A screening protocol for developing high-yielding upland rice varieties with superior weed-suppressive ability

In West Africa, upland rice (Oryza spp.) is mainly grown in low-input systems by resource-poor farmers. Weeds are one of the major constraints to rice production. Ideal rice varieties are high-yielding with strong weed-suppressive ability (WSA). WSA is the ability to suppress weed growth and reduce weed seed production and is determined by assessing weed biomass under weedy conditions. Development of such varieties requires simple screening protocols based on highly heritable traits, which can be evaluated in small plots under weed-free conditions. Then, in later stages of the breeding program, selected breeding lines could be evaluated with substitute weeds instead of natural weeds, as natural weed growth is heterogeneous. To develop such a protocol, I evaluated agronomic traits of 10 diverse rice varieties grown in plots with different numbers of rows (unbordered 1-row and 2-row, and self-bordered 4-row, in which the middle 2-rows were sampled) without weed competition and evaluated weed biomass of these varieties under competition with rice and cowpea as substitute weeds. I then examined whether agronomic traits in unbordered plots can predict yield and weed biomass in self-bordered plots. Broad-sense heritabilities of agronomic traits measured in weed-free conditions and weed biomass under weedy conditions were estimated. Among agronomic traits in unbordered plots, yield was positively correlated with yield in self-bordered plots (r = 0.88 and 0.96 in 1-row and 2-row plots, respectively, P < 0.01). Weed biomass estimated in competition with two rice varieties was more heritable than weed biomass in competition with cowpea. In 1-row plots, growth vigor at 63 days after sowing (DAS), total aboveground biomass at harvest, and yield were related to weed biomass in competition with cowpea and rice variety Aus 257. Growth vigor at 42 and 63 DAS was a reliable estimator for total aboveground biomass and number of panicles at harvest, as well as integrated information on height and number of stems (height multiplied by number of stems). I suggest that growth vigor at 42–63 DAS in unbordered, 1-row plots appears to be a useful selection criterion for developing high-yielding varieties with superior WSA, and WSA of selected varieties can be validated using rice varieties as substitute weeds. Future research is needed to validate this protocol with breeding populations, a wide range of genetic materials, and other important weed species in West Africa before it can be implemented in breeding programs