Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Madagascar ranks as one of the world’s top biodiversity hotspots because of its high endemism and high rate of habitat degradation (Myers et al. 2000; Ganzhorn et al. 2001). For primates, Madagascar has the highest conservation priority with 5 endemic families and 15 endemic genera (Mittermeier et al. 2010). Ninety-seven lemur species are now recognized of which 41% are threatened with extinction (8 critically endangered, 18 endangered, and 15 vulnerable), while 42 species remain data deficient (IUCN 2010; Mittermeier et al. 2010). Knowledge obtained from long-term field studies, such as the ones described here, is particularly valuable compared to short-term “snapshots” because long-term data can be especially useful to conservation management efforts. For effective management, park authorities can benefit from understanding the differences between normal fluctuations in population size and real trends over time, patterns which can only be detected with decades of data.

To date there are four long-term lemur research sites in Madagascar: Kirindy Forest in the dry deciduous forest of the west (Ganzhorn and Kappeler 1996; Kappeler and Fichtel 2012), the southern spiny desert reserve Berenty Private Reserve (Jolly et al. 2006; Jolly 2012), Beza Mahafaly Special Reserve (Richard et al. 2002; Jolly and Sussman 2006; Sussman et al. 2012), and Ranomafana National Park (RNP) in the southeastern rainforest (Wright and Andriamihaja 2002; Wright 2004). At each site, lemurs have been studied for several decades (Jolly 1966; Sussman 1974; Richard 1978; Jolly and Sussman 2006; Sussman and Ratsirason 2006; Kappeler and Fichtel 2012). In this chapter, we discuss a compilation of the findings of long-term studies of the 13 lemur species found in RNP.

1.1 History of Ranomafana National Park

Ranomafana National Park (RNP) is a rainforest park located in southeastern Madagascar (21°16′S, 47°20′E). The landscape is dominated by submontane rainforest, which receives a mean of 3,000 mm of rain per year during the December through March rainy season. The RNP project was initiated in 1986 with the goal of protecting the habitat of a then newly discovered lemur species, the golden bamboo lemur (Hapalemur aureus >), and the rediscovered greater bamboo lemur (Prolemur simus >; Meier et al. 1987). From 1986 to 1989, logging concessions were granted by the forestry department and selective logging for valuable hardwood trees was intensive. In 1991, 41,000 ha of the montane rainforest were designated as a national park (Wright and Andriamihaja 2002). The Namorona River and a parallel paved road (Route 25) bisect the park into northern and southern parcels with a third parcel on the western boundary to the north. Patricia Wright, then at Duke University, spearheaded the initial park project, an integrated conservation and development project conducted simultaneously with research on lemurs and other aspects of biodiversity (Wright 1997; Wright and Andriamihaja 2002). Management transitioned to Stony Brook University (SBU) when Wright moved there in 1991. While retaining management of research in RNP, SBU handed park management over to the Association Nationale pour la Géstion des Aires Protegées (ANGAP), the national park system, in 1998. In 2006, the System of Protected Areas of Madagascar (SAPM) replaced ANGAP and incorporated sustainable practices of resource use into park creation and management. In 2009, SAPM became Madagascar National Parks (MNP).

1.2 Infrastructure at Ranomafana

The first research station was built in 1989 near the entrance to RNP. This structure was a small, one-story log cabin. In 2003, the station was upgraded to a three-story stone facility adjacent to the park and overlooking the rainforest. The new research station, named the International Centre for Research and Training for the Valorization of Biodiversity (Centre ValBio), is located on Route 25 near the park entrance (Wright 2004). This hub of scientific research and education is managed by a consortium of universities headed by SBU and the Institute for the Conservation of Tropical Environments (ICTE). Founding institutions include SBU and the Universities of Antananarivo, Fianarantsoa, and Helsinki. Currently, the main building houses administrative offices, a small laboratory, a library, and a dining hall that serves 65 people. A second four-story building (15,676 square feet) will open in 2012, and will be equipped with high speed internet, modern hormone, parasite, genetics and infectious disease laboratories, an audio/visual/computer center, and living accommodations for 54 students and researchers. The Centre ValBio has authorization to do research from the Ministry of Forests and Environment of the Government of Madagascar, and works closely with the MNP, especially on conservation management.

Centre ValBio’s administration oversees the Departments of Research, Biodiversity Monitoring, Logistics and Management, and Community Outreach including Health and Education. Seventy-two local staff, many trained as lemur technicians and Malagasy biodiversity experts, work at Centre ValBio and live in the villages surrounding the park. The health and education team provides training and outreach programs to 22 nearby villages. Reforestation with native species and medicinal plant gardens managed by traditional healers are two important components of Centre ValBio’s outreach efforts (Wright et al. 2005a). Twenty conservation clubs foster appreciation for conservation and a cooperative of artisanal women weavers is a sustainable contributor to village economics.

1.3 History of Ranomafana Lemur Research

Ranomafana contains 13 lemur species of which 8 have been subjects of long-term research (Table 4.1). Seven of the 13 species have been subject to taxonomic revision during the 20-year period (Table 4.2). Five species have been redesignated based on genetics and morphology. One new species has been described (H. aureus), one taxon has been raised to a new genus (P. simus), and one remains to be identified (Cheirogaleussp.).

Table 4.1 List of RNP lemur species and references to research
Full size table
Table 4.2 List of RNP lemurs including recent taxonomic changes with activity pattern: nocturnal (N), diurnal (D), and cathemeral (C); weight in grams; IUCN status
Full size table

Twenty-six PhD dissertations and 26 Masters theses on lemurs have been completed since the park’s initiation, and an additional eight dissertations are currently in progress. Major foci of these long-term studies have been diurnal lemurs, including behavioral ecology, demography, life history, reproductive biology, stress and reproductive hormones, parasites, feeding and nutritional ecology, morphometrics, predation, communication, and cognition (Table 4.1). Nocturnal lemurs have been studied less intensively, with emphases on reproduction, hibernation, parasites, and vocalizations (Table 4.1). Moreover, research in ecosystem dynamics and conservation with emphasis on lemur seed dispersal, climate, and tree phenology are also ongoing (Dew and Wright 1998; Wright et al. 2005b; Dunham 2008; Dunham et al. 2008, 2010; Ganzhorn et al. 2009).

