50 CONDENSATION OF VAPOR AS INDUCED BY NUCLEI AND IONS. 



s=J.O 6.1 5.4 4.3 3.2 2.0 i.o 



S= 7-4 3-9 10-4 8.7 6.9 3.3 



s = 7-5 6.8 5.9 4.9 4.2 3.4 2.4 1.8 

 5= 2.0 7.6 9.1 4.8 5.8 5.8 2.9 



Leaving out the smallest coronas and those which are no longer normal, 

 the data 5 = 7.2 and 5 = 6.8 were taken as fair averages in the two 

 cases. The data for show that the first table (16) is somewhat over- 

 compensated, while the second (17) is undercompensated by the values 

 of 5 entered. The high value of [o_> 2 ] = i6.8 was accepted with mis- 

 givings, but there is no evidence against it. It is interesting to com- 

 pare with the above values of 5 those which may be computed from sub- 

 sidence data in the way given in equation 7. From this it appears that 

 5 = 1.7 for t = 5 seconds of subsidence of fog. Now, the time needed for 

 complete evaporation was about 15 or 20 seconds, whence it follows that 

 5 must be of the order of 5 to 7 , agreeing therefore very well with the 

 datum computed from coronas. For the very small coronas subsidence 

 is too rapid to enter into any correction of this kind. 



The selection of a constant = ^5 = 0.0032 is the weakest part of the 

 above deduction. It is based on the earlier memoir and obtained from 

 the subsidence of observed coronas. Since the theory of diffraction 

 for an angular radius <f) of the coronas gives 



sin ^=5/60 = 1-22 A/d (13) 



for the first minimum annulus of wave-length A, and ds = a, 



= 73. 2 A (14) 



whence = 0.0032 would correspond to blue violet. With an eye at but 

 30 cm. from the fog chamber, the equation for sin <j) is certainly not quite 

 true and a must be variable with A, except perhaps for the smaller normal 

 coronas, which are so closely packed that a mean value of A is suggested. 

 If m be taken as 3.2Xio~ 6 , equation 4 shows ^' = 190 s 3 . Equation 

 14 incorporated in equation 4 would imply for io 8 7^ = 3.2 



6m s 3 i c-6s 3 



n' = 0*036 s 3 M 1 = oo72 s 3 n 1 = 0'24os 3 



according as the first red, orange, or violetminimumwereused, data which 

 merely imply an order of values, as equation 13 is not fully applicable. 

 Tables 16 and 17 and figs. 12 and 13 show a satisfactory order of 

 agreement between the observed and computed values of s and the 

 corresponding data computed for n as far as 5 = 7 to io cm., where the 

 middle green coronas enter. The agreement thereafter improves again 

 until the higher green coronas are passed, when further divergence is 

 marked. I will not enter into this here, as the subject has been discussed 



