Control of human infectious disease has been promoted as a valuable ecosystem service arising from the conservation of biodiversity. There are two commonly discussed mechanisms by which biodiversity loss could increase rates of infectious disease in a landscape. First, loss of competitors or predators could facilitate an increase in the abundance of competent reservoir hosts. Second, biodiversity loss could disproportionately affect non-competent, or less competent reservoir hosts, which would otherwise interfere with pathogen transmission to human populations by, for example, wasting the bites of infected vectors. A negative association between biodiversity and disease risk, sometimes called the "dilution effect hypothesis," has been supported for a few disease agents, suggests an exciting win-win outcome for the environment and society, and has become a pervasive topic in the disease ecology literature. Case studies have been assembled to argue that the dilution effect is general across disease agents. Less touted are examples in which elevated biodiversity does not affect or increases infectious disease risk for pathogens of public health concern. In order to assess the likely generality of the dilution effect, we review the association between biodiversity and public health across a broad variety of human disease agents. Overall, we hypothesize that conditions for the dilution effect are unlikely to be met for most important diseases of humans. Biodiversity probably has little net effect on most human infectious diseases but, when it does have an effect, observation and basic logic suggest that biodiversity will be more likely to increase than to decrease infectious disease risk.