228 Transactions of the Boyal Microscopical Society. 
since near what was the surface all the cavities are empty, even in 
the centre of the thickness of the section. This loss of liquid 
is also to some extent dependent on the total relative hulk of the 
cavities, since, as might he expected, the solid amber has presented 
a greater obstacle to evaporation than that rendered comparatively 
porous by the enormous number of vacant spaces often met with in 
opaque bands. There is thus no difficulty in understanding why 
so many thin sections of amber show only empty cavities, as to 
have led some observers to conclude that this was their original 
normal condition, and to explain their formation by assuming that 
a gas had been evolved in the interior. Probably this has really 
occurred to some extent, but still the evidence is not perfectly 
conclusive. 
In the centre of the thickness of those sections in which the 
cavities retain their liquid, and surrounded by many cavities still 
quite full, a few empty cavities may be seen, and it seems difficult 
to understand how they and they alone can have lost their liquid, 
and it is far more probable that they were originally filled with 
a gas. 
The liberation of a gas in the midst of the amber appears to be 
established by the characters of such cavities as Figs. 8 and 9, 
which, however, are comparatively rare. Those like Fig. 8 at 
once remind an observer of a balloon with an attached car. By care- 
fully looking over many sections we have been able to find only one 
or two cavities more or less closely approximating to this form that 
were entirely filled with liquid. They are usually partly filled w4th 
liquid, and partly with gas, as shown by Fig. 8, and usually, if not 
always, the car, so to speak, is full of liquid, and the balloon itself 
mainly or entirely filled with gas. Such cavities are most emi- 
nently characteristic of amber, and totally unlike any seen in 
crystals. Their formation appears to have been brought about in 
the following manner. What we may call the car was originally a 
spherical fluid-cavity of the usual kind, and after the amber imme- 
diately surrounding it had become somewhat hardened, a gas was 
given ofi", either by some chemical change or by the pressure being 
relieved by a contraction of the general mass of the amber after its 
exterior had become somewhat hard. If the surface of the original 
cavity had been equally soft in every direction, the evolution of the 
gas would have merely dilated the surrounding amber, and given 
rise to a somewhat larger spherical cavity, partly filled with gas, 
like Fig. 3 ; but if one side of the original cavity could yield much 
more readily than the rest, the gas would be able to force its way 
out into the surrounding softer resin. The actual or relative pressure 
being thus relieved, the softer side of the original spherical cavity 
partially collapsed, and some of the liquid rose up from the car into 
the balloon. By these simple suppositions all the peculiarities of 
