Yield and Agronomic Traits of Grain Legumes in Organic and Conventional Cultivar-Testing Networks in Poland

Grain legumes are important in organic cropping systems because biological nitrogen fixation reduces dependence on external N inputs. We analysed 2022–2025 data from two separate Polish COBORU cultivar-testing networks conducted at 52 distinct experimental stations/trial sites across the main agricultural regions of Poland: an organic network (approximately five soybean and 7–8 field pea and narrow-leaved lupin locations annually) and a conventional post-registration network (27–33 locations per species). Trial soils spanned agricultural values of 52–94 points, pH 4.6–7.5, available P 40–190 mg kg−1 and K 60–240 mg kg−1; soybean and pea were tested mainly on wheat/wheat–rye complexes, whereas lupin was concentrated on lighter rye complexes. Because the networks were spatially independent, all comparisons are network-level and descriptive rather than causal. Mean soybean yield was 3.77 t ha−1 in the organic network and 3.50 t ha−1 in the conventional network (+7.7%), but the annual difference changed from +13.2% in 2022 and +22.8% in 2023 to approximately 0% in 2024 and −8.0% in 2025. Organic-network mean yields of field pea and narrow-leaved lupin were 21.6% and 11.6% lower than the corresponding conventional-network means, respectively. For soybean, very early/early, medium-early/medium-late and late/very late groups yielded 3.28 vs. 3.11, 4.04 vs. 3.64 and 3.93 vs. 3.83 t ha−1, respectively; the medium-early/medium-late group showed the largest positive network difference (+11.0%) and therefore merits further common-site validation. Field pea showed less lodging in the organic network (6.63 vs. 5.70; 9 = no/minimal lodging). Precipitation–yield relationships were non-monotonic, indicating that annual totals alone did not explain yield variation. Conventional fertilizer ranges were 16–36 kg N, 34–45 kg P2O5 and 59–80 kg K2O ha−1; permitted mineral P and K inputs in organic trials averaged only 1–3 kg P2O5 and 3–9 kg K2O ha−1. Synthetic herbicides were prohibited in organic trials, whereas harmonized herbicide-use intensity in the conventional network could not be reconstructed from the annual summaries. Indicative 2025 gross-margin scenarios were EUR 1424 vs. 407 ha−1 for soybean, EUR 264 vs. −59 ha−1 for pea and EUR −20 vs. −449 ha−1 for lupin (organic vs. conventional), with strong sensitivity to assumed organic price premiums.