J. AITKEN ON THE NUMBER OF DUST PARTICLES IN THE ATMOSPHERE. 7 



allowed to pass from M into the flask, till the water in M falls to the index line. The 

 air which has been displaced by this process from the flask G enters the tube E, and moves 

 in the direction of the filter. One or two strokes of the pump and corresponding con- 

 densations in the manner explained soon gets rid of this impure air. After all condensa- 

 tion has ceased, some water, say 1 c.c, is allowed to run from the burette N into M, and 

 one stroke of the pump is made, care being taken that both stopcocks were previously 

 closed. The stopcock H is now opened till the water in M falls to the index line, after 

 which it is closed. By this process 1 c.c. of air is displaced from the flask G and sent 

 into the tube E. The stopcock F is now opened, and, as we had previously made a stroke 

 of the pump, air rushes in from the filter and carries the 1 c.c. of dusty air along with it 

 into the receiver. The stage is now carefully watched through the magnifying glass, and 

 one stroke of the pump made, the eye being steadily fixed on a selected square on the 

 stage. At once a shower of very fine rain is seen to fall, and the drops that fall on the 

 selected square are counted. If the stage is too hot, the drops rapidly evaporate, and 

 the attention must be confined to one square ; but in certain conditions, difficult as yet 

 to maintain, the drops remain visible a considerable time, so the number on a great many 

 squares can be counted, and an average quickly got ; but in most conditions of the stage 

 the test must be repeated frequently before a satisfactory average can be obtained. The 

 measured quantity of water must again be run into M, and its equivalent of air displaced 

 from G into the tube E, and carried into the test receiver, condensation produced, the 

 drops counted for say ten times, and the average taken. 



Too much dusty air must not be sent into the test receiver at one time, or the drops 

 will be too close for counting ; and further, the whole of them will not be thrown down ; 

 some of them will remain as dust, and will fall if a second expansion is made. Any error 

 from having too many particles present need never take place, as in practice there is 

 a constant check on the number admitted in the following manner : — After expansion 

 has been made and the drops counted, the stopcock F is opened, and filtered air allowed 

 to enter. A stroke of the pump is then made to prepare the receiver to draw in the air 

 for the next test, and if, while this stroke is being made, we watch the stage, we shall see 

 whether any drops fall or not. If none fall, then the amount of dusty air admitted has 

 been within the correct limits ; but, if any drops appear, then too much has been used, and 

 less water must be run into M for the next test. If no condensation took place after this 

 stroke was made, the measured quantity of dusty air is again passed into the tube E, and 

 carried into the receiver for another test. A complete check is thus kept on each test, 

 ensuring that all the particles in each quantity of air tested have been thrown down. 



Let us now take an example to show the manner of estimating the number of particles 

 in a sample of air by means of this apparatus. In addition to the number of drops per 

 square millimetre, the quantities required to complete the estimate are, the capacity of 

 the receiver A and of the pump B. In the apparatus sketched, A has a capacity of 500 

 c.c, but, as there are 50 c.c. of water in it, its air capacity is reduced to 450 c.c. If 

 into this pure air we introduce 1 c.c. of the air to be tested, the dusty air will be, 



