Condensation, and on the contrast of Colors. 93 



The same method applied to the results obtained from the 

 double drum showed 2/10 4cm as the diameter of the particles in 

 the first exhaustions. The smaller size here corresponds to less 

 exhaustion. 



A final estimate* is obtainable from the size of the normal 

 coronas after say twenty or thirty exhaustions. Estimating this 

 as about twice as large in diameter as the ordinary lycopodium 

 corona, if the particles of the latter measure -0032 cm , one may 

 rate the residual water globule at -001 cm . Indeed Fraunhofer 

 and Kaemtz's measurements showed particles in lunar coronas 

 as large as *0017 to -0033 cm , depending on the season. Thus if 

 the final size is *001 cm . the initial size must have been -0003 to 

 •0002 cm , corresponding to twenty or thirty exhaustions. Similar 

 results are deducible from the drum and they are thus again in 

 agreement with the preceding estimates. Finally the normal 

 coronas may be measured absolutely and the dimensions com- 

 puted from the deviation on diffraction. This is the practical 

 plan with which I am now engaged, and I will therefore waive 

 further discussion. 



If the independent estimates just stated be summarized, the 

 data are clearly of an order ten times greater than would follow T 

 if the axial colors of the drum or the steam jet were produced 

 by interference of thin plates, granting that the old vesicular 

 theory of atmospheric condensation is disproved. I admit that 

 data as large as those found for the water particles are con- 

 trary to my expectations. 



11. Size of nuclei. — This has been variously estimated by 

 the aid of Kelvin's vapor pressure equation, successively modi- 

 fied by the younger Helmholtz,* and by C. T. R. Wilson. f 

 Naturally the nuclei are supposed to be of the same size and 

 the supersaturation carried far enough to condense water on 

 each. Helmholtz found 15/10 6 to 26/10 6cm as the size of his 

 nuclei. Wilson finds 8*7/10 8 for the case of rain-like conden- 

 sation, and 6*4:/10 8 and 5*9/10 8cm for cloudy condensation, 

 foggy and colored. The above data for the globe similarly 

 interpreted give 8/10 8 and 18/10 8 as the smallest radii of the 

 nuclei on which condensation takes place. 



The small size of the nucleus obtained in this way is start- 

 ling, but as the method involves a huge extrapolation from the 

 radius of capillary action (say 5 X 10~ 6cm ) almost as far as the 

 molecular diameter, it cannot be received with much confidence 

 and the size of the nucleus must be left in abeyance. 



On the other hand, the size of the water particle is suffi- 

 ciently large to admit of the application of Kelvin's equation. 

 Moreover, there is here a mere accretion of water upon water. 



* R. v. Helmholtz, Wied. Ann., vol. xxvii, p. 526, 1886. 



f C. T. R. Wilson, Phil. Trans., London, vol. clxxxix, p. 306-30*7, 1897. 



