Habitat suitability and threat analysis of Greater One-horned Rhinoceros Rhinoceros unicornis Linnaeus , 1758 ( Mammalia : Perissodactyla : Rhinocerotidae ) in Rautahat District , Nepal

The Greater One-horned Rhinoceros Rhinoceros unicornis has been listed as a Vulnerable species on IUCN Red List, Appendix I of CITES, and a protected animal under the National Parks and Wildlife Conservation Act 2029 B.S., 1973. In Nepal, it was found only in Chitwan, Bardia, Shuklaphanta and Parsa national parks, but it has recently been also reported from the forests of Rautahat. The main objectives of the study were to assess habitat suitability and threats for rhinoceros in Rautahat at an elevation range of approximately 300–1,000 m. Remote sensing data and geospatial modeling techniques were used to assess habitat suitability of rhinoceros. Vegetation assessment was carried out for tree, shrubs, and herbs of plot size 10m × 10m, 5m × 5m, 1m × 1m respectively for habitat suitability. Threat analysis was carried out using purposive sampling among local people and their perceptions were collected on the movement of rhinoceros and threats. The integration of nine explanatory variables showed that about 0.06%, 29.18%, 20.45%, and 50.31% of the study area was found to be most suitable, suitable, moderately suitable and unsuitable habitat respectively for rhinoceros. Out of 30 respondents, 37%, 23%, 20%, and 20% identified the main threat to rhinoceros to be unmanaged habitat, poaching, human-wildlife conflict and environmental factors, respectively. This study recommends parts of the Rautahat District to be extended as the habitat of rhinoceros and starting of immediate conservation initiatives in the area.


INTRODUCTION
Of the five remaining extant species of rhinoceros, three live in Asia: the Greater One-horned Rhinoceros Rhinoceros unicornis, Sumatran Rhinoceros Dicerorhinus sumatrensis and Javan Rhinoceros Rhinoceros sondaicus, and two are found in Africa: the White Rhinoceros Ceratotherium simum and Black Rhinoceros Diceros bicornis (Thapa 2016). In Nepal, the Greater One-horned Rhinoceros is found in Chitwan National Park (CNP), Bardia National Park (BNP), Shuklaphanta National Park (ShNP) and Parsa National Park (PNP), and it has recently been reported in the forests of Rautahat District. The Greater One-horned Rhinoceros (Indian Rhino), hereafter "rhinoceros", has been listed as a Vulnerable species on IUCN Red List of Threatened Species (Talukdar et al. 2008) and is listed in Appendix I of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES). Rhinoceros is listed as the protected animal under National Parks and Wildlife Conservation Act 2029 B.S., 1973 by the Government of Nepal.
Rhinoceroses are mostly solitary with the exception of mothers and calves and breeding pairs, although they sometimes gather at bathing areas. They are active mostly at night, early in the morning and in the late afternoon (Laurie 1978). In the middle of hot days they are commonly seen resting in the shade, or mud, wallowing and bathing in lakes, rivers, and pools. A recently published report by WWF Nepal showed that habitat loss and poaching are emerging as major threats to rhino conservation (Rookmaaker et al. 2016). Poachers kill rhinoceros for their horns, which are highly valued and used in Chinese traditional medicine to reduce fever and fear, and as an aphrodisiac (Crawford 1994).
Rautahat District is connected on the west to Bara District, which includes PNP. In the past few years Rhinoceros have frequently visited the area from PNP searching for suitable habitats, and the previous trends showed migration of rhinoceros from CNP towards the east via PNP to Rautahat. CNP is contiguous to PNP in the east and PNP, in turn, has some forest connectivity to Rautahat forests on the eastern side. Rautahat District is unique being outside the protected area and highly populated with diverse ethnic communities. Of the three rhinoceroses found in Rautahat, one was killed recently by poachers (Acharya & Ram 2017). Thus, it became necessary to find out the habitat suitability and threats to the rhinoceros in the study area for proper management.
Habitat suitability modeling for wildlife is currently gaining interest in wildlife conservation and management. To define habitat suitability, multivariate models are applied in combination with remote sensing (RS) and geographic information system (GIS). Remote sensing is an invaluable source of information and GIS is an excellent tool for creating land cover and habitat factor maps required for habitat modeling. Remote sensing has been used to produce land cover maps since the 1970s (Bradley & Fleishman 2008;Adhikari & Schneider 2012;Tripathi et al. 2012).
This study used remote sensing data and GIS technology with field study for analysis of habitat condition to predict suitable habitat for rhinoceros in Rautahat. Habitat suitability models have become wellaccepted tools to understand the habitat attributes of different organisms, evaluating habitat qualities and developing wildlife management and conservation strategies (Verner et al. 1986;Kafley 2008). Habitat models are based on the relationship between animal and environment (Kushwaha et al. 2005). The habitat suitability index (HSI) modeling assumes that the amount of habitat is related to the potential of the land to support individuals or populations of wildlife and that habitat designated as high quality are more suitable than those assigned lower quality ranking. HSI models are analytical tools for determining relative potential of an area to provide habitat for wildlife (Clevenger et al. 2002).
The main objectives of this study were (1) to assess habitat suitability, and (2) to do a threat analysis for rhinoceros in the study area using geospatial datasets on topography, climate, land use and statistical modeling at the landscape scale in Rautahat District.

Study area
The study area is situated in Chandrapur Municipality, Gujara Rural Municipality and Phatuwa Bijayapur Rural Municipality of Rautahat District in the central part of Nepal (Fig. 1). It is located between85.23-85.50 0 E and 26.73-27.23 0 N. Lower tropical zone lies below 300m and covers 64.4% of the total area of Rautahat and upper tropical zone covers 5.6% of area and elevation ranges from 300-1,000 m (District Report 2011). It covers an area of 112,600ha. Forest covered by Rautahat District is 29,400ha or 26.11% of the forest area including the central 'Charkoshe Jhadi' of Nepal. Charkoshe Jhadi is the broad strip of forests south of the Siwaliks from east land cover) related to habitat requirements of rhinoceros were used (Table 1). Remote sensing satellite data were used as a source of information, and spatial analysis of the data was performed in Arc GIS Desktop 10.2.2 to process the data. Weightages that influence the habitat of rhinoceros by these different variables were decided after expert consultation from PNP (Table 1).
A four level suitability was depicted on the map with reference to habitat used by rhinoceros. Areas away from human settlements and close to water bodies were categorized as highly suitable while areas near roads and human settlements but away from water bodies were considered as unsuitable for rhinoceros (Thapa & Lichtenenegger 2005).
Suitable habitat categories included the areas currently being used by rhinoceros and the areas that could be potentially used. Overlay process was carried out to produce suitable area map ( Fig. 4 (a)). Nine suitability maps were prepared based on the explanatory variables ( Fig. 4 (b-j)) used in this study.

Field measurement
The field measurements from a total of 26 plots (10m × 10m) were conducted between May-June 2017 and used in this study for habitat assessment Various quadrats of 10m × 10m were randomly assigned to tree species. Within a quadrat, 5m × 5m quadrats were allocated randomly in the corner for shrub species. Likewise, herbs were recorded from nested sampling of 1m × 1m quadrat within the 5m × 5m quadrat. The distribution of nested sampling within main quadrat (Mandal & Joshi 2014) is shown in Fig. 3.
All plant species within each quadrat were identified and counted. For the entire tree stems, diameters at breast height (DBH) at 1.3m were measured using diameter tape, and height of each stem was measured by a clinometer. A local parataxonomist and field guide identified the tree species. Leaves of unidentified tree species were brought to the faculty of forestry at the Agriculture and Forestry University (AFU) for identification.

Threat analysis
Field visits were undertaken to major places where rhinoceros encounters had been reported, and relevant staff of PNP and district forest office were interviewed.
A questionnaire survey was conducted among 30 respondents in the study area, including protected area managers, experts and community representatives; their  knowledge about rhinoceros and its habitat, threats to rhinoceros in the study area and possible conservation measures were documented. The vegetation data collected in the field were used to calculate the importance value index (IVI), density, frequency, and relative frequency of the tree species by using the following procedure (Smith 1980).
To calculate the prominence value (PV), the percentage cover of each species is assumed in each quadrat recorded in classes as follows: for high coverage = >50%, medium = 26-50 %, low = 0-25 %. These data were used to calculate prominence value for each species (Jnawali 1995) and it is used to calculate the availability of plants in the study area.

Habitat Suitability Mapping
Suitability map based on RS and GIS application showed that only about 0.06% (28.8ha) of the area was found to be most suitable, approximately 29.18% (13198.23ha) of the area was found to be suitable, 20.45% (9248.58ha) was moderately suitable and about 50.31% (22759.65ha) was unsuitable habitat for rhinoceros in the study area ( Fig. 4 (a)).

Threat Analysis
Almost all the respondents were well informed

INTRODUCTION
Information on the presence and distribution of species within a region is crucial for planning and evaluating conservation strategies (Tobler et al. 2008). This is especially true in sites where armed conflict has complicated conservation efforts (Hanson et al. 2009;Daskin & Pringle 2018) and impacted species populations and habitats. There is no general consensus as to whether conflicts have positive impacts on wildlife (through relaxing pressure on wildlife when people avoid combat zones or the decline of extractive industries; Hallagan 1981; Butsic et al. 2015) or negative impacts (through direct killing from the use of ordnance and chemicals or bushmeat hunting by soldiers; Orians & Pfeiffer 1970;de Merode et al. 2007;Beyers et al. 2011). Thus it is critical to assess the effects of conflict on biodiversity.
Manas National Park (MNP), spanning 500km 2 is located in the eastern Himalayan biodiversity hotspot. Falling within two administrative districts (Chirang and Baksa) of the state of Assam that are under the administration of the Bodoland Territorial Council (BTC), this region experienced intense ethno-political conflict in the late 1980s until 2003. During this period the population of Indian Rhinoceros Rhinoceros unicornis was poached out, necessitating a reintroduction program to repopulate the park (Barman et al. 2014). Preliminary studies and anecdotal evidence suggest that the conflict has severely impacted other wildlife species as well (Goswami & Ganesh 2014).
It is noteworthy that 80% of worldwide armed conflicts between 1950 and 2000 overlapped with biodiversity hotspots (Hanson et al. 2009). A more recent analysis from Africa highlights the fact that population trajectories of large mammals fell significantly below replacement levels (i.e., instantaneous rate of increase of population; λ less than 1) with an increase in conflict frequency (Daskin & Pringle 2018). Therefore, documenting species assemblages in the aftermath of conflict is critical to inform subsequent conservation interventions.
In this study we conducted camera trapping surveys across three administrative forest ranges (Panbari, Bansbari and Bhuyanpara) of MNP in 2017 with the aim to (a) document the mammalian species assemblage of the park, and (b) understand the influence of civil conflict on the mammalian assemblage. Given that there is no comparable data on mammal distribution prior to the conflict from the site, it was not possible for us to make direct comparisons of pre and post conflict effects on the mammalian assemblage. Therefore, we evaluated differences in photo capture rates of mammalian prey and large carnivore species between Panbari (a forest range with conflict until 2016) and Bansbari-Bhuyanpara (forest ranges that have been conflict-free since 2003). These two forest sections of MNP differ in their history of conflict but are similar in terrain, climate, vegetation communities, and faunal assemblages. Therefore, we assume our comparisons to serve as a proxy for the effects of conflict.

Study site
MNP, situated in the eastern Himalayan biodiversity hotspot, is also an UNESCO Natural World Heritage Site, a tiger reserve, an elephant reserve and a biosphere reserve. Contiguous with Royal Manas National Park (RMNP) in Bhutan, it is home to several endangered species. Located in the foothills of the Himalaya, MNP is predominantly flat, with the mountainous regions primarily falling within RMNP, Bhutan. The vegetation of MNP can be broadly classified into eastern wet alluvial grasslands, moist deciduous, and semi-evergreen forests (Champion & Seth 1968).
Spread over Kokrajhar, Chirang, Baksa and Udalguri districts of the Bodoland Territorial Areas Districts (BTAD) of Assam, much of the forests of the Manas Tiger Reserve (including core area of MNP) experienced large scale deforestation (i.e., conversion of forests to farmland and settlements) during the conflict period leading to the loss of over 40% of primary habitats (Sarma et al. 2008;Lahkar et al. 2012). While political stability was initiated in 2003 with the formation of the BTAD, since 2004, there have been several incidents of ethnic conflict in the region emphasizing the fragile socio-political environment around this site (Web data source: South Asia Terrorism Portal, Satp.org).
The forest ranges of Bansbari and Bhuyanpara have largely remained conflict free since 2003. Occasional conflict in Panbari until 2016 has resulted in our inability to conduct surveys within the forest range. Although we, in collaboration with the park management, have been carrying out long-term biological monitoring using camera traps since 2010 across Bansbari and Bhuyanpara, it was only in 2017 that surveys could be undertaken simultaneously across all three ranges of MNP (Panbari, Bansbari and Bhuyanpara).

