Ryan and O'Neill — Studies in the Diflavone Group. 169 



Again, dihydroxy-diehalkone (i) may exist in three stereoisonierie modifi- 

 cations (l'j i", l'") : — 



HC-C-H HO-C^ , C-OH H-C-CH H-C-CH HO-C^ C-OH H-C-CH 



56 II I I II II 65 II 6S I 1" II II 65 



C CO C C CO C-H H-C CO r C — CO C-H 



^*CH^ ^CH^ 



H-C-CH HO-C^ C-OH HC-C-H 



|| 6 5 | Y" || 56 H 



H-C CO C. C CO C-H 



^CH^ 



Since flavindogenides of the type iv are not formed by the condensation of 

 ortho-hydroxy-ketones with aldehydes in the presence of alkalis, although 

 readily formed, as we shall show later, in the presence of acids, it is unlikely 

 that any of the four isomers isolated by us can have that formula. "We are 

 therefore limited to a consideration of the remaining five formulae. 



The a, (5, and S isomerides on acetylation with sodium acetate and 

 acetic anhydride give oily acetates for acetate', which on addition of bromine 

 form the same tetrabromide of dihydroxy-dichalkone-diacetate, and there- 

 fore on treatment with alcoholic potash yield the same diflavone. 



The method by which the /3-isonieride was first obtained by us — i.e. the 

 action of aluminium chloride on dihydroxy-dichalkone-dimethylether — shows 

 that it must be dihydroxy-diehalkone : — 



C 6 H 5 CHO A1C1 3 

 C 6 H 2 (OCH 3 ) 2 (CO CH 3 ) 2 > C 6 H 2 (OCH,) 2 (CO CH : CH C 6 H 5 ) 2 > 



/0\ HC1 



AL C 6 H, CO CH : CH C 6 H 5 ) 3 — -> (HO). . C 6 H, . (CO CH : CH C 6 H 5 ) 2 . 

 \0/ 



The behaviour of the substance towards warm dilute alkali and towards 

 alcoholic ferric chloride is in agreement with this formula, as also is the fact 

 that it can be readily converted into the tetrabromide of dihydroxy-dichal- 

 kone-diacetate : — 



C 6 H 2 (0 OCCH 3 ) 2 (CO CHBr . CHBr . C 6 H 5 ) 2 . 



Moreover, since the a and S isomerides can also be readily converted into 

 the same diaeetate-tetrabroniide, it is evident that the latter substances are 

 either polymorphic or stereoisorneric modifications of the (5 compound. 



As the three substances crystallize from chloroform and alcohol in distinct 

 crystalline forms, and as each form retains its individuality even when a 

 saturated solution of it is inoculated with crystals belonging to the other 

 forms, the differences between the compounds can scarcely be due to 

 polymorphism. 



It seems to us, therefore, that the three substances are the three 



[2 5 2] 



