NUCLEATION OF AIR, FILTERED AND NOT FILTERED. 151 



its time variation. To guard against errors of interpretation, it was 

 necessary to install two fog chambers side by side, drawing from the 

 same filter and utilizing the same exhaustion system. In spite of 

 the fact that all appurtenances are apparently identical, the two 

 chambers do not show even approximately the same coronas or the 

 same nucleation. Each behaves as if it had its own specific coefficient 

 of radio-activity. Furthermore, the coronas for the same high press- 

 ure difference (8 p=^i.^) vary in the lapse of time as if some external 

 radiation were involved, though such a conclusion would not as yet be 

 trustworthy. It was shown above that the efficiency of the very pene- 

 trating gamma rays in producing nuclei is very marked, but the nuclei 

 here in question are very small in comparison with the cases examined. 



112. Nucleation of atmospheric air, not filtered Dust contents at 

 Providence, R. I. In the belief that a highly nucleated medium, no 

 matter whence the nuclei may arise, is a medium of special interest, 

 measurements of atmospheric nucleation have been in progress at this 

 laboratory since 1902. Four or more observations were usually made 

 by the coronal method per day, the details of which can not, however, 

 here be instanced.* If the mean of the daily observations be taken, 

 they make up a number of cotemporaneous series, the properties of 

 which are best shown graphically. Apart from details, for which 

 there is no place here, the things noticeable in the curves of successive 

 years are the extremely high winter, as compared with the summer 

 nucleations, the efficient of rain in depressing the nucleations, and 

 the totally different character of the curves for 1902-3 and 1903-4. 



These may be made even clearer by comparing the average monthly 

 nucleations, as are shown by the graphs in figure 103 of Chapter V, in 

 which the ordinates are again the nucleations in thousands per cubic 

 centimeter. (Cf. figs. 102, 104.) Here the degree of difference and 

 the similarities of the two curves are strongly brought out. As to the 

 latter, both tend to show sharp maxima near the time of the winter 

 solstice, and flat minima, much subject to rain, at about the time of 

 the summer solstice. It is clear from the enormous difference of 

 nucleation at the maximum and at the minimum that astronomical 

 causes can not be directly involved. The origin of the nucleation 

 must be in large part local, the nuclei themselves being the initially 

 ionized products of combustion. Nucleation is depressed by rain, and 

 possibly also (from the length of the summer day as compared with 

 the winter day) by light pressure. 



* Cf. Smithsonian Contributions, Vol, XXXIV, 1905. 



