This study analyzes the demand for food commodities in the United States in the postwar period using both time-series and cross-section data. Income-consumption relationships are based on data from 1955 and 1965 USDA household food consumption surveys. The analysis co cross section data emphasized: (1) effects of grouping observations, (2) choice between expenditures are quantities as the dependent variable, (3) effects of household size on income-consumption relationships, (4) shifts in the regression coefficients (intercepts and income elasticities) between 1955 and 1965, and (5) regional variations in the income-consumption relationships. A demand interrelationship matrix was developed for 49 commodities or commodity groups at the retail level. Commodities were classified into 15 separable groups and all direct and cross elasticities for commodities within a group were estimated directly. The cross elasticities corresponding to commodities outside a given group were estimated through assumptions of cardinal separability. The synthesis of demand interrelationships was achieved by the use of restrictions on demand equations for an individual consumer as suggested by Frisch (1959) and quantified by Brandow (1961). Consideration also was given to the measurement of time trends on consumption. Marketing margins were analyzed and demand interrelationships were developed at the farm level. Projections of 1980 consumption per capita were developed for individual commodities and group aggregates. These projections are based on a specification of constant real prices, exogenous projections of real income per capita, and continuation of past time trends for certain commodities.