358 JOHN AITKEN ON 
temperature there is no boiling and condensation, the change being perfectly 
continuous from the one state to the other, if under those conditions we can 
say there are two states. 
All the previously described experiments have been made at temperatures 
at which the condensed water was in a liquid state. It was now desirable that 
they should be made at lower temperatures, to see if the same conditions are 
necessary when the vapour condenses at temperatures below the “ freezing- 
point,” and passes from the gaseous to the solid state. The experiments were 
made with the air-pump arrangement of apparatus, the condensation being 
effected by the cooling produced by expansion in the receiver. In the first 
experiments the receiver was placed in a freezing mixture. They were, 
however, repeated under more favourable conditions during the severe cold of 
January last. The apparatus was removed to the open air and experiments 
made with it. The temperature at the time was 8° Fahr. The results were 
the same as at higher temperatures—cloudy condensation with unfiltered air, 
and no condensation when filtered air was used. ‘The amount of cloudiness 
produced was not so great as at higher temperatures. This is due to the 
smaller amount of vapour in the air at the lower temperature. 
I did not succeed in observing any of the optical phenomena produced by 
small crystals of ice in our atmosphere. This was probably due to the 
conditions under which the crystals in the experiment were produced. As the 
crystals were rapidly formed, there would not be time for the vapour molecules 
to arrange themselves in the simpler forms of crystallisation, but by being 
forcibly compelled to solidify, would form complicated shapes, which do not 
give rise to any peculiar optical phenomena. 
In the first part of this paper I have referred to the detection of small 
quantities of matter driven off by heat from pieces of iron, brass, and other 
kinds of matter. By the arrangement of apparatus then described, it was 
shown to be possible to detect the dust drawn off so small a piece of iron wire 
as the 745 of a grain. In later experiments in this direction, the apparatus 
has been entirely changed. In place of using the supersaturation produced by 
mixing steam and cold air, the air-pump arrangement of apparatus has been 
employed, and is found to work much more satisfactorily than the other. The 
impurities drawn off so small a piece of iron wire as the zo55 of a grain can 
with ease be detected with it. 
The arrangement of the apparatus for this purpose is as follows :—A glass 
flask provided with a tight-fitting stopper, through which pass two tubes, 
which rise to a short distance into the interior of the flask. One tube is 
connected to an air-pump, the other terminates in a stop-cock, to which is 
attached a cotton-wool filter. A piece of glass tube is introduced about the 
middle of the length of this pipe. Some water being placed in the flask, the 
