Fuente: PubMed "medicinal and aromatic plants" PeerJ. 2026 Feb 4;14:e20755. doi: 10.7717/peerj.20755. eCollection 2026.ABSTRACTOBJECTIVES: The objective of the study was to quantify seed flow consistency and in-row spacing accuracy when using a conveyor belt-style metering device under varying operational parameters, to develop mathematical models, and to optimize the uniformity of seed placement for black cumin seeding.METHODS: Seed flow properties and the uniformity of in-row seed distribution uniformity were evaluated through weighing tests and sticky belt methods, respectively. The uniformity of the flow was assessed using coefficient of variation (CV) values, while the in-row seed distribution uniformity was evaluated by using the variation factor (Vf) and the goodness criterion (λ). The experiments were conducted based on Central Composite Design (CCD) under laboratory conditions. The forward speed, seed rate, and seed falling angle were chosen as independent variables. The experiments were conducted at five levels of forward speed (1.01, 1.4, 2.0, 2.6, and 3.01 m s-1), five levels of seed rate (6.6, 10, 15, 20 and 23.4 kg ha-1), and five levels of seed falling angle (1.36, 15, 35, 55 and 68.64°).RESULTS: The polynomial functions were developed and the Vf and λ models were optimized. Optimization reduced the variation factor (Vf) to 0.53 and improved the goodness criterion (λ) to 91.67%, indicating a substantial enhancement in seed placement uniformity. For the Vf model, the optimum forward speed was found to be 1.05 m s-1, with a seed rate of 12.35 kg ha-1 and a seed falling angle of 35°; whereas for the λ model, the corresponding values were calculated 1.55 m s-1, 21.1 kg ha-1, and 28.5°, respectively. The developed models showed high predictive accuracy, with an R 2 value of 95.96% for the λ model. Based on the findings of this work, the seed rate was determined as the most important parameter for all models considered.CONCLUSIONS: The results of the experiments also revealed that the conveyor belt metering unit could be used for the seeding process of black cumin seeds with great success, without encountering significant problems. The comparison of R 2 value of the λ and Vf models indicated that the λ model had a better variable prediction. Therefore, the optimum values from the λ model may be more useful than those from the Vf model.PMID:41660104 | PMC:PMC12882728 | DOI:10.7717/peerj.20755