Extended time-series analysis of infectious diseases raises two issues: the spread of disease, and its persistence in space and time. Most studies are based on both data and models, corresponding to conditions encountered in developed countries. The present work sought to determine the impact of local heterogeneity on these two issues, regarding pertussis in tropical conditions. First, we tested the 'cities and villages' model in a small community of 30 villages in rural Senegal. Second, we focused on the impact of population size and density, as well as geographic distance, on population dynamics of pertussis. Results showed that pertussis initially arrived in urban centres, and then spread to surrounding areas. Both population size and density are implicated in the persistence of pertussis within the study area, whereas geographical distance between villages is not. This is the first study on pertussis in a developing country carried out on a very fine spatial scale. Furthermore, it confirms previous results for measles in England and Wales.