ON THE STRUCTURE AND MOTION OF GLACIERS 5 
re-attached itself; the pressure was continued, and in a few seconds 
the sphere was reduced to a transparent lens of the shape and size of 
the mould in which it had been formed. 
This lens was placed in a cylindrical cavity, two inches wide and 
half an inch deep, hollowed out in a piece of boxwood, C, fig. 3, as 
Fig. 3. 
Fig. 2. 
D 
hs 
before ; a flat plate, D, of the wood being placed over the lens, it was 
submitted to pressure. The lens broke as the sphere did, but the 
fragments attached themselves in accordance with their new conditions, 
and tn less than half a minute the mass was taken from the mould a 
transparent cake of te. 
The substance was subjected to a still severer test. A hemi- 
spherical cavity was hollowed out in a block of boxwood, and a 
protuberant hemisphere was turned upon a second slab of the wood, 
so that, when the protuberance and the cavity 
were concentric, a distance of a quarter of an 
inch separated the convex surface of the 
former from the concave surface of the latter. 
Fig. 4 shows the arrangement in section. 
The pins of brass, ad, fixed in the slab AB, 
and entering suitable apertures in the mould 
CD, served to keep the two surfaces con- 
centric. A lump of clear ice was placed in the cavity, the pro- 
tuberance was brought down upon it, and the mould submitted 
to hydraulic pressure. After a short interval, it was taken from the 
press, and when the upper slab was removed, a smooth concave 
surface of ice was exposed. By tapping the conical plug 4, this ice 
was lifted from the cavity, the lump having been converted by pressure 
into a hard transparent cup of ice. 
Fig. 4. 
