356 JOHN AITKEN ON 
ment in which I have been certain there might not have been some nuclei 
present. The first step in this direction was to test the action of the filter 
through which the air passed. All the cotton-wool was removed from the 
filter and a fresh quantity put im. At first only a thin layer was used, and 
its effect tested, noting the degree of cloudy condensation produced. More 
cotton-wool was then put over the first layer, and the improvement noted. 
Fresh quantities were added till no improvemeut was observed. Then double 
the total quantity was put in, and the filter was now considered to be doing all 
that cotton-wool could do to purify the air of the receiver from dust. 
The result was—when a small quantity of steam was blown into the 
receiver there was no cloudy condensation whatever ; the receiver remained 
perfectly clear. But when the steam valve was opened wider and more steam 
allowed to enter, although no effect was noticed at first, yet after a time the 
vapour became so supersaturated that it condensed and fell as fine rain. Ifa 
still greater amount of steam was blown in, then it was seen condensing on 
entering the receiver, and the falling rainy condensation was seen tossed about 
by the rush of the entering steam. 
Attention was now directed to the steam. It seemed possible that nuclei 
might be given off from the hot sides of the boiler, and from the hot parts from 
which the steam was rising. To prevent any nuclei which might be formed in 
this way from entering the receiver, the end of the steam pipe inside the 
receiver was covered with a cotton-wool filter, The result was, however, as 
before, with little steam, no condensation, with much steam, rainy condensa- 
tion. On account of the tendency of the cotton-wool to get wetted by the 
steam, the action of the filter did not seem satisfactory, some parts getting wet 
and stopping the passage of the steam, and throwing all the duty on the weak 
parts. The experiment was accordingly arranged in the following way :—The 
steam was generated in a glass flask. This flask, filled with water, was placed 
in a vessel full of water, kept boiling during the experiment. In order to make 
the water in the glass flask boil, or rather evaporate, under these conditions, a 
stream of filtered air was blown through it, and the mixture of air and vapour 
blown into the receiver. Again the result was as before—rainy condensation 
when highly supersaturated. By this last arrangement it seems impossible any 
nuclei could be given off from the vessel in which the water was boiled, and 
the fine drops given off by the bubbling of the air and the vapour in the flask 
are probably all caught on the sides of the pipes, because if they did enter they 
would form nuclei in very slightly supersaturated, as well as in highly super- 
saturated vapour. We may therefore conclude from these experiments that 
the nuclei of the rainy condensation in highly supersaturated vapour are either 
some fine form of dust which the cotton-wool cannot keep back, or are pro- 
duced by the vapour molecules combining together without a nucleus. 
