AbstractTransmission is central to our understanding and efforts to control the spread of infectious diseases. Because transmission generally requires close contact, host movements and behaviors can shape transmission dynamics: random and complete mixing leads to the classic density‐dependent model, but if hosts primarily interact locally (e.g., aggregate) or within groups, transmission may saturate. Manipulating host behavior may thus change both the rate and functional form of transmission. We used the ranavirus–wood frog (Lithobates sylvaticus) tadpole system to test whether transmission rates reflect contacts, and whether the functional form of transmission can be influenced by the distribution of food in mesocosms (widely dispersed, promoting random movement and mixing vs. a central pile, promoting aggregations). Contact rates increased with density, as expected, but transmission rapidly saturated. Observed rates of transmission were not explained by observed contact rates or the density‐dependent model, but instead transmission in both treatments followed models allowing for heterogeneities in the transmission process. We argue that contacts were not generally limiting, but instead that our results are better explained by heterogeneities in host susceptibility. Moreover, manipulating host behavior to manage the spread of infectious disease may prove difficult to implement.