International Journal of Primatology, Vol. 18, No. 2, 1997

How Many Red Ruffed Lemurs Are Left? Natalie Vasey1 Received July 11, 1996; accepted December 2, 1996

I surveyed the Anaovandrano River Watershed to assess whether large tracts of primary lowland rain forest still occur there and to what extent red ruffed lemurs (Yarecia variegata rubra) populate the interior of the Masoala Peninsula. These data permit an informed estimate of the remaining population size for this subspecies. I located large tracts of primary lowland rain forest in the interior, but fragmentation has begun. Red ruffed lemurs appear to be less dense in the interior than they are in remaining lowland coastal rain forest. I estimate the remaining population size to be between 29,682 and 51,115 individuals. Breeding population size is substantially smaller, probably < 14,841 individuals. KEY WORDS: Masoala Peninsula; Varecia variegata rubra; population and habitat survey. INTRODUCTION

If large tracts of primary lowland rain forest still occur on the Masoala Peninsula, the Anaovandrano Watershed seems most likely to hold them. This watershed flows at higher elevation and descends more steeply than other large watersheds on the peninsula (Fig. 1). It is not used by local people as an east-west thoroughfare and forms a natural barrier to foot traffic. In January 1995, I completed a survey of the Anaovandrano Watershed between the westernmost settlements of the east and the west coasts. I address three main questions: (1) Do large tracts of primary lowland rain forest still exist? (2) If so, to what extent do red ruffed lemurs populate the interior? and (3) How many red ruffed lemurs are left? Answers to these questions can set conservation management priorities for the red 'Department of Anthropology, Campus Box 1114, Washington University, St. Louis, Missouri 63130. 207 0164-029U97/0400-0207$12.50/D © 1997 Plenum Publishing Corporation

208

Vasey

Fig. 1. Selected features on the southern half of the Masoala Peninsula, Watersheds are labeled. Solid lines, rivers; broken contour lines, terrain 300-m elevation and higher; solid contour lines, terrain 200-300 m in elevation in the Anaovandrano Watershed; half-filled circles, temporary shelters; filled circles, cleared forest and shelters. Travelling east to west, the fifth half-filled circle depicts the location of both the fifth and the sixth shelters. The first filled circle on the south shore depicts the location of both the eighth and the ninth shelters. We slept at shelters 1, 2, 4, and 10. Therefore the distances between these shelters comprise the portions of the watershed we surveyed each day. The first day of the survey we hiked at 100 to 200-m elevation, the second day mostly at 300-m elevation with surrounding mountains rising 400-500 m, the third day we hiked mostly at 400-m elevation with surrounding mountains rising 500-600 m, the fourth day we hiked mostly at 400-m elevation with surrounding mountains rising 500-700 m, and the fifth day we hiked at 400 to 500-m elevation with surrounding mountains rising 600-860 m until we descended to the coast. The map is compiled from maps published by the F.T.M. (Institut National de Geodesie et Cartographie) in 1976, which are based on aerial photographs taken in 1957 and 1961. (Map drawn by the author.)

ruffed lemur: the distribution of red ruffed lemurs is nearly limited to the Masoala Peninsula and most of the coastal lowlands on the peninsula either are deforested (east coast) or are currently undergoing rapid transformation due to logging, slash and burn agriculture, and hunting (west coast). If red ruffed lemurs are sparsely distributed or absent in the interior, then

How Many Red Ruffed Lemurs Are Left?

