314 CARL HALLQVIST 

As to the seed colour the above mentioned whites and their 
progeny had all white seeds, the rest of the F,-plants examined with 
regard to seed colour, 25 in all, had earth-brown seeds. 20 of these 
were blue coloured, the other 5 tinged blue. 
The di-hybrid crosses just discussed, where one of the parents 
has been white follow the hypothesis exactly, and the results of the 
segregations are quite satisfactory. The group of di-hybrid crosses 
where both parents have been coloured, although not blue coloured, 
has now to be discussed (crosses 8, 9 and 10). | 
Cross 8 (tables 20—21). 
Tinged blue flowers, earth-brown seeds (RRBBVVff) X violet 
flowers, rust-brown seeds (RRbbVVFF). 
F, sligthly diluted blue, earth-brown seeds. 
F, should segregate in blue, violet, tinged blue and tinged violet 
in the ratios 9:3:3:1, according to the hypothesis. The result was: 
470 blue : 241 violet : 255 tinged blue (table 20). | 
Thus the expected mode of segregation has not been realized. A 
double recessive of the violet tinged type is lacking; mere chance is 
excluded on account of the great number of individuals. About 60 
plants of this type are to be expected; however, since none has 
been obtained, a deviation 8 times the standard error has to be 
assumed, a deviation so large as to lie far beyond the allowed range. 
It could be argued, perhaps, that the absence of the F-factor did not 
influence the violet colour. This seems very improbable, as the num- 
ber of the blue and tinged blue groups is abnormal, the former is 
too small, the latter too large to account for the ratios 9 blue : 3 tinged 
blue : 4 violet to be expected in such a case; the results obtained in 
F3 also invalidate this argumentation. If valid, mono-hybrid famil- 
ies together with di-hybrids should be obtained from blue coloured 
Fs-plants, and mono-hybrid families in addition to the constant fam- 
ilies should originate from tinged blue parent plants. This is not 
the case, however. The numerical relations showing the best cor- 
respondence with the numbers found are 2 blue : 1 violet : 1 tinged blue 
giving when applied to ‚the number of individuals in question: 
483 blue : 241,5 violet : 241,5 tinged blue 
13 0,5 13,5 
D/M = (),84 0.04 
= 
15,5 ER plc Ina 
