Page 235 - Breeding and regulatory opportunities, Renaud

General Discussion

that head weight and diameter and late maturing cultivars were positively

correlated with glucoraphanin. Head colour was positively correlated with the

carotenoids, and positively correlated with early head maturation (particularly

in the Spring). Because head and stem colour are important traits to processors,

a cultivar could be bred for segmented heads (where the inlorescence divides

into individual lorets, e.g. broccolini) for early maturation and darker colour.

Positive correlation between darker head colour and carotenoids may be

a function of chloroplast density while a correlation between early maturity

and darker heads may be related to N uptake and use eiciency. A breeding

strategy for carotenoids would need to take into account the implications of

known genetic correlations, contribution of seasonal inluence and nitrogen

use eiciency.

Selection environments

As stated in the introduction of this thesis, most studies that have investigated

traits needed for organic farming systems have focussed on ield crops such as

cereals (e.g. Murphy et al., 2007; Löschenberger et al., 2008; Przystalski, 2008;

Wolfe et al., 2008; Annicchiarico et al., 2010; Reid et al., 2009, 2011; Kirk et al., 2012;

Koutis et al., 2012). Only a few studies had been conducted on vegetable crops,

for instance for onion (Osman et al., 2008; Lammerts van Bueren et al., 2012),

and that remains the case to date. From the trials performed in this study, the

main efects of location and season described the largest source of variation in

broccoli trait performance. For example, Oregon trials produced higher average

head weights than Maine in both seasons, and the Fall trials produced higher

head weights than Spring in both locations (highest overall head weights in

Oregon Fall trials). Greater heterogeneity in the organic management systems

and genotype by management crossover interactions were observed on a

local per trial scale. This supports the idea that direct selection (under organic

management) would potentially be beneicial for the development of cultivars

for organic agriculture, particularly if the intent of the breeder is to develop

cultivars for local adaptation. Burger et al. (2008) concluded that direct selection

under organic conditions for complex traits such as yield is preferred, whereas

indirect selection can be very eicient for highly heritable traits. Burger et al.

discovered that although heritabilities in their trials with genetically broad

populations of maize were assumed to be lower under organic farming due to

higher experimental error rate, these were compensated by greater genotypic

217