234 Proceedings of Royal Society of Edinburgh. [sess. 
cumulus clouds; the near ones were always snowy white, while it was 
only the distant ones that were tarnished yellow, showing that the 
light came to these clouds unchanged, and it was only the air between 
the far distant clouds and the observer that tarnished them yellow, 
it required a great distance at this elevation to give even a slight 
colouring. It seems probable that the lurid light of thunder clouds 
owes its colouring to the dust and humidity of the lower air being 
excessive at the time, and that these clouds are really white, if we 
saw them near enough, it being the heavy moist air near the earth 
which changes the white to a reddish colour in its passage to the 
observer. We seem therefore to have good reason for supposing 
that the colouring at sunrise and sunset will be more brilliant when 
seen from the valley than from the mountain top. 
Dust and Wind. 
It may now be asked, Are there any other meteorological pheno- 
mena connected with the amount of dust in our atmosphere 1 It is 
evidently rather soon to begin drawing conclusions on the relation 
beween dust and many of the complex phenomena in nature, as the 
number of observations yet taken is far too small, and many of them 
are not free from local influences, which greatly reduce their value. 
There are, however, one or two points to which we may here refer; 
but it must be clearly understood that anything we have to say in 
this part of the paper cannot be taken as indicating anything 
final. They are given more as suggestions as to some of the many 
points in connection with our subject which await investigation. 
One of the first things a meteorologist is likely to ask is, Do you 
find any relation between the amount of dust in our atmosphere 
and the distribution of pressure ? Has the air in cyclonic and anti- 
cyclonic areas the same amount of dust 1 The answer to the question 
is easy and direct, but the interpretation of the answer is more diffi- 
cult. There is no doubt from the observations here given that, as 
a rule there is less dust in cyclonic than in anticyclonic areas. For 
instance, when the observations were made at Kingairloch, on all 
days on which the number of particles was low, a cyclone was 
passing near, and on all days the number was great the circulation 
was anticyclonic or complicated. 
To illustrate this, the diagram given on Plate No. 1 has been pre- 
' 
