J. AITKEN ON THE NUMBER OF DUST PARTICLES IN THE ATMOSPHERE. 15 



to cause condensation to take place on even the smallest particles in the air tested ; from 

 which we may conclude that the showers which unexpectedly took place from time 

 to time in the experiments described, where high expansions were used, were not due to 

 the presence of extremely small particles which had become active with the high degree 

 of supersaturation. 



As already stated, Sir William Thomson has shown that the phenomena of capillarity 

 give experimental proof that the vapour pressure at a concave surface is less than at 

 a flat one. In the experiment above described, we have what is, so far as I am aware, 

 the only experimental illustration of the complement to that conclusion, as it demonstrates 

 that the vapour tension at a convex surface is greater than that at a flat one, and also 

 that the smaller the body, that is, the quicker the curvature, the higher is the vapour 

 pressure at its surface. 



In describing the apparatus last used, it was stated that the air tested was drawn from 

 the gasometer. The object of this was that in making an experiment of the kind described, 

 it would have been an almost endless task to throw down all the particles in ordinary 

 air, by successive showers, in the manner described. The gasometer was therefore used 

 for mixing a small quantity of the ordinary air of the laboratory with a large amount of 

 filtered air. Working in this manner, we have samples of all the different kinds of dust 

 in the air, and a comparatively few expansions are sufficient to complete the series 

 necessary to throw down all the particles. 



In making experiments of this kind there is a necessary precaution we must refer to. 

 The air under test is kept saturated by the wet sides of the receiver. It is, however, 

 extremely difficult to keep the air from being supersaturated, because any difference in 

 temperature of the different parts of the receiver will cause supersaturation where the hot 

 saturated air mixes with the colder. Great care was therefore taken to keep the temper- 

 ature as uniform as possible, all radiation being cut off, as it is easy to see that if there 

 is any supersaturation due to difference of temperature, then a slight expansion would 

 have the effect of a greater expansion, and interfere with the correctness of our results. 

 In illustration of the effect of supersaturation from inequality in the temperature of 

 the different parts of the receiver, the following experiment may be made : — Introduce 

 into the receiver air in which there is only a very little dust. Now place the hand on 

 one side of the receiver to heat it slightly, and watch the result. Inside the receiver will 

 be seen a shower of condensed particles, which will keep falling so long as there is any 

 dust in the air and the hot side of the receiver keeps wet. After it stops we may make 

 an expansion by means of the pump, but no condensation will take place, the condensation 

 produced by the mixing of the hot and cold currents having previously thrown down all 

 the particles. 



Having then satisfied ourselves that a very small degree of expansion is sufficient to 

 give rise to a supersaturation great enough to cause even the finest dust particles to 

 become centres of condensation, we may dismiss the idea that the showers which gave 

 so much trouble by their appearance from time to time with high degrees of expansion 



