Introduction
Hoolock gibbons (genus Hoolock) occur in forested areas of north-east India, Bangladesh, Myanmar and southern China (Harlan, Reference Harlan1934; Groves, Reference Groves and Rumbaugh1972). Geographically the natural range of these apes extends from the Brahmaputra River east to the Salween River (Lwin et al., Reference Lwin, Geissmann, Aung, Aung, Aung and Hla2011). The genus Hoolock comprises two distinct species, the eastern hoolock gibbon Hoolock leuconedys and the western hoolock gibbon Hoolock hoolock, which are separated based on differences in fur coloration (Mootnick & Groves, Reference Mootnick and Groves2005; Geissmann, Reference Geissmann2007). H. leuconedys was known to be distributed east of the Chindwin River to the Salween River in Myanmar and south-west Yunnan Province in China, at 1,067–1,219 m altitude (Groves, Reference Groves1971), until it was observed in Arunachal Pradesh, India, between the Lohit River in the north and the mountains of Dafa Bum in the south (Das et al., Reference Das, Biswas, Bhattacharjee and Mohnot2006). More recently the species has been found in Sadiya Division, the easternmost part of Assam, south of the Dibang–Brahmaputra River system (Chetry & Chetry, Reference Chetry and Chetry2010). The population, distribution, ecology and behaviour of the western hoolock gibbon have been studied in India and Bangladesh by Alfred & Sati (Reference Alfred and Sati1990), Das et al. (Reference Das, Feeroz, Islam, Biswas, Bujarborua and Chetry2003), Mukherjee (Reference Mukherjee1986) and Islam & Feeroz (Reference Islam and Feeroz1992). Preliminary studies have been conducted on the eastern hoolock gibbon to explore its distribution and population status in Arunachal Pradesh and Assam, India (Chetry et al., Reference Chetry, Chetry, Das, Loma and Panor2008, Reference Chetry, Chetry, Ghosh and Singh2010; Chetry & Chetry, Reference Chetry and Chetry2010). Population surveys of the eastern hoolock gibbon have also been conducted in China (Fan et al., Reference Fan, Wen, Sheng, Sen, Can and Tao2011) and Myanmar (Brockelman et al., Reference Brockelman, Naing, Saw, Moe, Linn, Moe, Win, Lappan and Whittaker2009; Walker et al., Reference Walker, Molur, Brockelman, Das, Islam, Geissmann, Fan, Mittermeier, Wallis, Rylands, Ganzhorn, Oates and Williamson2009). The eastern hoolock gibbon is categorized as Vulnerable on the IUCN Red List (Brockelman & Geissmann, Reference Brockelman and Geissmann2008) and listed as a Schedule I species in the Indian Wildlife (Protection) Act, 1972. To inform future conservation and management planning we carried out a survey of the fragmented forest areas of the Lower Dibang Valley district, Arunachal Pradesh, India, to determine the current population and conservation status of H. leuconedys in this region.
Study area
The study was conducted at seven sites outside Mehao Wildlife Sanctuary, in Arunachal Pradesh (Fig. 1): six sites to the south-east and north-west of the Sanctuary (Injuno, Koronu, Delo, Hawaichapori, Horupahar, Iduli) and one site to the north-west (Chidu). These seven study sites lie within two administrative circles in the Lower Dibang Valley district: Koronu Circle and Roing Circle. The study sites are at 145–390 m altitude and are unclassified forests, not included in any management class and under the jurisdiction of the territorial forest officer. The main forest types recorded in the area are low hills and plains semi-evergreen forest, Assam alluvial plains semi-evergreen forest 2B/CIa and sub Himalayan light alluvial evergreen forest 2B/CI/ISI (Champion & Seth, Reference Champion and Seth1968; Kaul & Haridasan, Reference Kaul and Haridasan1987). The study area is occupied by the Idu-Mishmi and Adi peoples, who generally depend on agriculture for their livelihood. They cultivate cash crops such as ginger Zingiber officinale, maize Zea mays, mustard Brassica juncea and rice Oryza sativa.
