PHENOMENA CONNECTED WITH CLOUDY CONDENSATION. 215 



it undistuibed. If we examine tlie cloudy coudeusatioii after a time we 

 shall find that a considerable change has taken place in its appearance. 

 The change is due to two causes. Part is due to the gradual descent 

 of the particles, by which a clear space is formed in tlie upper i)art of 

 the vessel. But it will also be observed that the clouding- in the lower 

 partis much thinner than it was at first. Probably part of this thin- 

 ning is due to some of the particles having fallen to the bottom of the 

 vessel, but this is not the i)rincipal cause of the change. The thinning 

 is due mainly to a reduction in the number of particles in the air, by 

 the smaller particles gradually becoming absorbed by the larger ones. 

 Tliis is caused by the vapor pressure at the surface of small particles 

 being greater than at the surface of larger ones, with the result that 

 the smaller particles evaporate in air of the same humidity in which 

 the larger ones are condensing vapor. 



We are now in a position to understand our first point, namely, why 

 the degree of super-saturation by which the condensation is j)roduced 

 should have an eff'ect on the appearance of the condensed va^ior. For 

 the study of tliis point the condensation produced by expansion is the 

 most convenient, as it is more under our control than the condensation 

 in steam jets. Suppose we take a glass flask connected with an air pump. 

 If we wet the inside of the flask and then fill it with unfiltered air, the 

 slightest expansion of the moist air by the pump will cause condensa- 

 tion to take place. But the density of the condensation which can be 

 produced by any degree of expansion will depend on the rate at which 

 the expansion is made. If the expansion be made very slowly, thecloud- 

 iug is very thin, but if it be made rapidly, it is very thick. If the 

 expansion be done slowly, the amount of super-saturation is only slight, 

 and only the largest dust particles come into action as active centers 

 of condensation ; and after a particle of dust has once become a nucleus, 

 it has tlien, in virtue of its size, an advantage over the particles which 

 have not begun to have vapor condensed on them. The result of this 

 is that so long as the degree of super-saturation is very slight, these 

 large particles relieve the tension, and if by any chance other dust 

 ])articles become active any reduction in the rate of condensation allows 

 the large particles, after they have relieved the tension, to rob the 

 small ones of their burden of water, so that a slow rate of condensation 

 always produces a snuill number of drops and a thin form of clouding. 



But now suppose you cause the expansion to be made rapidly; the 

 super- saturation then becomes much greater, as there is not time for 

 the water molecules to select a resting jdace, and the small number of 

 large dust particles can not relieve the tension, and the result is a 

 uuicli greater number of nuclei are forced into action. And all these 

 nuclei continue to grow so long as the super-saturation is kept up, but 

 the larger ones grow most. After the tendency to condense has begun 

 to diminish, those particles which have accunuilated least are the first to 

 feel the change and cease to grow, while the larger ones are still accumu- 



