Mathematical modelling has been used to assess the level of malarial transmission, during the dry season, in Nouna province, in north-western Burkina Faso. The data used were collected, at four sites (one semi-urban and three rural), from 867 children aged 6-60 months who were randomly selected. Almost all of the children (850) completed the follow-up, which involved the active detection of malaria (i.e. febrile, smear-positive malarial infection) throughout a single dry season (December 2003-May 2004). Light traps were used to sample the local populations of Anopheles vectors, in order to estimate the daily biting rate. The mathematical model was then used to simulate the incidence of malaria, which was compared with the observed incidence. At all four study sites, new cases of malaria were observed throughout the dry season, although the level of transmission was low. The monthly incidence of malaria estimated using the mathematical model was very close to the observed incidence. The fit was sensitive to daily mosquito survival and daily human parasite clearance. In Nouna province, effective interventions to prevent malaria should not be confined to the rainy season but must continue throughout the year. The focus should be on the clearance of parasitaemias, by the use of effective drugs, and on decreasing vector survival, by the use of vector-control methods.