A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 59 



exhaustion an external vacuum chamber provided with a mercury gauge and 

 a wide stopcock suffices. A jet pump is suitable for evacuation. 



The condensation chamber is placed between the goniometer and the 

 source of light, nearer the former if large coronas are to be obtained. In my 

 experiments the distances were usually 84 and 250 cm., or about as 1/3. The 

 eye at the goniometer is focussed unconsciously on the distant source (334 

 cm.). 



The exhaustions must be made systematically in connection with a seconds 

 clock, to admit of allowance for the time losses. The time during which the 

 fog remains suspended is particularly important, and must be uniform and as 

 short as possible.' 



3. Color distributions. — In classifying the coronas, a statement of the 

 colors of the first two or three annvili, counted from the center, will usually 

 suffice. For the case of the electric light the central patch remains white, or 

 at least opalescent or bluish. With the Welsbach lamp a central disc of vivid 

 green or green-yellow, or even yellow, is frequently observed; but the use of the 

 electric light in parallel series shows this to be due to the absence of strong 

 complimentary blues and violet. 



For convenience in specifying color, the following abbreviations will be 

 used throvighotit : w, white; p, purple; c, crimson; r, orange-red; br, brown; 

 o, orange ; y , yellow ; g, green ; b, bltie ; v, violet. Mixed colors are written 

 together; thus bg is blue-gi-een, rv red-violet. An accent denotes an approxi- 

 mation to the color ; thus b' is bluish, which has been otherwise indeterminable. 

 A dot or capital denotes a deep or dark color; thus b or B is dark blue. 

 A mere line denotes a color ring too narrow or dark to be recognized. This is 

 the freqvient transition from red to green, marked w r g. 



Beginning with the most intense nucleation obtainable, i. e., with particles 

 of the least size producible, the following coronas appear in succession, at first 

 filmy and fleeting, but eventually brilliant and dense. The numerals attached 

 to the series are arbitrary. 



I. w' o' 



II. w v g'; b' b r'; w' g v; w y V bg'; w yo v g'; w c yg v' 



There is thus an obvious tendency for the colors succeeding white to follow each 

 other in the order of wave length, as the particles continually increase in di- 

 ameter. All intermediate gradations are represented. The second cycle is 

 nearly complete, the first ( ?) cannot be obtained except in the opalescent orange 

 tint, unless the steam jet is employed. The second anntilus of any corona 

 is apt to vary in width so as to be unequally important. 



The next series (III) for successively larger particles is a contraction of the 

 preceding. There is obviously much overlapping. The following types of 

 coronas may be cited. The colors are very brilliant. The second "green" 



' Smithsonian Contributions to Knowledge, 1. c. 



