Résumé
Double cropping of rice is constrained by tight cropping calendars. In mechanized production systems such as in the Senegal River delta of Africa, drainage date is important for yield, maturity and field trafficability. Randomized complete-block field experiments were conducted at the WARDA research station at Ndiaye in Senegal during the hot-dry and wet seasons to determine effects of drainage date on soil drying, grain yield and assimilate partitioning. The soil was a slightly acid, shrinking clay. Pre-germinated seed of I Kong Pao rice was sown broadcast and the plots kept flooded until drainage at four different dates after flowering (treatments). Topsoil and grain moisture content and the dry weight of panicles, stems, green leaves and dead leaves were measured on six dates between 1st flowering and 2 wk after maturity.
Soil moisture content decreased more rapidly under a green canopy than with a mature crop. Grain yield was reduced by 34 to 36% in both seasons when plots were drained 3 to 4 days after flowering (DAF) (
P < 0.01, compared to drainage at maturity) but were not affected when drained 12 to 14 DAF or later. Grain yield reduction was due one-third to poor grain filling and two-thirds to unfilled spikelets. Maturity (20% grain moisture content) was advanced 10 to 11 d by early drainage (3 to 4 DAF), compared to drainage at maturity. Apparent mobilization of reserves based on dry weight decrease of vegetative organs after flowering was 32 to 42% and constituted 46 to 84% of panicle growth, depending on treatment. Early drainage reduced dry matter production during ripening but did not proportionally affect grain yield because of compensatory mobilization of stem reserves. Models are proposed on the buffering effect of reserves if assimilation during grain filling is variable.