THE STRUCTURE 01<' THE NUCI,Ki;s. |21 



22. Tarhid ahiminvm nitrate. Concentration. Time losses. — Accordino- to 

 table 7, the uucleatioii is n = 197 for the larger concentration and ?/, = 81 for tlie 

 smaller, the dilution being again 1/10. This result is exceptionally large, recalling 

 ferric chloride, which was also turbid. Cf. figure 12. 



The time coefficients show an equally pronounced difference in the two cases 

 which, together with the preceding result, point out a change in passing from one 

 solution to the other. In the light of the other tables, the discrepancy is probauiy 

 not due to the concentrations, and one may sus{)ect that some subtle effect of siir- 

 face tension is involved. 



23. Calcie nitrate. Concejitration. Time Imse>^. — The chief results (c/. fio-ure 

 13) aimed at in table 8, are the effects of dilution. The original nucleation is 

 n = 256 ; for the dilution 1/100, n — 87 ; for the dilution 1/I0O0,to = 61 ; for the 

 dilution 1/10,000, w = 63. It is to be suspected that in the last instance the 

 impurities of the water itself mask further decieraent, and yet such an inference is 

 uot warrantably safe. These experiments indicate the enoimous dilution needed 

 to bring out the true nature of the concentration effect. In most of the above 

 instances the dilutions used are insufficient. 



The time coefficients are an equally interesting feature. They show further 

 progressive change as dilution increases, even after the number of })articles pro- 

 duced has leached an infei'ior limit. Thus, whatever causes the solution, on 

 indefinite dilution, to reach fixed conditions of nucleation is not al;)le in the same 

 degree to influence the persistence of the nuclei. This persistence is therefore to 

 be associated with the residue of calcic nitrate left in the successive cases. 



24. Alum. Concentration. Time losses. — The effect of the dilution of 1/100 

 is to decrease the nucleation over 50 fc. The time coefficients do not show this 

 effect because the interval for which the second applies is too small as compared 

 with the first. In such cases reference must be made to the rates, h. Cf. figure 13. 



25. Ammonic nitrate, oncentration. Time Closses. — The effect of the dilu- 

 tion 1/100 is here about 33 %. The time coefficients show corresponding varia- 

 tions. Cf'. figure 13. 



26. Sodic sulphate. Concentration. Time losses. — Three concentrations ai'e 

 given, the strengths being in the ratios of 1, .01, .0001. CJ. figure 14. The number 

 of nuclei are respectively //. = 453, 110, 48. 'These large numbers, appearing in 

 spite of the small initial concentrations, are a feature of this body. The work was 

 sufficiently often lepeated to vouch for an unmistakable I'esult ; and yet the nuclea- 

 tion is not stable in the lapse of time, as the repetitions at the end of table 11 attest. 



Time coefficients show the usual mai-ked variations. 



Data in the first part of table 11 give evidence that the precipitation may be 

 made complete for a pi'essure decrement of less than 4 cm. 



27. Repetitions.— Lat^r experiments with the same solutions (marked " Re- 

 peated " in table 11) did not rei)roduce the excessively high values of n, though 

 they reappeared at the higher concentrations tested. The new n in the lapse of 

 time showed a tendency to diminish in its turn. As was the case with HCl above, 

 so here some occult influence must be invoked to account for the undeniable differ- 



