————— i ee ee 
PRESIDENTIAL ADDRESS. 671 
or benzidine and subsequently with hydrogen peroxide, they show no sign of 
peroxydase neither in epidermis nor in bundles. Hence such flowers either lack 
peroxydase or else they contain a substance which inhibits peroxydase from 
exercising its oxidizing action on our oxydase-indicators. 
That oxydases may be inhibited in vitro has been demonstrated already 
by Gortner, who has shown that the addition of certain phenolic compounds— 
orcin, resorcin, etc.—prevents tyrosinase from exercising its characteristic action 
upon tyrosin. P ; 
Assuming that an inhibitor of peroxydase occurs in dominant white flowers, 
it may be supposed to act either by destroying peroxydase or by setting up 
conditions under which the activity of peroxydase is arrested. Assuming further 
that the inhibitor acts in the latter way, it follows that if means of destroying 
or removing the inhibitor be discovered and employed, the peroxydase released 
from the inhibitory grip should be free to effect the oxidation of our reagents. 
This train of reasoning gave us a point of departure for experiment. 
Starting from this point Dr. Armstrong and I have found in hydrogen cyanide a 
means of removing peroxydase-inhibition. Thus if dominant white flowers are 
immersed in a 0°4 per cent. solution of hydrogen cyanide for twenty-four hours, 
washed, and treated with either of our reagents together with hydrogen peroxide, 
pronounced peroxydase reactions are obtained, both in the epidermal and bundle 
tissues of the corolla. Carbon dioxide in aqueous solution produces a like, albeit 
a less pronounced, effect. 
Now, it so happened that we had at our disposal a race of Primulas, the 
flowers of which lend themselves peculiarly well to the purpose of confirming 
these observations. The race in question is characterised by blue flowers with 
fairly symmetrically placed paired white patches on each petal. We have 
reason to believe from the known ancestry of this race that these white patches 
are produced by a localised inhibitor. 
Corollas of these flowers treated with a-naphthol or benzidine become quite 
colourless. When, however, hydrogen peroxide is added the natural pattern 
is restored. The parts originally blue are stained lilac-blue or brown according 
to the reagent used, and the inhibitory patches stand out as in the intact 
flower as white areas on the coloured ground. 
lf instead of submitting the particoloured flowers directly to the oxydase 
reagent, they are treated first with hydrogen cyanide, and then treated with 
the reagent and subsequently with hydrogen peroxide, the inhibition located 
in the white areas is found to have been removed, and the peroxydase reaction 
is produced over blue and white areas alike. 
Hence the Mendelian hypothesis of the inhibitory nature of dominant whites 
is confirmed by biochemical methods. Moreover, these methods demonstrate thas 
the inhibitor acts not by destroying but by preventing the action of oxydase upon 
the chromogen. 
There are many other aspects presented by the phenomena of oxydase distri- 
bution in P. sinensis and other plants which we have investigated. Some I 
may enumerate, but lack of time must be my excuse for not dealing fully with 
any of them. 
The close proximity in the flower of the superficial and deep oxydases sug- 
gests that the latter may co-operate with the former in producing flower- 
pigments. This possibility entails the hypothesis of a translocation of oxydase 
from the region in which it is secreted to that in which it acts, and there are 
not a few facts which are in favour of this view; for example, the lines of deep 
colour which occur. along the veins of many flowers, the frequency with which 
the walls of cells appear to contain oxydase, the occurrence of oxydase 
in the mesophyll cells which adjoin the bundle sheath, and the evidence pro- 
vided by the mutual influence of stock and scion in grafted plants and in graft 
hybrids. Though these and other subjects must be passed over, I cannot resist 
giving what appears to me to be the most elegant mode of demonstrating the 
relation between oxydases and pigmentation which we have as yet observed. 
The plant which has served for this purpose is the Sweet William (Dianthus 
barbatus), and any of the old-fashioned races of this plant common in cottage 
gardens suffices, provided that it be an ever-sporting race. Such a race is 
known by the fact that it bears, on one and the same head, flowers of different 
