Road and railway construction on a landscape scale entails a transformation of ecosystems, involving changes in their composition, structure and functioning, including interactions between predators and their prey. Wildlife passages are currently built at these transport infrastructures to re-establish connectivity within the habitats intersected by them. However, it is unknown whether crossing structures influence predator-prey interactions, potentially resulting in prey exclusion effects or increased predation risks that may reduce the effectiveness of the crossings. Prior to this contribution only one study has thoroughly analysed the question for large vertebrates in the Boreal forests of Canada concluding that evidences point to a lack of such interactions. In case that predators attend the wildlife passages in search of prey or that prey avoid passages visited by predators, it could be hypothesized that the presence of any pair of predator and prey species in one such structure would deviate from randomness. Thus, we evaluated the co-occurrence patterns of predator-prey species-pairs at 113 crossing structures noting their coincidence at the same structure and/or on the same day. Monitoring involved recording tracks within a one-metre-wide strip of marble dust extending from side to side across each crossing structure. The patterns were analysed by means of two-species occupancy models employing presence-absence matrices for three prey groups (small mammals, rat-sized rodents and lagomorphs) and five predator types (small mustelids, Eurasian badger, cats, red fox and large canids) obtained during 2076 passage-days of monitoring. Such models allow the fitting of 'interaction parameters' that in case of significance can be directly interpreted in terms of deviation from randomness in the presence (daily or at a passage level) of any species pair. The value of any interaction parameter informs whether deviation points to two-species attraction or repulsion. The data set comprised 2,329 animal records in total: 52.2% of them being prey species and 47.8 % predators. The results indicate that predators and prey do not use passages independently. Attraction or segregation effects appeared within 20% of predator-prey species-pairs, and were detected in 67% of cases with respect to same-day use. Models in which small mammals were the prey show significant interaction parameters in all cases, revealing trends for both co-occurrence and avoidance. The probability of recording badgers or cats at passages also used by small mammals was significantly greater than chance by 10.5% and 5.0% respectively. These same trends were detected in same-day use patterns at structures, the probability of detecting the badger-small mammals pair increasing by 15.3% above chance and just 1.4% increase in the cat-small mammals pair. At the same time, the adjusted models only show significant interactions between rat-sized mammals and small mustelids and red foxes, the detected trend being avoidance in all cases. Finally, regarding lagomorphs, co-occurrence of badgers with them at particular structures was 59% higher than random and same-day co-occurrence was 125% greater than expected by chance. Similarly, the red fox-lagomorphs pair showed 11.9% higher coincidence in same-day use of structures than expected. Our results show that both predator and prey species used the same structures to cross fenced roads. However, the spatial and temporal patterns of crossing use strongly suggest that there were predators that attended crossings to hunt prey and that some prey species avoided using crossings in the presence of predators.
