Most of the required increases in food production over the next decades are expected to be achieved through increases in crop yield. As wheat is essential for food security it is worrying that its yield gains over the last two decades were small. To achieve further yield increases it is critical to continue increasing number of grains per unit area (GN m^-2), the trait best related to yield. In this context, it is relevant to identify the main determinants of GN m^-2 in response to genetic and environmental factors as well as the trade-offs between them. In the present study we compiled a large database across the literature to analyse the relative importance of components when affected by genetic or environmental factors, producing small or large changes in GN m^-2 and its components, either numerical (the number of spikes per m², SN m^-2; and the number of grains per spike, GN spike^-1) or physiological (spike dry weight at anthesis, SDWa; and fruiting efficiency, FE) determinants. The database included 367 papers published in: (i) Field Crop Research (FCR), (ii) European Journal of Agronomy (EJA), (iii) Crop Science (CS) and (iv) Crop and Pasture Science (CPS, formerly Australian Journal of Agricultural Research) between 1990 and 2020. The complete dataset was split into classes, depending on the source of experimental variation, environment or genotype and was normalised to remove the differences between experiments and determine the environmental and genotypic effects within each experiment. Normalised data showed that the responsiveness of GN m^-2 was similarly explained by changes in both SN m^-2 and GN spike^-1, but in terms of physiological components SDWa was more relevant than FE for explaining the variations in GN m^-2. Considering the numerical components of GN m^-2 genotypic and environmental factors modified more GN spike^-1 than in SN m^-2. On the other hand, physiological components were differently modified by genotype and environment: for genotypic effects FE was more critical than SDWa and the other way around for environmental factors. A trade-off between numerical and physiological components was observed although was greater between physiological than between numerical components.

预计未来几十年粮食生产所需的大部分增长将通过提高作物产量来实现。由于小麦对粮食安全至关重要，令人担忧的是，过去 20 年小麦的产量增长幅度很小。为了进一步提高产量，关键是继续增加每单位面积的晶粒数 (GN m ^-2 )，这是与产量最相关的性状。在这种情况下，确定 GN m ^-2的主要决定因素以响应遗传和环境因素以及它们之间的权衡是相关的。在本研究中，我们在文献中编制了一个大型数据库，以分析组件在受到遗传或环境因素影响时的相对重要性，从而产生 GN m ^{-2的小或大变化}及其组成部分，无论是数值（每 m ²的穗数，SN m ^-2；以及每穗的晶粒数，GN 穗^-1) 或生理（花期峰值干重，SDWa；和结果效率，FE）决定因素。该数据库包括 367 篇发表在以下领域的论文：(i) 田间作物研究 (FCR)、(ii) 欧洲农学杂志 (EJA)、(iii) 作物科学 (CS) 和 (iv) 作物和牧场科学 (CPS，原澳大利亚Journal of Agriculture Research) 1990 年至 2020 年间。完整的数据集根据实验变异的来源、环境或基因型分为几类，并进行了标准化以消除实验之间的差异并确定每个实验中的环境和基因型影响。归一化数据显示，GN m ^-2的反应性同样可以通过 SN m ^-2和 GN 尖峰^{-1的变化来解释}，但就生理成分而言，SDWa 比 FE 更适用于解释 GN m ^-2的变化。^{考虑到 GN m -2}基因型和环境因素的数值成分，比 SN m ^-2修改了更多的 GN peak ^-1。另一方面，生理成分因基因型和环境而不同：对于基因型影响，FE 比 SDWa 更关键，而对环境因素则相反。观察到数值成分和生理成分之间的权衡，尽管生理成分之间的权衡大于数值成分之间的权衡。