The optimization of sowing density requires simultaneous consideration of grain yield and technological quality, because genotype responses may differ when several correlated traits are evaluated together. The present study used principal component analysis (PCA) to characterize the multivariate response of three winter wheat (Triticum aestivum L.) cultivars—Glosa, Anapurna and Biharia—grown on chernozem soil at three sowing densities: 780, 650 and 530 viable grains m−2. The multivariate model included mean grain yield, crude protein content, wet gluten content, Zeleny sedimentation value and thousand-kernel weight (TKW). PC1 had an eigenvalue of 3.8835 and explained 77.67% of total variance, whereas PC2 explained 14.07%; together, the first two components accounted for 91.74% of the information contained in the five original variables. PC1 was strongly and positively associated with protein (r = 0.9420), TKW (r = 0.9376), wet gluten (r = 0.9143) and yield (r = 0.8546), while Zeleny sedimentation value showed a negative loading (r = −0.7422). The Biharia treatments occupied the positive side of PC1 at all three sowing densities, with the highest PC1 score recorded for Biharia × 650 viable grains m−2 (1.4381), followed by Biharia × 780 (1.2361) and Biharia × 530 (0.9862). In contrast, the Anapurna combinations had negative PC1 scores, while Glosa occupied mainly intermediate positions. PC2 was determined especially by Zeleny sedimentation value, which contributed 53.19% to this component. The results demonstrate that PCA efficiently separates cultivar × density combinations according to an integrated yield–quality profile and reveals cultivar-specific responses that are not evident when traits are interpreted separately. The multivariate approach can therefore complement univariate analyses when selecting sowing density for wheat production systems on chernozem soils.

winter wheat; principal component analysis; sowing density; grain yield; protein; wet gluten; Zeleny; thousand-kernel weight

agronomy

Presentation: poster