[Untitled]

The definition of a suitable breeding strategy in drought-prone environments is an important task for sunflower breeders. To achieve this task, reliable information on heritability and gene effects on yield and related traits under these conditions is necessary. Thirty six sunflower hybrids were produced by factorial cross of six male-sterile and six restorer lines. Parents and their hybrids were evaluated in eight environments. Six environments consisted of two adjacent trials in the experimental area, the first under irrigation and the second under dryland conditions, during 1987, 1988 and 1992. The other environments were: one early planting trial in dryland conditions, conducted during 1987, and a winter trial planted in January during 1988. Estimates of female variance (σf) were significant for seeds per head, seed weight, head sterile center, days to blooming and oil content. Female × male interactions (σ2fm) were significant for all characters except harvest index and index of susceptibility to drought. Estimates of narrow sense heritabilities, calculated with information from analyses combined across environments, were 0.65 for yield, 0.80 for seeds per head, 0.84 for seed weight, 0.81 for head diameter, 0.60 for sterile head center, 0.72 for oil content, 0.61 for harvest index, 0.72 for biomass, 0.94 for days to bloom, and 0.42 for drought susceptibility index. Heritability estimates for individual environments showed more variation for yield than for other traits. Estimates for heritability of canopy temperature were high (0.68–0.79). Rainfed yield was positively correlated with yield components and negatively correlated with canopy temperature and susceptibility index. It is concluded that an efficient breeding strategy for sunflower under moderate drought-stressed conditions is the simultaneous selection for seed yield in both rainfed and irrigated environments together with selection for canopy temperature and stem diameter.