2i6 POPULAR SCIENCE MONTHLY. 



from Dougherty Island and Peter Island in the southern Pacific. 

 Meteorological evidence supports the idea that the Antarctic is still a 

 continent geographically, and that the center of the land is not coin- 

 cident with the south pole, but is in the eastern part of the area. 



The available evidence appears to be decidedly in favor of Sir John 

 Murray's theory, though the question cannot be definitely settled until 

 the range of the land has been mapped. This task may be facilitated 

 by the guidance as to the probable trend and position of the coasts, that 

 is given by the principles of geomorphology. 



If the current theory of the structural unity of the Pacific ocean 

 be correct, then that ocean must be bounded on the south by a coast 

 of the ''Pacific type.' With one exception in Central America the 

 whole of the known coasts of the Pacific belong to what Suess has called 

 the 'Pacific type.' The main character of this form of coast is that 

 the trend is determined by mountain ranges rimning parallel to the 

 shore. In the South Pacific this type is well exemplified by iSTew 

 Zealand on one side and by the Andes of South America on the other. In 

 southern Patagonia the Andes are turned from their meridional course 

 and run eastward across Terra del Fuego. The tectonic line of the 

 Andes is then apparently bent suddenly southward and reappears in 

 Graham Land. It is probably continued round the southern Pacific, 

 meeting the known end of the New Zealand line near Mounts Erebus 

 and Terror. 



The theory of the structural unity of the Pacific is sufficiently 

 established to render it probable that Cook was close to land when he 

 turned back from his furthest south in the South Pacific (71°S. 

 123°E.), that the 'ice-barriers' of Ross and Bellingshausen are both 

 the fronts of glaciers flowing from highlands to the south; that there 

 is a land connection of the Pacific coast type running from Koss's ice 

 barriers northeastward to Graham Land; and that Victoria Land is 

 connected to Wilkes Land by a broad bight. 



There are no such data for predictions as to the distribution of 

 land and water in the German and Scotch areas of work. For Wilkes 

 Land and the lands that may extend thence westward towards Graham 

 Land are no doubt plateau countries bounded to the north by coasts of 

 the 'Atlantic type'; and the trend of such coasts is not determined by 

 simple continuous tectonic lines. That Wilkes Land and Geikie Land 

 repeat the structure of southern Australia is rendered probable by the 

 geological collections of all the expeditions from Wilkes to the ' South- 

 ern Cross.' The westward extension of this land line has probably the 

 same structure, and it is accordingly impossible to predict how far the 

 Weddell cuts into the Antarctic lands. 



The principles of geomorphology not only suggest the external 

 shape of 'Antarctica' but also its internal relief. It is probable that it 



