1Q3 
in the Cavities of Minerals, 
spot comes from the summit AB, and takes its station in the 
centre of the ring of new fluid abed. This brown tint some- 
times rises to higher orders of colours ; but disappears by the 
application of heat. That the coloured rings formed within V V 
are vapour, and not a film of the fluid itself, may be inferred 
from its never mixing with the fluid with which it is in imme- 
diate contact. It might, however, be a fluid substance, arising 
either from the decomposition of the fluid itself, or from the 
condensation of gaseous matter within the vacuity ; though this 
is not very probable, from its constant disappearance when it has 
accumulated to a certain degree, and its constant reproduction 
while the temperature remains the same. 
These views respecting the vaporisation of the expansible fluid, 
have been fully confirmed by the discovery of cavities, in which 
the expansible fluid occupies only one-third or one fourth of the 
cavity. These cavities are represented in Fig. 8., where AB is 
the cavity, V the vacuity in the expansible fluid m n o p> and 
A m n, B p o the second fluid. When heat is applied to this 
cavity, the vacuity V does not contract, as in ordinary cases, but 
expands, till its circumference coincides with the boundary m n op 0 
This unexpected effect might have arisen from the expansible fluid 
occupying the lower part of the cavity below Y, as in the section. 
Fig. 9. In this case cef d might have been the vacuity, and the 
surface of the fluid e f might have risen by heat, and gradually 
filled the vacuity V, while its boundary at c and d retired to m 
and n as e f ascended. In order to determine if this supposition 
was true, I placed AB vertically between two rectangular prisms 
of glass ; and having examined in succession the light reflected 
from the surfaces m p and n o, I found that it had suffered 
total reflexion, both from the side c d and the side g h of the 
vacuity, and consequently that the vacuity occupied the whole 
thicknesss of the cavity. After the heat was applied, the sides 
c d and g h continued equally luminous, and when c g and d h 
had retreated to m n andp o, as shewn in Fig. 10., it became 
quite manifest that the space m n o p was not filled with the 
expanded fluid , but with the fluid in the state of vapour . 
The coloured rings at first appeared both on the faces c d and 
g h , and when the whole was converted into vapour they dis- 
appeared, and the light reflected from both the surfaces m p , 
