200 
Proceedings of the Royal Society of Edinburgh. [Sess. 
with diffused light has not been identified. It does not seem to be carbonic 
acid gas; at least, that gas liberated from its solution in water has no 
effect. It may, however, be contended that a gas in that condition is, so 
to speak, dead, and not so chemically active as one freshly prepared. The 
other impurity may possibly be carbonic oxide ; but as there seems to be 
no method of preparing that gas free from nuclei, or other objectionable 
impurities, direct tests have not been made. The following experiments, 
however, were tried with the products of perfect and imperfect combustion, 
in the hope that they might throw some light on this point. The air 
entering the filter was drawn from the top of a flame of a bunsen burner. 
The supply of air to the burner was first set to give perfect combustion — 
that is, non-luminous. In that condition almost all the carbon would be 
in the form of C0 2 . These products were drawn into the apparatus 
through the filter, mixed with S0 2 , and sunned. On testing, only a slight 
condensation took place, such as was obtained without the products, the 
air on that day being very pure, as it came from the north-west. The air- 
supply to the burner was now cut off, so as to produce a luminous flame, 
and its products were now drawn into the apparatus, sunned, and tested 
as before, when a very dense condensation resulted. These tests would 
seem to indicate that the C0 2 of perfect combustion is not the impurity 
in the air required to increase the activity of the S0 2 and make it a nucleus- 
producer with filtered sunlight. The experiments also indicate that it is 
probably the CO of imperfect combustion that is the partner in the forma- 
tion of the nuclei. This conclusion is evidently not final, as there are so 
many other gases besides CO given off in imperfect combustion, and it may be 
one of them. The above experiment was made in another way for con- 
firmation. The windows of the room were closed and a bunsen flame 
burned for a short time and the products mixed with the air of the room. 
On testing this air with S0 2 it showed little or no increase in condensing 
power over the slight condensation it previously gave. Ordinary gas-jets 
were now lighted and allowed to burn a short time, and the air in the 
room mixed. On testing this air it gave a great increase in the density 
of condensation. Both these ways of testing show that the impurity 
required to make S0 2 active is some product of imperfect combustion. 
Radio-activity. 
Another outside influence which acts on S0 2 and makes it a nucleus- 
producer is radio-activity. If we keep the flask S in the dark and bring 
near it a small quantity of any radio-active substance, the acid becomes a 
