Dkobmbkr 12, 1902.] 



SCIENCE. 



951 



correspond to winds coming in general from 

 the north. The notched minima on the after- 

 noons of October 31 and November 1 and 2 

 are curious novelties, but for them there is no 

 other explanation than the possible haze effect 

 as observed on the 2d. 



INFERENCES. 



8. Variation. — Mere inspection of the chart 

 shows the extreme variability of atmospheric 

 nucleation. Only a small part of this can be 

 a local effect, since the changes correspond to 

 the weather, though observation in cities 

 where there is so much chance for pollution 

 of the air is doubtless less satisfactory than 

 work in the country would be. It is probable 

 that even the small variations of the chart 

 after October 15 are real. If the nuclei were 

 colored the atmosphere would look like mottled 

 soap, with the clear regions usually, but by 

 no means always, accompanying rain or lying 

 under clouds. 



9. Rain Effect. — The observation next in 

 importance is the occurrence of pronounced 

 minima during rain, as for instance on Oc- 

 tober 2, 5, 12, 14, 19, 23, 28. There seems 

 to be no exception to this rule. It implies a 

 faster removal of nuclei by precipitation in a 

 saturated atmosphere (the result of any fall 

 of temperature) than the supply of nuclei to 

 the same region, by either diffusion or sub- 

 sidence or other more occult causes. But 

 whether the deficiency is eventually made up 

 from the lower air strata in contact with the 

 lands and the seas, or from the higher air 

 strata of the atmosphere where solar activity 

 especially prevails, is left open. 



Rain minima never fall quite down to the 

 zero of nucleation, and are themselves quite 

 variable in value. 



One may note that the tendency of rain to 

 change the normal air potential from positive 

 to negative values is accompanied by a rela- 

 tive absence of nuclei. In other words min- 

 imum nucleation exists here cotemporane- 

 ously with maximum negative ionization. 



10. Cloud Effect. — The third important fea- 

 ture, and one which became particularly evi- 

 dent after the temperature correction was 

 applied (October 15), is the cloud minimum 



as seen on October 15, 17, 21, 23. Usually a 

 higher nucleation is again established after 

 the cloud train has passed over the sky, the- 

 phenomenon beginning and ending with. 

 periods of clear weather. At other times,, 

 however, the minimum remains, as on Oc- 

 tober 17. 



The only explanation of this result, as will. 

 presently appear, is at hand ; the air has moved 

 bodily with the cloud, the whole constituting 

 a region of deficient nucleation. The nuclei 

 may have been precipitated by rain elsewhere,, 

 and the cloud may even have vanished from 

 the region. To this extent, then, a region is- 

 identified by its nucleation. 



11. Solar Effect Alsent. — Since the nuclei 

 cannot enter a region by diffusion as quickly 

 as seen on October 21, for instance, one is- 

 tempted to believe that solar radiation is 

 the cause by which the nucleation of a defi- 

 cient region is reestablished. There is, how- 

 ever, no evidence for this in the observations,, 

 and much against it. Thus, on October 19,. 

 a remarkably low minimum is maintained 

 almost all day in full sunlight; there were 

 but few hazy clouds in the sky. This mini- 

 mum was apparently only a part of the cloud, 

 region with which the day closed, but sunlight 

 was powerless to replenish it. Similar refer- 

 ence may be made to the notched minima of 

 November 1, 2 and 3. By contrast the high, 

 nucleation which occurs on October 25 and 

 27 in spite of an overcast sky may be cited. 

 In the latter case, as on the 10th, the maxi- 

 miun is a precursor, as it were, of the storm, 

 which follows. Finally, though no observa- 

 tions were made after midnight, there are a 

 number between sundown and ten o'clock. 

 These show neither marked increments nor 

 decrements, but have usually a normal char- 

 acter. 



Hence there is no evidence, so far as these- 

 observations go, that ultra-violet light or 

 other solar radiation has any potency in pro- 

 ducing nucleation, and I have, therefore, ex- 

 plained the cloud effects, etc., as purely con- 

 vective. 



12. Origin of Atmospheric Nuclei. — Specu- 

 lation as to the origin and character of these- 

 nuclei is premature. Conclusions must be- 



