Distribution of biting sites on the human body by the malaria vector Anopheles arabiensis Patton (Diptera: Culicidae) was investigated near a source of mosquitoes in the Kruger National Park, South Africa. Eight adult male volunteers (2 teams x 2 pairs of subjects) conducted human bait collections while seated on camp chairs in the open-air, wearing only short trousers (no shirt, socks or shoes). Mosquito collections during 18.30-22.30 hours on five consecutive nights in April 1998 yielded a total of 679 An. arabiensis females biting subjects with or without their ankles and feet treated with deet insect repellent (15% diethyl-3-methylbenzamide, Tabard lotion). On subjects whose feet and ankles were smeared with repellent, 160 An. arabiensis females were captured biting in 60 manhours: 88.1% on the legs, 1.4% on the arms and 1.2% on other parts of the body, but none on the repellent-treated feet or ankles. On subjects without repellent treatment, 519 An. arabiensis were caught biting in 60 man-hours: 81.1% on feet and ankles, 16.4% on legs, 1.4% on arms and 1.2% on the rest of the body. For individual subjects, the reduction of An. arabiensis bites ranged from 36.4 to 78.2% (mean protection 69.2%). Results of this study confirm previous findings that, in this part of South Africa - inhabited only by wildlife - when people sit outside during the evening An. arabiensis prefers to bite their lower limbs: 97.5% below the knees. Overall, the number of bites by the malaria vector An. arabiensis was reduced more than three-fold (from 26 to 8/person/evening), simply by treating ankles and feet with a consumer brand of deet repellent. Whether or not this provides a satisfactory degree of protection against malaria risk would depend on the malaria sporozoite rate in the malaria vector population.