Current interest in the relation of diet to the maintenance of good health has stimulated the demand for food composition data. Values for energy, nutrients, and other food-based components are required to calculate intake, to provide an adequate food supply, to formulate and label new products, and to facilitate trade. The US Department of Agriculture National Nutrient Databank serves as the foundation for food composition databases used by the food industry, government agencies, and academia for these activities. To be accurate, a specific estimate in a database must be statistically representative of the universe of all values for a component in a food. Statistically based sampling strategies must be developed to identify and select appropriate units of specific foods for analysis. The determination of energy values for foods is a unique challenge for food sampling because energy values are calculated from values for total fat, protein, and carbohydrate fractions. To establish priorities for sampling, food consumption data can be used to rank the primary contributors of energy for the population of interest. Demographic and marketing data for foods can be used to identify specific products and locations for sampling. Users of food composition data should question the quality and representativeness of estimates for energy, including estimates for total fat, protein, and carbohydrate to be used for their applications.