Schizophrenia is a common, debilitating mental illness that has persisted over the generations. For a disease with a strong genetic component, such prevalence has been difficult to understand in evolutionary terms. A model for its prevalence as a phenotype is presented in this manuscript, based on reports of specific differences in gene expression, metabolite levels and historical epidemiology. The selective force that underlies the proposed model is tuberculosis, a scourge of huge proportions that itself evolved to interact with the human host in a manner ensuring both its long term persistence in the host and its transfer to other carriers prior to the hosts unfortunate death. The focal point of the interaction between humans and M. tuberculosis is hypothesized to be the de novo synthesis of NAD via activation of the kynurenine pathway. The strategy that M. tuberculosis employed to circumvent this aspect of the hosts response to mycobacterial infection, and how that strategy interacted with a poor diet to force human evolution towards increased risk for schizophrenia, forms the basic premise of this paper. The model has implications for treatment of both diseases and generates hypotheses to be tested.