209

the dense populations in the remaining west coast forests are our last reservoir and should be the main focus of a conservation management plan. METHOD Navigation and Charting. We used an altimeter, a Brunton international waterproof pocket transit, and the most detailed maps available (F.T.M. topographic maps at 1:100,000 scale) to navigate through the Anaovandrano Watershed. Declination is not printed on the maps so it was not possible to triangulate precise locations we passed using topographic landmarks. Peddie (1993) has determined 11.5° declination for this region. Due to dense vegetation, we could estimate our position relative to mapped landmarks (mountains, valleys, and rivers) only by climbing trees on mountain tops or when hiking alongside the river. To chart our route, I counted the number of mountains that we passed and judged the height, shape, and position of mountains relative to each other and to the river whenever possible and drew an approximation of the route taken onto the F.T.M. maps. We hiked between 45 and 60 km while crossing the watershed. We simultaneously surveyed as we hiked. Data Collection. I employed two methods to assess whether large tracts of primary lowland rain forest still exist. (1) I noted the presence or absence of two tree species which indicate human resource use: Canarium madagascariense and Ravenala madagascariensis. Local people use large mature Canarium madagascariense to construct dugout canoes. Therefore, large mature trees of this species are rare near human settlements and stumps are common. Ravenala madagascariensis is used to construct huts. Likewise, large mature Ravenala madagascariensis are rare near human settlements and stumps are common. This light-loving species proliferates in secondary forest but is also found in primary forest (Koechlin, 1972). Its hard underground seeds are protected against fire and can persist for many years in the forest until they receive sufficient light for germination, i.e., after land is cleared or where natural disturbances have occurred, e.g., along watersheds and after cyclone damage. Presence of large mature individuals of one or both species without accompanying stumps indicated the presence of primary rain forest free of human exploitation. Their absence could be due to human removal, natural floral geographic variation, or absence of natural disturbances for Ravenala madagascariensis. (2) Whenever we encountered traces of human resource use, I determined the direction from which the people who had left the traces had come and the types of resources they had harvested with information provided by my guides. I

210

Vasey

evaluated each type of human resource use as concerns threat to lemur populations and forest fragmentation. I employed three methods to estimate the extent to which red ruffed lemurs populate the interior. (1) I noted the presence or absence of two tree species that red ruffed lemurs appear to be highly dependent upon: Canarium madagascariense and Canarium boivini. At Andranobe, my longterm study site (Fig. 1), red ruffed lemurs fed extensively on the fruit of large mature Canarium madagascariense between March and June 1994, a time of relative fruit scarcity in the region (Andrianisa 1989, Rigamonti, 1993). They commonly fed upon Canarium boivini in October and November at Andranobe. Accordingly, the presence of large mature individuals of both plant species, particularly Canarium madagascariense, indicate habitat potentially suitable for a dense population of red ruffed lemurs. Contrarily, the absence of these two tree species indicate habitat capable of supporting less dense populations of red ruffed lemurs. (2) I searched for areas where lemurs might cross the Anaovandrano River via canopy foliage to assess the contiguity of lemur populations in the southern part of the Masoala Peninsula. The presence of such dispersal corridors suggests a larger breeding population. The absence of dispersal corridors indicates fragmented lemur populations each with smaller effective population sizes. (3) I listened for red ruffed lemur loud calls to assess their presence directly. Red ruffed lemur calls or sightings are poor primary sources to estimate population density. Population Size Estimates. I provide nine estimates of how many red ruffed lemurs are left by multiplying three estimates of population density by three different estimates of how much land on the Masoala Peninsula currently provides suitable habitat for red ruffed lemurs. The three population density estimates are based upon one community of red ruffed lemurs from Andranobe. Estimates of how much suitable habitat remains assume that (1) red ruffed lemurs occur at all elevations, (2) red ruffed lemurs occur mainly below 700 m, or (3) red ruffed lemurs occur mainly below 400 m. The estimates of remaining suitable habitat include only Masoala National Park Lands (49.2% of the Masoala Peninsula) because including forest outside the proposed park lands would grossly inflate any estimate of remaining red ruffed lemur population size. Proposed park lands exclude village territories and most land below 300 m in elevation and 20° in slope. Where forest still remains in these areas, red ruffed lemur populations are highly fragmented or absent due to hunting, habitat fragmentation or habitat loss (PCDI Masoala, 1995, personal observation). Proportions of park lands in each elevation category are provided by the PCDI Masoala (1995).

How Many Red Ruffed Lemurs Are Left?

