Résumé
The "dilution effect" hypothesis suggests that high host diversity can reduce the prevalence of certain zoonotic pathogens. Despite growing interest, its generality and practical relevance remain debated. In particular most previous studies have focused on large scale analyses or meta-analyses, limiting their potential use for zoonotic outbreak prevention. Here, we present an extensive field-based study of small mammals and their multiple pathogens to assess diversity-disease relationships at a local scale, revealing pathogen-specific patterns. While no significant association was detected between the prevalence of some pathogens and host diversity, others such as Orthopoxvirus and a strain of Mycoplasma haemomuris exhibited dilution effect patterns, whereas another Mycoplasma haemomuris strain and Mycoplasma coccoides displayed amplification effects. Crucially, we demonstrate that these contrasting outcomes are consistently explained by the host competence-abundance relationship. Our findings underscore the importance of host community composition and highlight the need to consider multiple pathogens when evaluating the ecological mechanisms underpinning the diversity-disease relationships.