Field and analytical methods
We conducted a camera trapping survey in the winter of 2016-17 from 28 December 2016 to 24 February 2017 covering the three ranges of Panbari, Bansbari and Bhuyanpara. We used 4km 2 grids to guide camera placement. Cameras were operational for 24 hours a day. We used Panthera (New York, USA) V4 & V5 digital white flash passive camera traps mounted on trees, on poles in steel cages customised specifically for the cameras to minimise the damage from wild animals. In total, camera traps were placed at 118 locations (26 in Panbari and 92 in Bansbari-Bhuyanpara; Fig. 1).
We first downloaded photographs from all the trap stations across the park at regular intervals (usually twice a week) and catalogued all captures using Camera Trap File Manager software (Olliff et al. 2014). During the cataloguing process species identity was confirmed based on expert knowledge. We also referred to Menon (2014) to confirm species identity.
The camera traps were operational for 24 hours a day and each day was counted as a trap-day. The trapping effort at different trap locations diferred due to time and days a camera trap was active. On average camera traps were operational for 52.3 trap-days. To calculate the photo-capture rate index (PCRI) of all species captured we first identified independent captures (i.e., captures that were 30-minutes apart for each station). We then divided the number of independent captures obtained at each trap by trap-specific effort (i.e., number of trapdays that a particular trap was active) and expressed the estimate per 100 trap-days (Carbone et al. 2001). Trap specific PCRI were then used to map the spatial variation in capture rates. All maps were created in the open source software QGIS (QGIS Development Team 2012). To assess the difference in PCRI of mammalian prey and large carnivores between Panbari and Bansbari-Bhuyanpara, we summarized species-specific PCRI and tested for differences using a two sample T-test assuming unequal variances. Given that we were conducting a series of significance tests on the same set of data, we set the false discovery rate to 10% and used Benjamini-Hochberg procedure (Benjamini & Hochberg 1995).

RESULTS
Camera trapping effort totaled 6,173 trap-days in 2016-17 spread across MNP. We obtained 21,926 photographs of mammals from which we identified 25 mammal species belonging to 13 families (Appendix 2). Of these, six species are Endangered and seven are Vulnerable as per the IUCN Red List of Threatened Species (Table 1; IUCN 2017).
In addition to 2016-17, using the data from long term monitoring study in MNP since 2010, we observed presence of number of other species which included Spotted Deer Axis axis (confirmed its eastern range limit in Panbari; Least Concern), Chinese Pangolin Manis pentadactyla (Critically Endangered), Marbled Cat Pardofelis marmorata (Near Threatened), Golden Jackal Canis aureus (Least Concern), and Painted Bat Kerivoula picta (Least Concern).
For mammalian prey and large carnivore species we mapped the spatial variation in photo capture rates across the Park (Figs. 2 & 3). In addition, we assessed the variation in capture rates between Panbari and Bansbari-Bhuyanpara (Figs. 4 & 5). In general our results indicated lower capture rates of mammalian prey species in Panbari as opposed to Bansbari-Bhuyanpara, while for four large carnivore species photo capture rates were higher in Panbari compared to Bansbari-Bhuyanpara. Significant differences in capture rates using a two sample T-test assuming unequal variances were, however, noticed only among four mammalian prey (Barking Deer, Sambar, Gaur and Buffalo) and one large carnivore (Wild Dog) (Figs. 4 & 5) (Appendix 1).

DISCUSSION
Our surveys confirm the presence of 25 mammalian species photo-captured in MNP, 13 of which are     It is observed that ethno-political conflict likely has some impacts on abundance and distribution of species and habitats. While the mammalian species assemblage in MNP appears to be intact, we detect differences among photo capture rates of several species between Panbari (a forest range with conflict until 2016) and Bansbari-Bhuyanpara (forest ranges that have been conflict-free since 2003). In general, prey capture rates were higher in Bansbari-Bhuyanpara compared to Panbari, and significant differences were noticed for four mammalian prey species (i.e., Wild Buffalo, Gaur, Sambar and Barking Deer; Fig. 4). Three of these (Wild Buffalo, Gaur and Sambar; over 175kg) are large prey species that are all threatened and particularly vulnerable to poaching (Wolf & Ripple 2016;IUCN 2017). In the case of large mammalian carnivores, however, species capture rates were higher in Panbari compared to Bansbari-Bhuyanpara, although significant differences were noticed only for Wild Dogs (Fig. 5). While it is possible that Panbari acted as a refuge for large carnivores as villagers may have avoided the combat zone, it is also possible that disturbances emanating from the conflict could have depressed large prey populations. Disturbances, however, were more of armed militants camping deep inside the Panbari range two to three years preceding this survey, rather than ethnic conflict as such or severe anthropogenic disturbances due to natural resource collection. Thus, the disturbances within the park during that period were mostly related to hunting (potentially ungulate species) for food by those camping inside as well as subsequent sanitization operations by government forces.
From our study it appears that RMNP in Bhutan situated immediately north of MNP, next to Panbari, likely acted as a refuge, particularly for long ranging carnivore species. This is evidenced by the fact that in 2017 our camera trapping data confirmed presence of eight individual tigers (five males and three females) in Panbari range of which three individuals were captured the previous year (2016) in RMNP (Singye Wangmo pers. comm. 22 January 2018). This also indicates that the large carnivores have taken the advantage of the progressively re-established security in the area and rapidly moved there. The animals probably began using that area as well but did not relocate there -perhaps their ranges are wide enough to use portions of both areas. This may, however, also negatively impact the herbivore population that are still recovering and thus, may take longer to re-establish themselves.
Ideally, long-term data on population trajectories are required to uncover the effects of conflict-related disturbance on populations.
MNP offers us the opportunity to compare capture rates of wildlife species across two study blocks that primarily differ in their history of ethno-political conflict. The contiguity within TraMCA (Trans-boundary Manas Conservation Area) certainly has a positive effect contributing to the repopulation of large carnivores in the aftermath of the conflict as RMNP has acted as a refuge for the animals displaced by disturbances in MNP. Ahmed et al. (2015) have highlighted the trans-boundary importance of the TraMCA based on data obtained through synchronized camera trapping exercises across the boundary. The present study further highlights the importance of large and contiguous conservation areas for the conservation of biodiversity.
Our study found camera trapping to be an effective method to document particularly rare and elusive mammalian species and their relative abundance across the park. Photographic capture-recapture methods could help assess the population trajectories of individually identifiable species such as tigers, leopards, clouded leopards and leopard cats. Additionally, the baselines we set through this study could be used to monitor future changes in the capture rates of several species, especially those which are not individually identifiable (e.g., Wild Dogs and Jungle Cats).
In conclusion, we present evidence that ethnopolitical conflict has likely influenced the spatial variation of several species in Manas National Park. It is critical, however, to note that more detailed studies assessing mammalian prey densities, distribution and density of large carnivores and correlation with specific 12015 Appendix 1. factors emanating from conflict are required to further understand the effects of conflict and peacetime conservation efforts on the species assemblage and abundances.

INTRODUCTION
The taxonomic status of Hipposideros pomona Andersen, 1918 sensu stricto has been confusing since its establishment. Knud Andersen's study on species allied to Hipposideros bicolor resulted in the description of two new species: H. pomona from southern India, and H. gentilis from northeastern India, Burma to west coast of Sumatra (Andersen 1918;Hill 1963;Hill et al. 1986;Douangboubpha et al. 2010) (Fig. 1). Andersen (1918) based his diagnosis of H. pomona in possessing broader than usual noseleaf, horseshoe and 'sella' (probably referring to posterior leaf) measuring 5.8mm and 5.2mm respectively, and forearm measuring 40.5 mm based on a single specimen from Coorg, southern India; while that of H. gentilis in possessing not broader than usual noseleaf, horseshoe and 'sella' measuring 4.5-5.5 mm and 3.7-4.8 mm, and forearm measuring 38.5-41.5 mm based on specimens from Masuri, Pegu (=Bago), Burma (=Myanmar). In addition to the nominate gentilis, Anderson (1918) described three other subspecies of H. gentilis-sinensis, atrox, and major. Hill (1963) included pomona, gentilis, sinensis, atrox and major as subspecies of H. bicolor. Later, Hill et al. (1986) while revising H. bicolor considered H. pomona as distinct species including nominate form, gentilis, and sinensis, and assigned atrox and major to H. bicolor. A trend followed by subsequent authors (Yenbutra & Felten 1986;Zubaid & Davison 1987;Corbet & Hill 1992;Simmons 2005 (Fig. 1).
Until now, there existed no information about the taxon H. pomona sensu stricto from southern India, as the type specimen (skin) was not traceable and the damaged skull (BMNH No. 18.8.3.4) in the British Museum of Natural History was the only material (Hill et al. 1986). During a museum study conducted by the authors, the type specimen in the type collection of the Zoological Survey of India, Kolkata was studied. Three more specimens from southern India (bearing ZSI No. 21535, and 7193 and 7196 collected from Travancore and Shevaroy Hills respectively) were found in the collection of ZSI, Kolkata by the first author. These specimens were originally labelled H. fulvus fulvus and have been catalogued as H. fulvus (Ghosh 2008). This species has been collected from southern India in the recent past, and the specimens are in personal collections of bat researchers in Tamil Nadu which have not been taxonomically studied. This recent discovery of additional specimens in National Zoological Collection of Zoological Survey of India provided us with an opportunity to conduct detailed study and compare this taxon with gentilis from northeastern India, Andaman Islands, and Myanmar.
Through this communication, we present evidence based on morphometric, bacular, and acoustic characters that prompt distinct specific status of the populations assigned from southern India as H. pomona, and northeastern India and Southeast Asia as H. gentilis. We also provide a detailed description of H. pomona as the original description of the taxon lacks detail.

MATERIALS AND METHODS
The present study is based on museum specimens in the National Zoological Collection, Zoological Survey of India, Kolkata and the Natural History Museum, Department of Zoology, Osmania University, Hyderabad, Telangana State. We examined a total of 10 specimens of H. pomona sensu lato (four vouchers including the type specimen (skin only) of H. pomona sensu stricto, and six vouchers including one voucher specimen of H. gentilis sensu stricto from Sagaing, Myanmar about 560km north of the type locality of gentilis). External and craniodental measurements were taken (to the nearest 0.1mm) using a digital vernier caliper (Mitutoyo make) following Bates & Harrison (1997) and Srinivasulu et al. (2010). Measurements of the museum specimens were compared with published data (from Douangboubpha et al. 2010;Bates & Harrison 1997). External measurements included FAforearm length, E -ear length, Tl -tail length, Tib -tibia length, Hf -hindfoot length, 3mt -third metacarpal, 4mt -fourth metacarpal, 5mt -fifth metacarpal, 1ph3mt -first phalanx of third metacarpal, 2ph3mt -second phalanx of third metacarpal, 1ph4mt -first phalanx of fourth metacarpal, 2ph4mt -second phalanx of fourth metacarpal. Craniodental measurements included GTL -greatest length of the skull, CBL -condylobasal length, CCL -condylocanine length, CM 3 -maxillary toothrow, C 1 -C 1 -anterior palatal width, M 3 -M 3 -posterior palatal width, ZB -zygomatic breadth, BB -braincase breadth,  Topal (1958), and later stored in glycerol. The information on the echolocation calls were sourced from published literature and our studies in Andaman Islands, India.

Materials examined
Hipposideros Abbreviations used: BMNH -British Museum (Natural History); ZSI -Zoological Survey of India; NHMOU -Natural History Museum of Osmania University.

