Bait stations are widely used to control Ceratitis capitata, even though the effect of intrusion from untreated areas decreases the efficacy of the technique. In this work, the edge effect of outer fruit fly populations from untreated areas and backyard orchards has been studied. For this purpose, a field trial was carried out during four years in a wide area of 10 000 ha located in Valencia (Spain). The main growing species were Prunus persica, Prunus domestica, Citrus sinensis, Citrus reticulata and Diospyros kaki. An area of 3,600 ha was treated with chemosterilant bait stations. Fruit fly populations were monitored during the 4 years of study inside and outside the treated area using Tephri-traps baited with trimedlure. A grid of 180 traps was placed in the field with an intertrap distance of 450 m. Multiple linear regression was applied to model population levels as a function of distance to untreated areas and backyards. The distance at which outer medfly populations influenced the treated wide-area population was about 1.3 km. This value reflects the distance at which a given fruit fly population is influenced by outer populations and suggests that C. capitata is able to move more than 1 km seeking for hosts. Thus, buffer areas in area-wide integrated pest management of Mediterranean fruit fly should be at least 1.3 km wide.