Number of seeds per silique (NSS) is an important determinant of seed yield potential in Brassicaceae crops, and it is controlled by naturally occurring quantitative trait loci. We previously mapped a major quantitative trait locus, qSS.C9, on the C9 chromosome that controls NSS in Brassica napus. To gain a better understanding of how qSS.C9 controls NSS in B. napus, we isolated this locus through a map-based cloning strategy. qSS.C9 encodes a predicted small protein with 119 amino acids, designated as BnaC9.SMG7b, that shows homology with the Ever ShorterTelomere1 tertratricopeptide repeats and Ever Shorter Telomere central domains of Arabidopsis (Arabidopsis thaliana) SUPPRESSOR WITH MORPHOGENETIC EFFECTS ON GENITALIA7 (SMG7). BnaC9.SMG7b plays a role in regulating the formation of functional female gametophyte, thus determining the formation of functional megaspores and then mature ovules. Natural loss or artificial knockdown of BnaC9. SMG7b significantly reduces the number of functional ovules per silique and thus, results in decreased seed number, indicating that qSS.C9 is a positive regulator of NSS in B. napus. Sequence and function analyses show that BnaC9.SMG7b experiences a subfunctionalization process that causes loss of function in nonsense-mediated mRNA decay, such as in Arabidopsis SMG7. Haplotype analysis in 84 accessions showed that the favorable BnaC9.SMG7b alleles are prevalent in modern B. napus germplasms, suggesting that this locus has been a major selection target of B. napus improvement. Our results represent the first step toward unraveling the molecular mechanism that controls the natural variation of NSS in B. napus.Elite Brassica spp. (majorly Brassica napus) with low erucic and glucosinolate content is the third leading source of both vegetable oil and oil extraction meal worldwide (Rondanini et al., 2012). Yield enhancement has been one of the most important goals of B. napus production and genetic improvement. As a complex quantitative trait, the seed yield of a B. napus plant is comprehensively determined by three components: number of seeds per silique (NSS), number of siliques per plant, and seed weight, all of which are typical quantitative traits (Clarke and Simpson, 1978). Although there is rich variation for each of the three components (Zhang et al., 2012a), it is critical to balance the three traits during the breeding program, because significant negative correlations among them have been extensively observed Zhang et al., 2011). Thus, it is highly desirable to identify the genes responsible for each component and subsequently, understand the genetic mechanisms underlying the natural variations. Toward this goal, a large number of quantitative trait loci (QTLs) associated with these yield components in B. napus has been identified in different mapping populations (Quijada et al.,