132 A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 



1.5 lit/ mill. The capacity of the trough, inckiding parts above the cloth lining, 

 was about 11.3 liters. Hence 3 to 5 minutes suffice to renew the air, and each 

 particle remains in the trough from 1.5 to 4 minutes. The time needed for 

 adjustments after the influx pipe was shut off and prior to sudden exhaustion 

 was about 1.25 minutes. Therefore the total time for saturation was 2.7 to 5 

 minutes, which should fully suffice even in case of diffusion alone. 



The conditions, however, are much more favorable as the influx and efflux 

 currents evoke considerable convection. The lightness of water vapor is itself 

 favorable to the same end. It is obser\'ed, for instance, on exhausting imme- 

 diately after the introduction of phosphorus nuclei, that filamentary condensa- 

 tion is in evidence, denoting currents upward axially and downward near the 

 walls of the receiver. ^ These fog strands are the inevitable convection currents 

 due to the relatively low density of the vapor. 



4. Miscellaneous tests. — If there had been under-saturation the coronas 

 found on condensation should have been larger in non-saturated and smaller in 

 more saturated parts, which was not observed. Sudden exhatistion immediately 

 after shutting off the influx showed a somewhat enlarged unifonn corona, but 

 even enlargement was not invariable. 



Whether the fast or the slow influx specified was adopted proved to be 

 without marked effect for reasonable differences of time, i. e., such as would not 

 imply time losses of nuclei. 



The effect of a long influx pipe (10 meters of ^^-inch lead pipe) and of a 

 short pipe i meter long could not be shai"ply differentiated, owing in a measure 

 to the cotemporaneous variation of atmospheric nucleation. So the presence 

 compared with absence of the coil in the water bath (24 turns each about 3.5 

 cm. in diameter) showed a negligible difference. 



Experiinents were made with regai'd to the usefulness of this coil for keeping 

 the influx air at room temperature both by filling the bath with abnormally hot 

 water (40° C.) and with broken ice. The effect on the apertures of the coronas 

 was in both cases of little importance. Hence, except on very cold days, the 

 water bath and coil may be withdrawn altogether. The atmospheric air after 

 traversing the 10 ineters of influx pipe is already sufficiently heated to be intro- 

 duced into the condensation chamber directly. 



Tests were also made with a U-tube loosely filled with wet sponges and 

 with a drying tube one meter long containing phosphorus pentoxide. In 

 neither case was a definite effect on the coronas ascertained. 



Freedom from leakage was finally tested by filtering the air. The coronas 

 on sudden exhaustion showed a gradual decrease to complete evanescence. 



5. Diffusion from a single surface. — The case is nattirally less favorable if 

 the upper wet surface (double cotton cloth) is omitted. The computation may 

 be made from an expansion of Kramp's integral, so that 



^Smithsonian Contributions, No. 1373, 1902. 



