Introduction

The red ruffed lemur, Varecia rubra, and the black-and-white ruffed lemur, Varecia variegata, inhabit the eastern rain forests of Madagascar. The most recent IUCN assessment treated them as subspecies, with the former categorized as Critically Endangered and the latter as Endangered (IUCN 2006), but they have recently been reclassified from subspecies to full species (Groves 2001, 2005; Vasey and Tattersall 2002). The two taxa currently exhibit a parapatric distribution, with the Antainambalana River in northeastern Madagascar considered the primary geographic boundary between them (Fig. 1; Petter et al. 1977; Tattersall 1982). Varecia variegata has a range extending from south of the Mananara River in southeastern Madagascar to the Antainambalana River, northwest of the Bay of Antongil (Petter et al. 1977; Tattersall 1977). Varecia rubra has a smaller geographic range occurring only in northeastern Madagascar. While it is known to occur on the Masoala Peninsula to the exclusion of V. variegata, there is little data on the western or northern boundaries of its range (Tattersall 1977).

A recent review of a wide variety of data, including historical documents and museum collections, has revealed several localities in northeastern Madagascar where more than one species of Varecia was collected in the late 19^th and early 20^th centuries (Vasey and Tattersall 2002). These localities span the three major river drainages north of the Bay of Antongil: from east to west, they are the Mahalevona, the Andranofotsy, and the Antainambalana Rivers (Fig. 1). Despite this area of historical overlap, hybridization appears to have been rare in the wild, although it evidently spanned all three river drainages. Only a handful of wild hybrids are known from northeastern Madagascar. All were collected in the late 19^th and early 20^th centuries, and diagnosed on the basis of their pelage patterns (Buettner-Janusch and Tattersall 1985; Vasey and Tattersall 2002).

Figure 1.

Masoala Peninsula and region north of the Bay of Antongil and its major river drainages. Small black arrows indicate the Mahalevona, Andranofotsy, Antainambalana, Vohimaro, and Sahantaha Rivers. Labelled dots () indicate towns or villages mentioned. Stippling () shows area of historical overlap between different species of ruffed lemur where hybridization has occasionally occurred, with hybrids collected at Mahalevona and Bevato (Vasey and Tattersall 2002). Dashed lines (----) mark boundaries of the Masoala National Park. Dotted lines (......) mark the boundaries of the Makira Protected Area and its link to the Masoala National Park near Sahavary. (Note that boundaries of the Makira Protected Area extend beyond the region depicted to the north, west, and south). Thick black line () running from Mahalevona to Andaparaty represents the route for the habitat survey, September 1998. Numbers along route indicate localities where Varecia were observed or heard, and correspond to geo-referenced localities in Table 1 (5 = Belampona River; 15 = Ampoantsatroka Village; 27 = Sahantaha Village). Inset shows the general distributions of black-and-white and red ruffed lemurs (hatching and black areas indicate approximate distributional limits of each species but do not imply continuous distributions within the indicated regions. Figure adapted from Vasey and Tattersall (2002).

Beginning in the late 1980s, habitats surrounding the Bay of Antongil were identified as priority areas for Integrated Conservation and Development Projects (ICDP). The region north of the Bay of Antongil, in particular, was considered an area of exceptional biological importance (e.g., Ganzhorn et al. 1997). Two protected areas have since been established there: the Masoala National Park, created in 1996, encompassing 230,000 ha, most of which lies on the Masoala Peninsula east of the Bay of Antongil; and the Makira Protected Area, established in 2004, encompassing 371,000 ha, mostly west and northwest of the Bay of Antongil (Fig. 1). North of the bay, in their narrowest sections, the two protected areas are contiguous, joined together by forest blocks that are linked by narrow forest corridors, thereby spanning all three river drainages mentioned above. The Masoala National Park harbors V. rubra, while the Makira Protected Area contains both V. rubra and V. variegata, including some zones where the two species overlapped historically.

