IN THE PRESENCE OF DUST-FREE AIR AND OTHER GASES. 
283 
With expansions greater than this, the density of the fog appears to go on 
increasing with great rapidity as the expansion is increased. It is now convenient 
to remove the condensing lens and examine the fog by looking directly through it at 
the gas-flame. Coloured diffraction rings make their appearance when is about 
1'38 and they increase rapidly in brilliancy and size as the expansion is made greater 
and greater. 
Before v.^jv^ reaches 1 '40, the region within the fii'st ring, which is whitish with 
smaller expansions, becomes brightly coloured. With greater expansions, the rings 
rapidly become so large that the colour corresponding to the central part of the field 
fills the whole tube. 
The colour phenomena beyond this stage are surprisingly definite. They are best 
observed by looking through the cloud chamber slighly to one side of the source of 
light, which ought now to be made as bright as possible, and have a black 
background. 
If be between 1‘41 and 1'42, brilliant greens and blue-greens are seen. At 
about 1’42 there is a very rapid change from blue to red through violet. The violet 
appears only for a very small range of expansion, a change of one or two millimetres 
in the initial pressure being sufficient to complete the change from blue to red. 
As the expansion is further increased, the colour passes from red through yellow 
to white. With expansions greater than about 1’44, the fog is always white with a 
greenish or bluish tinge. 
The whole of these colour phenomena, it will be seen, are confined to quite a narrow 
range of expansions. Below 1*38 the drops are too large and few ; and above 1’44 
they appear to be too small to produce the colours. 
Colours of exactly the same kind were obtained by Kiessling* and by AitkenI by 
expanding ordinary moist unfiltei’ed air ; but in the reverse order, pale yellow being 
the first to appear, followed with increasing expansions by a reddish colour, then by 
blue, and then green. 
The explanation of the difference plainly, is that in their experiments the number 
of the drops was determined by the number of “dust” particles present, and 
increased expransion caused a larger quantity of water to condense on each particle. 
Increasing the expansion thus increased the size of the drops. Now in the 
experiments here described, the grrjater the expansion, the smaller appear to be the 
resulting drops, which indicates that as the supiersaturation is increased, a larger 
number of nuclei come into pilay, so that each receivt^s a smaller share of the water 
which condenses. 
Similar phenomena are exhibited by light transmitted through a steam-jet under 
* Gotting., ‘ Naclir.,’ 1884, p. 226. 
t ‘ Proc. Roy Soc.,’ vol. 51, p. 422, 1892. 
2 o 2 
