ee) 
bo 
for) 
CARL HALLQVIST 


be di-hybrids. Thus only 90 families are left for the third sowing. 
As 4 plants from each family have to be sown 360 plants are needed. 
At least 90 % of these 90 families must be assumed to have showed 
di-hybrid segregation (the figure has not been determined exactly as 
the results of the previous sowings have to be considered, which com- 
plicate the calculations), and only 9 such families have not yet been 
identified. The remaining 20 seeds from each family are now sown. 
The maximum number, 32, is reached, increasing the total with 180 
plants. The following summarizes the results. 
30 families segregating 3:1, each with 32 individuals.... 960 plants 
1 sowing of 450 di-hybrid families, each with 4 individuals 1800 » 
> » » 256 » ‘ » » » 4 » 1024 . » 
3 90 > > a > 360» 
41"; 
9 » » > 320 » 180 » 
Total 4324 plants 
The total number of plants becomes less, viz. 3790, in the case of 
coupling where the distinction is to be made between the constant 
and the 3:1-segregating families, and where only simplex recessive 
plants are taken as parent plants. 
If */, blue, '/, violet and */, bluish red had been selected according 
to the numerical relations in F, instead of following the procedure 
just discussed a number of 9000 or 10000 plants had to be raised. If 
the number had been limited to 32 plants only in each family, and, 
furthermore, if parent plants had been grown from every phenotypic 
group without any portioning out of the investigation on different 
sowings a number of 15360 had to be raised. 
In the case of a repulsion of 1 : 30 only one double recessive is 
expected in about 4000 individuals, which is just about the number 
required according to the F,-method. This shows clearly the advan- 
tage of the F,-method compared to the F,-method. That it also ad- 
vantageously competes with the back-cross method is seen from the 
fact that about 1000 plants had to be raised to secure a fixing of the 
gametic ratio with the same degree of accuracy. The raising af such 
a great number of hybrids must be considered a heavy task in the 
case of poor seeders and difficult pollination. 
It is clear that the above procedure saves much labour, and that 
an F,-analysis of sufficient extent is made easier. This must be of 
great value in cases where back crossing is difficult or impossible to 
perform on a large scale as, for instance, in the families Leguminosae 
