62 



I. NAGAI : 



upon esparimeutal evident'?, tliiit the metal organic or complex compauncls of 

 rednced flavonol gliicoside 



/ 



(sugai-J 



3IeX 



O 



/\/\_./ v 



I 



V OH 



-OH 



J 



MeX. 



are the most important factor in the production of flower colom-s. The blue 

 anthocyanins ai'e complex compounds of reduced flavonol glucosides, which 

 possess several hydi-oxyl gi-oups belonging to the flavonol nucleus besides those 

 of sugar molecules, and the metal with which they are co-ordinated is probably 

 calcium or magnesium, for salts of these metals are always present in the 

 plant cells. The violet, violescent red or red pigments ai-e either the analogous 

 metallic complex compounds of flavonol glucosides, wliich contain fewer of the 

 auxoclirome hydi'oxyl gi'oups or are a mixture of the blue pigments and their 

 decomposition products bj- excess of acids, i.e., the red oxonium salts of 11. 



WiLLSTATTEE. 



It is likely to be inferred then that the purple and rod anthocyanins 

 formed in the awn and in other parts of Ori/za are the derivatives of the sarte 

 chi'omogenic flavonol glucoside and the pm-ple is the complex salt of the red 

 anthoc3*anin which is formed at first in the cell b}' the reduction of the 

 clu'omogenic substanc3. The latter part of the changes may be due to the 

 action of gene B. Indeed, it can be obseiTed that the extract of the shglitly 

 red colom-ed awn yielded a more intense red colour by reduction than that 

 produced by a simple addition of hydi'ochloric acid, and the faintly coloured 

 exti'act of red and purple anthocyanins that is due to isomerization, attained 

 a bluish hue by tlie addition of ZuCL and a reddish hue by CaCL. The 

 gene R' may therefore be i-efen-ed to the agencies which set free the reactions 

 leading to the formation of a pm-ple complex salt fi-om red anthocyauin ■with 

 tlie existing metallic salts in the cell. 



With respect to the colour of the paleas, tlie following hypothesis may 

 be pravide«.!. Let J] be the gene for the brown gixjuud-colom-, l> for tlij non 



