THE STUUOTUKE OF THE NUCLEUS. 71 



when the saturated water vapor is rephiced by the va[)or of some electrolytically 

 neutral liquid, like a hydrocarbon. I accordingly made a series of experiments 

 witli benzine, endeavoring at first to utilize benzine jet and color tube in the usual 

 way. In this I failed, for reasons without much lelevant interest here. I then 

 adopted the method of adiabatic cooling, partially exhausting a spherical receiver 

 (Coulier, Kiessling') about 23 cm. in diameter, illuminated by white light diverg. 

 ing from an external point. In this way not only were copious fogs obtained, but 

 the coronas produced were additionally availal)le as evidence. 



In the benzine jet, particles ai'e pi-obaltly cooled too suddenly, and at once 

 attain a size incompatible with axial color effects. Using tlie exhaustion method, 

 however, these axial colors appearing in benzine are not only of pronounced depth, 

 but they run into higher ordeis than in the case of moist ail' subjected to like ex- 

 haustions. Sequences })assing through blue, green, yellow, brown, puiple, etc., green, 

 brown, etc., may be seen in the axis of a column only 23 cm. long. The reason, uo 

 doubt, is associated with the lower latent heat of benzine,^ insuring the formation 

 of drops not less unifoim, but of a size, cset. par., regularly larger than for water 

 vapor. The fact that axial colors are producible botli with water and with a pro- 

 nounced insulator like benzine, is a result of fundamental importance in its bearing 

 on any theory adduced to account for the axial absoiption in (piestion. 



3. The exhaustion experiments'' were thus at once successful. (Jloudy con- 

 densation was as densely produced in benzine vapor as in water vapor, with 

 phosphorus, flame, and other nuclei. Care was taken to iusui'e diyness of vessel by 

 test experiments both before the benzine was introduced and after it had been 

 quite removed by eva[)oration. The exhaustion of about one sixth, say 13 cm., 

 seemed best adapted to bring out the following phenomena. When the receiver 

 was left standino- overnight no marked condensation occurred in the absence of 

 nucleation, or else the condensation was rain-like, a fine mist falling about 2 or 3 

 cm. per second. 



The iutrodiictory experiments were made with light nearly iu parallel, the sun's 

 image being used as a coronal center. The even dense tawny benzine fog after the 

 first nucleation was expected to develop on subsequent exhaustions (each followed 

 by an influx of filtered air) into the magnificent coronas which characterize this ex- 

 periment in the case of water vapor. On the contrary, however, the fogs wei-e more 

 fleeting, showing a more lapid descent than aqueous fogs, and the color fields ob- 

 tained'were not ring-shaped as expected, but sharply stratified horizontally, roughly 

 speaking, in altei'uations of green and red. 



Moreover, if the exhaustions were made successive without influx of air be- 

 tween each, the colors rose in sti-ata fi'om below, as they fell in strata when left 

 to themselves. On mounting, the strata grew successively wider and thinner till 



•For a digest of earlier observations, see my Report on Cloudy Conikusation, Bull. 12, U. S. 

 Weather Bureau, 1893. 



= Latent heat is not the only determinative factor, nor even the most important, as has been 



seen above. 



= Coronas afterwards produced hy paraffine nuclei in ben/.ol vapor (Chap. \', S 37) might have 



been effectively inserted here. 



