252 
Proceedings of the Royal Society 
that there will be an apparent motion of the bubble towards the 
hotter end of the cavity, — an apparent motion only, for in reality it 
is not one and the same mass of vapour which is travelling through 
the liquid. Any such identical bubble has only a momentary 
existence. It is continually being changed into a new bubble in a 
new position by the accretion of vapour on the side A, and by the 
restoration of vapour to the liquid state on the side B ; and the 
change of place of the existing bubble is in the direction from B to 
A, or in a direction opposite to that of the heat current. Such a 
motion, it is scarcely necessary to remark, agrees with that which 
is actually observed. 
The action thus set up in a vapour bubble is precisely that which 
takes place in Wollaston's cryophorus, where vapour, rapidly 
generated at the hotter, is recondensed into liquid and frozen at the 
colder end of the apparatus. Suppose a cryophorus to consist 
simply of a cylindrical tube placed vertically and cooled by a freez- 
ing mixture at its upper end. The cavity occupied by vapour will 
then suffer a continual displacement downwards ; for the surface 
of the water, which is its lower boundary, is being depressed through 
loss by evaporation, while the glass at the top, which was at first 
its upper boundary, is becoming coated with ice of constantly in- 
creasing thickness. The downward displacement of the cavity of 
such a cryophorus may serve to illustrate that of a bubble in a 
liquid heated from below. But it may seem that any movement 
thus produced would be far too slow to displace a bubble downwards, 
which was rising freely through a liquid. In considering such an 
objection to the proposed explanation of the motions of bubbles in 
the cavities of crystals, when these motions take place vertically, 
or otherwise than in a horizontal direction, it is to be borne in mind, 
