4IO SCOTT'S LAST EXPEDITION 
50 feet down, and rarely got a lead down beyond 
8 fathoms (50 feet). 
Let us compare an ice sheet over land with a similar 
sheet extending over water. In tJie first type one could 
not expect uncrcvassed areas of any size if the ice were 
moving over the land. (Though we must remember 
that the stationary ice over the Plateau is not crevassed.) 
Again, over a sea surface the crevasses would only 
extend for a limited distance, in fact to sea-level, where 
they would freeze over. This may account for the limited 
depth observed. 
Near the Barne and Shackleton Inlets the great lateral 
trench was filled with pools due to thaw waters, and this 
was 100 feet deep. If the sheet were 1000 feet thick, one 
would expect this ' rupture crack ' to be much deeper. 
IV. Temperature and Pressure, — The temperature in 
the crevasses seemed fairly constant near the land, but 
when farther away it seemed to rise with depth. This 
looks like the result of a subglacial sea. 
The atmospheric pressures as taken on the journey to 
82° S. varied very little from those at Hut Point : 
At 79°, + "045 difference 
„ 8o^ + -04^' 
„ 8i^ + -06^^ 
The barometric gradient probably rises as one goes 
south, so that one cannot use the barometer to obtain 
accurate levels. 
One could get equal (and useless) readings all the way 
