TOWARDS THE RELIABLE PREDICTION OF TIME TO FLOWERING IN 6 ANNUAL CROPS .6. APPLICATIONS IN CROP IMPROVEMENT

Variation in time from sowing to flowering (f) was examined for 44 cul tivars of soyabean, mungbean, black gram, ricebean, cowpea, chickpea, lentil and barley, when grown in up to 21 diverse environments obtaine d by making one or more sowings at each of six locations spanning trop ical, sub-tropical and temperate climates in Australia. The utility of simple linear models, relating rate of development (1/f) towards flow ering to mean photoperiod and temperature prevailing between sowing an d flowering, was evaluated. The models were highly efficient, explaini ng most (86.7%) of the variation observed across species, cultivars an d environments. They were particularly efficient in describing respons es where cultivars were relatively well-adapted, in agronomic terms, a nd least efficient where cultivars were exposed to unfavourable temper ature and, to a lesser extent, photoperiod. Opportunities for exploiti ng the models in applied crop improvement include their use in interpr etation of G x E interaction, genotypic characterization and selection of parental genotypes, selection of test environments, designing scre ening procedures, and more efficiently matching genotypes to target en vironments. The main strengths of these linear, additive rate models i n crop improvement are their wide applicability across species and gen otypes, their relative simplicity, and the requirement for few genotyp e-specific response parameters. Their main weakness is their lack of p recision in describing responses when plants are exposed to unfavourab le photothermal extremes, albeit in circumstances that are sometimes u nrealistic for cropping those particular genotypes.