530 H. T. Fi'irar—T/ie Antarctic Ice-cap. 



localities clear bright days predominate : under these conditions, when 

 no shadows are produced, it is exceedingly difficult to be certain 

 which was originally the upper surface of a disintegrated iceberg. 

 Tlie accompanying illustration shows an iceberg estimated to be 

 200 feet high which still has a portion of its original upper surface 

 intact. If this portion were removed it would be extremely difficult 

 to make sure wliether the berg had turned over or not. To argue 

 that such icebergs as those figured in a great number of publications 

 have all turned over, is hardly safe, especially as bodies with 

 diameters of unequal length appear, in the majority of cases, to float 

 with the longer diameter horizontal. It therefore seems highly 

 improbable that a berg would increase its height by turning over. 



A berg, however, may increase its height, and probably often does 

 so, in the following way. A newly calved iceberg usually has 

 vertical sides ; the waves beat against these and, undermining them, 



Fig. 1. — Au iceberg which has turned through nearly 90°. 



cause pieces to split off. This splitting will probably take place 

 along vertical planes, parallel to inherent planes of weakness, and 

 this undermining is accelerated by the temperature of the water at 

 the surface being raised by -radiant heat from the sun. The visible 

 part of the berg will thus be reduced in mass, and consequently the 

 berg will rise. That this is a common occurrence may be seen in 

 almost any of the narratives of polar voyages, where one constantly 

 finds references to the toe of an iceberg. 



For icebergs floating freely in the sea, Von Drygalski gives the 

 proportion of normal glacier ice above water to ice below water as 

 1 is to 4. Kink' makes the proportion as 1 is to 6 ; Steenstrup as 

 1 is to 7 ; Sir John Murray as 1 is to 7 ; and Scott - as 1 is to 5. The 

 cannonading performed upon an iceberg by H.M.S. '" Challenger," 



1 Eiuk : Danish Greenland, London, 1877, p. 35S. 

 "^ Scott: Geog. Journ., April, 1905, p. 356. 