Researchers have studied lemurs at four sites, each approximately 4 km2, within the contiguous forest of the park (Fig. 4.1). Three sites (Vatoharanana, Valohoaka, and Mangevo), each with bush camp facilities, are in undisturbed or minimally disturbed rainforest. The fourth is Talatakely located near Route 25, selectively logged by hand from 1986 to 1989, and now accessible to tourists (Wright and Andriamihaja 2002).

Fig. 4.1
figure 1

Map of Ranomafana National Park with long-term study sites and long-term transects marked

Full size image

Ad hoc transects and surveys have been conducted throughout the park since 1987. Since 2003, researchers conducted lemur surveys along eight 2 km transects, from the edges to the interior (Fig. 4.1). The surveys have led to the identification of new social groups and discovery of new species (Irwin et al. 2000, 2005; Arrigo-Nelson and Wright 2004; Andriantompohavana et al. 2007a, b; Wright et al. 2008b).

2 Long-Term Data Collection and Management

2.1 Long-Term Research on Focal Species

Most adults from 6 of the 7 diurnal lemur species at each of the main research sites have been marked with tags and collars for individual identification (Propithecus edwardsi, Hapalemur aureusand Prolemur simus, Eulemur rubriventer, E. rufifrons,and Varecia variegata). Many study groups have at least one member with a telemetry radio-collar for locating groups. Data collection on lemurs has been ongoing in Ranomafana for more than 24 years (Wright 2004; Table 4.1).

2.2 Weights, Measures, and Biomedical Data

Beginning in 1987, we established a protocol to obtain morphometric and health data on all seven diurnal lemur species. A trained team of Malagasy technicians capture individuals using remote injection techniques, whereby a Telinject blow gun or CO2-powered rifle is used to tranquilize individuals with Telazol administered with lightweight darts (Glander et al. 1992; Wright 1995). Researchers, technicians, and veterinarians measure, collect samples from, and mark adult animals with a nylon collar and individual identification tags. Since 1999, captured animals also received a subcutaneous microchip (AVID, HomeAgain®) for permanent identification. While animals are tranquilized, the team uses a checklist developed to record information on general physical condition, body mass>, and reproductive state>(Glander et al. 1992). Dental molds are taken; physical measurements are recorded; and hair, fecal, blood, and external parasite samples>are collected. We used data including body weights, canine eruption patterns, general tooth wear, female nipple length, and presence or absence of descended testes to assign age categories to all known study individuals (Johnson et al. 2005; Baden et al. 2008; Erhart and Overdorff 2008a; Wright et al. 2008a). Actual ages are known for most animals from recorded births. For individuals whose birth date is uncertain (i.e., individuals in the population since before 1986 or immigrants from other groups), Jernvall and King developed an accurate method to determine age by comparing year-to-year tooth wear (King et al. 2005). Beginning in 2003, veterinarians now also conduct detailed health evaluations>, compiling biomedical profiles of all lemur individuals (Junge and Louis 2005).

Microcebus rufus >and Cheirogaleus crossleyi >have been intermittently studied since 1990 (Wright and Martin 1995; Atsalis 1998, 2008). Beginning in 2003, subcutaneously placed microchips have allowed us to monitor individual mouse lemurs during long-term studies (Blanco 2008; Blanco and Meyer 2009; Blanco and Rahalinarivo 2010; Durden et al. 2010; Zohdy et al. 2010; Deppe 2011). Mouse lemurs are captured in Sherman traps baited with bananas, then weighed, measured, and released at the capture site. Over 300 individuals have now been marked. Dental tooth casts reveal ages, and repeated captures allow data collection on body mass fluctuations associated with torpor patterns (Atsalis et al. 1996; Atsalis 2008), reproductive status (Blanco 2008), parasite prevalence (Zohdy et al. 2010), and noninvasive behavioral experiments (Deppe and Wright 2009; Deppe 2011).

2.3 Long-Term Phenology, Climate, and Terrestrial Vertebrate Data

Tree phenology, daily rainfall>, and temperature>data have been recorded since 1987. Tree phenology>began with monthly monitoring of 100 trees of 25 lemur fruit species and was expanded in 1995 with monthly monitoring of trees from more than 71 species, representing 26 genera and 19 families (Clark and Clark 2006, 2010). Initially, tree diameter measurements (diameter at breast height, DBH) were taken every 5 years, but beginning in 2004, DBH measurements are taken every 6 months (Clark et al. 2003) to correlate tree growth with rainfall. Maximum and minimum temperature and rainfall data are taken at 06:00 h each day with a team of technicians responsible for accuracy. Camera trap data are taken in distant regions of the park to monitor carnivores and terrestrial vertebrates (Gerber et al. 2010).

2.4 Database Management>

Long-term data collection and management requires standardized protocols. The following are areas where protocols are in place: (1) permanent identification of individuals for behavioral observation and census, (2) lemur capture data including weights, measurements, and biomedical data, (3) focal lemur sampling, (4) tree phenological data, (5) daily temperature minima/maxima and rainfall, (6) ad libitum observations of predation, reproduction, and intergroup aggression, (7) GIS data for mapping, and (8) lemur fecal sample data. The phenology and climate data are archived and accessible to researchers at Centre ValBio and will soon be available on the ICTE/Centre ValBio website (http://icte.bio.sunysb.edu). Data from projects under the supervision of Dr. Wright are transcribed from field notebooks into Excel spreadsheets, by local research technicians with oversight by the CVB Chief Technical Advisor or US students at SBU with postdoctoral oversight. Data from other projects are the responsibility of the project’s principal investigators. Three years after a researcher has terminated data collection, the data can be transferred into the Centre ValBio central database and made available to CVB researchers (with proper citation assured).

