CHAPTER VII. 



I. MICROMETRIC MEASUREMENT OF FOG PARTICLES. 



Earlier Methods. 



1 . Before using the data computed for the dimensions of fog particles in the 

 reductions of my observations of atmospheric nucleation, it seemed expedient 

 to endeavor to obtain corroborative values by some straightforward method. 

 Aitken's dust counter had naturally suggested itself early in the course of my 

 work; but the results so obtained are essentially indirect, as the fog particles 

 are not themselves observed. It was necessary, therefore, to devise apparatus 

 by which the identical .fog particles of a given corona could be directly en- 

 trapped and held for examination. This was eventually accomplished in a way 

 admitting, apparently, both of the measurement of the diameters of the par- 

 ticles and of counting the number precipitated under known conditions. More- 

 over, the particles caught, however fine (even less than .0003 cm. in diameter), 

 can often be kept in place for observation for some time, so that microscopic 

 photography would appear to be at once applicable not only for the purpose of 

 obtaining size but number. 



Many investigations are thus suggested, as, for instance, a repetition- of 

 Thomson's method for determining the charge of an ion. Again, while the 

 corona gives merely the average size of the cloud particles, the microscope is 

 particularly available for indicating variations of diameter for the particles of 

 the same corona. In fact, the water particles of the coronas as caught on the 

 plate are not of a size; they are graded, and hence the nuclei are probably also 

 graded in size. 



2. Apparatus. Aitken's beautiful and highly ingenious instrument is well 

 adapted for the purposes for which it was designed. Apart from this, it will 

 furnish only an estimate of the dust contents sought. The droplets evaporate 

 too rapidly, and are often too numerous for exact counting. The need of 

 mixing atmospheric air with dust -free air with shaking and stirring is an inter- 

 ference with the nucleation. In fact, water nuclei may even be generated in 

 this way, possibly by the friction of air passing across damp surfaces. There is 

 the tendency of the plate after long use to fog permanently or to collect droplets 

 on its own account. Finally, it would be very difficult to remove the contents 

 of the coronal chamber to the dust -counter without reducing the nucleation 

 during the transfer. 