RESULTS
External measurements show that pomona is smaller than gentilis  with shorter ears (16.76-17.03 mm vs. 17.5-24.0 mm). In both the taxa the third metacarpal was the shortest, being shorter than both the fourth and the fifth metacarpals, with fourth being the longest of the three. Similarly, the first phalanx of the third metacarpal was longer than the second and was almost equal to the combined lengths of the first and the second phalanges of the fourth metacarpal (Table 1).
The cranial measurements show that the condylocanine length of pomona being smaller than that of gentilis (14.66mm vs. 14.6-16.3 mm). The zygoma are more widely placed in gentilis than in pomona. The anterior palatal width and the posterior palatal width are also greater in gentilis than in pomona (5.7-6.6 mm vs. 5.43-5.8 mm) signifying that the rostrum of gentilis is broad and long. The mandible measures shorter in pomona in comparison with gentilis (10.05-10.1 mm vs. 10.4-11.8 mm) ( Table 2). The sagittal crest in gentilis is well developed, and extends up to the parietal region of the skull, whereas in pomona it is weakly developed. The lambdoid crests are also well developed in gentilis. The first lower premolar (pm 2 ) is half to two third the height of the second lower premolar (pm 4 ). The pm 4 is one third the height of the tall canine. The lower molars are equal to or shorter than the pm 4 . The coronoid process and the condyle of the mandible are well developed (Image 1A-D). The structure of the noseleaf and the internarial septum in pomona varies from that of gentilis. The noseleaf of gentilis (Image 2C&D) is wider than long with the internarial septum being slightly narrow at the base, becoming parallel sided, and gradually narrows to a blunt tip (looking like a tapered triangle) almost touching the intermediate leaf. The internarial septum is separated from the walls of the anterior leaf by means of deep grooves due to which the narial lappets seem to be located away from the internarial septum. The nares are tear drop shaped and the well developed narial lappets are attached to the sides of the internal walls of the anterior leaf. The posterior and the anterior  (Srinivasulu et al. 2017).
Basing on the distinctness of the structure of the baculum of the southern Indian specimen from that of Assam and Southeast Asia (present study;Douangboubpha et al. 2010), and on differences in morphometrics of the southern Indian specimens from that of gentilis, the peninsular Indian population is considered distinct here. The rest of the specimens from Nepal, northeastern India, Andaman Islands to Myanmar, southern China, Lao PDR, Thailand, Viet Nam, Cambodia, and Western Malaysia (Corbet & Hill 1992;Bates & Harrison 1997;Simmons 2005;Francis 2008;Douangboubpha et al. 2010) are here considered as H. gentilis Andersen, 1918. Owing to the disjunct distribution of the two taxa ( Fig. 1), and considerable variations in bacular, morphological, and acoustic characteristics the resurrection of H. gentilis as a distinct species is well-supported.
Molecular phylogenetic studies to support the distinctness of Hipposideros pomona sensu stricto and H. gentilis sensu lato are in progress.

Diagnostic characters
A small to medium sized bat, with an average forearm length of 39.6mm (38.1-40.84 mm). The noseleaf is longer than broad, and covers the muzzle. There are no supplementary leaflets. The internarial septum is thick, parallel sided, becomes broader as it is nearing the proximal end and slightly narrows to a broadly rounded tip. The skull is slender and has an average condylocanine length of 14.4mm (14.2-14.66 mm). The first lower premolar (pm 2 ) is small, triangular in outline, and is one half of the height of the second lower premolar (pm 4 ). The pm 4 is about one half to two third the height of the tall canine. The baculum is long with a straight shaft, the proximal end of which is broad   4 (10.5-11.6) and bifid.

Descriptive characters
A small to medium sized bat with the forearm length ranging between 38.1-40.84 mm. Ears are large (16.76-19.0 mm) with ridges and rounded off tips. Feet are large (6.28-6.97 mm). The wing membrane and the interfemoral membrane are attached on either side of the ankles. In preserved specimens, the ventral fur is pale in colour; the fur along the flanks and toward the distal part of the body is pale at the base and light brown on the mid-portion, with pale tips; face has pale brown to darkish brown fur; while the dorsal surface is fawn to dark brown with pale hair bases. On the wing the third metacarpal is shorter than the fourth and the fifth metacarpal, with the fourth being the longest. The first phalanx of the third metacarpal is almost equal to the combined lengths of the first and second phalanges of the fourth metacarpal. The second phalanx of the third metacarpal is shorter than the first. The noseleaf is longer than broad and covers the muzzle (Image 2A&B). There are no supplementary leaflets. The internarial septum is thick, parallel sided and becomes broader as it is nearing the proximal end, and slightly narrows to a broadly rounded tip (Image 2A&B). The oval shaped nares are situated on either side of the tip of the internarial septum. The narial lappets are fleshy and are located on the sides of the nares attached to the internal walls of the anterior leaf. There is not much of a gap between the internarial septum, the narial lappets and the internal walls of the noseleaf. The intermediate leaf is short and small, shorter than the posterior and the anterior leaves (Image 2A&B). Due to its small size the gap between the internarial septum and the intermediate leaf is more. The posterior leaf is slightly wider than the anterior leaf and is divided into four cells by means of three septa (Image 2A&B). Penis is thin, long with a pointed tip. The baculum is long (1.367mm) with a long straight shaft, the proximal end of which is broad and bifid (Image 3A). The distal end is four pronged, and shows an inner curvature on the ventral surface (Image 3A).
The skull is slender (Image 1A) and has an average condylo-canine length of 14.4 mm (14.2-14.66 mm). The rostrum is short with the CM 3 measuring 5. 4-5.7mm. The palate is narrow as is seen by the posterior palatal width being 5.3-5.6 mm and the anterior palatal width being 3.15-3.4 mm. The zygoma are slender toward the anterior portion, with a low jugal process on the posterior portion (Image 1A). The sagittal crest is weakly developed, and is visible on the anterior part of the braincase and not present on the posterior part. The braincase is bulbous (8.84mm) and shows a slight depression over the parietal region in the lateral profile. The lambdoid crests are not well developed. The rostrum is bulbous with three nasal inflations, an elongated one located at the bottom, a kidney shaped inflation located on either side, and a round inflation on the top of the rostrum. The upper toothrow averages 5.5mm (5.4-5.7 mm). The first upper premolar (pm 2 ) is minute located between the cingulum of the robust canine and the anterior border of the second upper premolar (pm 4 ) whereby the canine and the pm 4 are not in contact. The first lower premolar (pm 2 ) is small, triangular in outline and measures one half of the height of the second lower premolar (pm 4 ). The pm 4 is about one-half to two-third the height of the tall canine. The first lower molar is subequal to slightly taller than the pm 4 while the second lower molar is equal to the height of pm 4 . The average length of the mandible is 9.9 mm (9.8-10.1 mm) while the lower toothrow (cm 3 ) measures 5.58-6.0 mm. The coronoid process and the condyle of the mandible are well developed.

Echolocation calls
The echolocation calls of H. pomona from southern India were recorded by Wordley et al. (2014) from Valparai Plateau, Tamil Nadu where in the average frequency at maximum energy (FMAXE) of the echolocation calls produced by this species was 126.337±1.25 (range: 123.7-128.2 kHz).

Ecology
All that is known about the ecology of H. pomona is

Conservation status
H. pomona sensu lato, according to Bates et al. (2008) included both the taxa pomona and gentilis, was assessed as Least Concern. In the light of new information and upgradation of these taxa as distinct species, with H. pomona being known south of Coorg in Western Ghats and H. gentilis being present from Nepal, northeastern India through Myanmar to Southeast Asia, there is a need to reassess these two species.

Remarks on misidentified specimens
A specimen (ZSI No. 19450) from Macherla, Andhra Pradesh labelled as H. pomona pomona, collected by B. Nath in 1962 (Bates & Harrison 1997;Ghosh 2008) is hereby identified as H. fulvus basing on the structure of internarial septum and morphometrics. Three specimens collected by Indraneil Das from Andaman & Nicobar Islands and labelled as H. larvatus leptophyllus (ZSI Reg. No. 24773-75) are hereby identified as H. gentilis based on noseleaf structure and morphometrics.

INTRODUCTION
Amphibians are important predator and prey species in both aquatic and terrestrial habitats, especially in the tropics where the diversity and abundance of taxa are high. According to Whiles et al. (2006) loss of one species is akin to loss of two species in the case of amphibians. Baillie et al. (2004) stated that among the vertebrates of the world, amphibians are the most threatened taxa and have the highest proportion of species on the verge of extinction. The most pervasive threats to amphibians are habitat loss and habitat degradation.
For the amphibians of the Western Ghats the species accumulation curve has not yet reached a plateau (Aravind et al. 2004). According to Nameer et al. (2015) & Das (2015) 90% of amphibians in Kerala are endemic to the Western Ghats and 33% belong to various threatened categories.
Generally, protected area networks are considered as the corner stone of biodiversity conservation efforts. Perfecto & Vandermeer (2008) commented that conservation strategies which focus on tropical forests while ignoring the multiple land uses in which they are embedded are failed strategies. According to Nair (2008), agroforests can be considered as potential oases for disappearing species, even though they cannot substitute for natural forests. Wanger et al. (2009) assessed the herpetological diversity in cacao agroforests of Southeast Asia, where they observed that certain habitat features like increase in leaf litter favour the richness and abundance of disturbance-tolerant species. Bionda et al. (2011) worked on the amphibians of various agroecosystems in Argentina and found that there are species which take advantage of the hydrology and hydroperiod of agroecosystems for their survival. Not all agroecosystems contribute equally towards conservation.
Hence, for a realistic conservation strategy one should evaluate the conservation value of these multiple land use systems such as agroecosystems. The present study is expected to shed light on the amphibian diversity and richness in agroecosystems of Kerala.

STUDY AREA
The study was conducted in the selected agroecosystems in Thrissur District, southern Western Ghats, . The chosen agroecosystems include, cashew plantation, coconut plantation, homegarden and rubber plantation (Fig. 1). The study area chosen mostly comes within the main campus of the Kerala Agricultural University, in Kerala, southern India. The campus has a total area of 391.44ha and is located very close to the Peechi-Vazhani Wildlife Sanctuary. The major habitats include gardens, botanical garden, plantations of rubber, coconut, plantain, cacao, and orchards of mango, jackfruit, sapota and guava. The whole area must have been under the forests about one and a half century or so and was subsequently converted mostly into rubber plantations. Later the land was handed over to the Kerala Agricultural University (KAU) in 1971, and the KAU had developed these areas into the different land uses as is explained above. The 14-year mean minimum temperature is 23.3 0 C and 10-year mean maximum is 31.9 0 C. The area receives south-west and north-east monsoons, the greater portion of the rainfall, however is received from the south-west monsoon between June and September. The mean annual rainfall is 2,803.4mm. The mean number of rainy days per year is 112 days (KAU Weather Station 2014).
The study was conducted in four selected agroecosystems such as cashew plantation which is 16 years old and spreads over an area of 90 acres, a coconut plantation, which is about 35 years old and has an extent of 50 acres, a homegarden covering eight acres about 40 years old, and a rubber plantation with an extent of 60 acres and is about 67 years old.

METHODS
The study was conducted from January 2017 to May 2017. Quadrat sampling was the primary method adopted for the study and was supplemented with visual encounter survey.

Quadrat Sampling
One hundred quadrats each having a size of 10m × 10m were deployed in the study area randomly with a minimum distance of 10m between two quadrats. The observations were taken between 19:00hr and 20:30hr. Two observers surveyed the quadrat from opposite corners approaching the center in a clockwise manner for thorough search of the ground-dwelling amphibians (Harikrishnan et al. 2012).

Visual Encounter Survey
Visual encounter survey was conducted in the study locations (Harikrishnan et al. 2012). This was done between 20:30-21:30 hr. LED torches and head lamps were used to spot the amphibians. During the survey, a range of possible microhabitats where the amphibians could be seen, such as rocks, marshes, fallen logs, tree holes, snags and water bodies were thoroughly examined.

Amphibian diversity of selected agroecosystems
A total of 10 anurans belonging to five families were recorded from the selected agroecosystems of central Kerala (Table 1). This includes two species each from the families Microhylidae, Ranidae and Rhacophoridae, three species from Dicroglossidae and one species from Bufonidae. Dicroglossidae represented 30% of the amphibians encountered during the present study, however, these are commensal and generalist species. These include Euphlyctis cyanophlyctis, Hoplobatrachus crassus and Minervarya keralensis. According to Gururaja et al. (2007), anthropogenic changes in land use type, canopy cover and hydrological regimes support the presence of more generalist amphibian species. Three species that are endemic to Western Ghats namely Minervarya keralensis, Pseudophilautus wynaadensis and Indosylvirana urbis were reported from the agroecosystems that were studied (Biju et al. 2004(Biju et al. , 2014Biju & Bossuyt 2009;Gururaja 2012;Das 2015;Sanchez et al. 2018). Pseudophilautus wynaadensis, an Endangered species (Biju et al. 2004) was found to be present in all the agroecosystems surveyed during the present study (Table 2). Rathod & Rathod (2013) reported highest diversity of amphibians from organic coffee plantations of Kodagu District where diversity of native rainforest trees was also high. Chemical contamination by the use of fertilizers and pesticides in the agricultural fields lead to incidence of abnormalities among common frog species as their breeding period coincides with the time of agrochemical application in the fields (Gurushankara et al. 2007;Kittusamy et al. 2014;Krishnamurthy et al. 2008). No abnormalities, however, were observed among the 569 individuals of amphibians recorded during the current study.
Quadrat sampling recorded only seven species from the study location. Though visual encounter survey also recorded seven species, it recorded three species that were not recorded through quadrat sampling (Table 1). It is potentially possible to obtain the complete species inventory of the sampled area when visual encounter survey is combined with another sampling technique (Eekhout 2010).