2.5 Research Highlights from Long-Term Data

Since 1986, the principal goal of primate research at Ranomafana has been to understand the factors driving the behavioral ecology of lemurs in a species-rich community. Our research has focused on particular species that have been the subjects of intensive study over 20+ years, as well as community level analyses to understand the effects of competition, predation, and habitat quality on species richness and on relative abundance. Further, the mixed history of anthropogenic disturbance has allowed comparative work to elucidate the effects of disturbance on lemur physiology, behavioral ecology, and community structure. However, for this chapter we will concentrate on population changes documented over decades of observation.

3 Population Ecology

3.1 Flexibility of Behaviors

Many social behaviors were not observed until after many years of study. Female takeovers of groups, male lethal aggression from other males, and simultaneous immigrations of related males into groups were not observed in the first 10 years of studying P. edwardsi >, but were observed in the next decade (Morelli et al. 2009). After years of losing track of focal study groups, continuous long-term observations confirmed that E. rufifronsgroups seasonally expand their range and travel over 4,000 m to find fruit (Erhart and Overdorff 2008a). Likewise, when P. simusgroup size began to decline in 2003, we began to observe patterns of male disappearance, as well as females, (and their offspring) migrations and wanderings for 3 months before returning to the original territory. This migration behavior was not seen in the first decade (CVB unpublished census data; Wright et al. 2008b). In addition, while breeding out of the birth peak had been observed early on in E. rubriventer(Overdorff unpublished data), it was not until almost 20 years of data were analyzed that it was determined that they have been observed to breed in eight different months of the year (Tecot 2010).

3.2 Lemur Group Size>, Composition, and Social Organization>

Social groupings have long been investigated in primates (Crook and Gartlan 1966; Eisenberg et al. 1972; Clutton-Brock and Harvey 1977) by correlating ecological factors to social organization (Janson 1992). For instance, it has been proposed that lemurs’ relatively small group sizes may be due to nocturnal ancestry (van Schaik and Kappeler 1996) or smaller crowned fruit trees (Wright 1999). We initially reported that P. edwardsihave a multimale, multifemale social organization in groups of 3–9 individuals (Wright 1995). However, by 2003 we had observed every type of social grouping in this species with potential for all different types of mating systems. Moreover, we learned that group sizes can get as large as 11 (Pochron and Wright 2003). In contrast, we consistently found E. rubriventerin socially monogamous pairs accompanied by immature offspring (Merenlender 1993; Overdorff and Tecot 2006; Tecot 2008). Although previously described as pair-living>(Tan 1999), H. griseusand H. aureusgroups can contain two breeding females in habitats with abundant bamboo (Grassi 2002; Wright personal observation). Using the older mating system terminology did not reflect this flexibility and, as stated by Kappeler and van Schaik (2002), these groupings are really patterns of social organization. With that conceptual framework it makes sense that a high frequency of pair-living and small groups among lemurs may be a response to food resource scarcity and unpredictability (van Schaik and Kappeler 1996; Wright 1999), and may be linked to female dominance (Dunham 2008).

3.3 Dispersal>

Understanding dispersal is critical to understanding population dynamics, but these data are difficult to collect. Long-term data on multiple groups make it possible to observe dispersal in long-lived primates. For P. edwardsi,we initially thought that only males emigrated from natal groups (Wright 1995). We have since observed emigration and immigration by both sexes (Morelli 2008), aggressive group takeovers by females (Morelli et al. 2009), and targeted aggression within and between groups (Wright 1995). Females disperse from their natal groups at a younger age than their male counterparts but male secondary dispersal is more common (Morelli et al. 2009). Male dispersal occurs in the 3–5 months before the breeding season. During this time, males visit other groups and male scent marking frequencies and testicular volume increase (Pochron et al. 2005; Pochron and Wright 2005). Testosterone levels also increase (Tecot et al. 2010). Depending on breeding opportunities, males undergo natal dispersal between 3 and 9 years of age. Most adult P. edwardsimales transfer at least three times in their lifetimes (Morelli et al. 2009).

In contrast, E. rufifrons >males in Ranomafana transfer only once in their lifetimes, typically at 3–5 years old and just before the breeding season. Immigrant males are often accepted without aggression and become social partners of one adult female for 3–6 years, mating preferentially with her (Overdorff 1993, 1998; Erhart and Overdorff 1998). Female dispersal has not been seen in E. rufifrons, but groups may fission along matrilineal lines (Overdorff et al. 1999; Erhart and Overdorff 2008b), a behavior which differs from Eulemurgroups in western Madagascar (Wimmer and Kappeler 2002; Ostner and Kappeler 2004).

Observations of E. rubriventer >revealed that both sexes disperse from their natal group at 2.5–3 years of age (Overdorff and Tecot 2006). Hostile replacement of resident adult females by nongroup females has been observed, but males have not been aggressively replaced. One dispersing female was seen with a new male in an adjacent territory 15 years later (Overdorff and Tecot 2006). Our data on the genetics of E. rubriventeroffspring revealed that the resident male consistently fathers the offspring, and there are no data that indicate extra-pair copulations (Merenlender 1993).

Molecular and behavioral data show that dispersal in Vareciais not sex-biased, as within-sex relatedness scores were similar for males and females (Baden 2011). Mark–recapture studies of brown mouse lemurs (M. rufus) have shown that male membership in the population changes at a higher rate than female membership (Atsalis 2000, 2008), and that males can disperse relatively long distances (Karanewski personal communication; Zohdy personal communication). In all species thus studied, dispersal patterns ensure heterozygosity; an advantage for conservation strategies (Merenlender 1993; Morelli et al. 2009, Bradley and Baden personal communication).

