Every taxon on Earth is experiencing negative effects from anthropogenic disturbance and to cope with these significant changes some animals are modifying their behaviours. However, the degree to which these behavioural changes mitigate against the novel risk from humans and the true impact that further infrastructural development, in particular, is to have on wildlife within their habitats is still not fully understood. In this thesis, I studied the risk that a road intersecting the habitat of multiple Zanzibar red colobus (Piliocolobus kirkii) groups at Jozani-Chwaka Bay National Park (JCBNP), Zanzibar, poses to individuals and examined whether colobus adopt behaviours that might mitigate the risk of primate-vehicle collisions when crossing the road. In my first chapter, I examined mammalian roadkill recorded over a 4-year period (January 2016 – December 2019) in the vicinity of JCBNP. I found that primates accounted for the majority of dead mammals on the road, and that rainfall did not influence roadkill occurrence. For the Zanzibar red colobus specifically, all age-sex classes were killed in proportion to their availability within the local population (in groups which ranged across or near the road). The rate at which colobus were killed on the road was similar to what other primate populations experience from natural predation. However, because car-induced mortality, unlike natural predation, did not affect predominantly the very young and old individuals, road-related mortality would have a distinct effect on population structure. By comparing mortality rates from the present study to estimates of colobus mortality estimated on the same road in the early 1990s, I also show that colobus roadkill has been significantly reduced by installing several speedbumps near the national park entrance. In the second chapter of this thesis, I examine whether colobus are able to perceive roads as risky, as well as to determine risk variance between two different road types. Specifically, I compared the main tarmac road at the southern edge of JCBNP to a smaller dirt road in the same study area. Using video recordings obtained during two research trips over a one-year period (July-September 2019 and February-March 2020) and opportunistic data, I quantified three behavioural indicators of risk awareness: hesitancy to cross, crossing time and mid-cross sitting occurrence. Analyses of video footage revealed that colobus did not seem to perceive the risk variance associated with the different roads. However, despite age having no effect on wariness when crossing roads, females showed greater wariness than males, and females with clinging infants even more so. Combining these observations with data from chapter 1, I suggest that this wariness by females is not enough to reduce the risk that roads pose to colobus as we there was no corresponding difference in mortality likelihood of females, by comparison to males in the roadkill dataset.