670 TRANSACTIONS OF SECTION K. 
benzidine are each adapted admirably for the purpose of locating oxydases. By 
means of these reagents we have been able to map out the distribution of 
oxydase and peroxydase in the flowers and other parts of various plants, and 
although the work is laborious and the technique as yet imperfect, the results 
afford strong confirmation of the current hypothesis of the mode of formation 
of anthocyan pigments. This confirmation, however, was rendered possible only 
by reason of the fact that we worked with races of plants bred on Mendelian 
Jines, and hence of known genetic constitutions. 
Our method of investigation is briefly as follows. The oxydase-reagent is 
used in weak alcoholic solution, the part of the plant to be tested is incubated 
in the solution for a suitable time, and if no oxydase action takes place—that 
is, if no characteristic coloration of the tissues occurs—the material is tested 
for peroxydase by the addition of hydrogen peroxide. The method may be 
employed for intact corollas or petals or for sections of plant-tissues. 
Tt is important to mention that the first result of immersing a sap-pigmented 
tissue in either reagent is the decolorisation of the tissue. For example, a 
corolla of a coloured-flowered race of Primula s:nensis loses its colour com- 
pletely after being immersed for an hour or two in either reagent. The 
decolorised corolla, which in the case of 7. sinensis remains colourless, is treated 
with hydrogen peroxide, with the result that a well-marked peroxydase reaction 
is obtained. The reaction is confined to the non-chlorophyllous parts of the 
corolla, and does not occur, except in the epidermal hairs, in the region of the 
yellow or green eye, the tissues of which contain chlorophyll. Indeed, there is 
good reason to believe that chlorophyll inhibits oxydase action. 
By treating similar flowers with each of our two reagents we find that 
the actions of a-naphthol and benzidine are, in a considerable measure, supple- 
mentary one of the other. Thus the lilac-blue a-naphthol reaction is confined 
or almost confined to the veins of the corolla, the brown benzidine reaction 
is exhibited by the superficial (epidermal) cells and also by the veins. In order 
to emphasise the facts of distribution we speak of the peroxydases of 
P. sinensis as epidermal peroxydase and bundle oxydase. The former occurs in 
the epidermis and in the epidermal hairs, the latter in the bundle sheath which 
accompanies the veins. 
Similarly, if sections of a stem of P. sinensis be investigated they are found 
to contain a superficial peroxydase and a deep-seated peroxydase. As the 
result of remarking the peroxydases, not of any unknown variety taken at hazard, 
but of the several varieties characterised by constant differences of depth and 
extent of pigmentation, we have been able to show that the distribution of 
peroxydase in any one race coincides broadly with the distribution of pigment 
in the most pigmented races. In other words, in P. sinensis the peroxydase 
framework for pigmentation occurs throughout the species, and the building 
of the several colour varieties is determined by the activity of the factor for 
‘chromogen production. If we conceive of this factor as administered in a series 
of doses we can form a fair picture of the mode of evolution of the series of 
varieties characterised by increasing or decreasing amount of pigmentation of 
their vegetative parts. ; 
On turning to investigate the peroxydases in white-flowered races cf 
P. s‘nensis, we shall expect to find from analogy with the peroxydases of the 
stem that these agents of pigment-formation are not lacking from the corollas 
of recessive whites. The application of our reagents shows that this expecta- 
tion is correct, and that those white-flowered races which lack the factor for 
‘colour contain epidermal and bundle peroxydase. Hence we conclude that 
the absence of colour from recessive white flowers is due not tothe absence 
of peroxydase, but. to absence of chromogen. This conclusion is in conformity 
with that arrived at previously by Mendelian methods; for, as we have noted 
already, these methods demonstrate that anthocyan pigmentation of the flower 
,0f P. sinensis depends on the presence of one factor only, and that the absence 
of pigmentation which is characteristic of recessive whites is due to the absence 
of that single colour-factor. 
The result of our investigation of the peroxydases of dominant. white 
flowers is, on the other hand, quite different from that given by recessive 
whites. _ When corollas of dominant white races are treated with a-naphthol 
