Mendelian Factors for Flower- Colour. 



307 



reducing agent reverses the reaction. With a decrease in amount of 

 water in the cell the reducing agent becomes active and the oxydase inert, 

 but with an increase in amount of water the oxydase becomes active and its 

 effect is greater than that of the reducing agent. Hence, when petals are 

 treated with strong alcohol the oxydases can no longer function, and the 

 reducing agent is then able to reduce the anthocyanin to a colourless leuco- 

 compound. On addition of water the oxydase again becomes active and 

 re-oxidises the leuco-compound. 



Such is the hypothesis, but we are not clear as to the explanation offered 

 by the authors for the reappearance of colour in the alcoholic solution apart 

 from the petals. Two alternatives offer themselves. First, that both oxydase 

 and reducing agent are extracted by 95-99-per-cent. alcohol and are present 

 in the alcoholic extract and that neither is affected by heating to 100° 0.* (in 

 spite of the fact that extraction by absolute alcohol and resistance to heat is 

 not characteristic of oxydases), and that, although the authors quote 

 experiments to prove that the oxydase can oxidise to some extent in 

 95-per-cent. alcohol, the reducing agent is more powerful in this medium. 

 Or, that the reducing agent alone is extracted by alcohol and its influence 

 is removed by evaporating the alcohol or by diluting, when re-oxidation occurs 

 merely on exposure to air. If the latter be the case, the presence of the 

 oxydase is superfluous to the recovery of colour in the petals themselves. 

 We must also conclude that the reducing agent is very widely distributed, is 

 unaffected by temperature of 100° O, and can only act in presence of alcohol. 



To us the reduction and oxidation hypothesis appears directly opposed to 

 essential experimental facts, although the original production of anthocyanins 

 in the plant is, in all probability, either partly or wholly due to the action of 

 an oxydase on a chromogen, most likely a flavone or xanthone. 



In our experiments, various coloured petals of Stocks were used, and these 

 were the flowers also used by Keeble, Armstrong, and Jones. 



Experimentally we found that the same results are given both by the 

 decolorised petals and by the alcoholic solution. 



We find that if a little acid is added to absolute alcohol containing 

 decolorised petals, the usual red colour reaction of acid with an anthocyanin 

 is obtained both in the solution and in the petals. Moreover the same result 

 is obtained equally well when dry hydrochloric acid gas or dry hydriodic 

 acid gas is passed through the alcohol. Also, contrary to the observations of 

 Keeble, Armstrong, and Jones, we find that prussic acid gas acts quite as 



* In a later paper (Jones, "W. N., "The Formation of Anthocyan Pigments. Part V. — 

 The Chromogens of White Flowers," ' Eoy. Soc. Proc.,' 1913, B, vol. 86, p. 318) the author 

 definitely states that oxydase is destroyed by boiling 50-per-cent. alcohol. 