The distribution of eastern hoolock gibbons Hoolock leuconedys and forest patches in the study area, in the Lower Dibang Valley district, Arunachal Pradesh, India.
Methods
The survey was conducted from September 2010 to June 2011. We estimated the minimum population, based on visual encounters and auditory data, and recorded group size and age-class compositions (Kakati et al., Reference Kakati, Raghavan, Chellam, Qureshi and Chivers2009). We categorized individual gibbons as adult, subadult, juvenile or infant, based on body size and coat colour (Gupta et al., Reference Gupta, Sharma, Dasgupta, Chakraborty and Hazarika2005; Table 1).
Distinguishing characteristics of male and female eastern hoolock gibbons Hoolock leuconedys in different age categories (Gupta et al., Reference Gupta, Sharma, Dasgupta, Chakraborty and Hazarika2005).
We covered a total of 109 km during 33 census walks across the seven study sites (Table 2). We sighted gibbons directly from the transect and after following their calls (Kakati et al., Reference Kakati, Raghavan, Chellam, Qureshi and Chivers2009). We tried to locate all calling groups estimated to be < 500 m from the trail. For direct sightings we also recorded the local vegetation type and the degree of human disturbance of the habitat. We used stratified random sampling of transects to collect tree data within 18,500 × 20 m (1 ha; Sykes & Horrill, Reference Sykes and Horrill1977; Kumar et al., Reference Kumar, Marcot and Saxena2006). We established the transects along existing forest trails and footpaths and recorded the local names and girth of all trees ≥ 30 cm in girth at breast height.
Survey data from seven study sites in fragmented forest areas of Lower Dibang Valley district, Arunachal Pradesh, India (Fig. 1).
*Infant + subadult
We estimated the minimum distance between gibbon groups, using the Distance Matrix tool in QGIS (QGIS, 2012). Using the geographical information system map we calculated the distance from each group to the nearest forest patch to examine the effect of forest degradation on group size. We counted the number of trees in the patches where gibbons were located, and assigned each patch to one of the following categories: single tree, single tree in bamboo patch, 2–5 trees, 5–7 trees, and > 7 trees. We compared the mean group size among the different categories. We analysed the change in forest cover, using satellite images of the study area from 1985 and 2010, with ERDAS Imagine v. 9.2 (Intergraph, Madison, USA). We recorded the feeding and ranging patterns of two gibbon groups, A and B, collecting data on various aspects of food and feeding for each group for 7 consecutive days in May 2011. We recorded observations every 5 minutes, with focal individual sampling (Altmann, Reference Altmann1974). Each group comprised one adult male and one adult female with an infant. We recorded the time spent by each individual on each food plant, and the parts eaten, along with the time spent at different feeding sites (trees, climbers and crops). Plant parts were categorized as young leaves, mature leaves, fruits/figs, flowers and seeds. The groups were habituated to the presence of humans. We followed them from dawn to dusk each day. We focused on each adult in turn, to ensure representation of all members of the group. We used SPSS v. 16.0 (SPSS, Chicago, USA) for statistical analysis, using Friedman and Mann–Whitney U tests to examine the differences in encounter rates and mean group sizes between the seven study sites.
Results
We estimated the total number of groups from the number of visual encounters and the vocal data. A total of 54 groups and three solitary individuals were recorded. The maximum number of groups was recorded at Iduli (17), followed by Delo (9), Horupahar (8), Koronu (7), Chidu (6), Injuno (5) and Hawaichapori (2). Of the 42 groups/individuals encountered directly, 38 groups were recorded in the Koronu Circle and four groups in the Roing Circle, at Chidu.
A total of 116 individuals in 39 groups and including 3 solitary individuals were recorded: 41 (35%) adult males, 38 (33%) adult females, 16 (14%) subadult males, 3 (3%) subadult females and 18 (15%) infants (Table 2).
