150 
Proceedings of Royal Society of Edinburgh. [sess. 
same number of drops, then the number of particles would have 
been one-tenth, or 10,000. 
We can test air having any of these quantities in it with 
the small cylinder measure. But suppose that, with 10 c.c. of 
the dusty air, the particles are too few for easy counting, then we 
will require to send more than 10 c.c. into the receiver. To do this, 
the stopcock F is turned against the stop H 3 , the stopcock of the 
receiver is closed and a stroke of the pump is made ; this takes a 
certain quantity of air out of the receiver, and, on opening the 
receiver stopcock, air from the source to be tested rushes in to 
supply its place. This is mixed with the air in the receiver, expan- 
sion is made, and the drops are counted, all as usual. If the number 
of particles has not been too great, all of them will have been thrown 
down, and the receiver stopcock may be again opened, and another 
quantity taken in, mixed, expanded, and counted, and so on till 
the required number of tests has been made to give a good average. 
We must now know the quantity of dusty air taken in at each of 
these tests before we can calculate the number of particles in the air. 
The capacity of the receiver is 350 c.c. and of the pump 150 c.c. The 
pump therefore takes out °f the whole air of the receiver — 
350 
3-33 
= 105 
c.c. 
Practically only 100 c.c. of air is required to replace that taken out 
by the pump. The reason of this is, the theoretical quantity does 
not really come out with each stroke of the pump, owing to the 
pressure required to open the pump valve preventing the escape of 
the last 5 c.c., and probably partly owing to the expansion producing 
a greater cooling effect on the larger mass of air in the receiver than 
on the smaller one in the pump, — the smaller mass more quickly 
recovering its heat from the enclosure. 
When working with the apparatus in this way, at each test there 
is taken into the receiver 100 c.c. of the air to be tested; this is 
mixed with 250 c.c. of pure air, and the whole is then expanded to 
500 c.c., so that the number counted on the stage, in this case, 
requires to be multiplied by only 5 to get the number of particles in 
the air tested. It may be as well to note that, after making the 
s troke of the pump, the piston must be pushed back again ; other- 