211

RESULTS, INTERPRETATIONS, AND DISCUSSION

Navigation and Charting. We crossed the watershed by foot from east to west in 5 days (January 22-26, 1995), after traveling up the river for 2 days by dugout canoe. When we could no longer continue up the river by canoe, the first mountains, rising 200 m in altitude, were visible (Fig. 1). The route surveyed and the distance covered each day are shown in Fig. 1. Guides were familiar with terrain covered only during the first 1.5 and the last 1.75 days. Timing and fair weather were key to the successful completion of the survey. January being the first month of the 8-month rainy season, the river was still low. Therefore, we were able to guide ourselves by it when hiking through terrain neither guide knew, and we were able to cross it when one or another shore was not navigable. January is also the first month of the cyclone season, which is exceptionally long in this region of Madagascar (Donque, 1972). Fortunately, we experienced no heavy rain or storms, which would have slowed us down, forced us to return the way we came, or trapped us in the mountains until fair weather and low waters returned. Indicator Species. Large mature Canarium madagascariense were common on the first and fifth days of the survey, present but not abundant on the fourth day, and uncommon on the second and third days. Mature Ravenala madagascariensis dotted even hills of low elevation on the first and second days of the survey and were common on the third and fourth days as well. We did not find stumps of either species except where land was cleared for agriculture. These results indicate presence of primary rain forest free of human exploitation on all 5 survey days. Uncommon occurrence of Cananum madagascariense and Canarium boivini on the second and third days appears to be due to natural floral geographic variation. It is not due to human removal. Indeed, much of the landscape surveyed was pristine. Compared to other watersheds on the Masoala Peninsula, the Anaovandrano Watershed is exceptional in that it still harbors large tracts of virgin low-elevation rain forest in its interior. Since Canarium madagascariense and Canarium boivini were uncommon on the second and third days of the survey, habitat suitable for supporting a dense population of red ruffed lemurs may not be present or may be patchily distributed in the interior of the Anaovandrano Watershed. Red ruffed lemur density may increase in the stretch of forest covered on the fourth day, as more Canarium occur there. However, red ruffed lemur density is probably greatest in the stretches of forest covered on the first day (low altitude) and on the fifth day (high and low altitude). It is unlikely that these findings are an artifact of surveying close to the river since proximity to rivers does not discourage growth of large mature Canarium.

212

Vasey

Absence of Canarium near the Anaovandrano River probably reflects its absence in surrounding mountains. In summary, red ruffed lemur population density may be relatively low throughout much of the watershed, particularly in the higher-elevation terrain covered during days 2, 3, and 4 (Fig. 1). Human Utilization. Aside from paths left by wild pigs, we found no existing trails between the 200-m mountains passed on the first day of the survey and the mountains near the village of Ambanizana, which we reached on the fifth day. Human trails are easily distinguished from pig trails by machete hackings on trunks and cut saplings. We located 10 temporary shelters beside the river during the first 4 days of the survey (Fig. 1). These shelters are constructed by lashing saplings together with vines for the frame and thatching leaves together for the roof. The first six shelters were used by local men while hunting wild pigs and gathering bilahy bark to make an alcoholic beverage, betsa betsa. The first shelter is of Anaovandrano (eastern) construction, the second of Fampotobe (southern) construction, the third and fourth were so dilapidated it was not possible to tell from which direction the maker came, and the fifth and sixth are of Ambanizana (western) construction. The remaining shelters (7-10) are of Ambanizana/Nandrahanana construction and are associated with forest cleared for agriculture (tavy) in 1993 or 1994. The cleared forest beside shelters 7, 8, and 9 had not yet been burned for the first time at the time of the survey. In summary, encroachment comes from eastern, southern, and western human settlements (Fig. 1). Men hunting wild pigs and gathering bilahy bark spend several days in the forest because these resources are located too far away from their villages for them to walk back and forth each day. During these forays into the forest, they may occasionally remove other resources such as wild tubers (Dioscorea sp.), but they do not clear land for agriculture or for setting up lemur traps (laly). Thus the types of human utilization observed up through the sixth temporary shelter do not pose any threat to lemurs or contribute to forest fragmentation. However, land cleared in the western part of the watershed clearly indicates forest fragmentation. On the Masoala Peninsula, laly are commonly cut in forest adjacent to tavy. Hence, the western encroachment poses a grave threat, as it will most likely result in lemur hunting and currently involves deforestation surrounding the headwaters. Natural Barriers and Corridors for Dispersal. I did not observe locations where lemurs could cross the Anaovandrano River via canopy foliage until the fourth day of the survey, when the river narrowed and the western headwaters were reached. The river appears to form a natural barrier preventing regular gene exchange between populations up until the western headwaters. Thus red ruffed lemur populations from the north and south

How Many Red Ruffed Lemurs Are Left?

