348 JOHN AITKEN ON 
drawn through for a long time, there was not the slightest sign of the air 
becoming purer. To make sure the fogging was due to the flame, the gas was 
turned off, and combustion stopped, while the circulation was kept up. In a 
very short time after this was done, the air showed a marked decrease in 
cloudiness, and after a time became pure. 
This method of testing the effect of combustion does not seem, at first 
sight, the best. The intention was to have, first, circulated the air till perfectly 
pure, and steam gave no cloudiness, and then to light the gas and see the 
effect. The difficulty of working in this way was that I could not light the 
gas without introducing a disturbing element. It was intended to have lit the 
gas by means of an incandescent platinum wire, but on testing the effect of 
the hot wire alone, it was found to make the air active, and powerfully fog- 
producing. By highly heating the wire, it was possible to make it less active 
at lower temperatures, but the temperature produced by igniting the gas would 
again make it active. 
I have great hesitation in coming to any conclusion from this experiment. 
At first sight it would look as if the small flame is very far from being a dust 
destroyer, and is on the contrary a very active producer of it. It will be 
remembered that the flame was fed with filtered air, and the result of the 
combustion of filtered air and dustless gas is an intensely fog-producing atmo- 
sphere, and that the fogging is due to dust cannot, I think, be doubted, as the 
products of combustion, when filtered, give no cloudiness when steam is added. 
Yet the question may be asked, Was the dust produced by the combustion ? 
It seems almost possible it might be the result of soda driven off by the heat 
from the glass jet. 
On the 8th and 12th of January this experiment was repeated. The glass 
jet at which the gas was burned being removed, and a platinum one put in its 
place. Platinum was selected because it was thought in the highest degree 
improbable that any nuclei could be driven off the platinum by the heat of 
the gas flame. After the jet was fixed in its place it was highly heated to 
thoroughly cleanse and make it inactive at the lower temperature produced by 
the flame. The gas was lit, and the receiver then put in its place, and the 
supply of filtered air drawn through the apparatus. The result was the same 
as before. Increase of fogging on the gas being lighted, and the fogging con- 
tinued so long as the gas was kept burning, and only stopped when the flame 
was put out. 
There seemed a possibility that the fogging might be due to some residual 
motes still remaining in the receiver getting into the flame and being broken up 
by the heat into a great number of parts. The experiment was accordingly 
varied to meet this. A fine platinum wire, which could be heated by a battery, 
was arranged so that the gas might be lit by it without opening the receiver, 
: 
