Abstract Crops respond similarly to drought and heat stress: life cycle is accelerated reducing photosynthetic capacity via restricted leaf area and duration. Metabolism is inhibited at temperature and water potential ranges outside those optimal for growth. Reproductive processes are impaired when stress occurs at critical developmental stages reducing seed set. Both stresses can be exacerbated by nutrient deficiencies and biotic factors while elevated CO2 levels may partially ameliorate stress in C3 species. Although stress adaptive traits - and consistent quantitative trait loci associated with them - are used to design new cultivars, the physiological and genetic bases of adaptation are only partially understood. Therefore, plant selection requires empirical approaches such as multi-location testing across representative environments, while detailed characterization of target sites permits genotype × environment interaction to be dissected, providing feedback into breeding and research. Precision phenotyping approaches assist by dissecting yield into its physiological components and have application in breeding and gene discovery. Examples of stress-adaptive traits which have been selected for in several species include deeper roots enabling plants to remain hydrated under drought and permitting canopy cooling under heat stress, transpiration efficiency, delayed senescence in sorghum, and synchronous flowering in maize. New traits and genes must be identified - perhaps among crop wild relatives or in model species - that permit cultivars to be buffered against temporal variation in water supply, adapt to higher temperatures without loss of water-use efficiency, and tolerate sudden extreme climatic events or combinations of stress factors. Examples of past successes and promising new approaches are discussed.