6 PHOTOCHEMICAL PRINCIPLES 



reasonably stable (for review, see Mustafa, 1952). In some cases the 

 dimer may be decomposed by heat to give monomers. Thus, anthra- 

 cene in solution readily forms dimers when illuminated with ultra- 

 violet light; the dimer is insoluble in hydrocarbons and is converted 

 to the monomeric form by heating. We have found that acridine dyes 

 form dimers on illumination with blue light; the photodimer absorbs 

 at longer wavelengths than does the monomer but it is unstable and 

 reverts to the monomer as soon as the illumination ceases (Oster and 

 Millich, to be published). 



Many dyes capable of cis-trans isomerization are capable of con- 

 version to one or the other form by suitably chosen incident light (for 

 review, see Wyman, 1955). Thus, cis-thioindigo (absorption maxi- 

 mum at 490 m^t) is converted by blue light into the trans isomer (ab- 

 sorption maximum at 560 nifi), and the latter is converted back to the 

 cis form by irradiation with yellow light (Erode and Wyman, 1951). 

 Heating also favors the trans form. Isomerization may also be operative 

 in some cases of reversible photoeffects of complicated organic mole- 

 cules at very low temperatures (Hirshberg, 1956). My own feeling is 

 that at least with bianthone derivatives the blue substance formed by 

 irradiation with ultraviolet which is destroyed by yellow light is a semi- 

 quinone form of the bianthrone. If so, it should exhibit a paramag- 

 netic resonance spectrum. Semiquinones are usually quite unstable 

 unless kept at low temperatures. They are highly colored and are 

 paramagnetic (for review, see Michaelis, 1940). It is possible that 

 far-red-absorbing semiquinones are produced by red light in photo- 

 periodism, but the required stabiUty of hours or even of days is not in 

 their favor. Semiquinones are generally unstable in the presence of 

 atmospheric oxygen, especially at room temperature. The possibility of 

 stable semiquinones of the anthocyanidines being the far-red pigment 

 should not be overlooked, however. 



A curious case of a hght-induced spectral shift is that of methylene 

 blue when bound to polymethacrylic acid (Wotherspoon and Oster, 

 1957). Only with this polymeric substrate and no other polymeric 

 acid is the dye converted with red light to a purple dye. The reaction 

 involves the conversion of methyl ammonium groups on methylene 

 blue to amino groups by some unknown process. Although this reac- 

 tion of itself has no relevance to photoperiodism, I mention it only to 



