47C) 



BULL SYSTEM TECHNICAL JOURNAL 



in close agrecmciit for all except the topmost of the cur\es, for which 

 the fundamental of the oscillator corresponded to wavelength 4.2 mm. 

 and that of the receiver to wavelength 8.4 mm. 



The lowest wavelength mentioned by Nichols and Tear as ha%ing 

 been manifestetl and visualized in this lucid fashion is 4.2 mm.; while, 

 replacing the two mirrors of the Boltzmann interferometer by a set 

 of eight mirrors forming an evenly-rising staircase or echelon, the>- 

 obtained curves which in one instance indicated a fundamental of 



) /. 2^ J i 5 S^ 7 8 9 



Spiegeryerscfiiebung Kun/e a, b, c 



Fig. 8 — Serrated ('iir\e^ 



ilicating Very Short-Waved Components of the Wave 

 Train. {Annalen der Physik) 



l.S mm. It would be a conser\ati\e, perhaps a too conscr\ati\f, 

 policy to regard this as the i)resent liinil nf tlif spiTtnuii of .irtiruial 

 electromagnetic waves. 



Whether we may believe that rays lying beyond this limit have 

 actually Ijeen generated depends upon the interpretation of certain 

 narrow sharp serrations obscr\'ed upon curves of the more uneven 

 sort; for example, those of Fig. 8 (Mobius) and curves A and C of 

 Fig. 9 (Nichols and Tear). If these are reliable indices of waves of 

 corresponding wavelength in the mixed radiation from the oscillator, 

 the frequencies in question must be considerably higher than the 

 fundamental frecjuencies of the oscillators heretofore made; wave- 

 lengths ranging down to 0.1 mm., corresponding to frequencies ranging 

 up to 3-10'-, have been inferred from such curves. If these are over- 

 tones emitted by the oscillator along with its fundamental, there would 

 be little objection to extending the spectrum to cover them (although 



