62 



A CONTINUOUS RECORD OP ATMOSPHERIC NUCLEATION. 

 TABLE i. SECOND SERIES. n = 190000; a=.o6i. 



1 Gap in coronal sequences probably due to accidental delay. Similar difficulties at end of series. 

 Timed fog intervals desirable. 



METHOD OF REDUCTION. 



8. Constants of the geometric progression. To determine whether the 

 factor of the geometric progression of successive nucleations, z, was to be com- 

 puted isothermally or adiabatically, a series of direct temperature measurements 

 was deemed necessary. These were made by aid of a thermocouple of ex- 

 tremely thin wires (.007 cm. in diameter), of copper and german silver. The 

 junction within the receiver was not soldered, but the flexible copper wire 

 looped once around the other. In this way the variation of the instantaneous 

 air temperature in the receiver could be closely followed. It was necessary to 

 use a sensitive astatic galvanometer, and the measurements are thus subject to 

 the fluctuations of the earth field. As it is the immediate purpose of these data 

 to determine about how soon the isothermal conditions are re-established by 

 radiation from without, the irregularities are of little consequence. 



Table 2 contains three series of results, the upper end of each row corre- 

 sponding to the period of exhaustion, the lower to that of (slow) refilling. 

 Readings were taken in intervals of half a minute. The table shows that after 

 the lapse of one minute following the sudden exhaustion the temperature has 

 been regained to within a degree. As the coronas can hardly be observed and 

 measured within this time, the exhaustion ratio may be computed isothermally. 

 I have therefore computed the density ratio of nucleation p/p' = n/n', before 

 and after exhaustion, as follows. 



Since p=Rp$ in the usual notation of Boyle's law, and p=P-p f where P 



