514 Dr. J. H. Vincent : Electrical Experiments 



the lamp took *075 ampere, and the extreme length of the arc 

 was 1 cm. With 1420 ohms, the current fell to '06, the 

 width of the beads increased to 1*8 cm v and the waists were 

 now *4 cm. long. The root mean square volts across the 

 lamp were 107. 



It is possible to employ glass capillary tubing in experi- 

 menting on the necklace effect. Initial trouble may be 

 experienced owing to the gases disengaged from the walls of 

 the tube. If these preliminary difficulties are patiently dealt 

 with, it will be found possible to run such glass lamps for a 

 considerable time. A soda-glass tube (7 cm., (rl mm., "45 mm.) 

 ran for an hour, giving a fairly steady light on 104 volts, 

 after which the experiment was discontinued. Soda-glass 

 lamps end their lives by cracking ; the cracks commence 

 internally and spread radially. Lead-glass tubes seal up, the 

 bore of the tube becoming filled with an apparently crystalline 

 black opaque mass. The cracking is not so marked as in the 

 case of soda glass. Quartz tubes do not deteriorate appre- 

 ciably when used in necklace lamps. Internal staining of 

 the tube sometimes occurs, which may disappear on further 

 use. This staining is less the purer the mercury, and is 

 probably due to the presence of impurities. 



Tubes of much larger internal diameter than can be 

 employed for the production of the necklace effect when 

 surrounded by air can be used if plunged into cold water. 



The Fan Effect. 



Having described some of the phenomena occurring in the 

 part of the diagram fig. 2 between the lines DE and AC, the 

 area below AC must be investigated. 



