THE STKUOTUKK OF THIO NUCLEUS. 59 



G. Repetitions. — Owing to the tendency of benzol to secrete i-esidual nuclei 

 iu strata near the surface of the lic[uid, a peculiarity of which I was not at fii'st 

 aware, it seemed advisable to repeat some of the results for the successive coronas. 

 Examples of the data are given in table 2. The persistent coronas are all normal. 

 Punk nuclei were used for comparison with the sulphur nuclei of the preceding 

 table, but the conditions of experiment were not materially altei'ed. The results 

 are of the same nature. 



7. Differences hetwee)i nuclei. — The different rates at which s diminislies in 

 case of punk nuclei and of sulphur nuclei, is notewoi'thy. To bring this out, both 

 series of results (new and old) ai'e constructed in the chart, figure 4, the abscissa 

 being the number of the exhaustion (all identically y = 58/76) ; the ordinate, the 

 aperture of the corona. Of these curves, a, b, c were obtained with sulphur 

 nuclei ; a, h being the results of table 1, where the li(piid was shaken between each 

 exhaustion, c being the new series made without continued sliaking. The curves 

 d, e, /'refer to punk nuclei. Shaking was irregulai'l}' resoi-ted to, and the effect has 

 impressed itself on the curves. The new and the old curves for sulphur nuclei 

 show about the same steepness, no matter whether tlie liquid was shaken between 

 the successive exhaustions or not. To some degree this is also true for the jDunk 

 nuclei, all the curves showing strikingly less steepness than the sulphur curves. 

 In fact, the decrement of aperture rfx is ])er exhaustion : 



For sulphur nuclei, 6 s = .14 cm. 

 For punk nuclei, S s =^ .09 cm. 



Since S d = .00144 6 s/s^ = 6s X (t/2/.00144), the result in terms of d is less simply 

 stated. 



Hence the coronas vanish more rapidly for sulphur nuclei than for punk 

 nuclei, and since by § 5 this removal is due to subsidence of cloud particles it, 

 follows that the benzol-sulphur particles are larger than the benzol-punk particles. 

 A corresponding i-esult is found below § 8 for sulphur-water particles. 



The accompanying results drawn to the same scale and obtained in the final 

 exhaustions (z = 20-30) with water vapor are of the same order of steepness and 

 have been similarly explained. Petroleum cloud particles, in view of the small 

 vaper tension, are much smaller and neai'ly free from subsidence, as figure 4 

 shows. This case cori-esponds to coronas of abnormal type; it has, therefore, 

 seemed reasonable, in the definite computations below, to ignore the effect of 

 subsidence altogether when the order of the exhaustion was sufficiently far removed 

 from the end of the series. Cf. § 11. 



Since d^ jY = dl and dd= (J2/.00144) 6s, if the first equation is differen- 

 tiated logarithmically and 6 d replaced by the second, 



A^ .00144 



Hence for sulphur nuclei, sN/JV= 290 d, and for punk nuclei, S N/N = 190 d, 

 equations which show the part of all the residual nuclei precipitated per exhaustion 

 at any mean diameter d, and under the time interval and the supersaturations 