12031
(Image 1) was encountered from cashew and rubber plantations. It was also sighted in different colour morphs (Images 1A-D) and all individuals were adults.

Euphlyctis cyanophlyctis (Schneider, 1799)
Skittering Frog (Image 2) was sighted only at the pond in the rubber plantation. In this case also only adult individuals were sighted.

Microhyla rubra (Jerdon, 1854)
Reddish Narrow-mouthed Frog (Image 5) was sighted from the premises of the pond in the rubber plantation during visual encounter survey. Interestingly, all the individuals recorded were juveniles of very small size.

Hydrophylax malabaricus (Tschudi, 1838)
Fungoid frog (Image 6) was detected only from the rubber plantation. It was found in the moist areas of the plantation.

Indosylvirana urbis (Biju et al., 2014)
Urban Golden-backed Frog (Image 7) was found on the rocky patches near the pond within the rubber plantation.

Polypedates maculatus (Gray, 1834)
Common Indian Tree Frog (Image 8) is one of the most common tree frogs in Kerala. During the visual encounter survey, a male and female of the species were recorded from the coconut plantation. After the first summer shower in May, one additional individual was also observed from the coconut plantation.

Pseudophilautus wynaadensis (Jerdon, 1853)
Jerdon's Bush Frog (Image 9) was the most common frog during the present study and was recorded from all the agroecosystems selected for the present study. We observed four colour morphs of the species during the study period (Image 9A-D). All the 25 quadrats in the homegarden detected the presence of this species. The highest count of the species was from rubber plantation where there were 30-40 individuals on a single herb of Rauwolfia tetraphylla near a moist area, due to a leakage in the irrigation pipeline that passed through the rubber plantation.
It may be noted that all the amphibians sighted within the rubber plantations were either located in a pond in the rubber plantation or from artificially wet areas.

abundance of amphibians
The null hypothesis is that there is no association between the type of agroecosystem and abundance of different amphibian species. The Chi-square test (Chisquare = 236.6, df = 27, P < 0.0001) suggested that there is a strong association between the type of agroecosystem and abundance of different amphibian species (Fig. 2). Correspondence analysis was performed to understand the association between selected agroecosystems and amphibian abundance.

Polypedatus maculatus and Uperodon taprobanicus
were only present in coconut plantation and this agroecosystem was not preferred by other species of amphibians encountered in the study (Fig. 2). All other species of amphibians prefered rubber plantation.
Amphibian abundance and richness was found to be higher in rubber plantation followed by homegarden (Fig. 3). This higher amphibian diversity and abundance in the rubber plantation could be an artifact because of the presence of a pond as well as the presence of a couple of wet areas due to anthropogenic interventions, in the rubber plantations. This could also explain the presence of species such as Euphlyctis cyanophlyctis, Hoplobatrachus crassus, Indosylvirana urbis, and Microhyla rubra only from the rubber plantation, even during the summer months. According to Neckel-Oliveira & Gascon (2006), presence of an aquatic habitat is crucial for the existence of certain species of amphibians. Rathod & Rathod (2013) explained that an open canopy can increase the temperature and evaporation and decrease the persistence of moist areas. This can be one reason for fewer encounters with amphibians from coconut plantation and cashew plantation.

CONCLUSION
Protected areas cover 18% of the Earth's land area and 8% of Kerala's geographical area, and they are considered to be corner stones of biodiversity conservation. It is a fact that most of these protected areas are virtual islands embedded within a matrix of multiple land uses. A large proportion of biodiversity coexists with humans in their managed ecosystems, which can hold minimum viable populations of rare and endangered native fauna and flora. The potential of such landscapes in conserving native biodiversity is still untapped. The present study showed that agroecosystems have not only the potential to conserve generalist species, but also help to provide suitable habitat for some threatened and endemic species of amphibians.  Rondani, 1875 andCyrtodiopsis Frey, 1928 to be congeneric based on molecular affinity provided by partial nucleotides alignments of three mitochondrial and three nuclear genes. Feijen (2011), however, disputed the single clade phylogenetic hypothesis by Baker et al. (2001) and preferred Teleopsis and Cyrtodiopsis to be paraphyletic and that is being widely followed in diopsid taxonomy and biology till date (Roskov et al. 2015).
The original description of Cyrtodiopsis whitei under the genus Diopsis by Curran from India in 1936 had inadequate morphometrical details and illustrations of diagnostic characters. The whereabouts of the type specimens (male holotype and a female allotype) collected on 2 May 1935 and originally deposited in the American Museum of Natural History (New York) is uncertain, and Shillito (1940) based his study of the species on a single specimen by the identical name collected on 21 October 1920 from a location in 'Jungle of Assam', northeastern India. Földvári et al. (2007), based on laboratory culture specimens from Malaysia, provided another brief description of the species without sufficient illustration of diagnostic characters, and without reference to the original description provided by Curran (1936). It is, therefore, doubtful if the Malaysian specimens are really C. whitei. Extensive uses of laboratory-reared C. whitei in behavioral (Lorch et al. 1993;Wilkinson et al. 1998;Al-Khairulla et al. 2003 Therefore, in view of the recent collection of both the sexes of C. whitei from a new location in northeastern India, the country of the type locality, it became necessary to provide accurate description of the species based on biometric data and supported by photographs and line drawings which include descriptions of new morphological characters previously unnoticed in this species. Another objective of this study is to provide a better understanding of C. whitei in its area of distribution and its relation to other species so as to ensure its accurate identification and a better taxonomic appraisal of the genus. To that end, an identification key to the known species of Cyrtodiopsis, sensu Feijen (2011), and a note on the habitat have also been provided.

METHODS AND MATERIALS
Live specimens of C. whitei were collected from the wild habitat using insect nets and these were transferred to killing jars. Biotic and abiotic features of the habitat of occurrence were recorded on each occasion of specimen collection. Dead and dry specimens were brought to the laboratory and kept in relaxation boxes for 36 hours to allow softening of external parts. Individual insects were spread to their natural posture and mounted on paper tips, pinned, labeled, and studied under Leica M205C zoom stereoscopic microscope fitted with Leica DFC295 digital camera. Biometry to the accuracy of 0.01mm and microphotographs were taken using Leica Application 3.8.0 version software. Images, so acquired, were transferred in Microsoft power point slides to write the names of characters. Abdomens of six males and a female were dissected for genitalia study. These were individually subjected to heating in glass vials at 60 0 C, first in 10% KOH solutions for 10 minutes for maceration, then for five minutes each at increasing concentrations of ethyl alcohol (70%, 80%, 90% and 99.99%) for dehydration. Dehydrated specimens were boiled for five minutes in the saturated solution of choral phenol for softening of cuticle and sclerotised structures (Feijen & Feijen 2011). Finally, the abdomens with the ventral side up were mounted individually on clean glass slides with the help of fine tip needles under the Carl Zeiss Stemi 2000-C microscope and studied under Carl Zeiss AXIO Lab.1 microscope under 10X and 40X objectives and for Camera Lucida drawings. Specimens of this study are deposited with identical accession numbers, as used in Tables 1 and 2, in the Insect Biodiversity Laboratory, Department of Zoology of Tripura University.

Genus Cyrtodiopsis Frey, 1928
Diagnosis: Thorax with a pair of infra-alar spines, supra-alar spines absent, scutellar spines slightly to strongly curved outward, hairy, and each spine with a terminal bristle; fore femora conspicuously constricted on inner side at apex with incrassate surface; a tubercle is present at the base of inner margins of fore tibia that seems to fit into the constricted apex of fore femora when the fly is in rest. Taxonomic status of the genus: Frey (1928) proposed Cyrtodiopsis from a collection of the Stalk-eyed Flies of Philippines on the basis of a distinct "peg and hollow structure of forelegs in certain males" with dalmanni (Weidmann) as the type species. Shillito (1940) provided the first illustrated account of the family with a key to the identification of eight genera with particular reference to five species of the genus Cyrtodiopsis. Shillito (1940) distinguished Cyrtodiopsis from its nearest taxonomic relative Teleopsis in the wings without an alula, the thorax with infra-alar spines but without supra-alar spines, scutellar spines strongly curved, hairy and with a terminal long bristle and, most important, fore femora with constricted apex ventrally and fore tibia with rounded tubercles present at the base of the ventral side. Földvári et al. (2007) added a new species Teleopsis thaii, which is considered to be a Cyrtodiopsis species in the present work. Habitat: The habitat of C. whitei in the Dhalai District of Tripura province is an evergreen primary virgin forest of low hills of 'Longtharai' (local name meaning 'deep valley') and is the catchment area of two rivers, each of which traverses through valleys forming wide and narrow to very narrow streams lined by rocky banks and verticals. Stalk-eyed flies were found to be active in the sunshine hours in decomposing mixed vegetation dominated with banana leaves that was either floating in streams or in organic mass formed near the bank of stream. Atmospheric temperature and humidity of the habitat at the collection sites were recorded to be 24.2-25.6 0 C and 74-77 %, respectively, and that of the microhabitat within 10cm of aerial distance of the collection points were found to be 21. [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22].8 0 C and 81-84 %, respectively.

