1904 - 5 .] Dr Aitken on Evaporation of Musk. 
897 
sufficiently large to allow of condensation taking place on them 
with a very slight degree of supersaturation. Now, it is well 
known that the vapour pressure at a convex surface is higher 
than at a flat surface, and that the quicker the curvature the 
higher is the vapour pressure. From this we see that the smaller 
the dust-particle the higher the supersaturation must be before 
it can become a nucleus of condensation ; so we would expect, 
if the solid particles were of all sizes down to molecular, that 
while the larger particles might become centres of condensation 
in very slightly supersaturated air, the smaller ones would 
require a higher degree of supersaturation, and higher in pro- 
portion to their smallness. There is, however, no evidence of 
this, the slightest degree of supersaturation being sufficient to 
cause condensation on even the smallest. A simple way of show- 
ing this is to take part of the cotton-wool out of the filter W shown 
in the figure, and only leave enough of it to keep back all but a 
few particles. By this process we get not only a few, but these 
few in all probability the smallest in the entering air, and yet they 
require but the very slightest expansion to make them visible ; and 
if we clear the air of them by making successive very slight 
expansions so as to carry them down as the nuclei of rain-drops, 
when all drops cease to make their appearance with these very 
slight expansions, we shall find that higher expansions produce 
no further condensation though accompanied by much higher 
supersaturations. The supersaturation may be made to go on 
increasing without any sign of condensation making its appearance 
till at a very high degree of supersaturation ; as C. T. R. Wilson 
has shown, condensation takes place on the ions in the air. As 
all the dust-particles become centres of condensation with an 
extremely small amount of supersaturation, and when got rid of 
no condensation takes place on dust-particles with higher degrees 
of supersaturation, we are entitled to conclude that the particles 
are not of all sizes down to the mblecular. 
There is another experiment which supports these conclusions. 
If we leave the apparatus above described for a considerable 
time— it may be for days — after filling the flask with ordinary 
dusty air, it will be found that the sifting process, due to 
gravitation, will remove all the larger particles first, till in 
PEOC. ROY. SOC. EDIN. — VOL. XXV. 57 
