Rice yields under Rhamphicarpa fistulosa-infested field conditions, and variety selection criteria for resistance and tolerance

•Under R. fistulosa infested conditions, rice varieties yielded 1.2–2.4tha−1.•Variety-specific Rhamphicarpa-inflicted yield losses range from 24% to 73%.•Rhamphicarpa biomass is a suitable variety selection measure for resistance.•Tolerant varieties combine high parasite biomass with low relative yield loss.•High yielding, R. fistulosa resistant NERICA-L-40, -31 can be promoted to farmers. The facultative parasitic weed Rhamphicarpa fistulosa, is a widespread problem in rain-fed rice production systems in Africa. Little is known about rice varietal differences in infection level and yields in fields infested by this root hemi-parasite. During three cropping seasons (2012–2014), an experiment was conducted to address these knowledge gaps and to identify suitable variety selection criteria for R. fistulosa resistance and tolerance. Sixty-four adapted lowland rice varieties, including all interspecific lowland varieties of NERICAs, their most common parents – IR64 and TOG5681 – and two locally popular varieties – Mwangulu and Supa India – were grown in an R. fistulosa-infested field in southern Tanzania. As expected from a facultative parasite, host plant varieties had no effect on R. fistulosa numbers. Consistent varietal differences in R. fistulosa biomass were however observed, with no Year×Variety effects, and this parameter can therefore be used to select for resistance. Parasite-free observation plots were established in the last year. Due to the facultative nature of the parasite, creating such R. fistulosa-free plots was simply obtained by regular early weeding. The presence of parasite-free control plots enabled assessment of worthwhile additional information such as parasite-free yield, parasite-inflicted yield losses (RYLR) and varietal differences in tolerance. Under R. fistulosa-infested conditions (3-season averages, no Year×Variety interaction effect), rice grain yields ranged from 1.2tha−1 for the worst performing variety (TOG5681) to 2.4tha−1 for the best performing varieties (NERICA-L-39, -20). Under R. fistulosa-free conditions (2014 only) rice grain yields ranged from 2.4 (NERICA-L-22) to 5.4tha−1 (NERICA-L-17). Tolerant varieties were characterized by a low RYLR and a high parasite biomass (e.g. Supa India, NERICA-L-20). The selection measures identified are effective, easy and practical under field conditions. They facilitated identification of thirteen varieties with high resistance, sixteen varieties with low RYLR and two varieties wit