A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 49 



conform more closely to i=d (\/d za/a) or s = s az and ds = a. For 

 present purposes this is near enough. I shall therefore lay off the aperture, s, 

 as a linear function of the number of the exhaustion, z, for which the observa- 

 tions show per unit of z, in case of sulphur nuclei, 6s = .28, and in case of punk 

 nuclei, 8s = .ig. The initial aperture computed herefrom as the mean of six 

 series, in each of which the nucleation was introduced independently, is for 

 sulphur, $0 = 3.4 and for punk, 5 =2.2. Hence = 84o,ooo in the former case 

 and w = 230,ooo in the latter. 



Since the pressure ratio was in each case 1.36, the nuclei in the influx air 

 passing over burning sulphur or glowing punk must have been 3.8 times more 

 numerous. Thus there were nearly 3,000,000 sulphur nuclei and nearly 900,000 

 punk nuclei per cubic centimeter in the laden air currents entering the con- 

 densation chamber. 



I shall show in Chapter VI that the equation applicable to the present 

 experiments is 



where n z is the initial nucleation, y the volume ratio on exhaustion, z the 

 number of the exhaustion, and 5 an appropriate subsidence constant. The 

 function 77 is a product of the terms (i S/sJ) (i S/s z+l ) . . . ( i S/s|_ t ) , 

 so that Z is the number of the exhaustion in which the first corona is seen and 

 77= i. When the particles are as large as is the case for benzol the subsidence 

 function is of prevailing importance and masks the exponential function as all 

 the observations for benzol show. I have carried this method out for water 

 vapor, obtaining consistent results throughout. The present observations for 

 benzol are scarcely systematic enough to make it worth while to compute S, 

 and the experiments should be such in which the diffusion and homogeneity of 

 vapor is insured by continued rotation of the vessel rather than by shaking. 

 But there can be no doubt that, with proper precautions in this respect, the 

 number of nuclei furnished per cubic centimeter by any given nucleator can be 

 determined with benzol vapor as the coronas are all normal, even for large 

 values of 6p, with certainty. 



4. Axial colors. It is because of the relatively great number of relatively 

 large particles in case of benzol and similar hydrocarbon vapors, that the axial 

 colors are seen, and may be traced into much higher orders than is the case with 

 water vapor. The yellows, browns, etc., of the first order may be easily ob- 

 tained with the steam jet, though they cannot be produced in the condensation 

 chamber by any means except by pressure differences causing intense spon- 

 taneous condensation in moist air. The subsequent violets, blues, etc., how- 

 ever, are here distinctly seen as far as the orange red of the second order, after 

 which the admixture of white light makes recognition of color more and more 

 difficult. With hydrocarbon liquids like gasolene, benzine, etc., the axial colors 

 are seen much farther along the series even through a short column, and they 

 are intense in the drum. The difficulty encountered in observation is due to 



