Page 150 - Breeding and regulatory opportunities, Renaud

Chapter 4

that heritabilities are higher in high input environments compared to poor

environments (Ceccarelli, 1994; 1996). In our trials, repeatabilities for head

weight, head diameter, hollow stem and overall quality were higher for organic

compared to conventional, while for the traits of maturity, head color, and

head surface, repeatability levels between management systems were equal

or near equal. It is recognized that these coeicients combine additive and

non-additive genetic variance, and it would be anticipated that they would

be much lower if the additive component was partitioned out. For the traits

of head shape, bead size, and bead uniformity, repeatabilities were higher

in conventional compared to organic, which could be explained by a more

variable organic management environment. The traits with repeatabilities

larger or equal in organic systems were those generally associated with growth

and productivity, and probably under similar genetic control, whereas those

with repeatibilities lower in organic compared to conventional are probably

under separate genetic control. Higher heritabilities under organic conditions

were also found by Burger et al. (2008) and Goldstein et al. (2012) for maize

and for faba bean (Vicia faba) (Link and Ghaouti, 2012). They supported their

indings with the following justiications, which can also explain our results: (1)

with heterogeneous soils found in organic management systems the precision

of experiments may be more impaired under stress (slow nutrient release) than

under conventional high input conditions; (2) genetic variance may be greater

under stress conditions than non-stress conditions, and (3) the high genetic

variance in organic trials compensated for the high experimental error which

produced comparable heritabilities between organic and conventional trials.

Trait repeatabilities alone are not suicient to determine the optimum selection

environment. Both estimates of genetic variance and repeatabilities are useful in

predicting the response to selection in organic and conventional management

systems. Estimates of the genetic correlation between performance of traits in

the organic and conventional management systems is an indicator for the extent

of G × M interaction. In our broccoli trials the genetic correlations between

organic and conventional trials for the traits head weight, maturity, head shape,

and plant height were high (>0.90) indicating that a diferential response of

the genotypes to the two management systems was largely absent. The ratio

of correlated response to direct response for all traits was close to but below

1.0 with the exception of bead uniformity. This would imply that in most cases,

132