36 Mr Skinner , Observations on the Minute Structure of Ice. 
furrows very little can be made out in the cast, and the faces 
appear to be quite continuous. However, the optic axes are in 
all directions, and not parallel to one another as in the cave ice. 
It might be expected that the actual surface of a glacier 
would present a similar weathering as in Figure 2 with the 
exception that the furrows were deeper and broader. It is, 
however, much more complex and a cast of the surface is repre- 
sented in Figure 3. The cast is 10 centimetres high. The light 
has been arranged so that the ridges stand out prominently on the 
left-hand portion of the cast, and especially towards the upper 
part the granules are evident as small spaces surrounded by walls. 
But in the faces of these granules there are other cavities into 
which the plaster has flowed. These cavities have rounded edges, 
and appear either as cylindrical drill holes, or as elliptically- 
shaped recesses. The former are probably due to fragments of 
dust which have absorbed heat from, the sun and so melted a 
passage for themselves into the ice granule. This action of small 
particles of dust can often be seen on a glacier. The latter are 
possibly the closed cavities of Tyndall 1 which have arisen as flat 
star- shaped vacuoles, which when the surface of the glacier melts 
are opened to the air. The drill holes and the cavities may 
occur anywhere on the surface of the glacier granule. They are 
absent in the cave ice where fusion by radiation is not possible. 
This forms a point by which the fusion in the two cases may be 
immediately distinguished. In cave ice the fusion is confined to 
the interstitial spaces between the crystals, whilst in the case of 
the glacier surface fusion occurs both in the interstitial spaces and 
at points over the surface of the crystal wherever dust has settled. 
Two days after this cast was made the same spot on the 
glacier was visited and a very irregular honey-comb structure of 
plaster was found on the ice. This consisted of the plaster which 
had run down into the deeper cavities, and had not been removed 
with the cast on account of its frailty. The general surface of the 
glacier had melted and left the plaster net-work standing up. It 
was about an inch high, and showed the form of the glacier 
granules and cavities, but no further peculiarities were noticed 
than those already described. 
We may sum the results of this paper in two conclusions : — 
(1) Forel’s Streaks are probably indications of slip-bands or 
twinning in ice crystals under stress. 
(2) The porous nature of the surface ice of a glacier is due to 
fusion in the interstitial spaces between the granules, to the 
formation of dust drill holes, and to the opening to the surface 
of Tyndall’s closed fusion cavities. 
Tyndall, loc. cit. 