The establishment of ICDPs, such as those surrounding the Bay of Antongil, depended upon information on the distribution, abundance, and habitat requirements of threatened species, and the size, condition of, and threats to forest remnants (e.g., Merenlender et al. 1998; Kremen et al. 1999). The accomplishments of ICDPs should not be undervalued, even though this approach is currently undergoing considerable revision—creating protected areas hand-in-hand with improving livelihoods for local people proved more difficult to implement than to plan (Randrianandianina et al. 2003; Kauffman 2006). The Masoala National Park, for example, was created to protect the largest remaining tracts of lowland rainforest in Madagascar and, although designed to maintain a large, relatively pristine core area, its management regime allows for experimental timber harvesting along a number of river drainages previously subjected to substantial disturbance in the form of slash-and-burn agriculture, referred to in Malagasy as tavy (CARE/WCS/TPF 1995). Thus, the protected forest blocks in the northern part of the Masoala National Park lie between river drainages, while the protected forest corridors run across (or near) their headwaters.

Surveys conducted in the Antainambalana and Andranofotsy river drainages in the mid-1980s appeared to validate the species boundaries for V. rubra and V. variegata established by Petter et al. (1977). Few animals were sighted, however, suggesting that Varecia populations were small and difficult to detect in these areas due to anthropogenic habitat degradation and hunting (Simons and Lindsay 1987). Recent ecological studies on Varecia suggest that this obligate frugivore is a denizen of the largest trees in primary forest and for this reason is the most susceptible of the extant lemurs to habitat disruption resulting from selective logging and forest clearing (Vasey 1997,2002; Balko and Underwood 2005). Ecologically sensitive taxa are prime candidates for population and habitat viability analysis because the conservation measures designed to protect them are likely to contribute to the survival of many other species residing in the same biotic community (see, for example, Soulé 1987).

The goals of this project were threefold: 1) to determine the extent of suitable habitat for Varecia populations in selected areas of the three river-drainages directly north of the Bay of Antongil; 2) to assess the extent of human exploitation of this taxon in the region; and 3) to obtain data on the northern and, particularly, the western boundaries of V. rubra. Our work thus provides information on the conservation status of ruffed lemurs in a critical part of their range—an area where the two Varecia species overlapped historically, would occasionally hybridize, and where periodic reassessments of their habitat in and near protected areas are warranted in order to further implement and adapt current conservation measures. The enduring value of the geo-referenced habitat survey and interview results presented below comes from their use as comparators for future population and habitat assessments in the region north of the Bay of Antongil subsequent to the establishment of protected areas.

Methods

Survey techniques and data collection

Our aim was to conduct a brief survey over a relatively large area north of the Bay of Antongil, integrating local knowledge en route. With the help of several local assistants we conducted interviews with local informants, completed habitat surveys, and monitored the presence or otherwise of Varecia. We worked among the three major river drainages that enter the bay, which are from east to west, the Mahalevona, Andranofotsy, and Antainambalana. For habitat surveys, we used existing trails, which varied in condition from well-used with open canopy to freshly cut trails within relatively intact (i.e., primary) or secondary forest. En route, we listened for Varecia vocalizations and monitored for any signs of lemur activity (e.g., fruit dropping from the canopy). At sites with intact forest we split into three teams of two, and monitored for lemur activity off trail for 50–100 m. Work was suspended when heavy rain reduced audibility and visibility in the forest. Most rainstorms were brief, however, and we were able to resume our surveys after a pause of 1–2 hrs. In total, we surveyed on every day between 2–14 September 1998, between 0700 and 1800 h. Total effort each day depended on the condition of the forest (i.e., intact or secondary) with an average of 4 hrs spent in each intact forest patch encountered.

Local people were interviewed for information on Varecia populations and asked under what circumstances they interacted with the lemurs in their community (n = 16, Table 1). Interviews were conducted in Malagasy by Marius Rakotondratsima. Only open-ended questions were asked. We searched for Varecia in the forests where informants indicated they occurred, and carried out assessments of the habitat, estimating canopy cover and the presence of dominant plant species. One of our guides (Jao Aridy) had worked extensively with researchers on a forestry project in the area and was familiar with intact and secondary forest communities. We assessed each area for evidence of recent anthropogenic activity, including the collection of wood or other materials. We used a Garmin GPS12XL unit to collect waypoints at 15-min intervals along the route. These were used to determine total distance covered and to note forest condition between waypoints. Intact forest in surveyed areas was quantified as the percent of the total linear distance surveyed in kilometers. This was repeated for each type of habitat observed (i.e., secondary forest, agricultural land). We also noted the presence of tavy and laly. Laly are long narrow swaths of forest cleared for the purpose of setting snare traps for lemurs. When Varecia were sighted, we noted the tree species in which they were located and whether they were feeding, resting, or engaged in other activities. We noted pelage color and patterns and photographed animals using a 300 mm lens. The other diurnal lemur that occurs in the survey area is Eulemur fulvus albifrons; when sighted we noted the location, but it was otherwise not a focus of our study.

