DESCRIPTION OF THE ISLANDS 153 



eruptions are deposited to the east: thus, in most of the islands, it is on the western side 

 that we find the steepest cliffs and the deepest off-shore soundings. 



Glaciation. The islands and portions of islands that have lost all volcanic warmth 

 are buried deep in glacier, and ice also covers the still active cones of Visokoi and 

 Saunders. This ice-cap is similar in type to that commonly seen on Antarctic islands, 

 but it conforms to the larger irregularities of the surface and rarely shows the perfectly 

 even shield- or dome-shaped contour which characterizes the full development of this 

 form of glaciation (Plates XXIII, XXV, XXIX). The ice, however, may reach a thickness 

 of as much as 260 ft. at its coastal margin. Rocks outcropping through the ice-cap were 

 seldom seen, and there were no bergschrunds ; but not infrequently a crevassed pro- 

 trusion in the glacier spoke of projecting rock beneath. The action of the ice is evidently 

 to abrade and reduce surface inequalities, rather than to emphasize their relief as in 

 mountain glaciers. 



The term "ice-cap" is used here in the sense defined by Hobbs^ and seems very 

 appropriate to this form of glaciation ; but in Priestley's more recent classification of 

 land-ice formations'- it has been abandoned and is replaced by "island-ice" and "high- 

 land-ice ". The first of these is applied when the contour is perfectly even, and the latter 

 when the conformation of the underlying land is not completely obscured. The differ- 

 ence between the two does not appear important, for it depends merely on the amount 

 of relief in the land surface and the thickness of the glacier. An island with high hills 

 and very thick glacier might still be "highland-ice ", while one of low ground might be 

 "island-ice" though its covering were very thin. In their denuding processes and in 

 their alimentation^ the two appear to be indistinguishable. 



A type of glaciation agreeing with that in the South Sandwich Is. has been described 

 by Holtedahl* from Bouvet and from some of the islands in the South Shetlands and 

 Palmer Archipelago. Holtedahl considers that Priestley's definitions of "island-ice" 

 and " highland-ice" do not accord with these formations and he has therefore proposed 

 a new term, "the Antarctic ice-mantle type". 



We prefer here to use " ice-cap " in the sense employed by Hobbs without attempting 

 any finer classification, but we should perhaps point out that this term is not uni- 

 versally applicable to island ice-formations in the Dependencies of the Falkland Is. 

 In some of the islands of the South Shetlands and Palmer Archipelago — Low and Snow Is. 

 are examples — the land is covered with a smooth evenly-domed shield of ice, without any 

 irregularities due to the underlying rock, while in others the formation is closely similar 

 to that found in the South Sandwich Is. For both these we may use the term "ice- 

 cap ", and in the first of them it is seen in its most perfect development. But there are 



1 Hobbs, Characteristics of Existing Glaciers, pp. 7, 8, 285, 286 (New York, 1922). First published in 

 191 1. 



- Priestley, "Glaciology", British Antarct. (' Terra Nova') Exped. 1910-13, pp. 147, 148 (1922). 



^ It is these characters that Hobbs uses for his primary division of land-ice formations into "inland ice", 

 "ice-cap" and "mountain glacier". 



* Holtedahl, "On the Geology and Physiography of some Antarctic and Subantarctic Islands", The 

 Norwegian Antarctic Expeditions, 1927-8, 1928-9, No. 3, pp. 121, 122 (Oslo, 1929). 