213

currently are contiguous in the west and form one large breeding population. However, the new tavy are located in the western headwaters where red ruffed lemur populations become contiguous and probably more dense. If land continues to be cleared surrounding the headwaters, red ruffed lemur populations will become fragmented and effective population size will diminish. Lemur Calls. We heard red ruffed lemur loud calls in the evening and morning adjacent to the river during canoe travel into the watershed. However, we heard no loud calls during the first 4 days of the survey. We sighted one red ruffed lemur on the second day. Previous survey work indicates that red ruffed lemurs are abundant in the region traversed on the fifth day (Vasey, unpublished survey data for the mountain range east of Ambanizana, 1990). Several sources of bias may contribute to these findings. (1) We did not replicate walks; wishing to reach Ambanizana while good weather held out, and due to limited food provisions and time, we had to cover as much terrain as possible each day. (2) When crossing terrain that neither guide knew (survey days 2-4), we stayed within earshot of the river, which may have drowned out the sound of calls; the risk of becoming lost was a consideration, as no one, either native or foreigner, had ever crossed the Anaovandrano Watershed. (3) Ruffed lemur density may be low or patchy in the interior of the watershed wherever Canarium, or other critical resources, are sparse. Population Size Estimates. Given the results of this survey, the most accurate estimate of how many red ruffed lemurs are left from among those calculated (Table I) is the one excluding habitat higher than 400 m in elevation. Masoala National Park Lands 0-400 m in elevation hold an estimated population of between 29,682 and 51,115 individuals. This estimate is probably high because (1) population density estimates upon which it is based are heavily weighted by newborns and yearlings with unknown rates of survivorship, (2) it is based on the highest reliable population density estimate known for the species (Pollock, 1975; Morland, 1991; White et al., 1995), and (3) the estimate assumes that this subspecies is uniformly distributed and forms one contiguous breeding population, yet red ruffed lemurs appear to have a patchy distribution even in primary rain forest (Vasey, unpublished survey data for the mountain range east of Ambanizana, 1990; this study). Population densities of ruffed lemurs at high altitudes in other regions of Madagascar are extremely low [1.5 and 16.2/km2 (White et al., 1995); < 4/km2 (Pollock, 1975)], supporting the claim that the population size estimate using forest under 400 m is most accurate. In adopting a conservation management plan, it is of great interest to know the number of reproductive individuals in a population that are con-

Vasey

214

Table I. Estimates of Remaining Population Size for Red Ruffed Lemurs Based upon Estimates of Population Density and Estimates of Distribution According to Elevation Population size based on" Elevation category (m) 0-maximum 0-700 0-400

Masoala National Park Lands (km2) 2,032.46 1,628.73 951.34

Max. pop. density (53.73 ind./km2) 109,204 87,512 51,115

Min. pop. density (31.2 ind./km2) 63,414 50,816 29,682

Pop. Density of breeders (15.6/km2) 31,702 25,405 14,841

"Population density estimates are based upon a 15-month study (Vasey, 1997) of one community of red ruffed lemurs in the Andranobe Watershed on the west coast of the Masoala Peninsula at a study site that ranges between 110 and 260 m in elevation. Maximum population density is based upon maximum community size during the study (31 individuals— January 1995), minimum population density is based upon minimum community size during the study (18 individuals— November 1993), and population density of breeders is based upon the number of individuals that bred successfully in 1994 (9 individuals, 6 females and 3 males). The increase in community size is due to birth of infants. Home range size (57.7 ha) was divided by the maximum and minimum community size and by the number of successful breeders to obtain ha per individual; 1 km2 (100 ha) was divided by ha per individual to obtain individuals/km .

tributing genetic variation to the gene pool, i.e., the effective population size (Lande and Barrowclough, 1987). However, effective population size cannot be estimated since we lack comprehensive data on dispersal distance, effective sex ratio, mating frequency across generations, variance in reproductive success, history of bottlenecks, and the extent of natural barriers to gene flow, such as rivers and patchiness of foods. As an alternative, I provide estimates of how many red ruffed lemurs are left based upon a population density estimate that includes only successful breeders from the Andranobe community. With this parameter, park lands below 400 m hold an estimated population of 14,841 individuals. This estimate is undoubtedly high for reasons 2 and 3 above. Conservation Management. One goal of conservation biologists is to assess which aspects of the biology of a population and its situation contribute most to its vulnerability, and therefore, which aspects might be most effectively targeted for management (Lacy, 1994). This survey and related research on red ruffed lemurs lead me to recommend the following. Two critical regions in the southern part of the Masoala Peninsula should be adequately protected: (1) the western headwaters of the Anaovandrano Watershed, where northern and southern populations become fully contiguous; and (2) the remaining intact west coast forests, which harbor the most

How Many Red Ruffed Lemurs Are Left?

