'Drought' has many meanings in relation to crop production. These range from: statistical (say, the lowest decile of annual rainfall) to a meteorologist; through yield being limited by too little water to an agronomist; to sudden severe water deficits to many molecular biologists. To a farmer, the corresponding management issues, respectively, are risk management (how best to manage a meteorologically drought-prone farm over several years), how best to match cultivar and agronomic operations to the developing growing season, and how best to minimize possible major damage to (say) floral fertility induced by severe water deficits during flowering. All these definitions and the issues they imply are relevant to improving crop production when water is limiting. How can scientists best help? The answers depend on the scales (temporal and spatial) being addressed. Agronomists and breeders, interacting, can help improve components of seasonal water balance in the field, for example, minimizing evaporative losses from the soil surface by better matching the development of a crop to its environment. Physiologists, biochemists, and molecular biologists can help by identifying ways of improving the competence of particular organs. A promising target is floral infertility resulting from water deficits, which results from lesions in tissue, and cellular and molecular processes. Choosing problems whose solutions will have implications in the field and be attractive to farmers requires knowledge of what is important in the field.