308 CARL HALLQVIST 
brown seed of the blue flowers are crossed with the rust-brown in 
the violet and with the white seed of the whites. All these combina- 
tions have been made (crosses 1, 2, 3, 4), and both the F.- and the F,- 
generations have given the results expected. 
Cross 1 (table 7). | 
White flower colour 4 | blue flower colour 
» seed colour | | earth-brown seed colour 
rrBBVVFF RRBBVVFF 
F, is blue and earth-brown, and the segregation in 
F:: 289 blue: 73 white, expected 271,5 : 90,5. 
The difference — 17,5 is 21 times the standard error (8,21) and, 
consequently, somewhat great, yet it is not so large as to lie beyond 
the limits prescribed. F, was spoiled. The inheritance of the seed 
colour was not followed in this cross. 
Cross 2 (tables 8—9). 
Violet flower colour, rust-brown seed colour (RRbbVVFF) X blue 
flower colour, earth-brown seed colour (RRBBVVFF). 
F, is blue with earth-brown seeds, and F, has given: 
215 blue: 73 violet (table 8) 
expected: 216 72 D/M = 1,0/7,35 = 0,14, 
3 violet F.-plants have given constant violet progenies in F; 
(table 9) with 247 individuals in all. Among 9 F,-families from blue 
parent plants 1 has given a constant blue progeny (22 individuals), 
and 8 have segregated in: 
456 blue : 154 violet 
expected: 457,5 : 152,5 D/M = 1,5/10,69 = 0,14. 
The results of the segregation in F, as well as in F, are, as seen, 
D 
pretty satisfactory, although the proportions between the constant and — 
the segregating F,-families might have been better. Instead of the 
expected 3:6 proportions 1:8 was obtained. The low number of 
plants makes the deviation allowable. The seed colour was determined 
in F, in 209 blue and 69 violets; all blue were found to have earth- 
brown seeds, all the violets rust-brown seeds. 
Cross 3 (tables 10—11). 
Bluish red X blue flower colour, no difference in seed colour. 
RRBBvvFF = RRBBVVFF 
F, is blue, and F, segregates in 
