792 



SCIENCE. 



[N. S. Vol. V. No. 125. 



body. Certain contradictions concerning 

 this variation which have been made by 

 different experimenters indicated that it 

 would be worth the trouble necessary to 

 study this phenomenon with care. Perrin 

 (C. K. 123, 351 and 878) states that the 

 rate of discharge is proportional to the den- 

 sity, while Benoist and Hurmusezcu (C. R. 

 122, 926 and 123, 1265) state that the rate of 

 discharge is proportional to the square root 

 of the density. The only data given in any 

 of these statements are those in one of the 

 articles by Benoist and Hurmusezcu. They 

 determined the rate of discharge by noting 



that of Perrin, in general it does not obey 

 either law. 



The experimental work will be described 

 more at length in the Physical Review, and 

 only a brief account of the results are here 

 indicated. My apparatus consisted prima- 

 rily of a charged zinc plate and an electrom- 

 eter. The plate was placed inside a box 

 of the same metal from which it was insu- 

 lated. The plate was connected to the 

 electrometer and the box was ground, and 

 both plate and box were placed in an air 

 receiver which was connected to an air 

 pump and a monometer. The rays struck 



Fig. 1. 



the time it took the leaves of an electro- 

 scope to fall from a given initial position to 

 a given final position. The voltage to 

 which the electroscope was charged is not 

 given, but presumably it was laige. The 

 curve marked B. & H. in Fig. 1 gives the 

 results of their experiments. 



In experimenting along this hue it was 

 soon found that the curve given by plotting 

 rates of discharge as ordinates and densities 

 of air as abscissae did not have the same 

 form under all conditions, and while it may 

 under certain conditions obey the law of 

 Benoist and Hermusezcu, and may possibly 

 under other conditions approximately obey 



the charged plate at nearly grazing inci- 

 dence. 



Care was taken to avoid ordinary elec- 

 trical leakage and electro-static induction 

 due to the action of the induction coil, also to 

 protect the electrometer and its connections 

 from the action of the X-rays, and from the 

 action of the air through which the X-rays 

 had passed. 



The zinc plate had an area of about 50 

 qc; the sides of the box were 2 cm. from 

 the plate. The capacity of the plate, the 

 connections and a small condenser placed 

 in multiple with the plate was about 400 

 electrostatic units. 



