G. WEBER 



89 



88°. Thus the simple jnctiue of two transitions at a right angle to 

 each other, as suggested by the character of the transitions themselves 



(19) , is experimentally confirmed. The maximum polarization (0.22) 

 observed at 270 m^u, and longer wavelengths is still considerably be- 

 low the theoretical maxinunn of 0.33 for the linear anisotropic os- 

 cillator, and it may be doubted whether this large difference can be 

 entirely due to the depolarizing effect of torsional molecular vibra- 

 tions, as Jablonski (8) suggested. 



The polarization spectrum of the fluorescence of indole and trypto- 

 phan (Fig. 5) is considerably more complex than that of tyrosine 



(28) . No negative polarizations are observed at wavelengths of ex- 

 citation of 220 m^ or longer. A region of positive polarization, with 

 a flat maximum at 270 m/A, is followed by a sharp decrease in the 

 polarization, with a minimum accurately corresponding to the spike 



(indole) or shoulder (tryptophan) observed in the absorption spec- 



POLARIZATION SPECTRA OF INDOLE, N-METHYL INDOLE 

 •25^ 



_ & TRYPTOPHAN 



•23 



•21 



•19 



17 



•15 



P. 13 



I I 



-09 



07 



•OS 



03 



O' 



o 



- Ol 



« » indole 



o — on methyl indole 



« -© tryptophan 



5" 



J L 



220 230 240 250 260 270 280 290 300 3IO 320 



A 

 Fig. 5. Polarization spectra of indole, N-methyl indole, and tryptophan. 



trum. The correspondence of the region of decreased polarization 

 with the spike of indole shows that this is an independent electronic 

 transition, with a transition moment at a large angle to the transi- 

 tion direction of the 270 mp. band. In Piatt's theory (19) of the elec- 

 tronic transitions of the aromatic hydrocarbons, the L^ and L^ 