3.4 Reproductive Hormones

P. edwardsi, E. rubriventer,and M. rufusfecal hormone profiles>have been developed (Tecot 2008, Tecot et al. 2009; Blanco and Meyer 2009). In P. edwardsiand E. rubriventer, estradiol, progesterone, testosterone, dihydrotestosterone, and cortisol are being measured to determine the ovarian steroid fecal metabolites that characterize reproduction. Progesterone levels can reliably indicate pregnancy in these species, and estradiol levels reliably indicate fetal sex (Tecot in press). As expected, patterns of change in fecally excreted steroid levels during the reproductive season in M. rufus >showed estradiol (E-2) levels were elevated around estrus, whereas progesterone levels were highest during late pregnancy and around parturition (Blanco 2008; Blanco and Meyer 2009). Blanco (2008) documented moderate estrous synchrony among female mouse lemurs, with clusters of females showing strong estrous synchrony. Two females showing signs of abortion or perinatal death of offspring also showed renewed vaginal swelling in late December, suggesting that some form of polyestry (i.e., as reproductive compensation for fetal loss) exists at RNP (Blanco 2008). With these baselines and proof of concept established, we can now investigate questions associated with development, sexual relationships, and seasonal breeding.

Hormonal studies of dominance rarely investigate inter-sexual relationships. To determine whether female dominance>might be mediated by hormonal levels, we investigated androgens (dihydrotestosterone and testosterone) in male and female P. edwardsi(Tecot et al. 2009). While DHT levels were higher in males than in females, there was no significant sex difference in testosterone levels. Similar testosterone results were found in M. rufus. These results differ from those found for all other mammals studied to date, in which male testosterone levels are consistently higher than female levels, with the exception of the female-dominant rock hyrax (Koren et al. 2006). In other masculinized mammals such as the hyena and ring-tailed lemur, androstenedione is elevated in females, but testosterone levels remain higher in males. This finding may have important consequences for understanding sex differences in lifespan and senescence. Maintaining high testosterone levels may explain why mammalian males frequently have shorter life expectancies than females, but testosterone burden may not explain the sex differences in lifespan in P. edwardsi. However, dispersal season testosterone levels increase significantly in both sexes, and if males continue to disperse throughout their lives and females do not, testosterone levels may still contribute to shorter male lifespan (Tecot et al. submitted).

3.5 Reproductive Success>

Obtaining lifetime reproductive success data from wild primates is possible for females who have been followed throughout their lifespans (Bronikowski et al. 2011). By coupling life-long observations with genetic evidence, calculating male reproductive success is now possible. Currently, we have lifetime reproductive success for two females and one male P. edwardsi. One female, killed by a fossa at age 32, gave birth to 13 offspring (7 males and 6 females). Five of these animals lived to reproductive age with two males living to emigrate from their natal group (Pochron and Wright 2003; King et al. 2005, 2011). A second female who died at 16 had 7 offspring (3 males and 4 females). Three males and one female survived to reproductive age; the males migrated to breeding groups and the female reproduced in her natal group after her father transferred. Genetic evidence provides a measure of lifetime reproductive success for one male who produced offspring in two groups before he was killed by an immigrant male during a group takeover at age 19. He fathered 14 offspring, 9 females and 5 males, in one group with 4 male and 3 female offspring surviving to reproductive age. Following a second transfer, he sired a 15th offspring, which disappeared, a probable infanticide, after his father was killed. Without continuous long-term data collection we would not be able to have these data on lifetime reproductive success in even these few individuals.

In many primates, heavier females have more surviving offspring (Altmann 1980; Terborgh and Janson 1986). We also see this trend in lemurs. P. edwardsifemales who were heavier during the previous mating season were significantly more likely to give birth in the following birth season than lighter females (Morelli et al. 2009). Habitat disturbance>appears to have a disproportionate impact on the body mass of female P. edwardsi. A comparison between Talatakely (logged) and Valohoaka (unlogged) revealed that adult females but not males living within the unlogged forest weighed significantly more than those females living in the disturbed forest (Arrigo-Nelson 2006). When males and females are compared within sites, significant differences in body mass were found only at the disturbed forest site. Given the climatic and reproductive synchrony of the two study sites, and the fact that body mass>is positively associated with reproductive success in some primates (Stevenson 2005), these data suggest that differences in P. edwardsifeeding behavior and nutrient intake may affect future reproductive success.

3.6 Health>and Parasites>

Although we follow many species daily, we rarely see signs of illness. Over the years, we have seen a wide range of effects of fighting and predation attempts. Wounds are relatively common during the breeding season and we have observed one or both testicles missing in individuals of E. rufifronsand P. edwardsi.In P. edwardsi, E. rubriventer, E. rufifrons, and M. rufusindividuals have been found functioning with sight in only one eye (Erhart and Overdorff 2008a). Older individuals have worn teeth and in two individuals of P. edwardsiwe have seen healed abscessed teeth (King et al. 2005; Wright et al. 2008a). In E. rufifrons, E. rubriventer,and V. variegatawe have captured very old individuals with teeth worn to the gums.

The diversity and prevalence of parasites has been found to influence health and fitness in other mammals (Hart 2007; Price and Kirkpatrick 2009). The variation among individuals and the transfer of parasites among lemur species is presently unknown, as is the incidence of disease that parasites cause. A variety of roundworms and pinworms have been observed in lemur intestines, but further study is necessary (Junge and Louis 2005; Junge and Sauther 2006; Irwin and Raharison 2009). Our initial studies suggest that one species of parasite may be found on many species of these sympatric lemurs, but there are differences in prevalence among species. For example, Makialgesspp., a parasitic mite, was abundant on P. edwardsi(80%), P. simus(67%), and H. aureus(83%), yet rare on E. rufifrons(3%) and V. variegata(14%), and absent on H. griseus(Wright et al. 2009; Hogg et al. in press). Large group size has been proposed as a factor for higher parasitism (Freeland 1976), but the largest groups in E. rufrifronshad the lowest incidence. Large body size might be more attractive for parasites than smaller body size (Freeland 1976), and indeed larger species had the most ectoparasites.

Additionally, lemurs may have species-specific parasites. For example, M. rufusis ecto-parasitized by three tick and one louse species. This louse, Lemurpediculus verruculosus, is likely a brown mouse lemur-specific parasite (Durden et al. 2010). A new species of wingless, bloodsucking hippoboscid fly, Allobosca crassipes,was recently described as a parasite of P. edwardsiand V. variegata(Vaughn and McGee 2009).

