An arbovirus surveillance was carried out in Burkina Faso from 1983 to 1986. It was based on crepuscular catches of mosquitoes on human bait in some wooded areas and in one town. The total collection was 228 catches with an average of 8 men per catch. The total number of mosquitoes caught was 44,956 among which 32,010 potential vector of yellow fever; all these mosquitoes were analysed for arbovirology. In the south-western part of the country (region of Bobo-Dioulasso), surveillance was conducted each year from August to November, whilst the circulation of Aedes-borne arboviruses is well known to be favoured. In 1983, 1984 and 1986, seven strains of yellow fever virus were isolated in circumstances remarkably similar. They came from selvatic areas and never from the town. They concerned only Aedes (Stegomyia) luteocephalus which is the very predominant potential vector of yellow fever in the region. They were obtained in low figure, between 1 and 4 per year. They occurred from 27th of October to 21th of November. These observations confirm that the southern portion of the Sudan savanna zone of West Africa is the setting of a customary circulation of yellow fever virus and therefore belongs to the endemic emergence zone. In 1986, two strains of dengue 2 virus were isolated. One concerned Ae. luteocephalus from the selvatic area, the other Ae. (St.) aegypti from the heart of town. These data suggest two distinct cycles for dengue 2 virus, one urban and one selvatic, which could coexist simultaneously in the same region. In the south-eastern part of the country (region of Fada-N'Gourma) a yellow fever epidemic occurred between September and December 1983; its study has enable to precise their entomological aspects. The entomological inoculation rate of yellow fever virus has been evaluated to 22 infected bites per man during the month of october, for a man living close to forest gallery. 25 strains of yellow fever virus strains was isolated from Ae. (Diceromyia) furcifer which is the potential vector the most abundant in this region: the main role of this species in an epidemic was confirmed. An investigation in September 1984 had not permitted isolation of the virus therefore it is suspected that the large epizootic circulation of virus in 1983 has not been renewed the year after. In total 59 viral strains belonging to 10 different viruses were isolated from 9 species of mosquitoes.(ABSTRACT TRUNCATED AT 400 WORDS)