Io6 A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 



As the rings are not easily produced with a very viscous oil, it is probable that 

 the droplet has here penetrated to the glass and that the oil film is drawn over 

 it by the capillary forces at the common edge of the three media. 



4. Preliminary data. — Before adopting the eccentric focussing device, 

 many experiments were made to ascertain the cause of the uncertainty in 

 catching the drops on the plate when ke]rt in place, seeing that sometimes the 

 precipitation was abundant, while at other times under the same apparent con- 

 ditions drops did not fall. Failures occuiTed both for high and for low nuclea- 

 tion. From the outset it was improbable that radiation from the outside could 

 afifect the result. It was eventually ascertained that on tipping the condensa- 

 tion chamber after the fog had formed, so that the subsidence would reach the 

 plate obliquely, a precipitate would usuall}^ appear. Again, an oblique current 

 within the chamber and passing across the plate ustially produced a deposit. 

 Hence the drops actually exist within the fog, and success in bringing them down 

 upon the plate is probably conditioned by very close isothermal adjustment of 

 the plate to its surroundings, added to the advantages gained froni incidental 

 and favorable air currents. Hence a little time must always elapse before the 

 drops persist at the plate, and hence the droplets from a shallow capsule do 

 not appear. Using a film of mica as a plate the result was the same, and it is 

 useless to attempt to enumerate the drops by this method. Those wiiich fall 

 are carried in by grazing air currents, while no drops are obtained from the 

 fiducial space tuider the objective. 



Nevertheless, the measurement of the diameters, d, of the dro]is obtained 

 by the above method without modification is an excellent test of the results 

 obtained elsewhere bv computation. The factor of the ocular micrometer 

 described above was .002 cm. per turn of the screw, or .00004 cm. per scale part 

 of the di-um divided into 50. The extent of plate covered by the breadth of the 

 spider lines was about .0003 cm. The finest particles are of abotit this diameter, 

 so that such measuren^ents must at best be much inferior to photograllh^• with 

 a scale attachment. The results are given in the following table, in which only 

 those results among many are inserted for which the observations were clear 

 and satisfactory. The coronal color with its diameter, 5 (chord of a radius of 

 30 cm.), are as observed when the eye and the source of light (Welsbach mantle 

 seen through a small circular hole) were at distances 85 cm. and 250 cm., re- 

 spectively, from the center of the condensation chamber. The exhaustion was 

 usually to a pressure difference of 17 cm., but this is of no significance when 

 diameters are alone to be observ'ed. The particles were collected by tipping the 

 chamber, sometimes in large numbers, but at other times sparsely distributed 

 without apparent cause. Nuclei were conveniently obtained from burning char- 

 coal. Both floating and fixed globules were examined with strong microscopic 

 illumination. It was difficult to retain a clear image without frec^uently re- 

 moving the plate, as the adjustment for focussing the plate within the chamber 

 had not 3^et been adopted. A table showing the results from coronal measure- 

 ments under the same circumstances is added. 



