B.—CHEMISTRY 55 
same petal pH); sweet-peas, all about 5-3; delphiniums, 5-6 (most 
violet shade), 5-8 (most blue shade) ; clematis (blue), 5:4; viola, 6-2 
(blue violet), 6-0 (reddish violet) ; lobelia (blue), 5-7. 
It must be emphasised that these variations of pH are quite insufficient 
in themselves to account for the colour changes and it is evident that the 
most important single factor for flower colour, given the nature of the 
anthocyanin, is the question of the condition of the pigment in solution, 
and it would appear that all blue flowers are coloured by colloidal solutions 
of their respective pigments. 
Methods for the determination of the pH of the cell-sap of flowers 
depending on the use of the flower colours as indicators may be sound, 
but only if it can be guaranteed that the colloidal condition of the pigment 
solution is not altered by the extraction with the buffered solutions which 
are employed. In any case, the results bear no relation to the colours 
observed in vitro using isolated anthocyanins and they cannot be trans- 
ferred from flower to flower ; the colour series depends almost as much 
on the other conditions in the cell-sap as on the pH and on the nature of 
the anthocyanin. Another aspect of pH of the cell-saps is that the higher 
values appear to be associated with the formation of delphinidin deriva- 
tives. The remarkable distribution in the tropeolum—Empress of India 
—is as follows: leaf, delphinidin diglycoside (pH 5-6); calyx, cyanidin 
3-bioside (pH 5:0); flower, pelargonidin 3-bioside (pH 4-5). On the 
other hand three scabious with anthocyanins based respectively on 
pelargonidin, cyanidin and delphinidin had all the same petal pH 5:0. 
We have already discussed elsewhere the influence of certain substances 
termed co-pigments on the colour of anthocyanin solutions ; these effects 
are to be detected in strongly acid solution and the presence or absence 
of these substances is undoubtedly a factor to be taken into consideration. 
The extent to which the co-pigment effect is bound up with colloid 
phenomenon is a matter for future experiment and discussion, but it is 
convenient to maintain the term co-pigment for the present. 
Dr. E. A. H. Roberts has observed the shift of the absorption bands 
of chrysanthemin and cenin chlorides on the addition of papaverine 
(strongly blueing effect) and narcotine (weak effect), and correlated this 
with a corresponding change (lowering) of the distribution number of the 
anthocyanin using amyl alcohol (demonstration). 
It seems clear that papaverine salts and cenin salts combine in solution. 
The relation between the distribution number of cenin chloride and the 
concentration of the pigment seems to require the assumption that the 
molecules of the anthocyanin are associated (2 mols.) in aqueous solution 
and free in amyl alcohol. Chrysanthemin and idzin behave similarly, 
also malvidin 3-galactoside. ‘This phenomenon appears to be related to 
that of co-pigmentation. 
The naturally occurring co-pigments include the anthoxanthins (flavone 
and flavonol saccharides, etc.) and tannins and some efficient substances 
not yet identified. 
The justification for assuming the operation of this factor can best be 
indicated by an example. Certain herbaceous phlox contain pelargonin, 
but have a much bluer-red colour than other flowers coloured by this 
anthocyanin. But the same observation applies to the extract in I per 
