251 
of Edinburgh, Session 1872 - 73 . 
therefore hotter than the spar. The freezing mixture at the con- 
clusion of the experiment had still a temperature of — 5° C. or 23° F. 
I found that the direction of motion of the hubbies was the 
same whether a heated copper wire was held above or below the 
Iceland spar, with the slide resting horizontally on the stage 
of the microscope. I fully verified Mr Sang’s statement re- 
garding the motion of the bubbles when the microscope is inclined. 
Placing its tube horizontally, so that the face of the stage and the 
glass slide were vertical, the bubbles, of course, all rose to the tops 
of their cavities. A hot copper wire or silver coin touching the 
surface of the spar at a point on a lower level than that of one of 
the cavities, instantly drew the bubble down to the bottom. The 
motion in a vertical plane with a tolerably hot wire seemed almost 
as brisk as it had been in a horizontal direction, so as to indicate 
that the effect of hydrostatic pressure, due to gravity, on the minute 
bubble was trifling as compared with the action set up by the heat 
current. 
Considering the enormous dilatability by heat of liquids, which, 
under ordinary conditions of temperature and pressure, are perma- 
nent gases, I at first thought the motions of the bubbles might be 
due to currents caused by unequal heating of the liquid on opposite 
sides of the cavities. The heat flowing from a piece of metal brought 
near a cavity would cause dilatation of the liquid on the nearer side. 
A current would then evidently flow along the upper surface of the 
cavity away from the heated metal, and carry a bubble resting at 
the top in that direction. But this is precisely the reverse of the 
motion actually observed ; so admitting, as we can scarcely doubt, 
the setting up of a current due to unequal heating, there must be 
some other and more energetic action at work, causing a real or 
apparent motion against any such current ; and this I take to be 
rapid evaporation and condensation of the liquid on opposite sides 
of the bubble. Suppose A B to be a bubble floating in a cavity 
through which a heat current passes in the direction A B. A state 
of equilibrium of the bubble is then evidently impossible. Liquid 
will evaporate from the hotter side A of the bubble, and vapour will 
condense into liquid on the colder side B. The liquid surface A 
will then, by continual loss, travel in the direction B A, and the 
surface B will by continual gain follow A in the same direction ; so 