Parasites may increase in primates living in forests with anthropogenic disturbance and be correlated with disease and decreased fitness (Dobson and May 1986; Chapman et al. 2009). We have some evidence that this trend holds true in the Malagasy rainforest. Wright et al. (2009) found that habitat disturbance may account for high ectoparasite loads in P. edwardsi. Endoparasite prevalence in M. rufusin 2007 was higher in more disturbed habitat than in the less disturbed habitat. In 2008 and 2009 this difference disappeared, and linking habitat disturbance with lemur parasites should be done with caution. Again, long-term studies allowed us to differentiate between minor fluctuations and the consistent correlations with factors such as climate, body size, group size, or habitat disturbance factors.

3.7 Mortality>: Adults and Infants, and Infanticide>

Adult mortality is generally caused by predation>, rather than by illness, wounds, or infections (see Sect. 4.4.1). Adult mortality for the lemurs at Ranomafana is low, as would be expected for long-lived primates (Erhart and Overdorff 2008a; Pochron et al. 2004). In contrast, infant mortality is high (overall approximately 50%) and food stress>due to environmental unpredictability may account for some mortality (Wright 1999; Richard et al. 2002).

Long-term observations have also allowed us to document infanticides and infanticide attempts. The killing of infants of up to 2 months of age has been observed in P. edwardsi, with both immigrant males and females as perpetrators (Wright 1995; Erhart and Overdorff 1998; Pochron et al. 2004; Morelli et al. 2009). Over 24 years, there have been 9 infanticides out of 60 births (15%) associated with immigration in 4 groups of P. edwardsi(Morelli et al. 2009). These infanticides have brought the mothers back into estrus a year earlier in a species that gives birth every other year, providing males with the opportunity to improve their reproductive success (Hrdy 1977; Erhart and Overdorff 1998). A potential infanticide attempt may have also been observed when a V. variegatamale approached and then entered an unguarded nest, knocking two young infants to the ground over 10 m below; neither infant survived the fall. To date, we have not seen infanticide in either E. rufifronsor E. rubriventer(Durham 2003; Erhart and Overdorff 2008a; Tecot 2008).

3.8 Lifespan>

Our long-term research has allowed us to document long lifespans (over 30 years old) in individuals of all the diurnal species of wild rainforest lemurs, regardless of body mass (Erhart and Overdorff 2008a; Baden 2011; King et al. 2011). For P. edwardsiwe documented a dramatic difference in the maximum lifespans of males and females. Since 1986, few old males have been observed whereas three females are known to have lived beyond 30 years of age. Therefore, males cease contributing genetically to the population after about 20 years, whereas we have no evidence that old females cease reproducing (Wright et al. 2008a; King et al. 2011). In E. rufifrons, males over 10 years old were peripheralized and replaced in breeding position by younger nonnatal males (Overdorff et al. 1999). Aged males were burdened with handicaps; one had only one eye, another only partial use of the right hand, another had lost both testicles, and two had visible limps (Overdorff et al. 1999; Erhart and Overdorff 2008a).

In E. rufifronsolder individuals have been seen, and one functioned with only one eye. These scars and wounds are male-biased and indicate violent male-male aggression, which may account for the shorter lifespan of males as observed in many primate species (Bronikowski et al. 2011). In E. rubriventerolder individuals have been seen, with one female a minimum of 17 years of age (Tecot 2008), though scars and wounds in this pair-bonded species are generally not evident (Tecot personal observation). In V. variegata, both older males and older females were observed in the population at Mangevo, suggesting that male–male aggression may not be as pronounced in this species (Baden 2011). These individual life histories add up, over time, into a better understanding of the evolution of social organization in each species.

With new dental technology that has become available in the past 5 years, Zohdy has documented that wild brown mouse lemurs survive up to 9 years of age and do not experience any of the physical symptoms of senescence>that are seen in captive congeners (Bons et al. 2006). On the basis of dental wear, we have found that many brown mouse lemurs survive past 5 years (the age of the onset of senescence in captivity) and these aged individuals represent 9% of those captured. It is possible that few mouse lemurs reach old age because of high predation rates (Goodman et al. 1993; Karpanty 2006; Karpanty and Wright 2007; Sefczek 2009; Deppe 2011).

3.9 Nutritional Ecology>

In contrast with many sympatric monkey diets>, rainforest lemur diets are very diverse (Terborgh 1983; Struhsaker 1997). Studying these species over the long term and in different environments revealed the flexibility in diet within certain constraints. Many lemurs have anatomical specializations such as a large cecum or a long foregut to better digest bamboo or leaves from other plants. For example, Avahiand Lepilemur >both eat leaves, but Avahi >chooses leaves with tannins, whereas Lepilemurchooses leaves with alkaloids (Ganzhorn et al. 1985). The three bamboo lemurs all consume nearly 95% bamboo, but two species can tolerate large amounts of cyanide in the shoots (Glander et al. 1992; Tan 1999; Ballhorn et al. 2009). Unlike any other lemur species, P. simuswith its strong jaws and big teeth has physical capabilities to open the tough culm of the bamboo and eat the pith (Tan 1999; Vinyard et al. 2008; Yamashita et al. 2009), and H. griseuseats primarily bamboo leaf petioles. Daubentonia >eats beetle larvae extracted from dead wood, a niche taken by woodpeckers in other continents (Cartmill 1974). Recently, a comparative study of the four diurnal frugivores revealed that there is much more specialization in fruit choice than previously thought, with the fruit of entire plant families exploited by only one diurnal lemur species in the forest (Wright et al. 2011).

