Undoubtedly the 

 correct explanation of this contrast is that proposed in 

 Part I; viz., that the ice-sheet, at least during the period 

 when the drumlins were forming, was deflected to the east- 

 ward by Boston Harbor, but the portion of the ice which 

 overcame the southern wall of the harbor regained very nearly 

 its south-southeast trend. Although it can be said that 

 a general agreement exists between the trends of the drumlins 

 and the glacial striae on the adjacent ledges, yet this agreement 

 is by no means perfect in all cases. The striae, as the following 

 table shows, depart, as a rule, less widely than the drumlins 

 from the normal direction of glacial movement ; indicating, 

 apparently, that they were formed chiefly during the period of 

 maximum glaciation, when the ice-sheet was less influenced by 

 the subjacent topography and had power to move the entire 

 thickness of the fjround moraine and abrade the solid ledo-es : 

 while the drumlins, it is generally conceded, must have been 

 formed, or at least finally shaped, during the waning of the ice- 

 sheet, when it was easily turned from its course and no longer 

 an eflicient agent of erosion, but, partly slipping over the 

 ground moraine and partly dragging it along, banked it up 

 ajjainst the ledo-es. This chanjje in the direction of the ice- 