Table 1.

Geo-referenced localities, habitat types, and interview results for the occurrence of Varecia north of the Bay of Antongil, northeastern Madagascar.

Results

In total, we covered over 120 km of trails in forest remnants. Table 1 provides the results of our geo-referencing and interviews, along with a summary of our findings on habitat for the 27 localities along our route. Habitats were significantly degraded throughout the region, with even intact forest showing evidence of human activity. Only 22% of the total linear distance surveyed was considered to be intact primary or secondary forest. Thirty-seven percent of the geo-referenced localities had one or more recent or active laly, and most areas (23 of 27) contained tavy (Table 1). In particular, areas thought to be relatively free of human disturbance during earlier surveys (CARE/WCS/TPF, 1995) were clearly used regularly for small scale tavy and wood gathering, and some showed evidence of laly as well (Table 1). Residents interviewed in 15 of 16 villages (94%) trap and eat Varecia despite knowing that it is prohibited by law (Table 1). Many villagers were aware of the declining numbers of ruffed lemurs, and suggested that their absence might be due to forest clearing and hunting. Villagers confirmed that V. rubra had disappeared from areas adjacent to villages where they had been present in the previous five years.

We sighted four V. rubra individuals in the region east of the Andranofotsy River, along the smaller Belampona River, prior to entering the Sahavary watershed (#5 in Fig. 1 and Table 1). These individuals were either resting or feeding on leaf tips or Pandanus seeds. They had large patches of white fur on the outer side of the ankle extending to the mid-thigh, a pelage pattern not seen in captivity. We later heard Varecia vocalizations on the west bank of the Andranofotsy at the village of Ampoantsatroka (#15 in Fig. 1 and Table 1). These two records indicate that Varecia still occurs between the Mahalevona, Andranofotsy and Antainambalana river drainages, and this is in agreement with interview results and the abundance of laly (Table 1). Most villagers indicated that Varecia were more active in December, a time when they are more frequently captured in laly.

At the village of Ankarongana, near the junction of the Antainambalana and Sahantaha rivers, villagers claimed that V. rubra was present to the north on Ambodi-Bilahay Mountain and to the east on Anjanaharibe Mountain (note that this is not the same location as the Réserve Spéciale d'AnjanaharibeSud lying further north). Residents of the nearby village of Sahantaha (#27 in Fig. 1 and Table 1) also stated that there were many V. rubra on Anjanaharibe Mountain. Furthermore, they related that it was not the “culture” of varignena (vernacular name for V. rubra) to cross the Antainambalana River, and that only varikandana (vernacular for V. variegata) is known from further west on Anjorompingotra Mountain.

Families living on both sides of the Antainambalana River at Sahantaha kept V. variegata as pets. The two pet lemurs we observed were captured as babies in laly, 1–2 km north of the village on the west side of the river, and both exhibited a subcincta coat pattern, the variety of V. variegata known from the west side of the Antainambalana River (Petter et al. 1977). These families regularly cross the river by canoe, taking chickens and other domestic animals with them, and presumably their pet V. v. subcincta arrived on the east side of the Antainambalana River in this manner.

Although we did not sight any wild Varecia between the Andranofotsy and Antainambalana rivers, villagers consistently referred to the ruffed lemurs in this area as varignena (i.e., V. rubra). Yet in the village of Manakana some informants described these animals as having a belt of white fur encircling the torso, as seen in the subcincta variety of V. variegata known from the west side of the Antainambalana River (Table 1; Petter et al. 1977).

