The Canadian dairy industry faces a changing market environment as processors react to apparent shifts in consumers' preferences, consumers react to an altered mix of products on retail dairy shelves, and industry adjusts to potential pressures of competition and the challenge of new market opportunities under the impetus of changes arising from international trade. The purpose of this study is to derive a set of updated and disaggregated estimates of demand for major dairy products in a manner consistent with the economic theory of consumer behaviour. These estimates are necessary for policy models, policy analysis and forecasting. Previously dairy demand estimates were only available for broad product groupings such as fluid milk, butter, all cheese and "all other dairy products". For this study, four weakly separable groupings of major dairy products and related foods are specified. These are milk and other beverages, fats and oils, dairy dessert and related products and cheeses and apparent substitutes. Skim milk powder is assessed not to be a member of any of these groups but is hypothesized to be a member of a fifth dairy subgroup of dairy protein products. Due to data limitations, it was necessary to follow a single-equation approach for this product. The appropriateness of each product grouping was assessed by a two-stage test. First, each subgroup was tested using non-parametric tests of the axioms of revealed preference, as a means of inferring whether or not choices within each subgrouping are consistent with constrained utility maximization. Second, parametric assessment of each subgroup gave further evidence regarding the appropriateness of the groupings in terms of whether the estimated demand parameters are relatively stable and plausible. Based on satisfactory performance in these tests, parametric analyses for each subgroup were conducted using the linearized version of the almost ideal demand system, incorporating appropriate seasonality and habit formation variables. Estimates of own-price, cross-price and expenditure elasticities of demand are derived and presented. In general these seem plausible. Signs on the own-price elasticity estimates are as expected; the magnitudes appear to be reasonable. As expected, the majority of the specified foods are price-inelastic. However, butter, cooking/salad oil and other cheese appear to be price-elastic. Yogurt, concentrated milk and ice cream are fairly expenditure elastic while the two cheese types and butter appear slightly expenditure elastic.