Veterinary cordon fences are used in Southern Africa to separate wildlife from domestic animals in order to prevent transmission of infectious diseases. Such fences are a control method recognized by the World Organization for Animal Health (OIE) for establishing disease-free zones in beef exporting countries. However, few studies have evaluated the ecological impact of these physical barriers or their effectiveness at a multispecies level. We examined the permeability of one such barrier, along 357 km of the western and southern boundary of the Kruger National Park (KNP) during 2007. Information was gathered using a semi-structured questionnaire implemented among 32 teams of fence maintenance workers. Data were analyzed to identify (a) the main causes of fence damage, (b) the seasonality, location and duration of fence repairs, (c) high permeability areas for elephant (Loxodonta africana), kudu (Tragelaphus strepsiceros), impala (Aepyceros melampus), buffalo (Syncerus caffer) and warthog (Phacochoerus africanus), and (d) the influence of fence electrification, rivers and elephant damage on the frequency of observation of wildlife species outside the KNP estimated during the year. Human and elephant damage were the most common reasons for fence repairs. Elephant and buffalo were the most and least common large mammal species reported observed outside the KNP (1076 and 162 reports/year), respectively. Cattle incursions into the KNP were also reported in 44% of the fence sections. Electrification of the fence was an important factor explaining differences in estimated wildlife species observations outside the KNP during the year. Correlations between estimations of observed species suggested that fence gaps created by elephants might be used by the other wildlife species. Estimated annual counts of kudu, impala and buffalo, but not warthog, were found to correlate with elephant observations. Negative binomial regression models were developed to explore the relationships between observed estimations of different wildlife species outside the KNP the fence, electrification of fence sections and the presence of watercourses, suggesting that kudu, impala and buffalo could use elephant induced fence damage to leave the KNP. The questionnaire was able to evaluate fence integrity and identify sections where integrity was sub-optimal for separating wildlife from domestic livestock species. If combined with more quantitative methods and applied on a routine basis, it could provide an efficient and cost-effective method for monitoring the effectiveness of physical barriers to contain wildlife within protected areas such as is the case with veterinary cordon fences in Southern Africa. (Résumé d'auteur