M. Wheldalb 121 



The results obtained were as follows. Broadly speaking, in all 

 coloured varieties, jjeroxidases were found in the epidermis and bundle 

 sheathes, the tissues in which anthucyanin is distributed. Flowers of 

 the dominant white varieties were found to give oxidase reactions only 

 after certain inhibitors were artificially removed from the flower. Albino 

 varieties gave the oxidase reaction as fully as coloured varieties. Hence, 

 if the hypothesis that colour in Primula is due to the action of an 

 oxidase on a chromogen be true, we must assume that albinism is, in 

 this case, due to the absence of chromogen. Since the chemical nature 

 of the chromogen is unknown in Primula, this assumption can, for the 

 present, be made, though all Primula albinos certainly contain a flavone. 



In the same way all albinos of the Sweet Pea and the Stock contain 

 both peroxidase and flavone, and here again, if the Keeble and Armstrong 

 hypothesis be con-ect, we must assume the absence of chromogen to be 

 the cause of albinism. 



But in the case of Antirrhinum, peroxidase is present in both the 

 ivory parent and the white parent which on crossing produce magenta. 

 Since the evidence is strongly in favour of apigenin being the chromogen, 

 some third substance or reaction must be postulated and the high 

 molecular weights of the anthocyanins lend support to the hypothesis 

 that there is some further process involving condensation. 



From the work of Keeble, Armstrong and Jones we can only at 

 present make the deduction that anthocyanin is formed in tissues 

 vi'here oxidation can take place, a;nd there is no evidence among the 

 well-known Mendelian cases of colour inheritance of a factor being 

 represented by a peroxidase. 



Quite similar results to those of Keeble and Armstrong have been 

 obtained by Atkins (2, 3), working with flowers of many species and 

 varieties of Iris. The author shows that the distribution of peroxidases 

 and inhibitors is correlated with the natural colouring of the flowers. 



The hypothesis that anthocyanin is formed by oxidation and con- 

 densation or by oxidation alone of the flavones is entirely refuted by 

 Combes (9, 10, 11, 12), who maintains that anthocyanin, on the contrary, 

 is formed by reduction fi'om flavones. His evidence may be summed 

 up as follows : 



(1) There are two pigments in Ampelopsis hederacea, a yellow 

 (flavone) and a red (anthocyanin). 



(2) The flavone can be converted into a red pigment, identical with 

 natural anthocyanin, by treating an alcoholic solution of the yellow 