Discussion

Our interviews with villagers suggest that the Antainambalana River currently remains the western limit of V. rubra's, range. The westernmost distribution of V. rubra that we were able to detect (through interviews) was near the confluence of the Antainambalana and Sahantaha rivers. Varecia rubra has been observed further northeast; several animals were sighted in the Besariaka forest corridor, about 20 km southwest of Andapa, in what is now part of the Makira Protected Area (Ranorovelohanta 1996). This is the northernmost confirmed report of V. rubra in recent years. Our sighting of V. rubra along the Belampona River was approximately 10 km east of a previous sighting near the village of Sahavary (Simons and Lindsay 1987). Limited observations of Varecia during our survey were likely due to their rarity and sparse distribution on account of hunting and habitat disturbance, as well as their relative inactivity at this time of year (V. rubra, for example, spends more time resting [∼60%] in June—September than at other times of the year [Vasey 2005; see also Morland 1993]). Early September is also a period of limited fruit availability for Varecia, when even their keystone fruit Canarium is unavailable (Vasey 2000).

Villagers in the survey area noted only the presence of V. rubra, not V. variegata. However, the pelage pattern of V. rubra described by villagers at Manakana, between the Antainambalana and Andranofotsy rivers, is unusual, and seems to resemble that of a specimen on display at the American Museum of Natural History which has been described as a hybrid (Buettner-Janusch and Tattersall 1985). This AMNH specimen has a belt of white fur encircling its torso, similar to the subcincta variety of V. variegata known from west of the Antainambalana River (Petter et al. 1977; Simons and Lindsay 1987). However, as we saw no Varecia near Manakana village, any intimation concerning natural hybridization between V. rubra and V. variegata in this locality remains speculative—it may simply be clinal variation, as seen in V. variegata along the east coast of Madagascar (Vasey and Tattersall 2002).

Only 22% of the area surveyed in the region north of the Bay of Antongil was considered primary or secondary forest. Furthermore, tavy and laly were ubiquitous along the survey route. From these findings we conclude that the natural habitat of the westernmost populations of V. rubra is disappearing at a dramatic rate. At the time of our survey, populations of V. rubra west of the Masoala National Park appeared extremely fragmented and were all under threat of extirpation by trapping and hunting. In fact, it appeared that populations of Varecia between the Andranofotsy and Antainambalana rivers (south of the east-west trail we surveyed at latitude 15°08′15″S) were completely isolated from other populations. The recent establishment of protected forest blocks and forest corridors in the region was critical (Hekkala and Rakotondratsima 1999), as these will provide the only connections between certain Varecia populations that appeared entirely isolated in forest patches surrounded by spreading agricultural land. The only links between the Masoala National Park and the north-south belt of eastern rain forest in Madagascar are these passages in the Antainambalana and Andranofotsy river drainages, now part of the Makira Protected Area.

The establishment of such corridors, in the sense described by Holloway (2000), includes the replanting of native trees to reconnect the forest blocks. Although Varecia are known to be highly sensitive to habitat disturbance, their densities can become quite high in coastal regions experiencing frequent natural habitat perturbation (Vasey 1996, 1997), and they have even incorporated large portions of pioneering, invasive species into their diet in protected forests that are being allowed to recover from cyclone damage and human degradation (e.g., Clidemia hirta, Cecropia peltata; see Ratsimbazafy 2002; Ralainasolo et al. 2005). Forest corridors created north of the Bay of Antongil, especially when planted with native trees, may ultimately prove successful in maintaining and connecting Varecia populations. The next round of surveys and in-depth field studies should focus on an examination of how Varecia populations are adapting to conditions in and near forest corridors.

The establishment of protected areas, however, can comprise only one part of an effective conservation management plan for ruffed lemurs. As food, Varecia has dual status, being both less expensive and more tasty (according to local palates) than domestic meats, such as chicken and beef (Golden 2005). In the Makira region, these two factors are paramount in reinforcing hunting activity among remote, rural people who do not have the means to either buy or produce domestic meats (Golden 2005). As indicated above, residents trap and eat Varecia in all but one of the 16 villages visited in the region north of Antongil Bay. Hence, conserving Varecia will require multi-pronged efforts involving the establishment and maintenance of forest corridors, monitoring of the protected areas, and enforcement of environmental legislation, as well as environmental education and development programs that capitalize on and are sensitive to traditional beliefs and practices (including palates). The enduring value of the geo-referenced interviews and habitat survey presented here is that they can be used as comparators for future population and habitat assessments in the region north of the Bay of Antongil subsequent to the establishment of protected areas and allied conservation measures.