The mean group size across the seven study sites was 2.89 ± SE 0.11 (range 2–4). For individual sites the highest mean group size was 3.0 (Koronu, Injuno, Hawaichapori and Chidu) and the lowest was 2.75 (Horupahar). There was a statistically significant difference in the group size and encounter rate across the seven sites (Friedman test, χ2 = 10.57, P = 0.005, n = 7). The ratio of immature (infant and subadult) to adult gibbons was the same at all sites. Group size also varied with habitat quality, specifically the number of standing trees (Fig. 2). Larger groups were found in forest patches (3.1) than in single trees (2.3). The distance of the group from the nearest forest patch also influenced the group size; smaller groups (2.5) were found at distances > 800 m (Fig. 3).
Mean group size of eastern hoolock gibbons in stands of different numbers of trees in the study area, in the Lower Dibang Valley district, Arunachal Pradesh, India (Fig. 1). The numbers above the columns are the total numbers of individuals counted.
Mean group size of eastern hoolock gibbons at various distances from the nearest forest patch in the study area, in the Lower Dibang Valley district, Arunachal Pradesh, India (Fig. 1).
The estimated male : female ratio among adults was 1.07 : 1 across the seven study sites. The infant : female ratio was highest in the Injuno area (0.67 : 1; Table 2).
The recording of gibbons in Horupahar and Hawaichapori was significant because both these fragmented forest patches are surrounded by agricultural lands and human habitation, leaving gibbons more vulnerable at these sites. There was a significant correlation between the encounter rate and the size of the study area (Mann-Whitney U, Z = 2.04, P < 0.05, n = 7); furthermore, a higher encounter rate was observed when the mean distance between sites was lower (Fig. 4).
The relationship between the encounter rate and the mean distance between groups of eastern hoolock gibbons in the study area, in the Lower Dibang Valley district, Arunachal Pradesh, India (Fig. 1).
In the 18 transects we counted 289 trees, of 26 species (19 identified to species) and 17 families (Supplementary Table S1). The most common species were Ficus spp., followed by Ailanthus grandis and Bischofia javanica. The least frequently observed species was Bombax ceiba.
Satellite images of the study area for 1985–2010 show that the loss of forest cover was mainly as a result of the expansion of agricultural land, which increased by 87% (147 km2). In the study area a 48% decrease in forest cover was documented during 1985–2010 (Table 3).
Land use/land cover change matrix for the study area during 1985–2010. The numbers indicate the change in area (km2); e.g. 171.62 km2 forested land was converted to cultivated land, and a total of 346.51 km2 of forested land was converted to other land uses.
Based on the total feeding time during a 7-day study period, the diet of the two focal groups of gibbons consisted of fruits and figs (29%), leaves (65%), seeds (2%) and flowers (0.5%). Gibbons showed a preference for young leaves (59%) over mature leaves (6%).
The main threats recorded at the study sites, through general observation and interaction with local people, were habitat destruction and hunting by local people. Habitat destruction was caused mainly by expansion of permanent agricultural practices, shifting cultivation, tea plantation, and construction of a national highway and permanent settlements. The eastern hoolock gibbon is hunted mainly as an alternative source of meat, and its bones, skin and fur are used for decoration and for making bags. Domestic dogs Canis lupus familiaris kept by local people to protect their property and livestock are potential predators of gibbons, particularly of immature animals. Dogs are opportunistic predators, capturing gibbons when they descend to the ground to move from one patch to another in search of food and secure sleeping places.
