220 
Proceedings of the Royal Society of Edinburgh. [Sess. 
condensation, so causing them to fall out of the air as raindrops. A few 
particles, owing to their affinity for water, will do this of their own accord 
if time be given. Many will fall out with a 1 per cent, expansion ; but 
it takes many expansions and a long time to remove them, as they are 
very small and fall slowly, and, further, many of them evaporate before 
reaching the bottom owing to the rise in the temperature of the air in the 
flask. A 2 per cent, expansion was therefore selected as a convenient and 
more practical lower limit to be used in the removal of the larger particles. 
We may now proceed in either of two ways to clear out the larger 
particles. We may set the expander to give a 2 per cent, expansion and make 
a number of quick expansions causing successive showers of very small drops, 
and continue this till no drops appear. But even with a 2 per cent, expansion 
tliis is a very slow process, and involves great waste of time. Much time can 
be saved, however, if we set the expander to give, say, an 8 per cent, ex- 
pansion ; and if, instead of making the expansion quickly, it be done slowly, 
the clearing-out process can be greatly accelerated. In this method of 
working, though we are making expansions of 8 per cent., practically no 
supersaturation is produced, because so long as there are plenty of large 
nuclei present there is no possibility of causing supersaturation ; the large 
amount of surface presented by the dense cloud prevents it. The advantage 
of the process is that we grow the cloud particles to such a size that they 
fall, and continue falling during the whole time the expansion is being 
made. To obtain the best result the expansion should begin with extreme 
slowness, and should then be quickened ; the very slow expansion picks 
out a small number of the largest particles, and these, becoming larger 
by the water deposited on them, continue to grow as the expansion is 
quickened. By this means fewer nuclei become active and grow larger, so 
ensuring their falling out. If a quick expansion is made at first, too many par- 
ticles become active, and so grow to a less size ; fewer therefore fall out before 
the air recovers its temperature. These slow, large expansions of 8 per cent, 
can be made till the number of nuclei becomes small, and may even be 
made at a certain speed till all are down ; but it is better to stop the large 
expansions when the particles are becoming few, and slide the stop on the 
expander down to 2 per cent. When this is done, a few quick expansions 
of 2 per cent, clear out all the remaining large particles, leaving only the 
smaller particles, which require higher supersaturations. During this 
process the test-flask must be frequently turned over to wet the interior, 
particularly when finishing clearing at each degree of expansion. 
While the clearing process is going on we get a rough indication of the 
number of large particles in the air tested. In these tests a record was kept 