Description of additional characters
Male: Length 4.56-5.69 mm; coloration generally brownish, head with antennae yellowish-brown to brown, thorax shining brown, scutellum darker than prescutellum or scutum; coxae and femora reddish-brown, tibiae and bases of tarsi reddish-brown to deep brown; wings pale brown with outer margins dark; abdomen with basal three segments pale brown and distal segments dark brown (Image 1). Head (Images 2,3; Fig. 1 ): sub-triangular, dorsum of central part yellowish brown (Image 2), conspicuously raised and bears three ocelli, one bigger in the front and two smaller on sides, with a pair of deep brown bristles having pointed apices ( Fig.  1), about 0.14-0.18 mm long; frons brownish, humped, projected forward, with a dorsal curved deep brown band and a distinct mid-suture, face concolorous with dorsum of head, slightly protruding, bear several long blackish bristles with pointed apices, 1-2 of these with bifid apices, about 0.20-0.24 mm long; vertex yellowish brown to brown, narrower in front, broader at base and with rounded edges, bearing long hairs of bifid apices on the posterior edges (Image 3a); eye stalks yellowishbrown, smooth, bears a row of sparse, thin hairs, facing outward, with pointed apices, curved gently to strongly, these gradually decrease in lengths from the origin of stalks in the central part of the head to the bases of antennae (Image 2), the longest ones at the base of eye stalk 0.12-0.14 mm long, 0.07-0.0.10 mm long in the middle of eye stalks, and 0.06-0.0.08 mm long at the bases of antennae, thus the longest ones at the base of eye stalk, on average, are 1.50-2.33 times longer than the shortest ones near the base of antenna; inner vertical bristles (IVB) and outer vertical bristles (OVB) at low tubercles, 0.51-0.66 mm and 0.25-0.37 mm long, respectively (Table 1) and 2.10-3.38 times the middle width of eye stalks, respectively (Table 2); eye span 3.98-7.52 mm long and 0.87-1.32 times the body length; antennae light brown, 3-segmented (Image 2), 0.32-0.36 mm long, the shortest basal segment with a dark bristle on inner side, called scape, 0.07-0.10 mm long, the middle segment, called pedicel, 0.15-0.19 mm long (Table 1) and about 2-3 times longer than the basal segment (Table 2), bear 2-3 dark bristles, and the third and last segment, called first flagellomere, nearly bulbous, densely covered with small hairs, 0.09-0.12 mm long, about twice the length of the middle segment, and bears a thick, long bristlelike structure with pointed apex on a raised base, called arista ( Fig. 2), 0.70-0.90 mm long (Table 1), and 1.30-1.52 times and 1.89-2.57 times the lengths of IVB and OVB, respectively (Table 2). Thorax (Image 3): collar glossy brown, V-shaped, scutum glossy brown, bi-lobed, smooth; scutellum shorter than wide, 0.18-0.25 mm long and 0.44-0.61 mm wide, with dorsum broad, glossy brown, densely pollinose in the centre of pronotum and mesonotum (Image 3b), bears many short thin hairs with pointed apices only and a few long and prominent hairs with pointed or bifid apices, the longest ones about 0.25-0.39 mm long; infra-alar spines yellowish, short, dorso-ventrally flattened, and with blunt apices (Image 3c), these about 0.24-0.32 mm long (Table 1); scutellar spines dark, slightly curved inward, 0.69-1.12 mm long, 3.75-4.73 times the length of scutellum, each spine bears on its inner side 2-3 small hairs with pointed apices and 3-4 longer hairs with bifid apices (Image 3d), the longest ones about 0.26-0.35 mm long, and a long apical bristle, about 0.44-0.49 mm long (Table  1). Wings (Image 4): 3.16-4.30 mm long, bases of fore wings leathery, rest membranous, dorsal surface densely covered with minute hairs; four distinct pale brown to brown bands present from base to the apex, the basalmost band paler than the other three bands, covers anal cell, discal cell, radial cell and subcostal-radial cell from lower to upper parts of the wing; the second band from the wing base broadest and darkest between R 2+3 and R 4+5 , and pale between costa and R 1 ; the sub-apical third band brownish, irregular, widest in the middle or radialmedial cell, and the fourth apical-most band narrowest, pale brown to brown in different specimens, extending from R 2+3 to M 1+2 from the apex and projects slightly to prominently towards the subapical band in the median cell; subapical and apical bands separated by three pale spots from apex to base of wings, with the median semicircular hazy spot in comparison to pale but distinct anterior and posterior circular spots; hind wings leathery, stump-like with a short stalk and attached to the raised bases. Legs: conspicuously hairy, longer hairs with bifid apices; fore coxae 0.78-0.85 mm long, swollen in the middle; fore femora much wider (0.33-0.37 mm) than mid-(0.13-0.16 mm) and hind femora (0.10-0.12 mm), basal ¼ part of inner margins smooth, rest ¾ margins incrassate, with a shallow constriction near the joint with tibia (Image 5); fore tibia with a low, rounded and dark tubercle in the apex that seems to fit in the constricted part of the fore femur on each side when the fly is in rest; tibiae dark, sparsely hairy on margins; tarsi 5-segmented, first segment the darkest and the longest, 0.66-0.74 mm long, densely hairy on posterior margins, next four segments paler, decreasingly smaller in sizes, the apical-most segment the smallest, 0.10-0.12 mm long, about 6.16-7.20 times the first tarsal segment, bear two dark, curved divergent claws. Abdomen black, clavate shaped (Image 1), first three segments fused, fuscus, with sparse long and thin hairs having pointed apices, tergites mildly pollinose, fourth, fifth and sixth segments with distinct inter-segmental sutures, wider than first three segments, gently deflexed ventrally, tergites and pleurites with hairs all over, segments 7 to 10 narrow to very narrow, condensed, covered with microtrichia and a few sparsely distributed long hairs; sub-anal plate triangular, heart-shaped; cerci clubshaped, apically rounded, about twice the length at base. Genitalia : In ventral view, epandrium rounded with sclerotised and smooth margins, with 18 pairs of long setae counted when mounted in slide; surstyli pale, broad basally with thin margins but brown, bulbous apically, bulbous ends sclerotised and proximate in the middle, covered with microtrichia, with four pairs of long setae, two pairs originate from the inner margins of the base and other two pairs originate from the outer margins of bulbous apex (     large, with broad base and nearly conical apices, about twice as long as broad at the base, with thin margins and five pairs of hairs, two pairs of smaller hairs projected outwards and three pairs of longer ones projected inwards; hypandrium flat brown, with smooth margins, hypandrial bridge glabrous, in ventral view smooth, pale brown, bridge brown, with rough surface (Fig. 4); aedeagal apodeme elongated, brown, connected basally to hypandrium, in lateral view aedeagus with somewhat rounded end, sclerotic, and with a well developed ejaculatory apodeme (Fig. 5).
Female: The single female in the collection is similar to males except in longer body (6.49mm), longer OVB (0.39mm), longer antennae (0.43mm), longer hairs on head and eye stalks (Table 1), longer scutellum (0.32mm) and scutellar spines (1.12mm), fore coxae, fore femora and wings (Table 1). Genitalia parts were damaged in course of slide mounting.
Taxonomic status: Curran (1936) described Diopsis whitei from a tropical dry deciduous forest in Jharkhand in eastern India. That description lacked illustrations or drawings of distinguishing characters of the species. Shillito (1940) transferred the species to Cyrtodiopsis because of the presence of characteristic apical incrassate constriction infore femora and low rounded tubercles in the fore tibia. Since then six males and one female specimen of this species from a tropical evergreen forest 3. Eye span, on average, 11.18mm long, 1.39 times the body length; scutellar spines up to 5.0 times the scutellar length ……………… thaii Eye span, on average, less than 8.0mm long, up to 1.25 times the body length; scutellar spines 3.50-4.0 times but never more than 4.30 times the scutellar length; pronotum and mesonotum pollinose, shining brown pleurally ….......……………………… whitei complex (i) OVB up to1.35 times and IVB up to 4.50 times as long as the width of eye stalks in the middle; eye span about 7.93mm long and 1.25 times the body length; cerci of male genitalia with several long, dispersed setae along their surface; habitat: laboratory culture specimens sourced from primary tropical rainforest in Malaysia ( found our specimens to show similarity with the single C. whitei specimen in his possession from Meghalaya in northeastern India. Detail examination of specimens used in this study, however, warranted description of new characters not described earlier; these include some of the hairs present on dorsal surfaces of head, thorax, scutellar spines and on femora and tibiae with bifid apices, presence of a row of progressively gently to strongly curved frontal hairs of decreasing lengths from the base of the eyestalk to the base of antenna, and in the structure of male genitalia which were not adequately described. None of existing literature on Diopsidae mention the occurrence of hairs with bifid apices and curved frontal hairs of eye stalks that were noted in specimens of C. whitei of this study. It is possible that earlier workers might have missed these characters in whitei or other taxa of Diopsidae, therefore, it may be premature to conclude that these characters are unique to C. whitei or that the sample of this study might represent a new population of a distinct species. The author was not able to access the type specimens of whitei or specimens of other species of Diopsidae from valid sources. Therefore, at this point, the specimens of this study from moist evergreen forests of northeastern India are considered to represent possibly a part of widely distributed populations of whitei complex in its geographic range extending from the dry deciduous forest of eastern India (Jharkhand, the type locality) to moist evergreen forest in northeastern India, and, possibly, further east in Southeast Asia (Malaysia included), and it is assumed that populations of whitei might show habitat/area-specific variations and this position may be maintained until such time future study reveals more information.

DISCUSSION AND CONCLUSIONS
Feijen (2011) considered Cyrtodiopsis to be a weakly defined genus from the oriental region due to inclusion of several unrelated species at different times but preferred its distinct identity sensu Shillito (1940) in view of distinctive morphological attributes (chiefly in having prominent incrassate constrictions on apex of fore femora and low tubercles on the inner bases of fore tibiae) absent in the species of other genera under the family. Earlier, cladistic study using mitochondrial genes made by Baker et al. (2001), Meier & Baker (2002) and Földvári et al. (2007) revealed phylogenic relationship between Teleopsis and Cyrtodiopsis but molecular distinctions between the two genera based on four marker genes, particularly between T. thaii and

12043
T. breviscopium, were not conclusive (Földvári et al. 2007). To date, the presence of sharp incrassate apical constriction in fore femora and corresponding rounded tubercles on the basal parts of the fore tibia are the most robust and unique characters to the species of Cyrtodiopsis. Also, supra-alar spines, characteristic of all Teleopsis species, are absent in Cyrtodiopsis species. The presence of several hairs with bifid or split tips on the body of C. whitei could possibly be another unique character of this genus till further study reveals its presence in other species of Cyrtodiopsis and possibly in other genera of Diopsidae. Despite prominent differences in morphometry between the two populations of whitei from geographically isolated locations of India and Malaysia, as evident from the taxonomic key above, we do not describe these specimens as a new species, because both populations share fundamental similarities in characters of fore wings, pollinosity distribution in thorax, body coloration, and general structure of genitalia. The observed differences in morphometry and presence of some of the body hairs with bifid tips in whitei from northeast India, among others, might represent the influence of differences in environments of the two habitats separated by several hundred miles. This study has founded the basis of future study to ascertain the prevalence of these characters when more specimens become available from these or nearby areas and also from the original type locality.

OPEN ACCESS
Journal of Threatened Taxa | www.threatenedtaxa.org | 26 July 2018 | 10(8): 12044-12055 Abstract: Benthic macroinvertebrate communities are frequently applied as indicators of aquatic ecosystem health as many species are responsive to pollution and abrupt changes in their surroundings. The qualities of benthic invertebrate communities greatly depend on habitat conditions. Thus the diversity in benthic community varies with different habitat conditions. This investigation on the structure of the benthic invertebrate communities was conducted on river Ichamati, a trans-boundary river between India and Bangladesh to assess the cumulative effects of water quality on the aquatic biota. The study period extended from February 2011 to January 2014 at three sites from Majdiah to Hasanabad (in West Bengal, India) a stretch of 124km. A total of 23 macrobenthic species belonging to three phyla, five classes and nine orders were identified. Fifteen species of benthic invertebrates belonging to Mollusca, three species under Annelida and five species under Arthropoda were found. The highest abundance density (3633.33 indiv.m -2 ) and species richness (18 species) were recorded up-stream (Majdiah) where marginal habitats covered by macrophytes were significantly higher than at other sites. Both the organic carbon (4.41±1.11) and organic matter (7.48±1.56) of soil at this site were the maximum thus influencing the richness of benthic macroinvertebrate communities. Hydrological variables, viz, dissolved oxygen, pH, alkalinity; hardness, salinity, nutrients, calcium, and magnesium were studied to determine their influences on the benthic community in the upper, middle-and down-streams of the river, respectively. Shannon's diversity index (0.95-2.07; 0.00-0.72; 0.00-0.64), dominance index (0.57-0.86; 0.00-0.44; 0.00-0.44), evenness index (0.72-0.95; 0.61-1.00; 0.00-1.00), Margalef index (0.72-2.23; 0.00-1.32; 0.00-0.28) of the upper, middle-and down-streams were calculated. Benthic macroinvertebrate density was correlated with hydrological variables which indicated that the abiotic factors had either direct or inverse influence on the richness and abundance; however, the abiotic factors did not correlate identically in all three sites.

INTRODUCTION
Benthic macroinvertebrates are sedentary or sessile aquatic fauna that exist in the bottom substrates of their habitats (Lenat et al. 1981;Victor & Ogbeibu 1985;Rosenberg & Resh 1993;Idowu & Ugwumba 2005) at least for a part of their life cycle. The benthic fauna perform a key role in nutrient cycling and are also used as food for other aquatic animals (Lind 1979;Milbrink 1983;Jana & Manna 1995). Further they play a critical role as a link in the aquatic food chain affecting bio-geochemical processes in the sediment (Wetzel 2001;Heck et al. 2003;Pokorny´ & Kveˇt 2004;Idowu & Ugwnmba 2005). Benthic invertebrates are difficult to sample especially in deep subsurface sediments. Thus, the species richness and functional importance of freshwater benthic invertebrates usually goes unnoticed until unpredicted changes occur in the ecosystems. Besides these organisms are used as bio-indicators as they frequently respond to pollution stress (Stanford & Spacie 1994;Gamlath & Wijeyartne 1997;Ikomi et al. 2005). The community structure of benthic macroinvertebrates is influenced by the physicochemical parameters of the water body (Timm et al. 2001;Johnson et al. 2004;Kagalou et al. 2006;Celik et al. 2010). Examination of parameters like richness, diversity, abundance, evenness and community composition are essential to determine the natural or anthropogenic changes with time (Mittermeier & Mittermeier 1997;Dudgeon et al. 2006;Srivastava 2007;Strayer & Dudgeon 2010;Jun et al. 2016). In riverine ecosystem macrobenthic invertebrates show an uneven distribution (Timms 2006). River Ichamati ('Icha' -fish and 'moti' -pearl), is one of the important trans-boundary rivers between Bangladesh and India, has variable biological, physical and chemical characteristics due to its irregular discharge pattern, diverse habitat arising out of abiotic and anthropogenic activities and both brackish and freshwater characters. Presently, this river is facing various environmental constraints due to siltation, discharge of organic debris from human settlements, production of macrophytic biomass, lack of sanitation and over-fishing (Das et al. 2012). Thus, it is ever more important to preserve the biodiversity of aquatic flora and fauna in this river to lower the risk of sudden unwanted consequences. A number of studies on macrobenthic community structure and hydrochemistry of various water-bodies are well documented (Degani et al. 1992;Jana & Manna 1995;Mancini et al. 2004;Moretti & Callisto 2005;Dolbeth et al. 2007;Sharma & Dhanze 2012;Basu et al. 2013;Mishra & Nautiyal 2013Nautiyal & Mishra 2013;Nautiyal et al. 2017).
To the best of our knowledge, information on macrobenthic fauna of river Ichamati is unavailable so far. This encouraged us to undertake the present study on the river to ascertain: (i) the structure and composition of the benthic macroinvertebrate species, (ii) the environmental factors (natural as well as anthropogenic) responsible for the community patterns, (iii) the present ecological status of the river and (iv) determine the quality of water by using benthic fauna to establish the pollution level of the river to create a base line data.

