182 SCIENCE PROGRESS. 



influence ; and we conjecture at present, as the result of 

 observation, that in summer when the leaves are fresh 

 positive electricity is most easily discharged, whereas 

 in late autumn negative seems to gain the advantage. A 

 geranium in a pot was the first thing tried (in June), and it 

 then discharged positive five times as quickly as negative. 

 In all cases it must be understood that the action is due to 

 the ultra-violet rays : glass cuts off the effect, quartz trans- 

 mits it. 



Garden soil was also tried, and when dry it has a very 

 fair discharging power, usually rather quicker for positive 

 than for negative. 



TENSION OR POTENTIAL. 



In a future article we may endeavour to discuss the 

 physical aspect of some of these facts from the point of 

 view of theory, but it is plain, in the experiments on metals 

 at any rate, how clearly a quasi-chemical origin is indicated 

 for the discharging power. That being so, it is essential 

 to ascertain whether the effect depends on the potential of 

 the electrified metal, or whether it is dependent on surface 

 density. It is easy to have a metallic surface at a high 

 potential without a trace of electric charge : such is the 

 inside of a metallic enclosure, or practically of any highly 

 charged hollow vessel ; its surface density and its electric 

 tension will be nothing, though its potential may be thou- 

 sands of volts. Conversely, it is easy to have a surface 

 charged to a high density and with an electric tension at 

 the point of bursting, while yet its potential shall be zero. 

 This is the case of an earth-connected surface exposed to 

 the inductive influence of a highly charged insulated body. 



It is a simple question for experiment to decide whether 

 the discharging power of light operates in one of these 

 cases and not in the other. Or, briefly, does the dis- 

 charging power of light on a surface depend on the tension 

 or on the potential of that surface ? 



The answer is distinct ; it depends upon tension. If 

 the surface is not actually charged with electricity, light has 

 no effect ; or at least its effect is limited to the feeble 



