Extreme treatments and data overinterpretation could lead to the unjustified conclusion that crop yield is source-limited during the effective grain filling

Understanding the physiology of crop yield is important to inform both agronomy and breeding. In grain crops, there is consensus in the interpretation of data, further supported by theory, to conclude that grain number is source-limited; this accounts for the strong correlation of yield and grain number, and the high phenotypic plasticity of grain number due to source limitation. However, whether grain weight during the effective grain filling period is source- or sink-limited remains debatable. This lack of consensus is commonly interpreted as variation associated with the interaction between genotype and environment. In this opinion paper, we argue that part of the inconsistency in the literature may stem from overinterpretation of experimental results, extreme treatments (e.g., 50–90 % shading), and the assumptions of linearity to conclude that grain weight is source-limited during the effective grain filling. A central flaw is the unjustified extrapolation of conclusions from manipulated plants to the unmanipulated real crop. We review the outcomes of both direct and indirect manipulations of source–sink ratios during the effective grain filling across grain crops with a focus on methods and interpretation of results. Indirect approaches that increase or reduce grain number to measure grain weight compensation (e.g., shading or thinning the plots during the critical period of grain number determination) are ill-suited because they influence potential grain size and grain size hierarchies, confounding interpretation of the grain weight–grain number relationship. Direct manipulations of source–sink ratio that do not alter grain weight (e.g., shading or de-graining plants during the effective grain filling), provide strong evidence that grain growth in the intact control is sink-limited. Conversely, when grain weight changes significantly in response to severe manipulation, the only valid conclusion is that the manipulated plants were source-limited; it is not justified to reach conclusions on the intact control crop. These considerations call for a more cautious interpretation of experimental data where direct manipulation of the source-sink ratio leads to a significant change in grain weight, and suggest a re-evaluation of experimental and analytical methods are needed to conclude on the nature of grain weight limitation.