Description of the study area
The river Ichamati is among the important transboundary rivers sharing the boundaries between Bangladesh and India. River Mathabhanga originates from the right bank of Padma at Munshigunj in Kustia District, Bangladesh. It bifurcates near Majhdia (Nadia District, West Bengal, India) creating two rivers, Ichamati and Churni. River Ichamati traverses a course of about 216km and finally discharges into the river Kalindi at Hasnabad in the district of North 24 Parganas and ultimately finds its way into the Bay of Bengal near Moore Island as a part of Kalindi-Raimangal estuary in the deltaic southern part of West Bengal. After about a 19.5km long journey in India it re-enters Bangladesh. It crosses the border again near Duttafulia in Nadia District (West Bengal, India). After a further 21km, it falls into the Bay of Bengal in Bangladesh near Hasnabad and Taki.
The stream at its origin is narrow and shallow clogged by macrophytes such as Eicchornia, Pistia, Lemna and Alternanthera. The middle and down reaches of the river are now facing problems due to siltation, high fluvial allochthonus discharges from the river banks, discharge of organic debris from the human settlements along the river, all domestic works such as bathing, washing clothes, utensils, bathing of cattle, lack of sanitation practices, boat ferry, immersion of idols during festivals etc.
The study period extended from February, 2011 to January, 2014 at three sites from Majdiah to Hasanabad (in West Bengal, India) a stretch of 124km. The locations of the sites chosen were (1) near the origin (Majdiah; up-stream, site I), (2) middle part of the stretch (Tetulia; middle-stream, site II), finally Hasanabad (down-stream, site III) before it reaches river Kalindi in the south (Mondal & Bandyopadhyay 2016).

Locations and characteristics of the sites
Locations of the sites (I, II and III) are marked in Images 1, 2. Physiological and geographical characteristics of the three sites are given in Table 1.

Sampling methods
Water samples were collected from two sampling points (140m apart) in each site in 1 L clean plastic containers between 06:00-08:00 hr during February 2011 to January 2014 twice a month and transported to the laboratory for chemical analyses.
Water temperature was recorded using mercury glass thermometer (0-60 0 C). Electrical conductivity, total dissolved solids (TDS) and pH were measured by ELICO Ion analyzer (Model: PE 138, India). All other water quality variables such as dissolved oxygen (DO), free carbon dioxide, total alkalinity, total hardness, calcium, magnesium, phosphate, nitrate, salinity and transparency, organic matter and organic carbon were monitored following standard protocol, American Public Health Association (APHA) (2005).
Benthic invertebrates were collected twice a month with a specialized box sampler having a dimension of 15 x 15 cm which can penetrate a maximum depth of 15cm (Paul & Nandi 2003). The samples were sieved with No. 40 mesh (pore size: 0.420mm) (Jana & Manna, 1995;Tagliapietra & Sigovini 2010). Considering the depth of the down-stream, desired samples were collected with the help of local fishermen. Collected organisms were preserved in 4% formalin. Benthic macroinvertebrate were then identified following Michael (1977) for the phylum Annelida, Barnes et al. (1988) and Rao (1989) for the phylum Mollusca whereas Arthropoda by the Zoological Survey of India, Kolkata, India. Benthic macroinvertebrates were quantitatively analysed by individual counting of each taxon and expressed in individuals/m 2 .
Taxonomic indices was subjected to univariate analyses for studying the benthic community structure using Margalef's richness index, Margalef (1968) for species richness (counts the number of different species in a community), Pielou's Evenness index (Pielou 1966) for species evenness (quantifies the relative abundance of species present in a community), Shannon-Weiner index (Shannon & Weiner 1964) for species diversity (reflects the types of species present in a particular area at a particular time) and Simpson's Dominance Index (Simpson 1949) for dominancy (quantifies the dominancy sharing species in a community). The data were computed using Paleontological Statistical software (PAST version 3.15). Pearson correlation (r) was applied to analyse the relationship between the benthic macroinvertebrates density and hydrological variables. The graphs were plotted with MS Excel Software.

RESULTS
The range and average of all water parameters were recorded in Table 2. In the Ichamati, 23 benthic

12047
macroinvertebrate species were found from all the samples collected from upper-, middle-and downstreams (Table 3). Of these, up-stream was the richest with species (18) followed by middle-stream (5) and down-stream (2). The maximum density (individual m -2 ) was found in the following sequence, i.e., up-> middle-> down-streams. Fig. 1 showed the monthly variations of total benthic macroinvertebrate community in three different sites of Ichamati. Benthic macroinvertebrate community was available throughout the year up-stream with peaks in the months of June and September (Fig. 1). Down-and middle-streams showed similar trends where the communities gradually increased from October and reached the maximum in May (Fig. 1). In down-stream, during monsoons (June-September) it was not possible to find and collect any benthic macroinvertebrate samples due to the dangerous rise in water levels and the highly turbulent character of the water. Perhaps, due to the same reason a low concentration of benthic macroinvertebrate was found mid-stream during the monsoons.
The results are presented separately for all three different study sites as follows:

(A) At upper reaches of Ichamati
In the upper-stream, 13 species of Mollusca belonging to class Gastropoda (three orders) and class Bivalvia (one order) dominated the community followed by Annelida (2 orders) and Arthropoda (one order). The population of benthic invertebrates was dominated mainly by three taxa of Mollusca: namely, Bellamya bengalensis Lamarck 1822, Bellamya dissimilis Muller, 1774 and Gyraulus convexiusculus Hutton, 1849 ( Table 3). The abundance of B. bengalensis increased to maximum density (322.22) in the pre-monsoon period then its population declined. In comparison, the B. dissimilis after attaining its population peak in pre-monsoons (255.54) drastically declined in the post-monsoon period (33.33). B. crassa was completely absent in pre-monsoon periods. On the other hand, species like Segmentina in monsoon and pre-monsoon periods and Melanoides in the monsoons were completely absent. Brotia and Bythinia were found in all seasons (Table 4).
During the investigation, one Bivalvia taxa (Lamellidens marginalis Lamarck 1819) was found exclusively in the pre-monsoons. Hutton 1849 found maximum Gyraulus convexiusculus (411.10) in the monsoons was another dominant species among Mollusca (Table 4). Further, two species of phylum Annelida (Glyphidrilus tuberosus Stephenson, 1916 and Pheretima posthuma Kinberg, 1867) could be detected both in monsoon and post-
Maximum species diversity (1.79) and Simpson's dominance index (0.79) were recorded in the monsoon period and minimum species diversity (1.58) and dominance index (0.75) in the pre-monsoons. Species richness, i.e., Margalef's index was found to be the maximum during monsoons and minimum in the premonsoon period. Pielou's evenness index was found to vary from 0.81 to 0.87 (Table 5).
Water temperature, transparency, free CO 2 , salinity, organic carbon, TDS and nutrients were positively correlated with the benthic macroinvertebrate abundance (Table 6). Dissolved oxygen and total alkalinity, two important parameters were negatively correlated with benthic macroinvertebrate density. These water parameters were additionally correlated individually with density of Gastropoda, Bivalvia, Annelida and Arthropoda (Table 7) in up-stream only.

12049
monsoons (Table 4). Only one Annelida, i.e., Neanthes was found maximum during pre-monsoons and declined at the onset of the monsoon. Pila globosa Swainson 1822, the Gastropoda were identified during monsoon and post-monsoon periods. One bivalvian species, Modiolus was recorded in maximum density (1965.33) during pre-monsoons (Table 4). Diversity index and dominance index recorded were the maximum in the monsoons and minimum in the pre-monsoons (Table  5). Evenness index and richness index was found to be maximum in the post-monsoon period.
In the middle-stream, all the water and soil parameters except DO, free CO 2 and phosphate were positively correlated with benthic macroinvertebrate density (Table 6).

(C) At down reaches of Ichamati
The benthic macroinvertebrate community ( Fig.  3c) was dominated by Mollusca (89.01%) followed by Arthropoda (10.99%; Fig. 3c). Two species of Arthropoda, Ocypode sp. were absent in the monsoons but were present in pre-and post-monsoon periods ( Table 4). Scylla tranquebarica was minimum (11.11) in the monsoons, but found in the other two seasons. One bivalvian species Modiolus was maximum (711.10) during pre-monsoons (Table 4). All the taxonomic indices determined (Table 5) were found to have maximum values in post-monsoon and minimum in monsoon periods.
In down-stream, total alkalinity, phosphate, organic carbon and organic matter were negatively correlated with benthic macroinvertebrate density, the rest of the parameters were positively correlated (Table 6).

DISCUSSION
The glory of river Ichamati has faded a lot with time. Ichamati now faces problems like forcible land occupation, weed infestation, different environment hazards due to lack of sanitation facilities, encroachment, ground water contamination etc. Destruction of aquatic flora and fauna in the river is the most serious problem regarding the ecosystem.
The important factors that affect the abundance of benthic macroinvertebrate fauna in a given community include the hydro-biology of water, substrate of occupants and food availability (Olenin 1997;Nelson & Lieberman 2002;Carlisle et al. 2007;Coleman et al. 2007;Basu et al. 2013).
The pH of water of all three sites indicated the alkaline nature of the water; the pH of the up-stream was the highest (7.86 ± 0.15) compared to the two other sites. The richness of diversity of benthic macroinvertebrates was found maximum in up-stream due probably to the alkaline nature and shallow depth of the river. Simpson et al. (1985), Feldman & Connor (1992) and Baldigo et al. (2009) also found that the site with the higher pH had a higher diversity of benthic macroinvertebrates.
Benthic macroinvertebrates density was negatively correlated with DO level as they could survive in poor DO conditions. In this study the low dissolved oxygen content observed in up-stream water might be due to the high organic matter decomposition from macrophyte vegetation and also bottom type which contained high percentage of mud (Sandin 2003;Williams & Gormally 2009;Jiang et al. 2010;Schultz & Dibble 2012;Zybek et al.2012). The high DO contents in middle-and downstreams were attributed to non-vegetation and strong water current characteristics of these two sites (Soszka 1975;Cogerino et al. 1995).
The low density of benthic macroinvertebrate was observed during the present study in all three sites particularly in middle-and down-streams. The species richness of benthic macroinvertebrate were found to be the highest in up-stream probably due to suitable habitat conditions, organically enriched soft bottom (Ingole et al. 2002), slow water current, shallow depth (Roy & Gupta 2010), bottom substrate (muddy and clayey) and the presence of macrophytes in marginal water (Kumar et al. 2013;Tall et al. 2016).
Molluscans were mostly associated with very low oxygen and lentic ecosystems (Spyra 2010). The upstream of Ichamati was enriched with molluscan density (13 species). The water in this region was motionless and had a shallow substratum with decomposed organic matter which facilitated the molluscans growth, especially Gastropoda (Principe & Corrigliano 2006;Zybek et al. 2012). The lowest concentrations of salinity, hardness and alkalinity of water may have enhanced the abundance of species in the up-stream, hence these parameters showed the negative correlations with benthic macroinvertebrate densities. This was

12051
further substantiated by the observation of Brucet et al. (2012). Two oligochaetes, Pheretima posthuma and Glyphidrilus tuberosus were present up-stream due to their preference for organically enriched polluted water bodies with low oxygen content, also noted by Barquin & Beath (2011). This was further corroborated by the negative correlation with dissolved oxygen, total hardness, total alkalinity and positive correlation with nutrients, organic carbon and organic matter (Table  6). Pheretima, though a terrestrial species, was found during monsoon and post-monsoon periods when the riverbank was flooded. Possibly because of inundation, they were found within 1m inside the river from the edge during these periods; however, Brraich & Kaur (

