soMi- c(>.\ I i-\tf'<>K.ii\y .inr.ixcns i.\ I'livstcs ir 



477 



it would Ix- (-(iiiiwilont to coiisidiTinn a tenor's range as extendinjj 

 In I lie hiv;hest overtone vvhiih roulil l)c detected in any of his notes, 

 whieh would certainly leail to asfonishinjj results). Or they may Iw* 

 radiated by oscillations within particles of metal torn from llie elec- 

 trcxles 1)\ the violence of the discharge — an idea suggested because 



Kig. •> — Serratctl Curves i.l and C) Indicating Very Short -Waved ("oinponents of 

 the Wave Train. (Physical Review) 



they arc prone to appear after the spark-gap has Ijeen widened and 

 the electrode-surfaces corrotled by a succession of sparks. Or they 

 might result from an excitation of molecules or atoms, in which case 

 they should be regarded as belonging to the spectrum of natural 

 ra>s, in the sense of my previous distinction.' Even so, if the serra- 



■ It is interesting to note that the opposite idea was put forward by Rubens — i.e., 

 that the rays of wavelengths O.t mm. and thereal)outs emitted by the mercury arc 

 might be due to oscillations in droplets of liquid mercury. 



