GROUND-ICE WEDGES 



637 



soon as slumping has taken place, erroneous conclusions may be 

 drawn as to the distribution of the ice, scattered outcrops being 

 interpreted as exposed parts of a single bed. 



The upper surface of the ground-ice is usually only a foot or two 

 under the surface of the tundra. Consequently in an area which 

 has discontinuous bodies of ice separated by masses of muck, etc., 

 there will be the least amount of material for slumping exactly 

 where the ice occurs. The ice melts back under the overhanging 

 turf, forming a cave, but at either side the muck will slump from 



ly/f/m^'f, 



W/Wi'B-wm 



Fig. 3. — Exposure of a bank showing an apparently continuous thick bed of ice 



^ /f(:^l,owut«v- Ickz. 



Fig. 4. — Structure of exposure in Fig. i 



the grass roots. Thus, wherever the surface of the bank is exposed, 

 ice is likely to be seen, and observers are led to beheve that a con- 

 tinuous body of ice underHes the whole area. 



In addition to the erroneous impression as to lateral extent, the 

 conclusions as to thickness are also often faulty. A winter's snow 

 drift against the foot of a bank may be covered by slumping and 

 later exposed, apparently showing a thickness of ice only Umited 

 by the height of the bank. The same may be true of river- or sea- 

 ice. An illustration is given of an apparently continuous bed of 

 heavy ice which the writer examined carefully for a quarter of a 

 mile before learning its true nature (Figs. 3 and 4). 