Our most comprehensive dietary studies have been conducted on P. edwardsi. Early work by Hemingway (1998) on the Vatoharanana population revealed that they ate leaves, fruits, and seeds in nearly equal proportions. More recent work by Arrigo-Nelson (2006) has added comparative data on sifaka populations at Talatakely (disturbed forest) and Valohoaka (undisturbed forest). Selective logging has altered species composition in the disturbed forest; in response, sifakas have altered their diet by consuming plant taxa in disproportion to their abundance in the forest and by relying more heavily on food from plant life forms other than trees. Disturbance limits the ability of sifakas to consume fruit and seeds, their preferred food and, as they appear to consume leaves in an effort to replace these missing foods, this creates a discrepancy in the nutrient intake of sifakas living within this habitat (Arrigo-Nelson 2006). As fruit availability was found to be lowest during the most climatologically and reproductively harsh months of the year, we hypothesize that this discrepancy may severely impact infant survival and, with it, the reproductive success of sifakas living in disturbed forest habitats.

Habitat differences>in diet are evident in E. rubriventeras well. In a 19-month study in Talatakely and Vatoharanana, Tecot (2008) found that during the scarce season, dietary overlap decreased and the proportion of the diet composed of fruits, flowers, and leaves differed between the two sites. Seasonal changes in behavior and diet were greater in the undisturbed site, indicating more flexibility in that site. Most notably, during an entire month of the scarce season, animals in the disturbed site spent 100% of their time eating unripe fruit from the invasive Chinese guava (Tecot 2008).

Atsalis (1999, 2008) conducted a 17-month feeding study on M. rufusand found that this species fed on a wide variety of fruits, mistletoe berries, and insects, especially beetles. The seeds of Bakerellaspp., an epiphytic semiparasitic mistletoes high in lipids, were present in 42% of fecal samples that contained fruit and this food was consumed year-round irrespective of habitat-wide fruit availability (Atsalis 1999, 2008). This abundant mistletoe is eaten by many lemurs and has been documented to be a fallback food in both disturbed forest areas and forest fragments (Arrigo-Nelson 2006; Irwin 2006).

4 Community Ecology>

4.1 Predators>on Lemurs

Predation>is a major selective factor in primates and major cause of mortality (Isbell 1994; van Schaik 1983; Zuberbühler 2007). Predators of Ranomafana lemurs include birds, mammalian carnivores, and snakes (Table 4.3). Two raptors, Henstii’s goshawk (Accipiter hensteii) and the Madagascar harrier hawk (Polyboroides radiatus) (Karpanty and Wright 2007) eat small-bodied nocturnal (C. crossleyi, M. rufus,and A. peyrierasi) and diurnal (E. rufifronsand H. griseus) lemurs. The lemurs preferred by raptors, A. peyrierasiand H. griseus, weigh approximately 1 kg (Karpanty 2006), suggesting that the body mass, not activity cycle, account for the preference.

Table 4.3 Known predation on lemurs in Ranomafana National Park
Full size table

A major predator of lemurs is the fossa (Cryptoprocta ferox), a viverrid carnivore that weighs between 8 and 12 kg, is agile in the trees, and hunts diurnally and nocturnally. The fossa has made a major impact on the population of the largest lemur in Ranomafana, P. edwardsi(Table 4.4; Wright 1995, 1998; Wright et al. 1997; Irwin et al. 2009) as well as on V. variegatain recent years (Baden personal communication). Over 20 years, it seems that predation events happen during temporally clumped periods that cut population densities by up to 50%. Fossa predation has caused the extinction of two long-term study groups and independent censuses of the study population support the tremendous impact on population density (Table 4.4; Irwin et al. 2009; see also Kappeler and Fichtel 2011). Galidea elegans, a diurnal mongoose weighing less than a kilogram, has been observed eating and stalking M. rufusand C. crossleyi(Wright and Martin 1995). Finally, boa constrictors (Sanzinia madagascariensis) have been mobbed by Microcebusand are known to eat Cheirogaleus(Wright and Martin 1995; Deppe 2011). Data on all of these rare events can only be accumulated during long-term studies.

Table 4.4 Changes in Propithecus edwardsipopulation size over time in relation to the observed predation events by Cryptoprocta ferox
Full size table

4.2 Interspecific Aggression>

Interference competition by close relatives has been cited as an important selective force in animal behavior (Case and Gilpin 1974; Terborgh 1983), and in a rainforest with 13 sympatric lemurs we would expect this to be the case. Results from observations over 25 years show that one strategy is for each lemur species to specialize on different fruit species (Wright et al. 2011). Yet interspecific aggression among lemurs exists, including agonistic vocalizations, chasing, and biting. These interactions occur during the season of ripe fruit availability (Overdorff and Tecot 2006; Wright personal observation). Since 88% of agonistic interactions among E. rufifrons, E. rubriventer, P. edwardsi,and V. variegataoccurred over ripe fruit during the high season for fruit availability (Overdorff et al. 2005; Overdorff and Tecot 2006), interference competition>for high quality fruit is the most likely driver of this agonism. E. rubriventerretreated in every case in the resource abundant season, but did not retreat in the few agonistic encounters which occurred with P. edwardsiand E. rufifronsduring the scarce fruit season (Overdorff and Tecot 2006). V. variegatawill often successfully defend fruit trees against all other lemur species (Balko 1998; Andrea Baden personal observation; Iris de Winter personal communication). The ruffed lemurs are most often the winner of any competition (the raucous barking of a whole group deters other species), even though the sifakas are twice their size. Congeners seem to compete most, and E. rubriventerand E. rufifronshave 0.04 aggressive interactions/hour, while V. variegataand E. rubriventercontest 0.006/h and P. edwardsiand E. rubriventercontest 0.003/h. The hierarchy established by % contests won is V. variegata(obligate frugivore), E. rufifrons, P. edwardsi,and E. rubriventer(Overdorff and Tecot 2006). During 3 months of fruit scarcity in 2010, E. rufifronsinitiated and won 7 out of 7 contests with E. rubriventer, and 6 out of 7 aggressive interactions were adjacent to fruit trees (Iris de Winter personal communication).

Interference competition may also be playing a role in the population dynamics of Eulemurspecies. Across three long-term study sites, population density changes are inverse between E. rufifronsand E. rubriventer; when E. rufifronspopulation densities increase, E. rubriventerdensity decreases and vice versa (Fig. 4.2). Further evidence that competition for food resources is driving interspecific behavior, not predation, is that polyspecific associations>are rare. This is contrary to observations of rainforest monkeys in Africa and South America where several species often feed together for protection against predators (Terborgh 1983; Holenweg et al. 1996).

