56 SECTIONAL ADDRESSES 
cent. hydrochloric acid, and moreover the presence of much anthoxanthin 
is noted. Hence, all the circumstances point to co-pigmentation of the 
pelargonin salt in the flower petal. 
Finally, we do not know whether or no traces of iron and other inorganic 
substances may affect flower colour. In this connection the case of the 
blue hydrangea is always quoted, and we have observed that when the 
stalks of red hydrangea flowers are immersed in very dilute aqueous 
ferric chloride, the flowers slowly become blue. .The ashes of many 
flowers contain 1-2 per cent. Fe,O,, and the anthocyanin test for iron is 
one of the most delicate known. 
Summarising, the main factors affecting flower colours are : 
(1) The nature and concentration of the anthocyanins and other 
coloured substances present. 
(2) The state of aggregation of the anthocyanin in solution—the pH 
of the cell-sap is one of the subsidiary factors affecting this, and naturally 
the presence or absence of protective colloids is another. 
(3) The presence or absence of co-pigments and, problematically, the 
effect of traces of iron and other complex forming metals. 
Time does not permit me to deal with other anthocyanin types such as ges- 
nerin, the leuco-anthocyanins, the yellow anthocyanin of Papaver nudicaule, 
or the nitrogenous beet-pigment and its analogues ; I will close, appro- 
priately I hope, on an experimental note by attempting a demonstration of 
some of the tests which we employ for the recognition of anthocyanidins. 
(1) The oxidation test—addition of 10 per cent. aqueous sodium 
hydroxide to a dilute solution shaken with air—petunidin and delphinidin 
are at once destroyed, the other anthocyanidins are relatively stable. 
(2) Extraction with amyl alcohol, addition of sodium acetate and then 
of a trace of ferric chloride. Characteristic colour reactions are observed, 
and in particular if cyanidin is present the violet amyl alcoholic solution 
changes to pure blue in the last stage. Pelargonidin, peonidin and 
malvidin give no ferric reaction. 
(3) Distribution between 1 per cent. aqueous hydrochloric acid and 
a mixture of anisole (5 vols.) and ethyl zsoamyl ether (1 vol.) containing 
5 g. of picric acid in 100 c.c. Delphinidin is not extracted by the organic 
layer, petunidin is taken up to a slight extent, cyanidin to a considerable 
extent, and malvidin, peonidin and pelargonidin are completely extracted 
if the solution is sufficiently dilute. 
(4) Distribution between 1 per cent. hydrochloric acid and a mixture of 
cyclohexanol (1 vol.) and toluene (5 vols.). Delphinidin and petunidin 
are not extracted at all; malvidin gives the organic layer a faint lilac 
tint ; cyanidin a pale rose tint ; peonidin, and still more pelargonidin, are 
extracted by the organic layer to a considerable extent. 
The deductions are confirmed by a study of the colour reactions of the 
anthocyanins. 
In conclusion I must express my very deep indebtedness to all co- 
workers in these fields, and especially to Dr. D. D. Pratt, Prof. A. Robert- 
son, Dr. W. Bradley and Dr. A. R. Todd on the synthetic aspects of the 
work, and to my wife, without whose co-operation a survey of natural 
anthocyanins could not have been attempted. 
