Page 219 - Breeding and regulatory opportunities, Renaud

General Discussion

In the ield trial and phytochemical testing component of this study (Chapters

4 and 5), the aim was to determine if commercially available broccoli cultivars

would perform diferently (by trait performance, cultivar ranking and trait

repeatability) in organic compared to conventional environments in order

to identify appropriate cultivars for organic growers and the best selection

environment for breeding for organic agriculture. Organic trial locations were

intentionally selected on farms under long term organic management as

less mature organic farms or those in conversion may more closely resemble

conventionally managed farms. Our organic trials produced comparable

head weight to the conventional trial locations, and therefore the level of the

environmental stress that we hypothesized would afect trait performance and

phytochemical content was minimal. For most traits, there was no management

efect across environments. Management main efect was only identiied

at the per trial level, demonstrating that each individual location/season/

year combination constituted a unique environment, and that genotype

by management system interactions resulted from diferent factors in each

environment. In the partitioning of variance components across all trials

location, season and its interactions were often the largest source of variation,

followed by genotype main efect. While we did not see the trait performance

diferentiation between production systems, we did observe some individual

varietal rank changes when performance of cultivars were compared between

organic and conventional management, including changes in stability of

performance (by head weight and by phytochemical concentrations) across

trial environments. Larger genotypic variances in organic environments for

horticultural but not phytochemical traits were observed, demonstrating the

innate heterogeneity in the organic agricultural system and varietal response

to such variation. Our results produced comparable or higher repeatabilities

under organic and the ratio of correlated response to direct response for all

traits was close to 1. The combined analysis of the repeatabilities and ratio of

correlated response to direct response would suggest that selection in organic

environments is equal or superior to selection in conventional environments.

As with horticultural traits, management main efect did not play a signiicant

role across trials in the phytochemical component of the trials (Chapter 5). At

the individual compound level, genotype main efect was most important for

glucoraphanin, neoglucbrassicin and the carotenoids, while glucobrassicin

201