DUST, FOGS, AND CLOUDS. 341 
other terminates in a stop-cock. To the other opening of the stop-cock is 
securely fixed a tube tightly packed with cotton-wool. Some water is placed 
in the flask to moisten the air. If now the stop-cock is closed, and one or two 
strokes are made with the pump, so as to cool the air by expansion, it will be 
noticed that a fog immediately appears in the flask. This fog is fine textured, 
close grained, and will scarcely settle. Now pump out a good deal of the air 
from the flask, and allow air, filtered through the cotton-wool, to enter in its 
place. After the temperature equilibrium is established, again make one 
or two strokes with the pump. The fog again appears, but is now open- 
textured and coarse-grained. Repeat the process, admittmg more and more 
filtered air each time, and it will now be observed that the dense light fog 
which at first appeared gradually gives place to one coarser and coarser in 
texture, till at last no fog appears ; but on looking closely a fine rain, as in the 
previous experiment, will be seen showering down inside the flask. If the 
process is continued still further the rain ceases, there being no more “ free 
surfaces ” to form nuclei for rain drops. 
These two ways of experimenting, as might be expected, give exactly the 
same result, the conditions being so similar. In one the condensation is pro- 
duced by the cold air mixing with the hot steam ; in the other the “ saturated ” 
air is cooled by expansion in the flask. These experiments show clearly that 
when there is dust in the air the vapour condenses out in a visible form, but 
when no dust is present it remains in a supersaturated vaporous state. That 
the air, when no dust is present, is really supersaturated, is evident from the 
fact that when the dust particles become few, the fog particles are not only 
few, but are much heavier than when they were numerous, and also by their 
increasing in size as they fall through the air. Each falling particle becomes 
a “free surface,” at which the supersaturated vapour can condense and increase 
the size of the drop. Another way of showing the supersaturated condition 
of the air is to allow unfiltered air to enter in place of filtered air. The 
unfiltered air will at once show itself by the vapour condensing on its dust. 
It will be seen rising from the jet into the pure air, falling over and spreading 
itself over the bottom like a fountain of some viscous cloudy fluid. 
It was in the autumn of 1875, when studying the action of “free surfaces ” 
in water when changing from one state to another, that I first observed the 
conditions necessary for cloudy condensation. I knew that water could be 
cooled below the freezing-point without freezing. I was almost certain ice 
could be heated above the freezing-point without melting. I had shown that 
water could be heated above the boiling-poimt, and that the nature of the 
vessel in which it was boiled had no influence on the boiling-point, and all 
that was necessary for cooling the water below the freezing-point and for 
superheating the ice, and the water, was an absence of “free surfaces” at which 
