THE BALLROOM FLOOR 105 



masses, as already mentioned, but even in the neigh- 

 bourhood of these we could not see any marked 

 disturbance. The first sign that the surface was more 

 treacherous than it appeared to be was when Hanssen's 

 leading dogs went right through the apparently solid 

 floor. They remained hanging by their harness, and 

 were easily pulled up again. When we looked through 

 the hole they had made in the crust, it did not give us 

 the impression of being very dangerous, as, 2 or 3 feet 

 below the outer crust, there lay another surface, which 

 appeared to consist of pulverized ice. We assumed 

 that this lower surface was the solid one, and that 

 therefore there was no danger in falling through the 

 upper one. But Bjaaland was able to tell us a different 

 story. He had, in fact, fallen through the outer crust, 

 and was well on his way through the inner one as well, 

 when he got hold of a loop of rope on his sledge and 

 saved himself in the nick of time. Time after time the 

 dogs now fell through, and time after time the men went 

 in. The effect of the open space between the two 

 crusts was that the ground under our feet sounded 

 unpleasantly hollow as we went over it. The drivers 

 whipped up their dogs as much as they could, and with 

 shouts and brisk encouragement they went rapidly over 

 the treacherous floor. Fortunately this curious forma- 

 tion was not of great extent, and we soon began to 

 observe a change for the better as we came up the 



ridge. It soon appeared that the Ballroom was the 

 VOL. II. 33 



