Phenological responses of Oryza sativa, O-glaberrima and inter-specific rice cultivars on a toposquence in West Africa

Phenological properties of rice cultivars, particularly crop duration, determine their yield potential, local agronomic suitability and ability to escape from drought. Crop duration of a given cultivar depends mainly on photoperiod (PP) and temperature, but is also affected by the crop establishment practice and environmental stresses. A sample of 84 ecologically and genetically diverse rice cultivars was sown on five dates between May and September 1997 on the flooded-lowland (transplant), hydromorphic and upland levels of a toposequence at 7 degrees 52' N in Cote d'Ivoire, in order to characterize the cultivars' phenological responses. In the upland, life-saving sprinkler irrigation was applied when drought symptoms were visible. A non-replicated design augmented with four replicated checks (four replications per ecosystem) was used. Phenology was characterized by date of emergence, first heading, 50% flowering and maturity. The period from emergence to flowering was subdivided into three phases following a simple model used at IRRI to characterize germplasm for photoperiodism. For each ecosystem and cultivar, the basic vegetative period (BVP) was estimated by subtracting 30 d from the duration to flowering at the sowing date associated with the shortest duration, and expressed in degree-days (dd), assuming a base temperature of 10 degrees C. The PP-sensitive phase (PSP) was estimated by subtracting BVP+30 d from the time to flowering. PP-sensitivity (PS) was calculated from the apparent change in PSP between 12.0 and 12.5 h mean astronomic daylength during the PSP, by regression across dates. Cultivars differed strongly in BVP (300 to 1200 dd) and PS (0 to 1000 dd). The BVP was generally longer in the lowland than in the hydromorph, and mostly longer in the upland than in the hydromorph, possibly due to transplanting shock (lowland) and drought (upland). Many cultivars, particularly upland-adapted japonicas, had a greater PS in the lowland than in the upland. Principal-component and cluster analyses based on BVP and PS in each of the three ecosystems established three large and three small groups of cultivars having common phenological responses. The linkage groups were associated with ecotypes (lowland vs upland, traditional vs improved) and genetic groups (O. sativa japonica and indica, O. glaberrima, inter-specific progenies). The groups were seen to represent past selection strategies by farmers and recent breeding strategies, with respect to achieving yield stability in the various ecosystems. For example, indigenous selection strategies for O. sativa upland rices seem to have favored a long BVP, whereas from O. glaberrima, which generally has a superior initial vigor, cultivars with a short BVP have been selected. The authors conclude that the modern upland rice breeding strategy for the region on the basis O. sativa, which aims at drought escape using a short BVP, is paralleled by existing indigenous O. glaberrima materials. Efforts to utilize these materials for breeding are ongoing.