216 Proceedings of Royal Society of Edinburgh. [sess. 
Number 
of Particles. 
Date. 
State of Air. 
550 
814 
1000 
1900 
15th, at 11.30 a.m. 
10th, ,, 
6th, at 10.30 A.M. 
15th, at 2 p.m. 
Clear. 
Medium. 
Thick. 
Thick. 
It will be seen that with a humidity corresponding to a depression 
of 4°, that if there were 550 particles the air was clear, and that 
as the number of particles increased the transparency decreased. 
Again, take the 12th and 13th; on both of these days the humidity 
was the same, namely, 5°, but the one was clear which had only 710 
particles in it, while the other was thick owing to there being 3000 
particles present Again, there was the same humidity on the 
morning of the 5th as at 1.30 p.m. on the 6th, but on the 6th the 
air was clear owing to there being only about 300 particles in it, 
while on the previous day it was thick owing to the presence of 
more than four times that number. 
These figures all tend to confirm the conclusion previously arrived 
at from the Colmonell observations, namely, that the transparency of 
the air depends on the number of particles of dust in it. We shall 
now see how far the Kingairloch observations support the other 
conclusion, that the amount of moisture in the air has also an effect 
by increasing the size of the particles, and so reducing the 
transparency. We shall use those Kingairloch figures and see if this 
affinity of the dust for moisture has practically any effect on the 
transparency of the atmosphere. 
When the observations began on the 3rd of July there was a 
good deal of dust in the air, about 3000 per c.c. This degree of 
impurity continued next day, but fell to about one-half on the 5th 
and to one-third on the morning of the 6th. Now, though the dust 
gradually decreased during these days, it will be observed that the 
thickness gradually increased from a haze on the 3rd to thick on 
the 6th. The reason for this decrease in the transparency was the 
increase in the humidity which took place with the decrease in the 
dust, the humidity falling from a depression of 13° on the 3rd to 
4° on the 6th. Here the decreased transparency produced hy in- 
crease in size due to humidity more than counteracted the effect of the 
decrease in numbers, showing how powerful the influence of humi- 
dity is, even when the humidity is some degrees from saturation. 
