219 
1916-17.] On some Nuclei of Cloudy Condensation. 
however, that, though the stroke may be made almost instantaneously, 
expansion is slowest at the end, while it is better that it should be quickest 
at that stage. Still, in spite of that defect, condensation can be steadily 
produced on ions with about a 25 per cent, expansion. Professor Wilson’s 
latest figure is slightly lower than this, but his apparatus is better 
arranged for giving expansions at great rapidity and at nearly constant 
speeds. 
In what follows it will, be frequently necessary to speak of the size of 
the particles or nuclei. So before proceeding further, and to save constant 
qualification, it will be well to say what is meant here by the term “ size.” 
As already stated, all particles have an affinity for water, in the same way 
as all large solid surfaces have ; but we do not know that this surface 
attraction is the same for all kinds of particles : we have therefore to 
assume that it is. When the particles are large they all become centres 
of condensation with the slightest supersaturation ; but when they are 
very small this does not happen unless the air is to some degree super- 
saturated, and the smaller they are the higher the degree of supersaturation 
required to make them active. When, then, we use the word “ size,” we 
mean their condensing power, which does not necessarily mean their 
relative dimensions, though probably not far from it. 
The first use to which the new apparatus was put was an inquiry into 
the size of dust particles in the atmosphere. In my early communica- 
tions it is stated that almost all the particles in the air become active and 
are carried down in raindrops with a supersaturation of 2 per cent. ; and 
other observers have found the same thing. We were wrong, however, 
though correct for the air we experimented with — that is, the air 
of a room in which gas is burning. On testing air drawn directly from 
the outside, or air of a room in which no gas was burning, a different 
result was obtained, and particles requiring a higher degree of expansion 
to make them active were always found. 
Before going further it will be as well to explain in greater detail the 
method of working when making a test. The tube K is connected with 
a pipe leading through the frame of the window to the outer air. The 
stopcocks J and K are opened, and a number of strokes made with the 
pump P to clear out the air in the test-flask T and replace it with the air 
to be tested. Stopcocks J and K are then closed, and T is taken out of 
its support and moved so as to wet its inside walls and then replaced. The 
flask being now full of the air to be tested, we can proceed to clear out 
the larger particles so as to ascertain if there are any small ones. The 
method of clearing out the larger ones is to make them centres of 
