1888 - 89 .] Mr John Aitken on Dust Particles. 
155 
length of the tube M is such that its elasticity keeps the diaphragm 
I at its top position, and against the stop R. The pure and impure 
airs having heen thoroughly mixed, expansion is made, a shower of 
rain produced, and the drops counted. 
We must now consider what is the best quantity of dusty air to 
he sent into the receiver for making a test. When beginning a 
new test, we have to be guided by experience as to what is likely 
to he the correct quantity. Suppose we think that 2 c.c. will he 
enough, that quantity is accordingly measured and tested, and from 
the density of the condensation produced by that quantity, we get 
an idea as to whether it is too little or too much. But what is to 
too little and what too much 1 A little experience soon settles this 
point ; but I may state that if more than 5 drops fall per square 
mm., there is too much dust, not only because when the number is 
much above 5 there is a difficulty in counting them before they 
evaporate, if the stage be slightly hot, hut also because, with so 
large a number, we cannot he quite sure that all the particles have 
heen thrown down. Suppose, for instance, that 10 drops fell per 
square mm.; if we now admit only filtered air, and again make an 
expansion, we shall find that some drops w T ill make their appearance, 
showing that some particles have escaped the first condensation. It 
has, therefore, been the practice to limit the maximum number of 
drops to 5 per square mm. With that number no drops appear 
on a second expansion being made. The lower limit, however, is 
not so definite ; there is nothing in the conditions limiting us here ; 
it is simply a question of convenience ; 1 per square mm. makes 
a fairly good lower limit. When working with that number, I 
generally use 4 squares instead of 1. By taking the number that 
falls on the 4 squares, we get a better average. The number so 
obtained is multiplied by 25 to get the number per c.c. We are 
not, however, limited to 1 or even 4 squares, occasionally 9 squares 
have heen used, and the number that fell on these 9 observed. 
Working with so large a surface requires some care. We have first 
to select a part of the stage where there are 9 squares all perfect 
and spotless. The eye is steadily kept on the square of 9 small 
ones, and a little practice enables us to count the number that fall 
on that area. I need not say that 9 squares are only used when 
the drops are very few, say 6 or 7, over the whole area ; if more fall, 