12052
In this study the positive correlation between nutrients and organic matter with benthic macroinvertebrate density supports the observation. It was interesting to note that species like Segmentina, Melanoides and Lamellidens, were not found during the monsoons, probably due to increased water levels and a relatively strong water current to unsettle the bottom substrate on which these species were attached (Koperski 2011). In this study it was observed that the species richness of freshwater Gastropoda depended on the type of bottom substrate and the richness of aquatic macrophytes (Lodge 1985;Perez 2004;Spyra 2010). The density of Gastropoda was positively correlated with phosphate, nitrate, total alkalinity, TDS, pH, organic carbon and organic matter (Table 6) and supported by Pip (1987) and Williams & Gormally (2009). In middle-stream, a very low species diversity comprising of two species of Mollusca (with one Gastropoda and one Bivalvia) and only one Polychaeta (Neanthes sp.) were observed. Polychaetes preferred fine to medium type of sandy bottom with moderate abundance of admixtures of silt and clay (Al-khayat 2005). The middle-stream had a very similar bottom type. Molluscan diversity was meager probably due to the high flow of river water and a particular bottom type (sand and clay). The Benthic macroinvertebrates experienced threats by the changes in its habitats associated with pollution and siltation. Moreover, the poor growth of bottom fauna could be associated with frequent water level fluctuations. The dependence of benthic macroinvertebrate fauna on a number of factors such as physical nature of the substratum, depth, nutritive contents, degree of stability and oxygen concentration of the water body (Barbour et al. 1999;Merz & Chan 2005;Braccia & Voshell 2006) was reflected by the findings of this investigation that in middle-stream -substratum, depth, nutrition and oxygen concentrations were not congenial for benthic macroinvertebrate diversity to flourish. This was supported further by the negative correlation between density and phosphate as well as oxygen concentrations studied. Presence of Pila globosa (Gastropoda) and Neanthes sp. (Polychaeta) indicated the freshness of the water (Perez 2004). Benthic macroinvertebrate in middle reaches were observed in the highest concentration level during pre-monsoons probably due to the maximum occurrence of Modiolus sp. (Bivalvia). It was likely that the species utilized the elevated concentration of calcium in the water during pre-monsoons contributing to the increase in the benthic macroinvertebrate density.
Meager Hydrological conditions such as extreme hard water and salinity alteration (due to freshwater inflow during monsoons) and food availability were major factors affecting the community dynamics of benthic invertebrates (Brucet et al. 2012).
Water temperature showed a positive correlation with benthos density.
During the pre-monsoon period (summer: March-June) density of benthic macroinvertebrates were higher than the post-monsoon period (winter: November-December) in all three sites presumably indicating that the temperature had a positive influence on the benthic macroinvertebrate community as noted by Hauer & Hill (1996) and Sharma & Rawat (2009).
Water transparency was positively correlated with the benthic invertebrates as also noted by Basu et al. (2013). A significant positive correlation was found between organic carbon content of soil, organic matter and benthic invertebrate density. The presence of aquatic vegetation in the study area supported the availability of more organic matter (Bath et al. 1999;Rosenberg 2001;Mikulyuk et al. 2011;Basu et al. 2013).
Community structure index is a measurement for two distinct aspects of biological community: (i) number of taxa (richness) and (ii) distribution of individuals among taxa (evenness). Diversity indices depend on the quality and availability of habitat (Barbour et al. 1999). Mason (1996) set diversity index <1 for highly polluted, 1-3 for moderately polluted and >4 for unpolluted water bodies. In up-stream the diversity index (Table 5) indicated moderately polluted water and the presence of a rich habitat. In middle-and down-streams the diversity index indicated more polluted water than upstream. In this study, the evenness indices of all three sites indicated that the taxa identified were consistently distributed (Table 5) in all sites.
The results pointed out that benthic macroinvertebrate diversity was very poor in middleand down-streams but had a moderate population in up-stream. Structure of macrobenthic population was mainly driven by seasonal variations, depth of water, water current, habitat type, riverbed characteristics and influence of anthropogenic interferences. The macrophyte vegetated marginal habitats supported greater species richness and abundance (up-stream) than non-vegetated habitats (middle-and downstreams). The Mollusca could be regarded as a bioindicator species thus indicated a good water condition of the river. It was evident from the investigations that the seasonal changes in the hydrological parameters influenced the community structure of the benthic invertebrates in river Ichamati. Referring to Townsend (1994) and Sukhorukov (2013), Aerva congesta is one of the two known endemic species of the Amaranthaceae in the Mascarene Islands (South West Indian Ocean) (Image 1). It is known from both Mauritius (Image 2) and Rodrigues (Image 3), but is not recorded on Réunion Island. The first specimen was collected by Balfour in 1879 where he indicated that the species grew as a small compact herb present on coralline limestone, in association with Abrotanella rhynchocarpa Balf.f. (= Rhamphogyne rhynchocarpa S. Moore) and Oldenlandia sieberi Baker var. congesta. According to Strahm (1989) A. congesta has not been seen or collected on Rodrigues since Balfour, and it is probably extinct there; extensive surveys made on mainland Rodrigues (e.g., Anse Quitor) and some islets (e.g., Ile Gombrani, Ile Chat, Ile Crabe) were unfruitful (Wiehe 1949;Cadet 1972Cadet , 1975Guého 1980;Smith et al. 2004a,b). On Mauritius, Strahm (1989) quoted that there were recent collections from one locality on mainland Mauritius, but only samples from Round Island, an outer islet northern of Mauritius (Image 2), were accessed at The Mauritius Herbarium.  where the species was found at Gris Gris.

Aerva congesta in the Mascarenes is extant only on
Mauritius and is believed to be extinct in Rodrigues. This

Evaluation of current conservation status (IUCN 2001)
The whole population of the species is considered under '2 locations' comprising two subpopulations.
The extent of occurrence (EOO) of the species cannot be calculated from two localities and is deemed to be small. Taking the topographic limitations throughout its geographic range, the Area of Occupancy (AOO) was measured as 8.0km² (800ha) and the number of mature The discovery of a new population at Gris Gris reduces the species' chances of extinction. However, the newly discovered population in Gris Gris is very small. Furthermore, there are signs of natural erosion of the cliff around where the species is growing, decreasing the number of suitable sites. The presence of alien invasive weeds, as well as fast-growing native species, like Stenotaphrum dimidiatum has potentially negative impacts on A. congesta at Gris Gris through interspecific interactions detrimental to native plant populations as shown elsewhere on the island (Baider & Florens 2011). To conserve the recently located population, we suggest setting up ex situ propagation by the institutions concerned like the National Parks & Conservation Service and the Forestry Service; avoiding mixing the plants from the two populations to preserve their eventual genetic distinctiveness.
A minor weeding would minimize the negative effect posed by alien species as well as the fast growing grasses like Z. matrella and S. dimidiatum; however, this needs to be well planned to minimize soil erosion. A restoration program should also be implemented to increase native cover of the site that contains other threatened native species like the fern Ctenitis maritima and the endemic liana Cissus anulata Desc., which is known only from this region of the island; taking into account not to create too much shade over the plant which requires substantial exposure to thrive.
Considering the failed augmentation on Round Island since 2004 and the failed introduction on Ile aux Aigrettes in 2013, we suggest setting up systematic research to better understand the ecological requirements of the species. Finally, it is advisable to survey similar habitats in Mauritius, Rodrigues and Réunion to try to locate eventual new populations. Rodrigues: no loc., August-December 1874, Balfour s.n. (holotype K 000243711; isotype M). elevation between 3-90 m. The species is mostly restricted to cliffs and predominantly brown soils. In association with the following native species of grasses and sedges (Zoysia matrella (L.) Menill, Stenotaphrum dimidiatum (L.) Brongn. and Fimbristylis cymosa R.Br.) as well as other native species (Dichondra repens J.R. Forst. & G. Forst., Ctenitis maritima (Cordem.) Tardieu and Selaginella obtusa Spring). In Gris Gris, the fire prone alien invasive grass Heteropogon contortus (L.) P.Beauv. ex Roem. & Schult. is also present. Key threats to the species observed in the study area are competition with fast growing native and alien grasses; sometimes competing for resources like light, water and nutrients and leading to poor plant growth.
System: Small prostrate perennial herb.

Information On Threats
Major Threats: The main threats to the two species population are: · Direct competition for resources such as water, light, nutrients with exotic grasses and other herbaceous weeds.

·
Increasing erosion, mainly during the dry season where there is less ground cover due to die off of exotic grasses like Heteropogon contortus and where the ground is more vulnerable to erosion. Additional threats: Drought; fire; cyclones; diseases; pest attack; littering (at Gris Gris) burrowing by Shearwater birds (on Round Island) and thus leading to damage of plants are other possible threats.

Use And Trade Information
Use: Not known. Livelihoods and sustenance: Not reported. Trend in off take from the wild: Not reported. Trend in off take from cultivation: Not reported.

Conservation actions: ·
The Mauritian Wildlife Foundation is propagating the species through seeds and cuttings for augmentation on Round Island and introduction on Ile aux Aigrettes. Forms part of the Round Island Plant Restoration Plan · Weeding and planting in localized patches

Research in place: ·
Optimising choice of sites for the augmentation, re-introduction and introduction programmes to ensure that the correct conditions for the species, including microclimatic requirements are fulfilled.

·
Investigating the translocation of Aerva congesta to other offshore islets and Rodrigues.

·
Investigating the fate of native herbaceous community on Round Island and carrying out weed management in localized patches.

Research needed: ·
Setting up systematic research to better understand the ecological requirements of the species.

·
Determining the success of different planting techniques regimes on plant survivorship.

Monitoring in place: ·
Protecting planted areas against Shearwater burrowing · Sowing seeds into restoration areas · Treating restoration planting and seed sowing as field trials in order to gain valuable information to aid future restoration work on Round Island and elsewhere.

INTRODUCTION
Phenology of tropical trees has attracted much attention nowadays from the point of view of conservation of tree genetic resources as well as forestry management, and for a better understanding of the ecological adaptations of plant species and community-level interactions.
The study of tree phenology provides knowledge about the pattern of tree growth and development as well as the effects of environment and selective pressures on flowering and fruiting behaviour (Zhang et al. 2006). Phenology of vegetative phases is important, as cycles of leaf flush and leaf-fall are intimately related to processes such as growth, plant water status, and gas exchanges (Reich 1995). Studies of phenology are of great importance in determining the temporal changes that constrain the physiological and morphological adaptations in plant communities for utilization of resources by fauna (van Schaik et al. 1993). The sunshine hours, temperature, and annual precipitation have been recognized as the main environmental indications for leafing and flowering in the tropics. In many evergreen species, leaf flush and flowering occur close in time on the same new shoot. Variation in flowering time relative to vegetative phenology, induced by a variety of factors (significant rain in winter/summer, decreasing or increasing photoperiod, or drought-induced leaf-fall), results in a number of flowering patterns in tropical trees (Borchert et al. 2004). Phenological processes are significant constituents of plant fitness, since the time and duration of vegetative and reproductive cycles affect the capability of a plant species to establish itself in a given site (Pau et al. 2011). Singh & Kushwaha (2005) suggested that climate change forced deviations in the length of the growing period, and competition among species may change the resource use patterns in different species. Global climate change may force variations in timing, duration, and synchronization of phenological events in tropical forests (Reich 1995).
Although a few research works have addressed the population dynamics of the species in homegardens, northeastern India (Saikia & Khan 2013), an attempt has been made to study the phenology of A. malaccensis, which could contribute towards the conservation and management of the species, considering its almost extinct status in the wild (Anonymous 2003). Therefore, the present study is aimed to assess the phenological behaviour of Aquilaria malaccensis in a secondary tropical evergreen forest to understand the response of climatic variables and the periodicity of seasons.

