284 THE BORDERLAND OF SCIENCE. 



parts of it all the year round. I have encountered 

 them myself as high as the parallel of thirty-seven 

 degrees, . . . and navigators on the voyage from 

 the Cape of Grood Hope to Melbourne, and from Mel- 

 bourne to Cape Horn, scarcely ever venture, except 

 while passing Cape Horn, to go on the polar side of 

 fifty-five degrees.' As he justly remarks, 'the nursery 

 for the bergs to fill such a field must be an immense 

 one ; such a nursery cannot be on the sea, for icebergs 

 require to be fastened firmly to the shore until they 

 attain full size. They, therefore, in their mute way, 

 are loud with evidence in [favour of Antarctic shore- 

 lines of great extent, of deep bays where they may be 

 formed, and of lofty cliffs whence they may be launched.' 

 It is remarkable, however, that Maury fails to notice 

 that the evidence of these enormous icebergs is opposed 

 to the theory of an Antarctic continent, or is, at least, 

 by no means in favour of that theory. It might at 

 once be objected, indeed, to the inferences derived by 

 Maury from the Antarctic icebergs, that similar 

 reasoning would show the unknown parts of the 

 Arctic regions to be mainly occupied by land masses. 

 But, apart from this, all that we know of glaciers 

 teaches us to recognise the fact that they are formed 

 only in regions where vast mountain ranges exist, and 

 where the lower levels are reached by ravines and 

 valleys gradually diminishing in slope as they descend. 

 Now, wherever this is the contour of the land, we have 

 in the surrounding regions one or other of the three 

 following conditions : either (i.), flat land regions 



