with Mercury contained in lubes. 525 



a lamp into a tank of cold water while it is giving the 

 necklace effect. This experiment, of course, can only be 

 tried with quartz tubing with any chance of continued 

 success. If a lamp is running with a high resistance in 

 series and giving a necklace effect, the ribbon is temporarily 

 produced, in some lamps, by suddenly decreasing the re- 

 sistance. When the ribbon breaks, it may give rise to the 

 fan effect or to a broken necklace according as the tube is 

 longer or shorter. With a tube which is known to give a 

 good necklace with a certain voltage and resistance, it may, 

 on first lighting up, give the ribbon instead, which becomes 

 unstable as the tube gets hot. Or perhaps delay in 

 assuming the necklace form may be a purely mechanical 

 result consequent upon the oscillations requiring time for 

 their amplitude to reach a steady v;due. An example of this 

 was met with when performing experiments from which the 

 lines AB and AC were drawn (PL VII. fig. 6), the following- 

 observations being taken with an electromotive force of 

 200 volts. 





External 

 Resistance 



E.M.S. 

 volts on 



Current 

 in 



Width 

 of waist 



Width 



of bead 



Mean 

 arc length 



Drop of 



volts 

 per cm. of 





in ohms. 



lamp. 



amps. 



m cms. 



in cms. 



in cms. 



mean 

 arc length. 



ecklnce 



j 1290 

 - 1-1820 



125 



•07 



•4 



2-75 



157 



80 



Limits , . 



88 



•07 







1-9 



•95 



93 



Ribbon 1610 162 '025 1*25 130 



The ribbon persisted for 23 minutes when it was replaced 

 by the necklace arc, the current rising and the volts across 

 the lamp falling. The rate of consumption of energy in the 

 necklace effect is almost twice that of the ribbon effect for 

 the same mean arc length. This suggests that an appreciable 

 fraction of the electrical energy supplied to the necklace 

 Limp is used in maintaining the oscillations of the viscous 

 mercury. 



With a fixed pressure, the external resistances on which 

 a given tube will run as a ribbon lamp are confined to a 

 region on a resistance-electromotive force diagram similar in 

 shape tothat found for a necklace lamp. On such a diagram (see 

 fig. 15, PL VII.) there is an area ABA'C, indicating the ranges 

 of volts and ohms within which the ribbon is produced. If 

 a ribbon lamp is alight and the resistance is increased so that 

 the tracing point crosses the line AB, the arc shortens and 

 finally goes out on the contact of the mercury columns ; 

 meanwhile the volts across the lamp fall and the current 



