A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. I05 



away n^oisture), rciroving the excess while hot by filter paper and when cold 

 submerging the plate in petroleum for transparency. With solution of vaseline 

 in benzine, etc., I have been much less successful. Damar varnish and tuqien- 

 tine was much used in the final work. When drops are to be counted by the 

 method given below, the oil film m.ust be practically solid; othei-wise the capil- 

 lary forces produce an immediate and often startUng redistribution of the pre- 

 cipitated granules, though they Ijut seldom coalesce. 



3. Behavior of the precipitated droplets. — In case of a petroleum film on the 

 plate, the water droplets were sometimes seen to fall and float on the film, which 

 is positive evidence against spurious droplets. They are usually black and cir- 

 cular in outline, but when the light is intense and axial, they are bright. Fixed 

 globules are apt to be larger and more irregular. Particles m.ay sometimes be 

 seen to coalesce on collision, but this is rare. 



On tipping the microscope so that the light does not penetrate the vividly 

 colored drops axially, they seem to cast shadows in opposed directions for 

 symmetrical inclinations on both sides; but in view of the aplanatic properties 

 of spheres, the phenom.enon is probably a case of refraction, with the shadow 

 beginning at the edge of a caustic. Similarly, on moving the lamp horizontally 

 to either side from the position corresponding to axial illumination, the globules 

 becon^e opaqtie, and l(_)ok like round shining steel beads. The diameter of the 

 beads has but little effect. If the lamp is moved vmtil the field is dark, the 

 plate looks like the starry heavens. These stars seem to be above the drops. 



(4) 



® © ® 



(5) 



O O 9KV>)9 



(6) 

 Figures 4, 5, 6. — Diagrams Showing the Behavior of Fog Particles. 



After remaining in the plate for some minutes the fixed droplets often 

 become rosette-shaped (ap]iarently) , at first showing a mere black spot in the 

 center of the color disc, which gradually enlarges to a ring-shaped appearance 

 sloidy moving radially outwards. As a rale, the color ' is eventually the same 

 on the inside and the outside of the enlarged ring, the ring itself appearing red 

 with black demarcations in the surrounding green field, as shown in the figure. 

 On influx of air the structure becomes washed. This ring-shape may be merely 

 apparent, but the small globules when at first deposited never show the same 

 color within and without, the former being uniformly red and the latter white. 



• ' The colors observed were afterwards found to be due to chromatic aberration of the 

 microscope. 



