THE STRUCTURE OF THE NUCLEUS. 119 



For slight partial exhaustions, therefore, an increasingly greater number of 

 particles escape pi-ecipitation by evaporating their loads as far as the persistent 

 nucleai- stage. The repeated [>i-ecipitatioii at these low pressure decrements (allow- 

 ing each fog to subside) nevertheless quite i-eiuoves tiie nuclei, as is shown by the 

 tests subsecpiently made by a[)plying large pressure decrements. 



Assuming that dp ^^ 2 cm. are needed for condensation, the size of the nucleus 

 may be estimated by the approximate method of computation, explained elsewhere, 

 the equation being 



2/ T 



''.'ip— j^e (.02176-1) ' 



where condensation takes place for the small pressure decrement Spid the pressure 

 J) and the absolute temperature 6. If in this equation — 293", p — 76 cm., 2'= 80, 

 Ji = 4.6 X 10" for water vapor, 



/■= 1 X 10-B cm., 

 which should be an inferior limit. This result agrees with the order of quantities 

 estimated in § 2 ; but as the value of the surface tension, 7,' accepted is certaiidy 

 too large for such excessively small particles as the nuclei, /• must lie indefinitely 

 below 10"" cm. 



18. Effect of concentration and of hulk. Time losses. — The third [)art of the 

 table 4 contains data for solutions of half the preceding strength. Furthermore 

 solutions of double the preceding bulk as well as the usual bulk (500 cub. cm.) 

 were shaken in like manner for comparison. 



The first important result is the effect of quantity. When 1000 cub. cm. are 

 shaken under otherwise like conditions, fully twice as many nuclei are produced 

 as for the case of 500 cub. cm. It would appear from this that the nuclei arise in 

 the watei- itself. Practically as well as theoi'etically this observation deserves 

 attention. Cf. figure 11. 



As usual, the number of nuclei produced at half the usual pressure decrement 

 (^Sp = 8 cm. instead of 16 cm.), is somewhat smaller on the average. 



The time coefficient for this doubled bulk of liquid is 6n/ St = —2.6, or about 

 twice as large as for the other solution. But as the effect of the difference of 

 concentration involved is insignificant (this is shown by all the cases below), and 

 as the number of nuclei has been about doubled for the greater bulk, it follows 

 that the time coefficients increase proportionally to tlie number of nuclei produced. 

 This therefore is direct evidence in favor of the constancy of the absorption 



velocity, h. 



19. Fert-ic cUoride. Effect of concentration, bulk, 6fc.— The average number of 

 nuclei produced in the normal case is 234, or about as many as in the preceding solu- 

 tions {Cf. table 5). -Valency, therefore, can have no appreciable effect on the genera- 

 tion of nuclei. The effect of pressui'e decrement is also like the above; fewer nuclei 

 are entrapped by precipitation at the lower pressure decrements than at the higher. 



.When the concentration is reduced one fourth the nucleation falls from 234 to 

 184 so long as the same bulk (500 cub. cm.) is in cpiestion. It is probable that the 

 true decrease of nucleation is less than this. 



