Discharge of Electricity from Cylinders and Points. 791 



The source of supply was a Wimshurst machine W, driven 

 by a small motor, and the insulated system was charged by 

 the discharge from a point, P, to a plane. The rate of 

 supply could be regulated by altering the distance between 

 the point and the plane. 



The potential was determined by an electrostatic voltmeter, 

 Y, connected to the insulated system, and the cylinder, 

 which was permanently connected to earth. 



A table galvanometer, G, o£ the moving-coil type, was 

 used to measure the current from the wire to the cylinder. 

 It was sufficiently sensitive to detect a current of about 

 2xl0~ 9 ampere per centimetre length of the wire. The 

 galvanometer was insulated from the table by blocks of 

 paraffin wax; one terminal of the coil was connected to the 

 case and to the voltmeter, the other directly to the wire 

 inside the cylinder. Under these conditions the insulation 

 was quite good and the system would remain charged for a 

 considerable period. In many of the experiments leyden- 

 jars, C, were connected to the insulated system to nullify 

 irregularities in the supply. 



It was, however, found that, in certain cases, more steady 

 currents could be obtained without the extra capacity. This 

 was particularly the case when working with low pressures 

 and the wire negatively charged. In these cases there was 

 a tendency to a " spark discharge " as distinct from the 

 "glow discharge.''' The effect of the spark discharge was 

 to discharge the system to a potential considerably lower 

 than that at which a discharge commenced, thus rendering 

 the discharge discontinuous. By cutting out the leyden- 

 jars the capacity of the system was reduced to a very small 

 quantity, and there was not enough electricity immediately 

 available to supply a current sufficiently large to make a 

 large change in the distribution of electric force. The 

 discharge then became sensibly continuous. 



3. The potential necessary to produce the spark was found 

 to be the same as that sufficient to maintain the smallest 

 current detectable with the instruments, and so either was 

 taken as the sparking potential. In all cases except where 

 special mention is made the air used was dried by drawing 

 it slowly through three drying-tubes of about two feet each 

 in length, the first filled with calcium chloride and the other 

 two tightly packed with phosphorus pentoxide. Just before 

 entering the cylinders the air was drawn through a plate 

 of unglazed porcelain to free it from dust. 



Irregularities frequently occurred in the negative dis- 

 charges which have been attributed to dust settling on the 



3 G 2 



