252 THE ANTARCTIC. 



about the texture of these two kinds of ice. Shots were 

 fired at a berg from a 12-pr. gun, first at the blue ice 

 near the surface of the sea, with the effect that large 

 masses were splintered off and hurled into the water, 

 thus proving that the blue ice was hard and brittle. The 

 second shot entered a broad white stratum near the upper 

 surface of the berg without producing any effect whatever, 

 whence it is inferred that these strata consist of soft, un- 

 resisting material, in fact, of snow or firn, that had under- 

 gone little alteration. Now, the question arises : How 

 does the snow from the deep hollows change into strata 

 of blue, hard ice, which at first are thin and gradually 

 increase in thickness? The best solution of this problem 

 seems to be supplied by Nansen's observations on the 

 inland ice of central Greenland. Here the climatic con- 

 ditions of the south polar regions are repeated, apparently 

 aided by the considerable elevation above the sea-level. 

 The snow falls in the shape of a very fine crystallised 

 powder all the year round, but mostly so in summer. 

 Although the conditions are more favourable, still the 

 sun is at that elevation no longer able to do more than 

 slightly melt the snow and moisten its surface ; this 

 freezes again at night, so that no part of it filters through 

 and converts the snow into coarsely-granulated firn-ice. 

 Nansen, too, found at the end of August, the first and some- 

 what thicker, hard crust beneath the thin surface crusts 

 formed in the same summer, at a depth of three to four 

 feet, and sometimes at a less depth. We are led then to 

 assume for the Antarctic regions also that in the height 

 of summer the sun is able to melt the surface of the snow 

 often marked by light undulations produced by the wind. 

 If, then, such a stratum of one year's melting is com- 

 pressed by the superincumbent strata of many years, the 

 volume of the loose snow between the two hard crusts must 

 be reduced by the contraction of the air-containing pores. 

 I conjecture, however, that the layer of snow placed over 



