340 JOHN AITKEN ON 
to take place at the “freezing” and “ boiling points.” At present we do not 
know at what temperatures these changes take place when no “ free surfaces ” 
are present. Indeed, we are not certain that it is possible for these changes to 
take place at all, save in the presence of a “ free surface.” 
Returning now to the condensation of the water vapour, we see from the 
experiments given that precisely the same conditions are necessary for the 
condensation of a vapour as for its formation. Molecules of vapour do not 
combine with each other, and form a particle of fog or mist ; but a “ free surface ” 
must be present for them to condense upon. The vapour accordingly condenses 
on the dust suspended in the air, because the dust particles form “ free surfaces ” 
at which the condensation can take place at a higher temperature than where 
they are not present. Where there is abundance of dust there is abundance 
of ‘free surfaces,” and the visible condensed vapour forms a dense cloud ; but 
where there are no dust particles present there are no “ free surfaces,” and no 
vapour is condensed into its visible form, but remains in a supersaturated 
vaporous condition till the circulation brings it in contact with the “free 
surfaces” of the sides of the receiver, where it is condensed. 
We see, then, that each fog particle in the experiment was built on a dust 
particle. This indicates an enormous number of dust particles in the air. We 
must not, however, suppose that the particles of that dense fog we saw in the 
receiver A represented all the dust particles in the air experimented on. The 
experiment indicated an extremely foul state of the air indeed, but it does not 
tell the whole truth. Those fog particles only represent a small part of the 
dust particles present. That this is really the case is easily shown in the 
following way :—Let as much steam be blown in as will form a dense fog. 
Now allow this fog to settle, but do not allow any dusty air to enter. After 
the fog has settled blow in more steam. Again you will find a dense fog con- 
densed on the dust which escaped the first condensation. Allow this again to 
settle, and repeat the process a number of times, when you will find, after many 
repetitions, that there is still fog forming. But it will also be noticed that after 
each condensation the fog becomes less and less dense, till at last it ceases to 
appear as fog; but on closely looking into the receiver the condensed vapour 
will be seen falling as fine rain. When the steam was blown in the first time 
the fog was very fine textured; each particle was so small it floated easily 
in the air. After each condensation the fog became less dense; it at the same 
time became more coarse-grained and heavier, and was seen falling slowly. 
Near the end, no fog was visible, and nothing but a fine rain to be seen 
falling. If the air was still further purified, even the rain seemed to cease. 
This experiment may be made in another way. A large globular glass 
flask is provided, having a tight-fitting indiarubber stopper, through which 
pass two pipes. One of these pipes is connected to an air-pump, and the 
