206 C. Barus— Changes of Colloidal Nucleation. 



they lie distinctly above the graph of the figure no matter 

 whether \f>p/p\ or [8p/p], is in question. Hence it is proba- 

 ble that something else besides mere variation of the barome- 

 ter is in question, and is not accounted for in the correction. 

 Thus it is next necessary to inquire into the effects of vapor 

 pressure. 



4. The effect of pressure. — The other limiting case 



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

 terms of {^—(^ — ^^/(p — ir) as the variable abscissa, for 

 comparison. Remembering that the total variation of pressure 

 to bring out the coronal phenomenon does not much exceed 

 3 em and that the observations below will be made within a 

 single centimeter, the precipitation of moisture may be treated 

 as dependent on t/tj, the ratio of the initial and final absolute 

 temperatures of adiabatic cooling, if the former is nearly con- 

 stant and if the same medium is retained ; though the case is 

 in reality more complicated. The data of figure 1 compared 

 in this way showed even greater coincidence and their mean 

 value is given by the dotted line ; but later and earlier series 

 (over ten of which were investigated) differed from them and 

 from each other in a way that was far too large to be referred 

 to errors of observation. Leaving these details for considera- 

 tion elsewhere, 1 conclude that by far the greater part of the 

 dependence of the colloidal nucleation upon the barometer is 

 the necessary result of the thermodynamics of the case ; but 

 that conclusive evidence of the absence of the effects of other 

 causes either within or without the fog chamber, on the time 

 variation of its nucleation, is outstanding. 



5. New data for colloidal nucleation in the lapse of time. — 

 Results of the same nature as the preceding were now col- 

 lected by choosing a particular {hp — {'K — '7r^)/(p — 'Tr) = - 320 and 

 reducing all data for n to this value. The results so found 

 ( M .32oXlO~ s ) should be independent of atmospheric pressure, 

 etc., and respond to external radiation as such exists. The 

 data are shown in figure 2 ; they do not follow the barometer. 

 The correction of «,is about 1*7 per *001 of the pressure ratio ; 

 but it is uncertain in this region since the graphs are of pro- 

 nounced curvature. 



Hence in succeeding experiments a larger fiducial value 

 ($p— {7T — frj)/(p — 7r) = -335 was selected in turn, as the graphs 

 in this part of the field are more nearly straight. At the out- 

 set complete series of results were investigated as in figure 1; 

 subsequently but three observations in the neighborhood of 

 the abscissa -335 sufficed. 



The data for n- m sometimes follow the barometer, sometimes 



