212 Proceedings of Royal Society of Edinburgh. [sess. 
moisture before tbe air is cooled to the dew-point, that is, condense 
vapour in unsaturated air. 
It is evident, therefore, that if the ordinary dust of our atmosphere 
has an affinity for vapour, and causes condensation to begin while 
the humidity of the air is some distance from the dew-point, that 
the condensed moisture will increase the size of the particles, and 
by so doing cause a decrease in the transparency of the air, under 
conditions which at present are not understood. It will be therefore 
necessary for us to bring a considerable amount of evidence to bear 
on this point, as the opinion generally held is that though condensa- 
tion may be caused to take place in unsaturated air by the artificial 
introduction into it of certain vapours and gases, yet condensation 
does not as a rule begin in our atmosphere till the air is saturated. 
The question then comes to be, Is there any evidence to show 
that condensation really does take place on the ordinary atmospheric 
dust, before the air is saturated, and if so at what stage of relative 
humidity does it begin to take place 1 The dust-counting apparatus 
evidently places in our hands a method of investigating this point; 
indeed, without it, it would be nearly hopeless to attempt its solu- 
tion. During the investigation this is one of the points to which 
attention was closely directed, and the conclusion come to, from 
all the evidence that has been collected so far, is that we have 
been far too much in the habit of supposing that condensation can 
only begin just at or immediately before saturation is reached. 
This investigation distinctly points to condensation beginning long 
before saturation is attained, and while the air is still dry enough 
to give a difference of some degrees between the wet and dry bulb 
thermometers. 
In order to get the necessary information to enable me to come 
to a conclusion on this point, it was customary to enter in the note- 
book, along with the number of particles observed, as many of the 
meteorological conditions as possible. Amongst other things noted 
were the direction of the wind, the temperature, the humidity, and 
the clearness of the air. These have been entered in the table, 
given with this paper. As already stated, the records of humidity 
and clearness are not very satisfactory, owing to the want of proper 
screens for taking the wet and dry bulb observation at the different 
stations and for the want of some satisfactory method of measuring 