Fig. 4.2
figure 2

Population densities of diurnal lemurs in Ranomafana National Park from 2001 to 2009 in three different study sites (Talatakely, Vatoharanana, and Valohoaka) within the park

Full size image

4.3 Seasonal Breeding>

Seasonal breeding is one female strategy to limit extra-pair copulation>by males (Wright 1999). In the first decade, Ranomafana researchers suggested strict seasonality in lemur breeding (Wright 1999). Long-term results revealed that in P. edwardsi, P. simus, V. variegata,and E. rufifronsall females in a group and usually all those in the [study] population come into estrus within the same week for a day or two, with older females breeding first. In contrast, Tecot (2010) found that reproduction in E. rubriventeris not strictly tied to photoperiod, with births occurring in eight different months. However, only infants born in the “seasonal breeding” window (with food abundance prebreeding) actually survived.

Another selective advantage to infants born at the same time is that synchronous births could be a successful strategy to satiate predators (Boinski 1987). Varecia variegataare strict seasonal breeders (Baden 2011) and use a boom or bust strategy (Ratsimbazafy 2006; Baden 2011). At Mangevo, a population of over 80 individuals, V. variegatareproduction was only observed once in 6 years of continuous observation. In 2007, 6 out of 7 adult females in the Mangevo population gave birth to twins or triplets (Baden 2011). This synchronous breeding at 6–8 year intervals has been observed in Pongo pygmaeusand was associated with fruit masting (Knott 1998). Preliminary analysis of plant phenological data from Ranomafana suggests that fruit availability is also the driver of V. variegatasynchronous breeding>(Wright unpublished data).

M. rufushas a breeding season from mid-October to mid-November, during which females have estrous periods of 5 days or more (Blanco 2008; Blanco and Rahalinarivo 2010), although some variation in the timing of breeding exists. Most females are gestating by mid- to late November (Atsalis 2008), although some are not gestating until December (Blanco 2008). The onset of estrus thus may not be completely controlled by photoperiod (Perret and Aujard 2001). There is no evidence for more than one litter a year (Atsalis 2008; Blanco 2008; Blanco and Rahalinarivo 2010).

4.4 Habitat Disturbance

To understand the effects of habitat disturbance>on lemur demography, we compared life histories of lemurs at sites within the park with high and low levels of human disturbance. The high level disturbance site (Talatakely) was selectively logged for valuable hardwoods between 1986 and 1989, with the intrusion of invasive plant (Chinese guava) and animal (Rattus rattus) species (Arrigo-Nelson 2006; Brown and Gurevitch 2004; Laakkonen et al. 2003). The intermediate disturbance site (Vatoharanana) had less than 1% of the study area trees removed during the 1986–1989 period (approximately 1,000 trees; Balko 1998). The low level site (Valohoaka) is considered sacred by local villages and no known timber extraction has occurred (Herrera et al. 2009; Balko 1998). In contrast to the forests north of the park boundaries, there has been very little hunting within Ranomafana forest since the park was established (Lehman et al. 2006; Golden 2009).

The strictest frugivore, V. variegata, did not occur in the high disturbance level site (Talatakely) where the big canopied fruit trees were removed (Balko and Underwood 2005). Compared to forests in South America and Africa, Madagascar rainforest tree growth is slow (Struhsaker 1997; Ganzhorn et al. 1999; Clark and Clark 2010), and this delayed regeneration of the forest may impact lemur demography>for decades. In the 20 years since selective logging, no one group of V. variegatahas since arrived in the high disturbance site. Similarly, 10 years postselective logging, the effects on the physical structure of the forest, its species composition, and availability of P. edwardsifoods have continued at the highly disturbed site (Arrigo-Nelson 2006). Sifakas in the disturbed forest consumed foods from tree taxa in disproportion to their abundance and relied more on vines and epiphytes than counterparts at the unlogged site. As a consequence, intake of fats and sugars was lower for sifakas at the previously logged site. These differences in food availability>and nutrient intake>are reflected in significant differences in the body weights of female sifakas between sites and significant male–female differences in body mass in previously logged forest. Sifakas in the disturbed forest spent significantly less time interacting socially and significantly more time feeding and self-grooming than animals in the undisturbed forest. That all of these differences were greatest during lean season only makes their potential impact on the sifaka population of greater consequence. In the long term, these differences may lead to differences in group cohesion, survival, and/or reproductive success (Arrigo-Nelson 2006).

To investigate the impact of habitat disturbance on E. rubriventerdemography, Tecot (2008, in press) collected data simultaneously on the red-bellied lemur populations in heavily logged and minimally logged sites. Results again indicate that logging has reduced the structure, species composition, abundance, and predictability of red-bellied lemur foods within the disturbed forest site. Additionally, red-bellied lemurs at the disturbed site were less active (Tecot 2008), bred out of peak season, and had higher infant mortality (Tecot 2010, in press).

High levels of cortisol>have been implicated in reduced fitness (Bonier et al. 2009), and the effects of stress>on lemur demography is being investigated (Tecot 2008). In a comparative study of stress hormones in adult E. rubriventerin selectively logged versus minimally logged sites, patterns of cortisol excretion were similar in both sites, but those in the undisturbed site showed little response to variation in food availability and rainfall. In contrast, at the disturbed site, fecal cortisol levels were significantly higher when fruit was scarce (parturition and early lactation) compared with when fruit was abundant (prebreeding season). Contrary to the Cortisol-Fitness Hypothesis (Bonier et al. 2009), cortisol levels were higher in the undisturbed site. Lower cortisol levels, minimal changes in hormones and behavior (Tecot 2008), and higher infant mortality (Tecot 2010) in the disturbed site indicate that there may be down-regulation of the cortisol stress response where environmental stress is prolonged (Tecot 2008, in press).

