Context: Conserving large carnivores can be challenging because of conflicts with human land use and competition with humans for resources. Predation on domestic stock can have negative economic impacts particularly for owners of small herds, and tools for minimising carnivore depredation of livestock are needed. Canids use scent marking to establish territories and avoid intraspecific conflict. Exploiting scent-marking behaviour may provide a means for manipulating canid movements. Aims: We hypothesised that human-deployed scent marks (i.e. ‘biofence’) could be used to manipulate the movements of grey wolves (Canis lupus) in Idaho, USA. Methods: We deployed 65 km of biofence within three wolf-pack territories during summer 2010 and 2011 and used location data from satellite-collared wolves and sign surveys to assess the effectiveness of biofencing. Key results: Location data provided by satellite-collared wolves and sign surveys in 2010 showed little to no trespass of the biofence, even though the excluded areas were used by the packs in previous summers. We also opportunistically deployed a biofence in between a rendezvous site of a resident pack and a nearby sheep grazing allotment; the pack was not implicated in any depredations in summer 2010, even though they had killed sheep every year since 2006. Location data provided by satellite-collared wolves in summer 2011 showed that wolves did trespass biofences. Conclusions: Biofencing effectively manipulated the movements of wolves in the first year of our study, but not the second. Implications: Our work suggests that biofencing may be most limited by the apparent necessity to maintain a continuous presence once the biofence is established. The inherent labour and costs associated with such efforts may limit the usefulness of biofencing. Our work can be improved on through further testing that maintains biofencing over a longer timeframe (>3 months), samples several animals per treatment pack, and uses a treatment and control design.