In the grainbelt of south-western Australia, which experiences Mediterranean-type climatic conditions, 3 field experiments with wheat were sown in autumn, 2 at Site A over 2 years and 1 at Site B in the first year only. These experiments related both activity of aphid vectors ( migration into and colonisation of wheat) and the spread of infection with Barley yellow dwarf virus (BYDV) serotype PAV to wheat grain yield and quality. Incidences of BYDV serotype RMV and Cereal yellow dwarf (CYDV) were mostly low and BYDV serotype MAV was not distinguished. Rhopalosiphum padi was the predominant vector species but small numbers of R. maidis and Sitobion miscanthi were also present. Repeated insecticide spray applications began at different times in the different experimental treatments. These sprays killed or repelled aphid vectors, thereby preventing further virus spread from the time they were first applied. At both sites, migrant aphids were caught flying into the wheat throughout the winter period. Peak numbers of colonising aphids ranged from 0 to 99/0.5-m transect of crop. BYDV-PAV incidence ranged from 0.1 to 58% of plants and yields ranged from 1.9 to 8.6 t/ha. First aphid arrival was earlier, and virus spread and resulting yield losses greater at Site A. At this site, in treatments where repeated insecticide sprays did not start until 8 weeks after crop emergence (WAE), virus incidence and subsequent yield losses were significantly greater than when the regular applications started at emergence. However, delaying the start of sprays beyond 8 weeks had no further effect on virus spread. Since aphid numbers were very low up to 8 - 10WAE, yield losses were due entirely to virus infection of plants during this early growth period. Variation in BYDV-PAV incidence explained 81 or 91% of the variation in yield gaps in the 2 years at Site A where, for each 1% increase in virus incidence, there was a yield decrease of 55 or 72 kg/ha. It also explained the variation in seed weight (88%) and protein content (69%), but not in seed screenings. At Site B, virus spread started too late to cause significant yield or quality losses. These results show that wheat yields are decreased substantially in a Mediterranean-type environment, when aphids immigrate early into wheat crops and remain active throughout the winter-growing period, spreading virus infection at young plant growth stages.