the Atmosphere over the Ocean. 257 



In this, as in other experiments, it was clear that the 

 presence o£ smoke decreased the number of ions detected by 

 the apparatus. A few puffs of tobacco-smoke blown into an 

 electroscope will decrease the natural leak as much as 35 

 per cent. The ionization current due to *3 mg. of radium 

 bromide placed near an electroscope was decreased from 4*2 

 to *7 scale-divisions per minute by the introduction of a small 

 quantity of tobacco-smoke. These effects are similar to those 

 first observed with an electrometer by Owens. It is clear that 

 the presence of smoke, haze, or fog may cause a diminution in 

 the number of ions observable by an Ebert apparatus. 



III. 



On the Effect of Spray, 



It was natural to expect that an artificially created mist 

 would also diminish the number of ions detected by the 

 instrument. But under the conditions of the following expe- 

 riment the reverse was the case. A hand-sprayer was 

 obtained such as is used to sprinkle roses with fluids in order 

 to destroy aphids. A horizontal current of air is driven by a 

 condenser-pump towards the orifice of a small tube leading- 

 down to the fluid employed. In this experiment water was 

 used and driven off in an exceedingly fine mist in all directions 

 around, but not towards, the Ebert instrument. The electri- 

 fication in the air, both positive and negative, was greatly 

 increased by the spray. 



March IS, 190(5. 



n + 



'U_. 





1640 



1770 

 9300 



With spray 



7070 









March 20, 1906. 



+ • 



! No spray .. 

 '. With spray 



1390 

 5350 



When the hand-sprayer was used without water there was 



