The first recorded outbreak of yellow fever in Kenya occurred from mid-1992 through March 1993 in the south Kerio Valley, Rift Valley Province. We conducted entomologic studies in February-March 1993 to identify the likely vectors and determine the potential for transmission in the surrounding rural and urban areas. Mosquitoes were collected by landing capture and processed for virus isolation. Container surveys were conducted around human habitation. Transmission was mainly in woodland of varying density, at altitudes of 1,300-1,800 m. The abundance of Aedes africanus in this biotope, and two isolations of virus from pools of this species, suggest that it was the principal vector in the main period of the outbreak. A third isolate was made from a pool of Ae. keniensis, a little-known species that was collected in the same biotope. Other known yellow fever vectors that were collected in the arid parts of the valley may have been involved at an earlier stage of the epidemic. Vervet monkeys and baboons were present in the outbreak area. Peridomestic mosquito species were absent but abundant at urban sites outside the outbreak area. The entomologic and epidemiologic evidence indicate that this was a sylvatic outbreak in which human cases were directly linked to the epizootic and were independent of other human cases. The region of the Kerio Valley is probably subject to recurrent wandering epizootics of yellow fever, although previous episodes of scattered human infection have gone unrecorded. The risk that the disease could emerge as an urban problem in Kenya should not be ignored.