Toward a metabolic scaling theory of crop systems.

S table and secure food production is essential to civilization. Any methods that improve our understanding of crop yields have the potential to reduce human suffering and help provide the caloric needs of an expanding world population. Similar hopes apply to commodity crops such as cotton and those used to meet energy needs. Given the vast amount of agricultural research and experience, it might be expected that we have a thorough understanding of the major crops upon which our civilization depends. As is evident from all-too-frequent crop failures, we clearly do not have a very effective means to predict yields accurately. The assumption is that year-to-year yield variability in crop yields arises from factors outside our control such as varying weather conditions. Although weather-impacted yield variability may not be feasibly and economically constrained by management practice, the expectation is that increased knowledge of crop systems will enhance the efficacy of management. Deng et al. (1) point out how to obtain relatively simple macrodescriptors of crop growth that are applicable across diverse species, and provide the surprising result that the single variable of maximal size at maturity can be used to project major crop properties such as maximal yield. Many of the key questions driving plant science arise from agriculture (2), and it is therefore appropriate to consider a panoply of approaches to enhance knowledge of crop systems. As with much of the life sciences, the tendency in recent decades has been to delve further and further into details of the genetics and cell physiology of important crops, providing a much greater understanding of biological processes at the molecular and cellular level. This reductionist approach has created opportunities, and concerns in some quarters, for genetically modified crops to significantly reduce the impacts of disease and weather on crop yield.