RESUMO
Three communities separated by 1.5-7.0 km, along the Matapí River, Amapá State, Brazil, were sampled monthly from April 2003 to November 2005 to determine relationships between seasonal abundance of host-seeking anophelines, rainfall and malaria cases. Out of the 759 821 adult female anophelines collected, Anopheles darlingi Root (Diptera: Culicidae) was the most abundant (56.2%) followed by An. marajoara Galvão & Damasceno (24.6%), An. nuneztovari Gabaldón (12.4%), An. intermedius (Chagas) (4.4%) and An. triannulatus (Neiva and Pinto) (2.3%). Vector abundance, as measured by human landing catches, fluctuated during the course of the study and varied in species-specific ways with seasonal patterns of rainfall. Anopheles darlingi and An. triannulatus were more abundant during the wet-dry transition period in June to August, whereas An. marajoara began to increase in abundance in February in two villages, and during the wet-dry transition in the other village. Anopheles nuneztovari and An. intermedius increased in abundance shortly after the rains began in January to February. A generalized linear mixed model (GLMM) analysis of 32 consecutive months of collections showed significant differences in abundance for each species by village and date (P < 0.0001). Correlations between lagged rainfall and abundances also differed among species. A strong positive correlation of An. darlingi abundance with rainfall lagged by 4 and 5 months (Pearson's r = 0.472-0.676) was consistent among villages and suggests that rainfall may predict vector abundance. Significant correlations were detected between numbers of malaria cases and abundances of suspected vector species. The present study shows how long-term field research may connect entomological and climatological correlates with malaria incidence.