196 Proceedings of the Royal Society of Edinburgh. [Sess. 
was used. In this apparatus the successive condensations always take 
place in the same air. The air is enclosed at the top of a vertical tube by 
a water seal, the same air being always subjected to the successive ex- 
pansions and condensations. When working with this apparatus, a little 
of the solution of S0 2 was put in the water to keep up a supply of 
sulphurous vapour. After all dust particles had been got rid of by 
successive expansions and condensations, so that no condensation took place 
in the dark with slight expansion, the tube was sunned, when, as was 
previously found, condensation took place on expansion. But most un- 
expectedly the condensation was observed to decrease in density with every 
succeeding shower, till at last all condensation ceased ; though there was 
plenty of S0 2 in the air, the sun no longer had any effect on it. Why this 
difference ? In the other apparatus condensation always followed after 
sunning. The conditions in the two methods of testing are, however, quite 
different. In the flask arrangement, fig. 1, fresh supplies of filtered air were 
constantly supplied, but in the tube apparatus the same air was used over 
and over again ; so that it looked as if something else was necessary to 
make the S0 2 active with sunshine, and as if this something was all used 
up after a number of showers had been made. The stoppage of all action 
seemed to indicate that the repeated showers had removed something from 
the air, and that the sun was unable to form nuclei from S0 2 in air so purified. 
If the above explanation was correct, then it might be possible to take 
this something out of the ordinary impure air and make it also incapable 
of forming nuclei with S0 2 and sunshine when using the apparatus shown 
in fig. 1 — that is, with successive quantities of filtered but in this case 
purified air, in place of simply filtered air. For the purpose of purifying 
as well as filtering the air, a cotton-wool filter saturated with caustic soda * 
was tried, and it was found that it purified the air; the S0 2 no longer 
* I have previously advocated the filter method of purifying gases in place of bubbling 
them through solutions, as its effects are far more powerful than those of the ordinary one, 
which has a comparatively small action unless the gas has a very strong affinity for the 
contents of the solution ; and even then it is far from being as perfect as the filtering process, 
and is practically useless for the experiments here described. Caustic solutions are not 
pleasant to work with, and the fingers may suffer if one attempts to pack the filter with wool 
soaked in caustic. The following method of preparing such a filter, however, presents no 
difficulties. The filter is first packed with cotton-wool, or any other suitable substance, in 
the dry state. The narrow end of the filter is then fitted into an india-rubber stopper which 
fits any bottle ; an air-pump is also connected with the bottle. Some of the caustic solution 
is now poured into the upper open end of the filter and the pump worked. This draws the 
solution down through the cotton and thoroughly wets it. The pump should be worked 
quickly at the finish to draw out as much solution as possible, otherwise there will be bubbling 
produced by the filtered air, and nuclei made. These, however, may be checked by a second 
filter. The solution in the filter can be renewed when necessary in the same way. 
