THE FERRAR GLACIER 89 



roughly hexagonal, similar to the ice photographed at Clear Lake. When the 

 hexagonal ice has been removed by ablation, snow-tabloids and bubbles of various 

 size are common, and occasionally the secondary ablation-formed ripple surfaces so 

 common at Cape Royds are developed. At other places there appears to be a 

 development of what is either coarse granular neve or minutely hexagonal ice. 



As the ice-falls were approached the bare surfaces became fewer, and finally 

 disajjpeared until the ice was covered with several feet of snow, the upper foot more 

 or less hardened into a crust, whilst underneath this crust was at least 3 feet of 

 loose, largely-granular snow. At this place the pressure from the glacier above had 

 caused considerable undulations in the surfaces, and we passed several crevasses. 



We had a fairly clear glimpse down one of these crevasses, and the structure 

 of its walls was very puzzling. Instead of the uniform sheet of ice that was to be 

 expected, the portion of the walls exposed to our view seemed to consist of flattened 

 lenticles of ice with layers of snow in between, and it was not until the return 

 journey that any explanation of this peculiar structure was suggested. 



On the way back to Butter Point we passed this part of the valley during a 

 very hot summer's day, when the thaw was most powerful, and almost the whole 

 of the surface for several miles below the ice-falls was occupied by a series of pools 

 of thaw-water full of drift snow, and it seems probable that this thaw explains both 

 the ice surface we passed over and the structure of the crevasse walls. Under 

 the ice formed in this manner there may be a substratum either of glacier ice or 

 of sea ice, but every piece of evidence points to the probability of the first few 

 feet at any rate being formed In the way just described. 



Above the ice-falls the true glacier ice set in, but very little of ice or neve 

 was exposed for some miles, as the surface of the glacier was hidden under a uniform 

 coating of snow. 



For some time before any surface moraine was reached the presence of an 

 englaclal one was indicated by numbers of circular patches of hexagonal Ice indicat- 

 ing where the boulders of the moraine had sunk beneath the surface. Always we 

 found that, when a moraine had persisted for some distance without being fed 

 from other moraines or from the cliffs under which it had passed, the number of 

 boulders above the surface became less and less, until the moraine had become 

 wholly englacial. It Is an interesting fact that the hexagonal structure In the Ice 

 was never present when the boulder had just sunk beneath the surface, but was 

 only superinduced after the frozen thaw-water had been formed a long time. It 

 seems therefore to be a secondary structure, due either to a molecular change In 

 the thaw-water ice itself, or to the re-crystallisation of the lower jjortlons of the 

 snow-drift collected in the depi"ession usually left where the boulder had disappeared. 



Another feature very well marked at this time was the recementation of old 

 cracks and narrow crevasses by the thaw-water, with a distinct and very fine 

 lamination parallel with the crack. The same thing occurred in the lakes at Cape 



