А 
* 
167 
End, in very narrow valleys between the drumlins, that is, in 
the most sheltered spots, where we would least expect erosion 
by waves. Still another difficulty in the way of regarding them 
as ordinary shore-marks is the fact that they occur only on the 
sides of the drumlins, the scarps never, so far as observed, 
changing their directions so as to cut across the ends of the 
hills. hat is, they notch or break the transverse, but never 
the longitudinal, profile of the drumlins. We may safely 
conclude, then, that they have not been formed in or by &tanding 
water. 
It has been suggested to me that the terraces may, perhaps, 
be regarded as original features of the drumlins, owing their 
formation to the molding action of the moving ice, the drumlins 
having been grooved or fluted by the ice on alarge scale. This 
explanation would account for their uniform parallelism with 
the major axes of the drumlins; and I am inclined to believe 
that some of the less distinct terraces are really gigantic glacial 
grooves, formed, of course, while the ice still covered the 
drumlins, and not by the edge of the ice, since it is probable 
that when the ice had melted sufficiently to uncover the tops of 
the drumlins its motion had practically ceased. Starting with 
the normal type of drumlin, we recognize departures from it in 
several directions. Thus some are long and narrow or ridge- 
like, while in others the longitudinal and transverse axes are 
nearly equal, and still others, though presenting the normal 
proportions of length and breadth, are unusually flat-topped, 
like Booth Hill and many of the lower drumlins, including 
Hoop Pole and Green Hills. "These variations are probably 
original and due in part to the contours of the underlying 
ledges and the character of the till and in part to the varying 
thickness and motion of the ice ; and no reason is apparent why 
the latter cause, at least, may not give us departures from the 
regular curvature of the transverse profiles. 
Asa rule, however, the terraces are too deeply incised and 
much too sharply defined to be explained in this way. It is 
