134 A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 



If the nuclei are derived from very dilute solutions, like river water, an 

 average value k = . i may be taken, whence if 



t = i mm. n/n, = .96 



3 -89 



5 -83 



10 .6S 



5 -IS 



IOO .02 



The reduction within 3 minutes will not exceed 10 per cent., which would 

 usually lie within a given type of corona. The datum k = . i , moreover, corre- 

 sponds closely to the values found for phosphorus and other nuclei. Hence, if 

 the type of corona changes after i or 2 minutes' waiting, it may be considered 

 certain evidence that the air is not saturated and the diffusion error predomi- 

 nates. Owing to the difficulty of avoiding either insufficient saturation or 

 excessive time losses, some of my observations in Chapter IX contain data for 

 two different aspirating currents, the faster corresponding to about 3 minutes' 

 sojourn of the nuclei in the receiver, the other to a time longer than 5 minutes. 

 In this way the effects of under-saturation which are most to be feared are 

 guarded against, while the faster current gives data falling short of the absolute 

 nucleation by not more than 10 per cent. In the course of time this also was 

 abandoned as superfluous. 



7. Effect of pressure difference. It is next to be considered whether the 

 pressure difference, dp, used in the exhaustions is pronounced enough to catch 

 all the nuclei. This is of particular interest as a safeguard against low numbers 

 in the nucleations obtained. 



The usual value, dp = 17 cm., corresponds to the following pressure ratios 

 and adiabatic temperature reductions in air, ((T6-p')/'j6-p'-6p'))- 4 '2'j^ S' if 

 p' is the vapor pressure of water and S' the reduced absolute temperature. 



10 Pressure ratio, 1.292 ' = 254. 7 



20 1-297 263.4 



30 1.341 268.8 



For comparison data were gathered with a larger pressure difference, dp = 

 22 cm., for which the values are 



10 1.414 245.5 



20 1.416 254.1 



30 1.432 261.5 



Clearly, the coronas for the larger temperatures and temperature differences 

 must be smaller, cast, par., in view of the greater quantity of moisture precipitated. 

 The data for m, the quantity of moisture precipitated per cubic centimeter of 

 saturated air, have been computed by the method of C. T. R. Wilson and J. J. 

 Thomson and are given in the following table 2. Here t, is the initial tempera- 

 ture , t, the temperature before and t after condensation. 



