Density‐dependent trade‐off between competitive ability and population yield in wheat is mediated by reproductive allometry

Despite efforts by plant breeders to develop higher‐yielding cereal genotypes, genetic gains in major crops have decelerated. Genes identified in controlled environments as having the potential to increase yield often fail to do so in the field due to undetermined genotype × environment interactions. We hypothesize that this reflects plant‐plant interactions and allometric allocation to reproduction. To test this, we grew two modern wheat ( Triticum aestivum L.) cultivars—the recently released LD211 and the established check cultivar LT19—along a density gradient of 45–788 plants/m 2 in field experiments at two locations. The modern cultivar LD211 exhibited weaker competitive ability (lower relative yield in mixture) but higher monoculture yield at densities above 450 plants/m 2 than LT19. At the more productive site, LD211 reached a maximum estimated yield of 844.5 g/m 2 at 561 plants/m 2 , while LT19 peaked at 820.8 g/m 2 at 483 plants/m 2 . The yield advantage of LD211 was driven by a higher reproductive allometric exponent ( α ), which was significantly higher than that of LT19 ( p < 0.05), particularly at densities ≥225 plants/m 2 . LD211's α increased with planting density ( p < 0.01), whereas LT19's remained stable. A strong negative relationship was found between α and competitive ability ( p < 0.001). Total aboveground biomass did not differ between the genotypes. We conclude that LD211 achieves higher population yield by reducing individual competitiveness and adjusting resource allocation to reproduction as plant size increases, a density‐dependent trade‐off mediated by reproductive allometry. Allometric analysis of allocation in the context of plant‐plant interactions may offer a pathway to breaking yield ceilings in cereal crops.