• fragmentation
• lemurs
• Madagascar
• perceived predation risk
• primates

The littoral forest of South-eastern Madagascar is one of the most threatened ecosystems in the world, due to the high levels of deforestation and habitat fragmentation. The red-collared brown lemur, Eulemur collaris, is the largest frugivorous remaining in this habitat, where it plays a fundamental role in forest regeneration as seed disperser.
Red-Collared Brown Lemurs show a high social and ecological flexibility, by adopting several behavioural and ecological strategies to cope with the unpredictability and degradation of their seasonal and fragmented environment. Despite this, due to the increasing loss of suitable habitats for this species, IUCN recently listed this lemur as Endangered.
In fragmented habitats where predators are still present animals are expected to be more exposed to predation because of the modified habitat structure, the small group size or the overall vulnerable state due to nutritional stress. Thus, this aspect is a crucial one to understand whether these animals can cope with the new condition.
In the present study I evaluated the simultaneous effect of habitat degradation, group size, habitat use, and physiological stress on perceived predation risk on two groups of Eulemur collaris living in the degraded forest of Mandena and two groups living in the less disturbed area of Sainte Luce, south-eastern Madagascar. I also evaluated whether animals of different sex or different rank show any difference in predation risk perception.
The lemurs were observed for a one year using an Instantaneous Focal Sampling Method every 5 minutes and an All Occurrences Method for rare behaviours. Perceived predation risk was evaluated indirectly in terms of frequencies of anti-predator strategies (alarm calls and scanning) used by the lemurs during feeding and resting sessions. I used as predictors the exposure in tree crowns, the exposure within the forest canopy, the party size, the distance travelled, and physiological stress. Perceived predation risk, the response variables, and all the predictors were modelled using a General Linear Model.
Furthermore, in order to elucidate the function of the unusual adult sex ratio in Eulemur collaris, which shows a multimale-multifemale social structure, the possible anti-predator role of the “extra-males” both for vigilance and as “sentinels” to better defend the offspring was tested by analysing the effect of sex, rank and presence of infants both on scanning frequencies and alarm rates, by using a Repeated Measures ANOVA model.
A significant difference in terms of scanning rate was found between groups during feeding. The exposure on the crown and party size had a significant effect on alarm rates during feeding. No significant effect of the exposure in the canopy was found in terms of perceived predation risk. Daily path length had a significant effect during feeding both on alarm rates and scanning frequencies. Physiological stress had a significant effect on alarm rate during feeding.
This findings seems to confirm the hypothesis that in fragmented habitats animals have to accurately evaluate costs and benefits of their survival strategies, in order to cope simultaneously with their perceived predation risk and with scarcity and patchily distribution of food resources, by making a choice accordingly to their constraints. This is further supported by the negative correlation found in the present study between time spent feeding and time spent for predator vigilance: indeed, in order to maximize their food intake, E. collaris seems to minimize the scanning frequencies during feeding, preferring to rather rely on some "sentinel" individuals that perform anti-predator alarm calls.
No one of the social correlates considered in the analysis (sex, rank, presence of offspring) had a significant differences in terms of alarm rates or scanning frequencies. This finding does not support the possible anti-predator role of the extra-males for a better defence of infants against predators.
Primate conservation strategies should take into account the role of predation in fragmented habitats. The present study is consistent with the most recent researches which underlined the necessity to evaluate the simultaneous effect distribution and availability of resources and lemurs' landscape of fear in fragmented habitats.
In fact, even when the animals are able to survive on available resources, potential extirpation due to predation events may have profound consequence for the ecosystem being recolonization less likely in forest fragments.