300 PEOFESSOE TYNDALL ON THE ’SEINED STELCTILRE OF GLACLEE ' 
Fig. 19. 
I have given in fig. 19 a sketch of a piece of ice taken from the end of the great 
Allalein glacier, on the Swiss side of the Monte Moro. 
On reference to M. Agassiz’s figure, it will be quite mam- 
fest that we are both dealing with the same phenomenon , 
we have the division of the ice into “ angular fragments,” 
the flattening of the “ bubbles,” and the non-parallelism 
of their directions in the different fragmente. 
Fig. 20 is a sketch of a piece of ice which showed the 
veined structure. The line AB was parallel to the veins, 
and it will be seen that the “ bubbles” are inclined to this 
line at different angles, and in different azimuths. The 
circles indicate, of course, that the “bubbles” were there 
parallel to the horizontal face of the slab, while the lines 
indicate that they were perpendicular. In one case the 
bubbles are seen in plan, in the other case in section. The eUipses show the bubb s 
foreshortened where their planes are oblique to the surface of the slab. ^ 
Associated with the air-bubbles, and usually beyond comparison 
more numerous in ice taken from the “ends” of glaciers, were the 
round liquid disks which I have described in my paper on the Phy- 
sical Properties of Ice. Associated with each liquid disk was a 
vacuous spot, which shone with exceeding lustre when the sunbeams 
fell upon it. That the spots were vacuous, and not bubbles of air, i 
proved by permitting them to collapse under warm water; the col- 
lapse was complete, and no trace of air arose from them. 
These, I doubt not, are the “bubbles” observed by M. Agassiz 
“ near the end of the glacier,” and which were “ so flat that they 
might be taken for fissures when seen in profile.” 
These “ vacuum disks,” as I have usually called them, were in- 
valuable as indicators of the planes of crystallization. When a con- 
densed sunbeam was sent through the mass, the six-petalled flowers, 
which always indicate the planes referred to, started into existence 
parallel to the disks. Consequently, as the beam passed through 
different fragments, flowers were formed, in different planes, along 
the track of the beam. 
True air-bubbles, associated with water, also occurred in these 
masses of ice, and such composite cells were always flattened out in the planes of e 
'’'TlTfct to is that many of the so-called ah-bubbles are not air-bubbles at all, and 
that the so-called flattening" is in reality no flattening at all; and that “ 
the sense hitherto conceived, has had nothing whatever to do with the shape of the e 
bubbles In glacier ice, as in lake ice, their shape is determined by the crystalhue 