215

dense population of red ruffed lemurs known to date. Future work should include efforts to locate stands of resources that are critical to red ruffed lemurs, in both the remaining coastal forests and the interior. Preserving forested links between these stands may be key to the survival of red ruffed lemurs. ACKNOWLEDGMENTS

This work was conducted under an International Interuniversity Accord between the Department of Paleontology and Biological Anthropology, University of Tananarive, and the Department of Anthropology, Washington University, St. Louis, with permission from the Ministry of Universities, the Ministry of Water and Forests, and the Ministry of Scientific Research, Madagascar, the Dean of the Faculty of Sciences, University of Tananarive, the Graduate School of Arts and Sciences, Washington University, and the Masoala ICDE Funding was provided by the Wenner-Gren Foundation, the L. S. B. Leakey Foundation, the National Science Foundation, Primate Conservation Inc., the Boise Fund of Oxford University, and Sigma Xi. Special thanks go to W. L. Kaufmann for sending freeze-dried food to Madagascar. I would like to acknowledge the four local men who accompanied me: Morazava, from the westernmost settlement in the Anaovandrano Watershed, served as a guide; the latter's nephew helped carry supplies and equipment; Pere Moi'se, from the west coast of the peninsula, also served as a guide; and my assistant, Velo Norbert, translated information provided by my guides concerning traces of human occupation found en route. REFERENCES Andrianisa, J. A. (1989). Observation Phtnologiques dans la Reserve Sptciale de Nosy Mangabe, Maroantsetra, Unpublished Memories de fin d'e'tudes, University d'Antananarivo Madagascar. Donque, G. (1972). The climatology of Madagascar. In Battistini, R., and Richard-Vindard, G. (eds.), Biogeography and Ecology in Madagascar, Dr. W Junk, The Hague, pp. 87-144. Koechlin, J. (1972). Flora and vegetation of Madagascar. In Battistini, R., and Richard-Vindard, G. (eds.), Biogeography and Ecology in Madagascar, Dr. W Junk, The Hague, pp. 145-190. Lacy, R. C. (1994). What is population (and habitat) viability analysis? Primate Conserv. 14-15: 27-33. Lande, R., and Barrowclough, G. F. (1987). Effective population size, genetic variation and their use in population management. In Soule, M. E. (ed.), Viable Populations for Conservation, Cambridge University Press, New York, pp. 87-123. Morland, H. S. (1991). Social Organization and Ecology of Black and White Ruffed Lemurs (Varecia variegata variegata) in Lowland Rain Forest, Nosy Mangabe, Madagascar, Unpublished Ph.D. dissertation, Yale University, New Haven, CT

216

Vasey

PCDI Masoala (1995). Proposition des Limites du Pare National Masoala, Unpublished document, Masoala Integrated Conservation and Development Project. Peddie, N. W. (1993). The magnetic field of the Earth, 1990 declination chart, USGS Geophysical Investigations Map GP-1004-D. Pollock, J. I. (1975). The Social Ecology and Behavior of lndri Indri, Unpublished Ph.D. thesis, University of London, London. Rigamonti, M. M. (1993). Home range and diet in red ruffed lemurs (Varecia variegaia rubra) on the Masoala Peninsula, Madagascar. In Kappeler, P. M., and Ganzhorn, J. U. (eds.), Lemur Social Systems and Their Ecological Basis, Plenum, New York, pp. 25-39. White, F. J., Ovedorff, D. J., Balko, E. A., and Wright, P. C. (1995). Distribution of ruffed lemurs (Varecia variegata) in Ranomafana National Park, Madagascar. Folia Primatol, 64: 124-131. Vasey, N. (1997). Community ecology and behavior of Varecia variegata rubra and Lemur fulvus albifrons, on the Masoala Peninsula. Madagascar, Ph.D. thesis, Washington University, St. Louis, MO (in preparation).