RESUMO
How rodents perceive predation risk may alter their seed foraging behaviour and therefore potentially influence the recruitment of tree species. In this study we used two methods to investigate the effect of predation risk on habitat use by the African giant pouched rat (Cricetomys sp. nov) in Ngel Nyaki forest reserve, Nigeria. The first method was 'giving up density' (GUD), an index of perceived risk of predation at an artificial food patch, and the second was the 'spool-and-line' approach, whereby unravelling spools attached to rodent bodies are used to trace their tracks. For our GUD experiment, we chose four major sites in the forest; two representative of core habitat and two at the forest edge. Additionally, three characteristic microsites were used in the GUD experiment: dense understory, open understory and near-burrows. We hypothesised that GUDs would be lower on every succeeding observation day as rats learn to use the food patches, higher GUDs would be observed in the forest edges and open microsites, and rats would show preference for the microhabitats with least exposure to potential predators. In support of our first hypothesis, we found that GUDs were highest on the first experimental nights of every session. We also found that GUDs in the forest edges were higher than GUDs in the forest core. Lower GUDs were observed close to the rat burrows and in dense understory microsites, even though these differences were not statistically significant. Tracking of rat movements using the spool-and-line method overall revealed an even use across microhabitats, with a weak preference for those with logs, dense understory or exposed ground. Overall, our results suggest that vegetation density on a microhabitat scale has little or no effect on the perception of predation risk by African giant pouched rats.