22 



CONDENSATION OF VAPOR AS INDUCED BY NUCLEI AND IONS. 



In table 9 I have, therefore, arranged the data for n with reference to 

 the corresponding values of dp/p, both for the cases of I, II, III, and V, 

 where the upper pressure limit of the drop dp (curve [dp]'), and cases 

 I', II', IV, where the lower pressure limit of the drop dp (curve [dp]^, 

 are varied. This result is also given in the chart (figs. 5 and 6) arid the 

 mean results of the latter are suggested by the dotted line in the former. 



In fig. 5 the curious result is obtained that the data for [dp]' are now 

 liable to lie even above those for [o/?] 1 which is the inversion of the former 

 case. As a whole, however, and with due regard to the subtleties in- 

 volved, the two sets of data practically belong to the same curve, for 

 the departure of either in the long run is seen to be positive as well as 

 negative. The results of fig. 5 (as has been stated) were obtained in a 

 single series of observations, all at the same temperature. If they be 

 compared with fig. 5 (dotted line), containing observations made at 

 other times, they lie distinctly above the graph of the figure, no matter 

 whether [dp/p] f or [dp/p] t is in question. Hence it is again probable 

 that something else besides mere variation of the barometer is in question 

 and is not accounted for in the correction. Thus it is next necessary to 

 inquire into the effects of vapor pressure, which in series I and II would 

 differ from series I' and IF, though in series I, II, III, IV, and V the 

 temperatures are so nearly alike that shifting of graphs due to this 

 disturbance is not appreciable. 



TABLE 9. 



14. The effect of vapor pressure. The second extreme limit, 



T/Tj-^I (3p(x1tS) I(P*YI 



may now be used and the data for nucleation, n, expressed in terms of 

 dp (TT XI)/(P TT) as the variable for comparison. Remembering 



