108 NUCLEATION OF THE UNCONTAMINATED ATMOSPHERE. 



of observations at one station, particularly as the prevailing winds 

 are land winds. Data from places where the northwest wind blows 

 over no cities, and where the northwest wind is warm and the south- 

 east wind dry, would be interesting. 



Again, the daily averages of the nucleation appear to vary inversely 

 with those of the temperature and vapor pressure, while in most cases 

 they all vary directly during the day, rising to a maximum near noon. 

 This would seem to suggest an overbalancing effect of the sun in pro- 

 ducing nuclei. As regards the relative influence of temperature and 

 vapor pressure, one might look for an increase of nucleation after a 

 sudden fall of temperature, but would hardly expect nuclei to be pro- 

 duced because of continued cold. It seems quite possible, however, 

 that, with a high vapor pressure, there is continual condensation of 

 moisture on the nuclei present, with a persistent tendency to drag 

 them to the earth. The inverse agreement between the nucleation 

 and the relative humidity is even more general, for the latter more 

 often descends to a minimum during the day. The percentage of 

 saturation, moreover, ought to be some index of the condensation 

 taking place, remembering that condensation may be in progress at 

 higher levels when the air at the surface is not saturated, and that, at 

 those planes, small variations of temperature may be producing satu- 

 ration. The absorption of the sun's rays, if they produce nuclei, 

 would also have some effect, but probably to a less degree. 



In the case of sunshine and cloud, likewise, the effect is obscured 

 for sudden changes by variation at the same time in the vapor press- 

 ure. The cloud effect is partly the opposite of that attributable to the 

 sun, which has been assumed to produce nuclei throughout the day 

 in the upper atmosphere. Clouds catch these if they come within 

 range, and prevent the formation of others near the surface. Clouds, 

 however, are at the same time apt to be an index of a region denucle- 

 ated by rain. 



The persistence of nuclei is shown in two ways. Readings, usually 

 less than 10,000, are often obtained through several days of cloud, 

 snow, fog, and even light rain. These have either penetrated through 

 the clouds or they have been brought from cities, in which case the 

 land effect here must be small. Persistence, and therefore the sup- 

 posed cumulative action of the sun, is shown in the day curve, in which 

 the 3 o'clock reading is nearly as high as that at noon, although the 

 sun is then as low as at 9 o'clock. Similarly, the lag of the uuclea- 

 tion in building up after rain, or after a persistent cloud effect, would 

 seem to indicate the cumulative action of the sun. 



