346 LOCK : STUDIES IN PLANT BREEDINC4 
a separate colour. The expectation is then of only 1 white 
among 16 individuals. 
De Vries (69) describes a more definite case, and his des¬ 
cription, though following that of Mendel, can be made 
somewhat clearer. We may, therefore, pass on to it at once. 
De Vries crossed Antirrhinum vulgaris having red flowers, 
by a variety the flowers of which were white. In ¥ 1 all the 
plants had red flowers like the dominant parent. In F 2 the 
colour of the flowers on the different plants varied consider¬ 
ably, but de Vrieswas able to distinguish among 49 offspring 
of 1 cross the following proportion :— 
51 per cent, red, 16 per cent, flesh-coloured, 31 per cent. 
“ Delila,” 2 per cent, white ; and one of these red forms 
(a heterozygote) gave in F 3 :■—58 per cent, red, 17 per cent, 
flesh, 20 per cent. Delila, 4 per cent, white, which approaches 
the proportion 56*25 : 18 75 : 18*75 : 6*25, or 9 : 3 : 3 : L 
characteristic of a Mendelian dihybrid. 
De Vries explains this case by regarding red as a combi¬ 
nation of flesh and Delila —- F + D and white as the 
absence of either—W + W'. 
Red x white gives FDWW' (gametes FD, FW', WD, 
WW') giving the combinations— 
(F + 2 FW + F), (D + 2 W'D + W') whence the zygotes 
FD + 2 FDFW' + FW' 
2 FDWD + 4 FDWW' + 2 FW' WW 
WD + 2 WDWW' + WW'. 
Each of the terms containing F and D is red, those contain 
ing F, but not D, are flesh ; D, but not F, Delila. 
There result— 
(a) 1 constant red 
... 
FD 
(b) 2 red which will give 
in F s red and flesh 
2FDFW' 
(c) 2 ,, ,, 
red and Deliia • ... 
2FDWD 
(d) 4 
red and F, D, and W.... 
4FDWW' 
(e) 1 constant flesh 
... 
FW' 
(f) 2 flesh giving in F s flesh and white 
2 FW f WW' 
(g) 1 constant Delila 
WD' 
(h) 2 Delila giving in F 3 
Delila and white . ... 
2 WDWW' 
(i) 1 constant white 
... 
WW 
