128 
B. IKENO : 
from white-II x orange, and since we see there the complete linkage between 
the factors jB and 13, it would not be probable tliat it has had that composi¬ 
tion from the very beginning ; in all probability it has been at first of the 
constitution CC1011313, changed into CCJRBBb by a “loss-mutation”, 
and then undergone the segregation above stated. 
In all cases under I above enumerated (and in some cases under III) I 
have assumed that the reversion takes place during the formation of gametes, 
but it would he equally possible that this process of mutation occurs, not 
duiing the gametic formation, but some time before in the somatic cell. To 
cite one instance for illustrative purpose, in No. 6 (Table VIII) where 
white-I ( ccrvbb ) gives rise to orange of the constitution Ccrrbb the 
process may be as follows :—a somatic cell of the constitution ccrrbb gets 
that of Ccrrbb by reversion ; such cell produces during each succeeding 
cell-division the cells Ccrrbb, so that in the reducing division the gametes 
of the constitution Crb and crb are formed. (Compare the discussion 
in p. 124 . If what is above described be true the wdiole process of reversion 
in I (and also in some cases under III) is essentially identical with that in 
II. To determine exactly in each individual case which alternative will be 
realised, i.e. whether the reversion occurs in the somatic cell or first in the 
formation of gametes would be an almost impossible task in the present state 
of our knowledge. 
IY. Each of the F, magenta parents in Nos. 16, 17 and 18, derived 
from the cross white-II X magenta has segregated into magentas and whites, 
and produced in addition 1 orange, 1 flesh-coloured and 1 pseudo-white. As 
this segregation has given rise to 90 magentas and 42 whites in total, i. e. 
99±4-97 and 33±4 - 97 respectively on 3:1 expectation we may in all 
probability regard each magenta parent in our case to be of the constitution 
CcHHBB. The question, however, to what kind of mutation will be due 
the production of flesh-coloured and pseudo-white plant in this case would be 
quite unexplainable, especially as the exact genotypic composition of the two 
latter is yet unknown. 
In respect to Nos. 19 and 20 we may consider that the F, magenta 
parent is homozygous, i.e. of the constitution CCMRJBB and has bred true 
