344 CARL HALLQVIST 
of the poly-hybrid segregations are facts of interest still uninvestigated. 
In the crosses analyzed, comprising about 800 families and 27000 indi- 
viduals, segregation with regard to earliness was also found. The most 
surprising uniformity prevailed in flowers as well as in vegetative parts 
for the rest. It is true, the colour of the seed varies remarkably 
with regard to the intensity and the distribution. The variations are, 
no doubt, only modifications, as the colour of the seed is very sensi- 
tive to environmental conditions. New mutations have not been 
observed. 
On this occasion I wish to express my sincerest thanks to Mr. 
H. WEIBULL, director of the plant breeding institute at Weibullsholm, 
Landskrona, for his great kindness of promoting these studies in dif- 
ferent ways. 
SUMMARY. 
1) Seven different types of flower colours and five types of seed 
colours have been genetically analyzed. 
2) One fundamental colour factor has been demonstrated (pure 
red). A synthesis of blue colour has been obtained from crosses 
between bluish red and violet flower colours. One »dilution» factor 
has been found to be present. 
3) Pleiotropic correlation has been demonstrated between certain 
flower and seed colours. 
4) Three flower colour factors have been found to form a linkage 
group. The linkage between two of the factors is very close, if not 
complete. The other linkage value represent a crossover percentage 
of about 22 %. 
5) The importance of the »F,-analysis» in the determination of 
the linkage values in cases where back crossing is technically difficult 
has been demonstrated. 2 
6) A method, that greatly facilitates the carrying out of the F;- 
analysis, has been proposed. 
7) The colour scale blue-violet-red (blue as the wildtype), now 
investigated genetically, is phenotypically analogous with the colour 
scale in the genus Centaurea a. 0. chemically analyzed by WILLSTÄTTER. 
