Short Spark-Discharges. 705 



The rate at which the potential gradient falls depends 

 somewhat on the drying substances used. Calcium chloride 

 gave the slowest tall, and was much hotter used alone than 

 when used with either phosphorus pentoxide or sulphuric acid, 

 or with both together. In the results already given an 

 elaborate system of drying-tubes and large reservoirs partly 

 rilled with fresh anhydrous calcium chloride was -done used. 



The drying-tubes and discharge-chamber were at one time 

 filled with illuminating-gas in order to vary the experimental 

 conditions. Table B gives the results obtained. It is seen 

 that the potential gradient fell, but no more than might have 

 been expected had air alone been used. 



Table B. 



Data taken December 17, 19C4. — Ten seconds was the 

 time-interval used between changes of potential. 



Air was used in the discharge-chamber. 



Ar V - 



\i. 



a. 



y Hours after 



4=100*.. f^s 



r r chamber. 



40 40-2 



3-9 ••'-•4 

 4r, 500 



2 



1 





 10 



Average... 



35-0 15 10 



345 30 10 

 38-0 3*5 10 



3 



Illuminating-gas was used in the discharge-chamber. 



41 313 20 

 43 351 20 



8 

 6 



Average... 



27-0 140 

 290 160 



15 



10 



10 | 



In order to vary the condition at the point of discharge, 

 the surfaces which had been polished in the same way each 

 time were now polished by a jeweller. They were thoroughly 

 cleaned with caustic potash, nitric acid, water, and finally 

 washed in alcohol., but not rubbed on the buffing-wheel 

 again. The tests Bhowed a medium value for the potential 

 gradient. Xo important change in the action of the surfaces 

 ibserved. 



Considering all of the observations taken, it cannot be said 



that tla- potential gradient approached any constant value. 



>uld be brought to any value within the wide range of 



12^ volts per I'm,,,, to 83 volts per loo 



/',". 



This 



range 



