The ratio of plant carbon gain to water use, known as water use efficiency (WUE), has long been recognized as a key constraint on crop production and an important target for crop improvement. WUE is a physiologically and genetically complex trait that can be defined at a range of scales. Many component traits directly influence WUE, including photosynthesis, stomatal and mesophyll conductances, and canopy structure. Interactions of carbon and water relations with diverse aspects of the environment and crop development also modulate WUE. As a consequence, enhancing WUE by breeding or biotechnology has proven challenging but not impossible. This review aims to synthesize new knowledge of WUE arising from advances in phenotyping, modeling, physiology, genetics, and molecular biology in the context of classical theoretical principles. In addition, we discuss how rising atmospheric CO₂ concentration has created and will continue to create opportunities for enhancing WUE by modifying the trade-off between photosynthesis and transpiration.

A core set of genes involved in starch synthesis has been defined by genetic studies, but the complexity of starch biosynthesis has frustrated attempts to elucidate the precise functional roles of the enzymes encoded. The chain-length distribution (CLD) of amylopectin in cereal endosperm is modeled here on the basis that the CLD is produced by concerted actions of three enzyme types: starch synthases, branching and debranching enzymes, including their respective isoforms. The model, together with fitting to experiment, provides four key insights. (1) To generate crystalline starch, defined restrictions on particular ratios of enzymatic activities apply. (2) An independent confirmation of the conclusion, previously reached solely from genetic studies, of the absolute requirement for debranching enzyme in crystalline amylopectin synthesis. (3) The model provides a mechanistic basis for understanding how successive arrays of crystalline lamellae are formed, based on the identification of two independent types of long amylopectin chains, one type remaining in the amorphous lamella, while the other propagates into, and is integral to the formation of, an adjacent crystalline lamella. (4) The model provides a means by which a small number of key parameters defining the core enzymatic activities can be derived from the amylopectin CLD, providing the basis for focusing studies on the enzymatic requirements for generating starches of a particular structure. The modeling approach provides both a new tool to accelerate efforts to understand granular starch biosynthesis and a basis for focusing efforts to manipulate starch structure and functionality using a series of testable predictions based on a robust mechanistic framework.