THE ANTARCTIC REGIONS. 287 



land and water are so distributed that, while the 

 shore-lines are of great extent, there is very free com- 

 munication with the open Antarctic Ocean. In other 

 words, it seems reasonable to conclude that there are 

 many large islands within the Antarctic circle, that 

 these islands are separated from each other by wide 

 passages, and not by straits readily blocked up and 

 encumbered with ice in such sort as to impede ',the 

 outward passage of the great icebergs. And nothing 

 which has been ascertained by Antarctic voyagers is 

 opposed to this conclusion. It is indeed very easy to 

 fall into the mistake of inferring otherwise from the 

 study of an ordinary chart of the Antarctic seas. If, 

 for example, we look at the chart in Maury's Physical 

 Geography of the Sea, we are apt to imagine that the 

 boundary-line indicating the limits of Antarctic explora- 

 tions, points to the existence of a continuous barrier of 

 ice, the advanced line of defence, as it were, behind 

 which lies as continuous a barrier of precipitous shore- 

 line. But a very slight study of the records of 

 Antarctic voyages will suffice to show how erroneous is 

 such an impression. We find that long before coast- 

 lines have been seen, the hardy voyagers have found 

 themselves impeded and often surrounded by masses 

 of floating ice. Wilkes,^ Ross, and d'Urville, when 

 struggling to advance towards the southern pole, were 

 repeatedly compelled to retreat without seeing any 

 signs of land. Land has not been seen, indeed, along 

 more than one-sixth part of the circuit of the Antarctic 

 barrier, and it has only been in the neighbourhood of 



