i»4 THK STRUCTtJRK OK THE NCCLECS. 



2. Order of siz^ of the nuclei. — It will conduce to clearness if a [ireliininary esti- 

 mate of the order of size of the niK-lci [jrodiiced by the present method is inserted 

 at the outset. They may be considered first with reference to the minimum time 

 of subsidence, which would be admissible, if gravitational subsidence of nuclei 

 occurred at all. If the viscosity of the air through which the nuclei are su[)posed 

 to fall Ije taken as .00019 (it will be seen below, that in relation to the actual size 

 of the nucleus, the term viscosity no longer has any definite meaning), the sub- 

 sidence equation reduces to ?• = 9.3 X lO'^i'y, centim. To fall about 15 centim. 

 pel' hour, where 15 cm. is about half the diameter of the spherical receiver, the size 

 of the particle would have to be aliout /• = 58 X 10'" cm. Thus the following 

 estimates may be assumed for reference: 



Time of falling out, i*" Radius of particle, r = .000058 cm. 



4 29 



16 14 



25 7 



Inasmuch as the nuclei obtained by shaking solutions may in some instances 



be found to persist over a day, they must be small as compared with the last datum. 



A[)art from this, the coronas remain annular, until, with the total removal of 



nuclei, they fade from sight as a whole and regular figure. Hence there is here no 



evidence of subsidence whatever. Not oidy is this true, but the nuclei obtained 



by shaking .solutions actually diffuse against gravity with a velocity which, in 



spite of the difficulties of nieasuiement, is none the less chai-acteristic and definite. 



A fortiori, therefore, the pai'ticles here in question are very small as compared with 



10"^ centim., and it is from this point of view that the following results are to be 



undej'stood. A corresponding estimate is given in paragrajih 17. 



It is cuiious to note in passing that a rough estimate of the size of the.se nuclei 



made on the electronic hy[)othesis brings out a plausible result. Thus, if in case of 



the nuclei produced by shaking solutions, the excess of vapor pressure at a surface 



of radius, /■, ])e regarded as eventually equal to the electrical surface pressure ; if 



moreover, the nucleus is considered to be so constituted that a central electron 



forms a closed field with a charge (possibly due to friction) uniformly distributed 



over the surface of the nucleus, the e(piation may be written, 



2 T p 2 n e" 



r ~ idn^r^K ' 



whence r'^ = e'/(\(i ^ Tp K), where e is an electron or 7/1 U' <> electrostatic units, 

 7' the surface tension of water (81), /a the density of its vapor (.000017), K\i% 

 specific in<luctive capacity (80). From this follows the radius of the nucleus, 

 /• = 4.5 X 10"'', which, in the light of other values to be derived inferentially 

 below, seems not to Vje an uni'easonable estimate; but one must note that if in 

 solutions T, P, K me all increased, ;-3 is coi'respondingly deci'eased. 



3. Apparatus. Methods. — The expei'iments themselves are sufficiently simple 

 and roughly satisfactory with the exception of one element, the method of shaking, 

 which has not yet been brought under control. In figure 1, 72 is the spherical 

 receiver, about '.V.\ cm. in diameter, at the bottom of which is placed the solution. 



