m\. C. T. R. WILSON ON CONDENSATION OF WATER VAPOUR 
070 
Aitken* was able to ]'emove all the dust particles from saturated air by repeatedly 
increasing its volume by - 5 -^) of its initial amount. An even smaller expansion was 
found in these experiments to be sufficient for that purpose, but the time taken for 
the removal of the dust was naturally much shorter when larger expansions were 
used, on account of the larger size and more rapid fall of the drops in the latter case. 
If, after the dust has been removed in this way the successive expansions be made 
greater and greater, no visible effect is produced till the ratio of the final to the 
initial volume, is equal to about 1’252. When v^/v-^ exceeds this value a shower of 
drops is invariably produced. The drops are not very numerous, even with consider- 
ably greater expansions, yet, however often we expand the air, no diminution in the 
number of the drops can be detected. 
Now, when, owing to the presence of dust particles, a shower of similar density is 
produced with a smaller expansion, all the dust particles appear to be carried down 
with the water drops, and the next expansion produces no condensation. 
Thus, the nuclei which enables condensation to take place when the expansion 
exceeds the limit mentioned, are only present in small numbers at any given time, 
but as fast as they are removed they are replaced by others of the same kind. 
Expansion rripiired to produce Rain-like Condensation in Dust-free 
Saturated Air. 
A large number of observations must generally be made in order to obtain 
within narrow limits a single determination of the ratio of iq to when condensation 
just takes place. 
When expansions of comparatively small amount had ceased to cause condensation, 
each increase in volume was made considerably greater than the preceding, till a 
shower of drops was observed. Then an observation was made with the apparatus 
adjusted to give a rather smaller final volume (the initial volume remaining practically 
constant), and perhaps no condensation seen. By making in this way a series of 
alternately greater and smaller expansions of gradually diminishing difference, 
a stage was at length reached when the smallest measurable difference in the final 
volume was sufficient to determine whether condensation should result or not. 
A large number of experiments were made during the summer of 1895. Only the 
results of the last series of measurements then made are given in the table which 
follows. The apparatus has been improved from time to time and the later experi¬ 
ments were carried out exactly as described above. 
The mean of the results previously obtained for the critical value of tq/i’i, however, 
is practically identical with that given below, and all the determinations of this ratio 
have results lying between 1‘24 and 1'26. 
* ‘ Ediii. Traus.,’ 35, p. I. 
