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



Introduction to the Method of Counting. 



I shall here give an account of a method I have devised and developed for counting the 

 number of dust particles in any sample of air ; and the figures which have been obtained 

 will, it is hoped, help to give deflniteness to our ideas as to the numbers of the particles 

 floating in the atmosphere, under both natural and artificial conditions. 



In approaching the solution of this problem, different ideas suggested themselves as to 

 the manner of counting the particles, but they all depended on the fundamental prin- 

 ciple of makiug them visible by supersaturating with vapour the air in which they were 

 floating. Under these conditions, each dust particle becomes a centre of condensation, 

 and what was invisible by itself becomes quite visible by its effects, as each invisible dust 

 particle becomes a centre of condensation, and grows into an easily visible drop. 



The simplest method of carrying out this idea was to put inside a glass receiver the 

 air to be tested along with some water, and to cause supersatu ration by reducing the 

 pressure in the receiver by means of a pump. We know that under these conditions a 

 foggy condensation takes place, the whole receiver becoming packed with small white par- 

 ticles, so small that, though visible to the naked eye, they float in the air. Each of the 

 fog particles so produced has a dust particle for a nucleus. Now, it might be possible to 

 devise some means of making an estimate of the number of these fog particles, by some 

 sort of microscopic measurement of the average distance between their centres. But, 

 supposing we had done this, we should by no means have solved the problem. These 

 fog particles do not represent all the dust particles present in the air ; only those having 

 the greatest affinity for the vapour being visible, a large proportion having no vapour 

 condensed on them. We might, after counting them, allow these visible particles to 

 settle, and then again produce supersaturation, which would reveal some more particles. 

 This process would, however, require to be repeated a great number of times before the 

 last of them would become visible, and as we cannot prevent the evaporation of some of 

 the drops, many of the particles would be counted twice ; altogether, the whole process 

 seemed too complicated and unsatisfactory to be worth attempting. 



The most hopeful suggestion was to put into the receiver only a very small and 

 measured quantity of the air to be tested, mixed with a large and known amount of per- 

 fectly filtered dust-free air, so that the particles would be so far apart that on super- 

 saturation being made, all of them would become centres of condensation, and thus by 

 one treatment every particle would become visible and be counted. On trial, the first 

 part of this plan was found to be of easy accomplishment ; by simply increasing the 

 proportion of dust-free air to the dusty air, a stage was arrived at in which every particle 

 was made visible by one supersaturation. 



Turning now to the method adopted of counting the drops. In experimenting on con- 

 densation in supersaturated air having dust particles in it, I had noticed that when the 

 1 (articles are few, the drops are large, that is, large comparatively to what they are when 

 the condensation takes place in ordinary air ; so large are they, that they are easily seen 



