380 
but also to those incident upon it after reflection at another 
place. On the other hand, rays falling upon the water will 
be either absorbed or reflected regularly, so as to pass clear of 
the ice, except possibly the actual wall of the hole. Hence the 
ice in the immediate neighbourhood of the hole will receive less 
radiation, and therefore be less melted than the rest of the 
surface, and there will be a tendency to the formation of a 
raised rim surrounding the water-hole. 
(3) As fast however as this rises, those portions of it which 
are on the north, east, and west sides of the hole will be melted, 
as they will at some part of the day receive the sun’s rays not 
only on their upper surface, but on the vertical face towards 
the hole. The east and west sides however will be most 
attacked, for reasons given above in (1). 
The final result will therefore be an oval hole with its 
major axis east and west, with a marked hummock of ice on its 
south side, and sometimes traces of one to the north. 
(4) The most serious difficulty in the way of this explanation 
seems to be in the local nature of the phenomenon. Why is it 
not produced wherever there is a level surface of glacier? The 
ice where the phenomenon is conspicuous is of a peculiar soft 
nature, full of minute air-bubbles, which give it an unusually 
white appearance. Ablation probably takes place rapidly over 
the surface, so that phenomena depending upon differential 
ablation are conspicuous. The peculiar kind of irregularity of 
the surface would favour the action described in (2). 
[Communicated Nov. 15, 1875.] 
Since the notice of my paper read on Oct. 18th, ‘Further 
remarks on the water-holes of the Gorner Glacier, was pub- 
lished, my attention has been called to a passage in Agassiz’ 
‘Nouvelles études sur les Glaciers,’ &c., Paris, 1847, p. 101—2, 
in which a similar phenomenon is described as having been 
observed by Dr Ferdinand Keller. The account of the phe- 