Discussion
Gibbons are known to occur in the tropical evergreen, tropical wet evergreen, tropical semi-evergreen, tropical moist deciduous and subtropical hill forests of South-east Asia. The geographical range of the eastern hoolock gibbon in India is contiguous with its range in Myanmar through the Chukan pass in Changlang district (Das et al., Reference Das, Biswas, Bhattacharjee and Mohnot2006). The Chindwin River and the mountains on the India–Myanmar border are physical barriers to the distribution of the species. After the first distribution surveys of the eastern hoolock gibbon in India (Das et al., Reference Das, Biswas, Bhattacharjee and Mohnot2006; Chetry et al., Reference Chetry, Chetry, Das, Loma and Panor2008) a population survey was carried out in Mehao Wildlife Sanctuary, in the Lower Dibang Valley district of Arunachal Pradesh (Chetry et al., Reference Chetry, Chetry, Ghosh and Singh2010). Gibbon occurrence in the area is dependent on the presence of the local Adi and Idu-Mishmi tribes; the Adi hunt gibbons, whereas the Idu-Mishmi do not. In this study we recorded gibbons in isolated trees and small patches of forest surrounded by agricultural land. The conversion of almost 50% of forest to agricultural land in the study area has led to canopy discontinuity, which prevents dispersal of the gibbons from one patch to another unless they descend to the ground. It also makes them more easily sighted by hunters and potential predators such as domestic dogs and predatory birds such as the mountain hawk eagle Nisaetus nipalensis. Fragmentation also causes deterioration of the habitat and the elimination of suitable territories (Kakati et al., Reference Kakati, Raghavan, Chellam, Qureshi and Chivers2009) and reduces the availability of suitable mates and the genetic variability of the population (Jiang et al., Reference Jiang, Luo, Zhao, Li and Liu2006). The disappearance of gibbons from small forest fragments has been documented. Alfred & Sati (Reference Alfred and Sati1990) recorded the disappearance of western hoolock gibbons from 168 forest patches (0.14–2.7 km²) in jhum (slash-and-burn agriculture) matrices in the Garo Hills of Meghalaya. This occurred mainly because the jhum cycles (Kushwaha et al., Reference Kushwaha, Ramakrishnan and Tripathi1981) had shortened to < 10 years and gibbon dispersal corridors (secondary forests and old-growth bamboo) were no longer available.
Little information is available on the populations of the eastern hoolock gibbon in India. Other surveys have estimated 51 groups (168 individuals) in Namsai Forest Division, 10 groups (24 individuals) in Koronu Circle, 157 groups (> 88 individuals) in Mehao Wildlife Sanctuary, in Arunachal Pradesh, and 23 groups in Sadiya Forest Division, in Assam (Das et al., Reference Das, Biswas, Bhattacharjee and Mohnot2006; Chetry et al., Reference Chetry, Chetry, Das, Loma and Panor2008, Reference Chetry, Chetry, Ghosh and Singh2010; Chetry & Chetry, Reference Chetry and Chetry2010). Our study has added to the knowledge of the Indian population of the species and highlighted the conservation importance of the fragmented population in the non-protected forest areas of Arunachal Pradesh.
The mean group size varies between habitats and depends on the level of anthropogenic disturbance. The highest mean group size for eastern hoolock gibbon (3.37) was reported in the Namsai Forest Division, Lohit district (Das et al., Reference Das, Biswas, Bhattacharjee and Mohnot2006), followed by 3.14 and 2.4 in Lower Dibang Valley district in Arunachal Pradesh (Chetry et al., Reference Chetry, Chetry, Das, Loma and Panor2008, Reference Chetry, Chetry, Ghosh and Singh2010). The smaller mean group size (2.89) in our study area relative to other parts of the species' range (e.g. 3.9 in China; Brockelman et al., Reference Brockelman, Naing, Saw, Moe, Linn, Moe, Win, Lappan and Whittaker2009; Fan et al., Reference Fan, Wen, Sheng, Sen, Can and Tao2011) may be a result of forest fragmentation and human disturbance. We found the ratio of infants to adult females to be similar in all seven study sites (i.e. breeding rates were similar) but the absence of juveniles indicates that their survival may be affected by fragmentation and lack of food. Female gibbons in forest fragments probably suffer high lactation costs because of an inadequate diet (Kakati et al., Reference Kakati, Raghavan, Chellam, Qureshi and Chivers2009), both in quality and quantity. In such a situation dependent infants are more likely to die when they stop suckling. When resources are scarce females may not survive to breed again (Moir, Reference Moir, Chivers and Langer1994).
