Abstract.In preparation for a Filariasis Control programme in Samoa, during 1978 monthly larval surveys of the vector mosquitoAedes polynesiensiswere carried out in four study villages in the main island of Upolu. A more extensive survey of larval habitat distribution was then made in twenty‐two villages of Upolu and eighteen of Savai'i island, to determine the importance of habitat types according to their abundance, volume of water and whether their productivity was permanent or seasonal.Ae.aegyptilarval densities and habitat distribution were also monitored and the occurrence of predatoryToxorhynchites amboinensislarvae in northern Upolu was recorded from forty‐one collections.AedesBreteau and container indices fluctuated with the pattern of rainfall in two coastal villages and an inland bush village, but not in a coconut plantation community. The five mainAedeslarval habitat types encountered were: 200 litre water‐storage drums, discarded tins and bottles, coconut shells, automobile tyres and treeholes.Aedesimmatures occurred perennially in drums and tree holes, but breeding discontinued in tins, bottles and coconut shells during the driest month of July.ForAe.polynesiensisin Upolu the Breteau and container indices of 104.5 ± SD 80.9 and 35.3 ± 12.4 respectively were significantly higher than those in Savai'i: 33.1 ±25.0 and 24.3 ±20.0 respectively. Likewise forAe.aegyptithe Breteau and container indices of 50.8 ± 32.5 and 23.9 ± 15.6, respectively, were also significantly higher than those in Savai'i: 12.7 ± 17.1 and 9.4 ±13.2 respectively.Habitat types greater or lesser importance were determined by plotting the percentage of each type of container utilized forAedesbreeding against the percentage of each type amongst all larva‐positive containers.Ae.polynesiensispreferred tree‐holes but not water‐storage drums.Ae.aegyptipreferred drums and tyres; mixed populations of larvae of both species were commonest in these two types of habitat.Ae.polynesiensisoccurred in every village.Ae.aegyptiwas encountered in all twenty‐two villages surveyed in Upolu and nine of eighteen villages in Savai'i.Total larval surveys revealed that drums and tree‐holes contained the highest numbers ofAedeslarvae. The study provided criteria for plann