Previous mixture models have assumed that members of a population of chemoreceptor cells are homogeneous in type, i.e., with either single shared or multiple independent receptor sites. In reality, many chemosensory systems actually consist of a heterogeneous population of receptor cells, consisting of both highly specific cells as well as more broadly and variably tuned cells. A mixed receptor composition model for binary mixtures is described which can be applied to chemosensory systems with heterogeneous receptor cell compositions. The model incorporates information on a) the number of receptor sites/transduction processes per cell, b) the specificity of receptor cells, and c) the contribution of the magnitude of response of each receptor cell to the overall response magnitude of the population of all receptor cells. The predictions of this model can be compared to behavioral responses of animals towards binary mixtures, or at any level of sensory processing which involves the input of the receptor cell population, in order to detect possible mixture interactions.