FLOWER AND SEED COLOUR IN LUPINUS 333 

that the deep blue colour is the result of the co-operation of the two 
factors B and V, as both factors seem to cause alkalinity and as, 
further, the effect of the V-factor apparently pervades the whole flow- 
er. It is, indeed, of great interest that the neutral colour violet 
genetically is an intermediate between red and blue, between acid and 
alkaline colour. However, the schematic outline of the hypothesis 
of WILLSTÄTTER has its limitations; it was found that the red pigment 
of the rose-coloured form of Centaurea was not a cyanin variety, 
although expected, but a related still different pigment called pelargonin. 
This shows that the generalization made in the above must be looked 
upon with reservation. 
A genetical analysis of a plant with similar colour scale has not 
yet been made, so far as I know. Deep blue colour together with 
violet and red do not occur, as a rule, in the plants more fully investi- 
gated in this respect (Antirrhinum, Matthiola, Mirabilis, Papaver etc.). 
In other cases a variety is found to be blue, being in these cases recessive, 
and not the type (Lathyrus, Phlox, Anagallis). The phenotypic dif- 
ference between Lupinus and the latter genera would answer to the 
lack on the part of these of one link in the factorial chain of Lupinus. 
Purple in Lathyrus would correspond either to bluish red or violet 
in Lupinus, and Lathyrus would lack one or the other of the Lupinus 
factors B and V, which explains the absence of the blue compound 
form BV. 
The necessity of assuming a fundamental factor for the appearance 
of colour on the whole is seen from crosses 6 and 7. White has been 
crossed in both these crosses with bluish red and violet respectively. 
F; is blue in both cases. With the knowledge of the constitution of 
the blue colour in mind the white parent plant must be assumed to 
have had at least one of the complementary factors for blue colour. 
We have, then, to assume that the white colour was due to the absence 
of a fundamental factor necessary for the development of bluish red 
as well as of violet, and thus also of the blue colour. Cross 5, white 
X tinged blue, favours also the idea of such a fundamental factor. 
However, it is not necessary to explain the blue colour in F, in this 
way. The tinged blue parent has indeed all determiners for blue co- 
lour and lacks only the full colour factor, which is introduced, however, 
by the white parent. It would be possible to assume that the white parent 
was devoid of all colour factors, and that it showed white colour be- 
cause of this circumstance and not because of the absence of a funda- 
mental factor only. The segregation in F,, however, would then have 
