KENNETH V. THIMANN AND (lEORC.E M. CURRY 653 



ill hexane is seen to agree exceptionally well with the observed ac- 

 tion spectrmn (Curve A) in the visible; and Curve C shows that 

 such a carotenoid is indeed cxtractable from the tip. However, the 

 carotene has no absorption in the near ultraviolet. It appears that 

 such absorption is a Junction of the spatial configuration of the chain 

 of conjugated double bonds. Presence of one or more ceVconfigura- 

 tions introduces a band, usually between 300 and 350 m^^. Curve D 

 shows the absorption spectrum of 9-9' cis ^-carotene, synthesized by 

 InhofFen, Hohlmann, and Rummert (1950); the ratio of absorption 

 in the ultraviolet to absorption at the main peak is just over 0.5. 

 This curve could be considered very acceptable agreement with the 

 action spectrum were it not for the location of the <:?5-peak at 340 

 instead of 365 m^. To date, in hexane solution none of the ciVpeaks 

 recorded in the literature has been found to approach within 140 m^ 

 of the long wavelength peak of the corresponding Sill-trans form. 



In 1949, however, Galston and Baker suggested that the photo- 

 receptor for phototropism might be riboflavin; this proposal was 

 leased on (a) the general agreement between the zone of absorption 

 in the visible by riboflavin and the range of wavelengths of photo- 

 tropic sensitivity, and (b) the ability of riboflavin to sensitize the 

 photoinactivation of indole acetic acid in vitro. The absorption 

 spectrum of riboflavin in water is shown in Curve E; it is evident 

 that although it does show a peak at about 365 m^^, this has almost 

 the same height as that of the main peak at 450 m^, while in the 

 visible the curve differs greatly from the action spectrum for photo- 

 tropism. Most of the known flavoproteins, including at least four 

 pure flavin-enzymes, have absorption spectra very similar to that of 

 riboflavin. However, in one or two cases, notably the dihydro- 

 thioctyl dehydrogenase of Searls and Sanadi (1959), the usually 

 smooth curve in the visible is replaced by an asymmetric one, with 

 the main peak at 456 m^u,, and a clear shoulder appearing at about 

 485 rufi. This evidence of fine structure becomes greatly magnified 

 in the ribose-free derivative lumiflavin, and especially in 3-methyl 

 lumiflavin; when these substances are dissolved in benzene the 

 shoulder at 485 m/^ actually appears as a second peak, and in addi- 

 tion a new shoulder appears at 425 m^a (curve F of Fig. 5) . Har- 

 bury et al. (1959) have shown that this behavior is a function of 

 the solvent; water or other hydrogen-bonding solvents suppresses 

 the separation of the peaks, making the absorption in the visible 

 spectrum approximate to a smooth curve, but at the same time they 

 shift the peak in the ultraviolet up from 340 nifj., in benzene, to about 