Chetry et al. (Reference Chetry, Chetry, Das, Loma and Panor2008) reported forest loss and fragmentation in Lower Dibang Valley district as a result of extensive agricultural practices (e.g. tea, ginger, corn and mustard cultivation). We identified activities such as firewood collection, selective tree felling, and encroachment for permanent settlement as indirect threats that may affect the gibbon population. Almost 50% of the total forest loss has occurred during the last 25 years, restricting the gibbons to small fragments of forest. Single-tree groups are most at risk as they are smaller than groups in forest patches. Although hunting is not widespread in Idu-Mishmi tribal areas, even a low level of hunting can significantly affect the gibbon population (Fan & Jiang, Reference Fan and Jiang2007). Hunting pressure has apparently led to the local extinction of the gibbon in the Adi-dominated area. In a tribal state such as Arunachal Pradesh, where hunting is part of the local culture, it is difficult to implement wildlife laws and protect species listed under the Indian Wildlife (Protection) Act, 1972. In addition to hunting and poaching, attacks by domestic dogs are a threat (Panor, Reference Panor2011). According to local people, domestic dogs sometimes attack gibbons, particularly immature animals that descend to the ground to move between forest patches.
Hoolock gibbons have been characterized as predominantly frugivorous (Chivers, Reference Chivers1974; Tilson, Reference Tilson1979; Alfred & Sati, Reference Alfred and Sati1994), with fruits constituting up to 70% of their diet (Chivers, Reference Chivers, Preuschoft, Chivers, Brockelman and Creel1984; Islam & Feeroz, Reference Islam and Feeroz1992; Ahsan, Reference Ahsan2001; Bartlett, Reference Bartlett, Dohlinow and Fuentes1999). However, our study indicated a more folivorous diet comprising 65% leaves and only 29% fruits and figs. Other studies have also recorded a significant reduction in the fruit content of the diet (Mukherjee, Reference Mukherjee1986; Kakati, Reference Kakati1997). Although it has been noted that folivory in hoolock gibbons increases during the wet (monsoon) season, when fruit tends to be less abundant (Gittins & Tilson, Reference Gittins, Tilson, Preuschoft, Chivers, Brockelman and Creel1984; Alfred & Sati, Reference Alfred and Sati1994), our study was carried out in May, when fruits are normally more readily available than at other times. A study of the species' diet across seasons is needed to assess annual dietary variation.
Fan et al. (Reference Fan, Wen, Sheng, Sen, Can and Tao2011) recommended long-term population monitoring to determine whether hoolock gibbons can disperse between fragmented forest patches. For effective conservation management of gibbons and their habitat ongoing evaluation of their status is necessary, particularly in areas that hold a comparatively large proportion of the total population in India (Struhsaker et al., Reference Struhsaker1975; Wilson & Wilson, Reference Wilson and Wilson1975). We recommend habitat improvement through reforestation, and construction of canopy bridges to connect remnant forest patches. The translocation of gibbons from the Delo area to forested areas in the Mehao Wildlife Sanctuary as a part of a conservation action plan that has been initiated by the Wildlife Trust of India, in collaboration with the Forest Department, Arunachal Pradesh, must focus on saving isolated populations or groups that are at risk of extirpation. However, the foremost requirement for better conservation of the species is increased awareness and involvement of the local communities in addressing the threats of forest cutting and agricultural expansion.
Acknowledgements
We thank the PCCF (Wildlife), Itanagar, for permission to carry out this research. We also thank Dr C.L Sharma, Dr Prabal Sarkar, Mr Raju Barthakur, Dr Jyotishman Deka, and Mr Biranjay Basumatary for their support and guidance, the NRDMS Division (DST) for the financial support for this research, and the anonymous reviewers for their comments.
Biographical sketches
Kuladip Sarma studies the behavioural ecology and conservation biology of the eastern hoolock gibbon in India. Murali Krishna studies the habitat structure and resource quality for the eastern hoolock gibbon in India and has also carried out research on the flying squirrel of Arunachal Pradesh. Awadhesh Kumar is a primatologist and his research focus is on the ecology and conservation biology of primates in the Eastern Himalayan Biodiversity Hotspot.
