160 WILLIAM H. HOBBS 



ridge extending from the outer margin of the bay (the peninsula 

 of Picnic Point) across the wide portion of the lake to the opposite 

 shore, and about this section no ramparts are developed (see 

 Fig. i). 



Ice ramparts can thus form only on shores of lakes which have 

 relatively small size or on small bays of larger lakes, though 

 a width of at least half a mile is probably necessary in order to 

 secure sufficient dilatation of the ice cover to make ramparts of 

 appreciable size. 



Anything which tends to deform the ice cover from a perfect 

 plane will effectively destroy its competency as a girder, and then 

 no ramparts will form. Mr. Tyrrell has shown in his valuable 

 paper that young lake ice will support, without bending, less than 

 its own thickness of dry snow, and that the ice on Canadian lakes 

 is bowed down under its load of snow to such an extent that water 

 comes to the surface through cracks and further increases the 

 bending. 



To sum up, the heavy snow cover alone would by blanketing 

 the ice, but probably even more by bending it, effectually prevent 

 the formation of normal ice ramparts. As already stated, such 

 ramparts may actually be seen in process of formation during a 

 warm wave in an} 7 favorable winter about Lake Mendota at 

 Madison, Wisconsin. 



It is fully realized that rafts of floating ice drifted by the winds 

 at the time of the spring "break up ,? do also produce small bowlder 

 ridges on shores which bear a close resemblance to some of the 

 types of normal ice ramparts. 



