This study was carried out to determine rodent species composition and abundance, the interaction between them, and the possible implication in plague dissemination to humans. Over 2000 rodents were captured, identified and the relative species abundance determined. These animals belonged to six species, namely Mastomys natalensis, Arvicanthis nairobe, Lophuromys flavopunctatus, Grammomys dolichurus, Mus sp. and Praomys sp. They were distributed in two principal habitats, namely fallow land and forest. The distribution of the species overlapped, indicating interaction between them, but their abundance varied considerably between the habitats. Three species of fleas were collected from rodents. Of these, Dinopsylus lypusus was most abundant, followed by Leptopsylla aethiopica and Nosopsyllus fasciatus. Rodent population densities declined rapidly in August and September and were followed by outbreaks of human plague in October. The observations made in the current study suggest that declining rodent population abundance leads to more ‘free’ fleas which probably seek alternative hosts, including humans. This consequently facilitates an increase in the transfer of plague from rodents to humans. The study further indicated that M. natalensis and A. nairobe form a continuum between forest-inhabiting rodent species and peri-domestic premises which therefore creates an avenue for transferring the disease from a potential forest reservoir to the human population. The presence of specific anti-plague immunoglobulin (IgG and IgM) antibodies in blood sera of rodents was tested by enzyme-linked immunosorbent assay (ELISA). The presence of Yersinia pestis DNA was tested by polymerase chain reaction (PCR). Both tests revealed that M. natalensis, A. nairobe, Rattus rattus (captured in houses) and L. flavopunctatus were the potential rodent reservoirs of plague in the western Usambara Mountains. Grammomys dolichurus and Praomys sp. tested negative for plague, but more specimens will be tested to confirm this finding.