Abstract. Focal control of malaria vectors, a potentially cost-effective alternative to conventional control, requires a spatio-temporal understanding of the mosquitoes. Trapping of African malaria vectors has generally been limited to inside houses making distribution estimates dependent on the location of dwellings. The development of tent-traps to sample outdoor biting mosquitoes has enabled more independent estimates. Here we describe both temporal and spatial variation in mosquito movements in an irrigation project village in southern Mozambique. Six hundred and ninety-three tent-trap collections (525 of which were paired with light-trap collections), 552 exit collections and 391 collections of mosquitoes resting inside houses were undertaken from March 2005 to April 2006. Fifteen species of mosquito were collected (five exclusively as larvae). Mansonia africana was the most common finding, numbers being greatest away from the village. Only Anopheles funestus, An. tenebrosus and Culex quinquefasciatus were collected in greater numbers in light-traps compared to tent-traps. Among the common mosquitoes, correlations in numbers of mosquito collected in paired tent and in light-traps were significant for all but An. tenebrosus. Inverse distance weighting was used to produce raster density maps of the most common mosquitoes. All species, with minor variations, in both hot and cool seasons, were collected in greatest numbers close to the edges of the village where water suitable for larval development was available. All exophilic anophelines species tested negative for sporozoites. It is suggested that focal control of larvae, applied by the villagers themselves, could be a suitable alternative to conventional control in this and similar villages.