Answers are grounded in this paper and cite their sources.
Millet ( Pennisetum glaucum (L.) R. Br.) is a strategic cereal for food security in Sudano-Sahelian areas due to its tolerance to harsh climatic conditions. This study aimed to identify elite local millet accessions through multi-environment evaluation of growth and yield traits in Côte d’Ivoire. Forty accessions collected across various regions of the country were evaluated using an alpha-lattice design with three replicates, at two contrasting sites: the CNRA station in Ferkéssédougou (2023 rainy season) and the UPGC experimental site in Korhogo (2024 rainy season). Eighteen agromorphological traits were measured on 600 plants. Results revealed high phenotypic variability among accessions, with coefficients of variation ranging from 11% to 86%. Combined analysis of variance showed highly significant genotype × environment interactions for 17 of the 18 traits, including grain yield (F = 2.82, p = 8.79 × 10⁻⁸), accounting for 14.3% of total variance in grain yield. Grain yield ranged from 189.8 to 2122.6 kg/ha (average: 1133.92 ± 367.50 kg/ha). Principal component analysis showed that the first three axes explained 57.85% of total variability, with panicle- and grain weight-related traits contributing most to the observed structuring. K-means classification grouped accessions into three distinct classes, mainly discriminated by grain yield and number of productive tillers, with the best-performing class averaging 1317.27 kg/ha. A yield-components decomposition based on multiple linear regression, with predictor contributions expressed as standardized regression coefficients (Fig. 8), explained 99% of yield variability and confirmed the significant positive effects of productive tiller number (NombTP, β = +207.46, p < 0.0001), total panicle weight (PoidsTEpis, β = +0.35, p = 0.040), and 1000-grain weight (Pds1000, β = +97.13, p < 0.0001); because these three components are mathematically linked to grain yield, the analysis is interpreted as a decomposition of yield structure rather than an independent predictive model. Based on yield performance and cross-environment stability (Wricke’s ecovalence and Finlay-Wilkinson regression), the accessions BOUA 3, GUEF, BOLO, KATL1, KORO 2, and TENG 2 emerged as broadly adapted elites, promising for deployment across northern Côte d’Ivoire. In addition, DENA and SAN 1, which ranked among the best accessions at CNRA Ferkéssédougou but showed high ecovalence values and marked between-site yield declines, were identified as specifically adapted to that environment and are proposed as superior genotypes for targeted regional recommendation in the Ferkéssédougou area.
No takes yet. Share an insight, caveat, or question.
Kouyaté et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: