Water deficit is the major constraint to rice (Oryza sativa L.) production in many rice growing areas. Yield is generally low under water limited conditions as a result of a combination of phenological, morphological and physiological responses to water stress. Identification of physiological and morphological traits for screening criteria would provide significant tools to improve rice yield for drought tolerance.
A field experiment consisting of three trials; high nitrogen - high density, high nitrogen - low density and low nitrogen - high plant density was conducted at Redland Bay, South-East Queensland, Australia to determine the ability of twenty one rice cultivars grown under upland conditions to tolerate or avoid drought in different water stress environments. Genotypic variation was examined for drought score (leaf death), leaf rolling score, leaf water potential, relative water content and epidermal conductance under water stress conditions.
In each trial, water stress was imposed 48 days after sowing and water was withheld for the next 36 days. Differences among cultivars were found in drought score, leaf rolling score, leaf water potential and epidermal conductance. No difference among cultivars was observed in relative water content.
Water stress development was fastest in trial one (high nitrogen) followed by trial two (low nitrogen) and slowest in trial three (low density). Thus the density treatment had a larger effect than the nitrogen application rate on rate of development of stress. Significant cultivar by trial interactions were found for drought score at 31 and 36 days after imposing stress and leaf water potential during the stress period. However, no significant interactions were observed for light interception, leaf rolling and epidermal conductance throughout this experiment.
Genotypic variation in leaf death (drought score) was associated with variation in radiation interception and leaf water potential; cultivars lost more leaves when they intercepted more radiation and had lower leaf water potential. However, when the variation in radiation interception and leaf water potential were removed statistically by covariance analysis, genotypic variation in leaf death remained. It is also apparent that cultivars with high leaf water potential tended to have high epidermal conductance.
The results indicate that leaf rolling score may be a useful selection index in a plant breeding program for the improvement of drought resistance in rice since analysis of these responses showed that genotypic differences were consistent across environments. Drought score can also be used as an index, provided that the difference in plant size (light interception) among cultivars is taken into consideration.