Malaria vector control by indoor residual spraying (IRS) in South Africa currently relies on the use of DDT to control pyrethroid-resistant populations of the major malaria vector Anopheles funestus. Pyrethroid resistance in An. funestus in South Africa and Mozambique is primarily based on the over-production of monooxygenase P450 genes. Preliminary evidence suggests that the pyrethroid-resistant phenotype is primarily expressed in adults, raising questions as to how resistance was selected for in the field given that there was no malaria control targeting adults in southern Mozambique when this resistant phenotype was first discovered there. The aim of this study therefore was to select for pyrethroid resistance at the larval or adult stages in southern African An. funestus and then assay the pyrethroid susceptibility of subsequent life stages using samples of progeny from the pyrethroid-resistant selections. There was no significant variation between the lethal permethrin concentrations inducing 50% mortality (LC50s) of larvae from three An. funestus colonies, but highly significant variation in the lethal dosages inducing 50% mortality (LD50S) of adults. It was not possible to enhance adult resistance/tolerance to permethrin by one round of selection at either the larval or adult stages in any of the colonies, although larval tolerance to permethrin can be enhanced by selection at the larval stage. It is concluded that monooxygenase-based pyrethroid resistance in southern African An. funestus is primarily expressed in the adult stage and that this particular resistance haplotype can only be selected by exposing adult mosquitoes to pyrethroids. This has important implications for malaria vector control in southern Africa because it suggests that the pyrethroid resistance in An. funestus populations in South Africa and southern Mozambique was selected for by private pyrethroid use directed against adult mosquitoes.