Vicia faba germplasm multiplication – floral traits associated with pollen‐mediated gene flow under diverse between‐plot isolation strategies

Abstract

AbstractContamination, by pollen, from neighbouring accessions is a risk during Vicia faba seed stock multiplication. In insect‐pollinated plants, floral traits are expected to play an important role determining pollen‐mediated gene flow variation. The aim of this study was to identify floral traits associated with interplot gene flow that could be used as effective predictors of gene flow potential to improve between‐plot isolation strategies in the field for seed multiplication. The isolation strategies tested were a barren zone, two trap crops, a faba bean male sterile variety and a tetraploid genotype and a non‐pollinated crop, a Vicia narbonensis population. Four genotypes, fixed for alternative isozyme alleles allowed identification of between‐plot hybrids through progeny testing. G‐test was used to analyse the independence of isolation zone and genotype. Principal component analysis was used to obtain information on the usefulness of faba bean traps as barriers to prevent interplot gene flow. Paternity analysis in conjunction with multiple regression allowed the determination of floral traits that accounted for the largest proportion of variation in pollen‐mediated gene flow across each of the four isolation strategies. The floral traits studied were grouped into functional categories and were used as independent variables with gene flow as the dependent variable in a forward stepwise multiple regression procedure. There were large variations in gene flow level and patterns among the genotypes. It was shown that gene flow was isolation zone and genotype dependent. Multivariate regression models showed that differences in floral traits accounted for more than 70% of the variation in interplot gene flow. The relative importance of various floral traits depends on the isolation zone. Flower advertisement appeared important in predicting gene flow in plots surrounded by a barren zone. With regard to plots surrounded by a V. narbonensis population, the role of a reward trait, pollen production, was established. In contrast, in plots surrounded by faba bean trap crops, ovary length played the most important and consistent role in accounting for variation in gene flow. Our results show that interplot gene flow is largely dependent on floral traits, and thus confirm the utility of floral traits for generating hypotheses about cultivar outcrossing or inbreeding behaviour. Thus, it could help to devise more efficient procedures for seed multiplication.