2S Sizes of Water Particles producing Cloudy Condensation. 



corresponds to a march of the field from yellows to reds and 

 purples. The colours thus follow each other around the 

 circle, apparently diametrically opposed in position. The 

 axial yellows, oranges, etc., of the first order, seen so 

 splendidly in the steam-jet, were not obtained with the drum 

 in any case. 



The data found may be sustained by independent estimates. 

 It may be noted that the order of size of particle is about ten 

 times larger than would follow if the same (axial) colour 

 were produced by interference in thin plates. 



When the coronas begin to vanish, the increased rapidity 

 of removal of particles from gravitational subsidence is 

 noticeable. In case of sulphur nuclei there were about 33 

 particles per linear centim. initially, and 5 particles per 

 centim. finally ; in case of punk nuclei, 44 per centim. 

 initially and 4 per centim. finally. The attempt to use 

 subsidence methods was abandoned. Such data are seriously 

 complicated by the evaporation at the surface of the fog and 

 by the fact that the air cleared by the warmer sides of the 

 vessel rises convectively to the top. Both causes produce 

 apparent but false subsidence. Again, the use of Kelvin's 

 vapour- tension equation breaks down quantitatively for the 

 present purposes, in practice. 



With particles as near together as they are intially (say 

 •02 to *03 centim.) one might anticipate some optical effect 

 due to their mutual action. The normal or white-centred 

 coronas do not appear until the particles are somewhat less 

 than one millimetre apart. The edge of the cube containing 

 one particle and the diameter of the particle are proportional 

 quantities, the ratio being N -1/3 d — l , or about 87 for the 

 globe. The density of suspended water is thus about 

 •6 X 10- 6 . 



For the drum the uniform successive exhaustions were 

 smaller and the ratio of diameter of particle to the edge of 

 the cube was about 113 ; but the length of the column (now 

 187 centim. as compared with 30 centim. for the globe) more 

 than makes up for the difference. Therefore a cube whose 

 edges are *04 centim., at the outset contains one particle 

 •00036 centim. in diameter ; and (113) 2 cubes of this kind 

 on end would, virtually, fill up the square content normal to 

 the line of vision to the first order of small quantities. In 

 other words, about 5 metres of column would be needed to 

 effectively blot out the white light. At the end of the 

 experiment, a column '096(113) 2 = 1280 centim. in length 

 would be needed. In the above experiments with axial 

 colours, the column was nearly 2 metres long, and this proved 



