A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 225 



its coincidence with December 22 not improbable. The summer minimum is 

 much more prolonged and uncertain, a result to be anticipated from the marked 

 rain effect which prevails at this season. 



The identification of maxima with the winter solstice and of minima with 

 the summer solstice is alltiring : for in these cases the earth is respectively nearest 

 and farthest from the sun. At the same time the orbital velocities are re- 

 spectively greatest and least, so that the path volume of the earth would have 

 corresponding values. Both causes are qualitatively in harmony with the 

 observed results, if the nucleation comes in great part from the sun. Quantita- 

 tively the results are less convincing. 



If the data for 1903-04 (which are more nearly absolute) be taken, the 

 ratio of greatest and least nucleation is about 3, seeing that December 22 corre- 

 sponds to about w = 66000 and June 22 to w = 22000. The case of 1902-03 is 

 even more accentuated. On the other hand, the gi"eatest and least values of the 

 radius vector of the earth's orbit are 1.0168 and .9832, in terms of the mean 

 radius. The same effect, approximately, is attributable to the differences in 

 velocity, making only about 6.7 % by which the winter nucleation should ex- 

 ceed the summer nucleation, for the case of linear distribution. No easily dis- 

 cernible distribution of nuclei from the sun outward would account for the 

 three- to four -fold winter nucleation as compared with the summer nucleation, 

 seeing that the decrement of about 2 % of distance corresponds to an increment 

 of 100 % of nucleation. Thus the reasonable siu-mise which makes the density 

 of the solar output vary as the inverse square of radius, while the path volume 

 per unit of time varies as the inverse radius, gives a compound law of the inverse 

 cube of radius, which, however, is quite inadequate. 



Unless some occult law of distribution is at the root of these cases, mere 

 change of distance and velocity does not account for the facts. On the other 

 hand, if the influence of the sun is atmosp)heric and productive of dissipation of 

 nuclei (by breaking them into fragments beyond the lower limit of nuclear 

 size, for instance), or if there is greater tendency toward supersaturation of 

 the atmosphere when the nights are longest with the consequent pre- 

 cipitation of local nuclei, the effect dtte to greater length of days in the 

 stunmer as compared with the winter will much more nearly correspond 

 to the nucleations observed. The effect postulated is, however, wholly 

 conjectural. 



38. Conclusion. — On the other hand, the two graphs showing the distri- 

 bution of mean monthly nucleation are not compatible with the postulate of a 

 purely local or incidental origin of the nuclei. Rains in general would make 

 the winter nucleation higher than the summer nucleation, while the products 

 of combustion still further increase the former. But neither of these inci- 

 dental effects is so distributed as to give rise to the salient maximum in Decem- 

 ber, reproduced independently and sharply by both curves. 



Briefly, then, while local and incidental effects enormously modify the 

 seasonal variation of atmospheric nucleation, it is not im.probable that 