4.5 Trends in Population Densities>of Lemurs

Over the past 20 years, we have documented population size changes across three sites with different histories of anthropogenic disturbance (Table 4.5; Fig. 4.2). These data allow us to determine changes, rather than trends, in population size over time. Our results indicate that changes in population size vary across sites, as would be expected of meta-populations in a varying landscape (Hanski and Gilpin 1991; Table 4.5). Oscillating population densities of E. rufifronsand E. rubriventersuggest that these changes in population density may reflect some degree of congeneric competition (Overdorff et al. 1999; Erhart and Overdorff 2008a). In comparison, in Vatoharanana, the density of E. rubriventerand V. variegataincreased over the study years from 15 to 25 individuals per km2and from 2 to 10 individuals per km2, respectively, while the density of P. edwardsiremained at 11 individuals per km2. Similar trends have been observed in the critically endangered V. variegata. In the low disturbance site, their population density has oscillated but increased slightly overall. Population density in the intermediate disturbance site increased gradually between 1990 and 2003, but seems to have declined again by 2008. Overall population densities are low (2–10 ind/km2). In P. edwardsiand V. variegata, we have seen an increase in population densities at Valohoaka, but population densities appear to be declining at the intermediate and high disturbance sites, which can perhaps be partly attributed to predation events or fruit scarcity (Arrigo-Nelson 2006; Irwin et al. 2009). Long-term data show that the species with declining populations in normal circumstances have larger home ranges and larger group sizes (Morriss et al. 2009). Theoretically this suggests that food constraints are more important than predation, as larger group size provides more eyes and ears for detecting predators (Hamilton 1964; van Schaik 1983; van Schaik et al. 1983).

Table 4.5 Estimated population size of lemurs in Ranomafana National Park
Full size table

5 Conservation>

5.1 Successes and Ongoing Problems

RNP has been designated as a conservation priority (Kremen et al. 2008) and in 2007 was declared a World Heritage Site. We have successfully integrated education, health, and economic assistance programs with biodiversity research and habitat protection goals (Wright 1997; Wright and Andriamihaja 2002; Lovejoy 2006). An evaluation of the educational impact has shown that local people have experienced a change in attitude (Korhonen 2006). Attempts to correlate human impact on lemur populations have shown that human immigration into the park’s peripheral zone is correlated with increased deforestation rates (Brooks et al. 2009). Villages which are the farthest from the road tend to encroach further into the park than do the on-road villages (DeFries et al. 2010). Moreover, the park itself has maintained edges with minimum invasion by exotic plants (Brown et al. 2009). Eco-tourism to visit the park has been a great boon to the local economy. However, the 30,000 tourists do have negative consequences on habitat and breeding birds (Razafimahaimodison 2003). Better management of tourism is in process. Satellite photos suggest that there is minimum forest destruction since 1991, when the park was gazetted, but the recent political instability is worrisome.

5.2 Implications of Climate Change>

Long-term data enable better understanding of the effects of climate change on rainforest ecology and lemur populations. Indeed, lemur observers with long-term research projects were among the first to gain evidence of the effect of climate fluctuations on mammal populations (Gould et al. 1999; Wright 1999, 2006). Using Madagascar climate data and the Centre ValBio long-term rainfall and temperature database, we discovered that dry seasons have become longer, and cyclones more frequent (Wright 2006). In particular, the November temperatures in 2007, 2008, and 2009 were over 30°C, much higher than ever previously recorded, and the gap between minimum and maximum temperatures has increased. But does this have an effect on lemur demographics?

Long-term demographic data show that older P. edwardsimothers lose infants in years with extended hot, dry seasons (King et al. 2005). Further analysis shows that the average fecundity of lemurs was over 65% lower in El Niño years (Dunham et al. 2008). As El Niño years>become more erratic and frequent (Fedorov and Philander 2000) this could lead to more extreme weather and increased impacts on biodiversity. The southern oscillation (ENSO) related to El Niño is known to cause drought (Thomson et al. 2003) and changes in vegetation indices (Ingram and Dawson 2005). Dunham et al. (2010) found that cyclones, ENSO phases, and rainfall levels affected the reproductive rates of P. edwardsi.Overall fecundity (defined as the number of offspring per female per year surviving to 1 year of age) was negatively associated with cyclone presence during gestation, and positively associated with colder ENSO phases during the second 6 months of life and during the period faced by mothers preceding conception. Wet season rainfall and intensity during gestation was negatively related to birth rates, and the number of drought months during lactation was negatively associated with first year survival. Finally, fluctuations in lemur stress hormones show an elevation of cortisol during the dry season (Tecot 2008), which may ultimately impact reproduction. Longer dry seasons in consequent years could impact lemur populations negatively (Wright 2006; Dunham et al. 2008, 2010). The effects of ENSO events on population dynamics has also been seen in many species of New World monkeys (Wiederholt and Post 2010).

6 Summary and Conclusions

Most of the lemur species in Ranomafana were data deficient before we initiated our first studies in 1986. The virtue and vice of long-term research is that it is never complete. Here we have compiled a list of essential findings, the product of long-term research that has many times resulted in the reevaluation of earlier findings. Thus, over the years we established the number of species residing in RNP through the rediscovery of P. simusand the discovery of H. aureusand a potentially new species awaiting description: a high altitude species of Cheirogaleus. We have established demographic changes through time including life history events, mortality, lifespan, and dispersal patterns for P. edwardsi, E. rufifrons, E. rubriventer, V. variegata, H. aureus, H. griseus, P. simus,and to a lesser extent M. rufus. We have documented variability of social organization in each species, and we have described how populations recover after cyclones and droughts.

Furthermore, we have confirmed that lemur population densities vary over time, and that predation by raptors and mammalian carnivores can have a major impact on local lemur populations. We know that many lemur species are important to seed dispersal and thus to forest dynamics. Monitoring and measuring the long-term effects of habitat disturbance on lemur populations, we have evidence that selective logging may negatively impact population densities of E. rufifrons, V. variegata, and P. edwardsi, even a decade after the last logging disturbance. We have also determined that fertility of P. edwardsifemales decreased during ENSO years and infants of older P. edwardsifemales died in years with extended dry months.