Study area
The phenological study was conducted in a secondary tropical evergreen forest located at Sonachera in Cachar District of Assam, northeastern India (Fig. 1). Secondary forests are those forests that regrow largely through natural processes after significant anthropogenic disturbance of the primary forest vegetation at a single point in time or over an extended period of time, and place prominently a major change in tree diversity and/ or species composition with respect to nearby original forests on similar sites (Chokkalingam & de Jong 2001). The studied secondary tropical evergreen forest covers an area of 5 hectares. The geographical location of the study site is 24.36˚N latitude & 92.44˚E longitude and altitude range from 73 to 102 m.
Topographically, the area is characterized by typical terrain and hillocks that harbour diverse biological diversity. The climatic condition of the study area is subtropical, warm, and humid. Maximum precipitation occurs during the months of May to September, which is

12066
mainly controlled by the south-west monsoon season. The mean annual rainfall of the study area during the study period (2013-15) was about 2055.8mm, most of which (94%) occurred during April-September. The mean annual minimum and maximum temperatures were 19.9 0 C and 31.6 0 C, respectively (Fig. 2). The mean annual relative humidity was recorded at 75.9%. The forest is categorized as "Cachar Tropical Evergreen Forest (Champion & Seth) (reprinted) (2005) (1B/C3)" type dominated by Chrysophyllum roxburghii, Maniltoa polyandra, Memecylon celastrinum, Mesua floribunda, Palaquium polyanthum, and Pterospermum lanceifolium. The selected secondary forest site is more than 35 years old. The secondary forest of this region is relatively unexplored and harbours a rich plant diversity.

Study species
Aquilaria malaccensis Lam. (Thymelaeaceae) is one of the most important species of commercial products in the world and is valued for its fragrant resinous dark-coloured wood known in trade as agar. Agarwood is formed by a complex plant-microbial interaction of a parasitic ascomycetous fungus known as Phaeoacremonium parasiticum (Ng et al. 1997). Phytogeographically, the distribution of A. malaccensis comprises the region of India, Myanmar, Sumatra, Peninsular Malaysia, Singapore, Borneo, and the Philippines (Chua 2008). In northeastern India, it occurs mostly in the foothills of Arunachal Pradesh, Assam, Manipur, Meghalaya, Mizoram, Nagaland, Sikkim, and Tripura (Saikia & Khan 2012a). In Upper Assam, the species is commonly cultivated in home gardens in association with other valuable plants for its high commercial demand (Saikia & Khan 2012c). The bark of the plant is also used as raw material for preparing a writing paper called 'Sanchi pat' for writing religious scripts (Nath & Saikia 2002). The agarwood is imported locally and exported internationally due to its wide use in the incense and perfume industry (Manohara 2013). Agarwood oil is a valuable component and is used as a digestive, sedative, analgestic, antiemetic and antimicrobial agents in tradition medicine (Cui et al. 2013). In the past few years, large-scale harvesting has caused rapid depletion of the stock in the natural forests. According to the IUCN Red List, the species is globally Vulnerable A1cd (Ver 2.3; IUCN 2017) and has been included in The World List of Threatened Trees (Oldfield et al. 1998). The species is also listed in Appendix II of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES 1994).

METHODS
A total of 35 individuals with girth size ranging from 12.7 to 81.7 cm were selected for phenological observations. The selected trees were marked with a metal tag for phenological observations. Using binoculars, phenological observations were made on leaf initiation, leaf-fall, flowering, and fruiting of the marked individuals. Phenological observations were based on phenological score: zero for no phenophase, one for less, two for moderate, and three for high (Broadhead et al. 2003). Detailed observations were carried out at 15day intervals over a period of two years from February 2013 to January 2015.

Data analyses
Data analysis was performed using statistical software (MS Excel 2010) and (SPSS 21) version. Spearman rank correlations were performed to investigate correlations between monthly phenophase activity and environmental variables such as temperature and rainfall following (Zar 1984). The duration was classified as short (<2 months), intermediate (2-5 months), or extended (>5 months) based on the mean number of months in which the phenophase occurred (Luna-Nieves et al. 2017). Circular statistical analyses were performed to determine whether the vegetative and reproductive phenophases were homogenously distributed throughout the year. For this purpose, months were converted into angles with intervals of 30°, and then the mean angle or mean date (α), the circular concentration (r), and the circular standard deviation (SD) were calculated. (α) indicates the time (month) of the year in which the largest number of individuals of a given species presented a phenophase, while (r)

12067
indicates the degree of dispersion or concentration of the observations (Zar 1984). To determine the significance of the angle, a Rayleigh test (z) was used. There is a seasonality in the phenophases if the average angle is significant. The intensity of circular concentration (r) values varies from 0 (phenological activity uniformly distributed throughout the year) to 1 (phenological activity concentrated in a particular period of the year) (Morellato et al. 2010). We used the program ORIANA 3 (Kovach 2007) for these analyses. The meteorological data of the study area are presented in (Fig. 2).

Leafing activity
A. malaccensis initiated leafing during the premonsoon period (March-April) and continued up to a warm monsoon period throughout the favourable season (July-August) (Fig. 3A). The degree of circular dispersion or concentration (r=0.99) indicates that the phenophase leaf initiation was concentrated in a particular period of the year (Fig. 5). The phenophase leaf initiation was seasonal (Rayleigh Z, p < 0.01), and it occurred once a year. Peak leaf initiation was observed during March-April (Fig. 4). During the years 2013-14 and 2014-15, temperature registered its influence on leaf initiation significantly whereas rainfall displayed its impact in 2014-2015 (Table 1). The combined effect of temperature and precipitation, rather than their individual effects, more strongly influenced leaf initiation. Leaf-fall occurred during November-March with peak fall during January-February (Fig. 4). Rainfall and temperature presented a negative slope in correlation with leaf-fall due to decreasing day length and rainfall (Table 1).

Flowering
Flowering occurred during April-June (Fig. 3C &  4). The degree of circular dispersion or concentration (r=0.97) indicates that the phenophase flowering was concentrated in a particular period of the year (Fig. 5). The flowering phenophase was intermediate (Rayleigh Z, p < 0.01), and the open flower lasted one month. The duration of flowering phenophase ranged from 30 to 85 days with an average duration of 58.05 ± 6.35 days during 2013-14 and 32-90 days with an average duration of 61 ± 6.37 days during 2014-15, and it varied greatly among the individuals with a coefficient of variation (C.V. % =20.62). Flowering was significantly influenced by temperature and rainfall while in one-month lag period only rainfall was significantly correlated with flowering in 2013-2015 (Table 1). The flowers are yellowish-green and produced in umbels (Image 1a & b); the fruit is a woody capsule (Image 1e & f).

Fruiting
The fruiting phase extended over the monsoon period (April-September) with a peak during May (Fig. 3D & 4). The degree of circular dispersion or concentration (r=0.98) indicates that the phenophase fruiting was concentrated in a particular period of the year (Fig. 5). The fruiting phenophase duration was intermediate (Rayleigh Z, p < 0.01), and the unripe fruits lasted for two months. The duration of fruiting phenophases ranged from 28 to 65 days with an average duration of 46.57 ± 5.24 days during 2013-14 and 30-72 days with an average duration of 51 ± 5.26 days during 2014-15, and fruiting duration varied greatly among the individuals with a coefficient of variation (C.V. % =23.37). Fruiting presented the correlation distinctly with temperature and rainfall while in one-month lag only rainfall was significantly related with fruiting ( Table 1). Availability of seasonal water had a strong impact on fruiting indicating that there was a significant relationship between onemonth lag rainfall and fruiting. Fruits mature by the end of July. The fruit is a single seed which remains hanging through a small thread-like structure (Image 1f) for a few days before dehiscence. Each seed bears a conspicuous crimson red, fleshy caruncle at the tip.

DISCUSSION
The phenological observations on the species and climatic characteristics of the study site suggest that the A. malaccensis is a seasonal flowering and fruiting tree species. Correlation of phenological characteristics with naturally occurring climatic events may be best documented by the pattern of leaf-fall. The greatest tendency of leaf-fall practice coincides with the relatively dry season during January-February. The timing of leafshedding is strongly correlated with a gradual increase in day-length, temperature, and solar insolation. This finding is in conformity with Mishra et al. (2006) who stated that maximum leaf-fall occurs during the dry period in tropical forest trees. Further, leaf-fall during this

12069
period appears to be an inherent strategy to minimize water loss and maximize photosynthetic activity during monsoon season (Rivera et al. 2002;Hamann 2004). Leaf production and flushing of A. malaccensis start towards the end of the dry season. Short dry period, maximum temperature, and increased day length triggered the emergence of new leaves during the premonsoon period. The advantages of peak leaf initiation during pre-monsoon period could possibly be explained by the fact that it was to take advantage of the long rainfall period by the fully expanded foliage on trees (Singh & Kushwaha 2005). Maximum temperature and photoperiod as driving factors for leaf initiation have been reported for other tropical trees (Rivera et al. 2002;Singh & Kushwaha 2005). Saikia & Khan (2012b) observed that leaf flushing in A. malaccensis in home gardens starts in March and continues up to October. Species that produce leaf during the rainy season tend to have shorter periods of leaf production because this period of abundant water will normally last only for a few months. The species that greatly depend on rainfall for initiation of the leaf would also be expected to show rapid leaf growth in order to maximize photosynthetic activity during the rainy season (Reich 1995), and this type of behavior is quite common in plants growing in seasonally dry environments (Wright et al. 2002). Leaf initiation in the early rainy season is attributed to the end of the long dry season and also due to the joint action of increasing day length and temperature (Kushwaha et al. 2010). Rivera et al. (2002) have implicated that increasing day length acts as the inducer of flushing which is relevant to the leaf phenology in A. malaccensis.
Flowering during the pre-monsoon season can be viewed as a strategy to make flowers more visible to pollinators and supply food sources during the poor periods of floral resources (Murali & Sukumar 1993). Species flowering during the pre-monsoon period can be capable of storing water in sufficient quantities to permit flowering even in the absence of rainfall (Borchert 1994). A. malaccensis as a dry season bloomer showed a significant positive correlation with photoperiod as observed for trees in tropical dry forests by Borchert et al. (2004). Soehartono & Newton (2001) reported flowering and fruiting in A. malaccensis growing in botanical gardens of Indonesia from April-September. Beniwal (1989) found flowering in March and fruiting in the middle of June in plantations of Arunachal Pradesh, northeastern India. Saikia & Khan (2012b) observed initiation of flowering from mid-February to May following fruiting from May and ending in August.
A. malaccensis concentrated peak fruiting during the wet season, producing dry fruits with small seeds. In tropical forests, fruiting during the rainy season may have evolved to ensure dispersal of seeds when soil water status is favourable for seed germination, seedling growth, and survival (Kushwaha et al. 2011). The requirement of moisture level for the proper development of fruits indicates that the decrease of soil water status reduced the rate of enlargement and final size of these fruits. During the wet season, availability

12070
of high moisture level also favours germination and establishment of seeds. The flowering phenology observed in A. malaccensis is reported in Psidium guajava and Vatica lanceaefolia growing in the home gardens of Barak Valley, northeastern India (Das & Das 2013). Synchronization of flowering during a particular season appears to be under the control of the prevailing climatic condition of that season (Singh & Kushwaha 2005). Maximum flowering activity during the premonsoon period may be related to the high insect population as pollen vectors in tropical forests. Further, seasonal flowering strategy observed in A. malaccensis may be a strategy to escape from seed predation on a timely basis. In A. malaccensis, fruiting initiation during the rainy season is indicative of a close relationship between rainfall and fruiting, as the rainfall factor acts as a cue for reproductive phenology, especially in dry tropical forests as stated by Griz & Machado (2001). Further, autochorous seed dispersal in this species is also probably related to the humidity factor as it has an influence on fruit dehiscence. Fruit dehiscence during the monsoon season may enable the plant to escape from seed predators and produce seedlings for continued survival (Hamann 2004). Fruiting during the rainy season in tropical forests evolved to ensure dispersal of seeds and this could be attributed to utilization of available soil water for seed germination and seedling establishment (Singh & Kushwaha 2006). Tropical trees have adopted a systematic strategy so that there is adequate development time from flowering to seed dispersal so that seeds are released during the rainy period (Stevenson et al. 2008) when germination is most likely to be induced and seedlings start growing with a low probability of drought.
In the present study, the duration of flowering was longer (59.52 ± 6.36 days) and fruiting was shorter (48.78 ± 5.25 days) in A. malaccensis during the two years of study. This longer duration of flowering can be viewed as a difference in time taken for the formation to the maturation of buds. The short duration of fruiting is advantageous for the plant to mature fruits during the rainy season due to the availability of highest precipitation. This flowering and fruiting duration does not agree with the reports on the same in the tropical montane evergreen forest of southern India (Mohandass et al. 2016). Further, the duration of these two phenophases appears to be influenced by the changes in day length, temperature, sunshine hours, and precipitation associated with the season (Bawa et al. 2003).

12071
The present study indicates that the vegetative and reproductive phenological events in A. malaccensis display a general annual flowering and fruiting pattern with a peak in these events during the pre-monsoon and monsoon seasons. Temperature and precipitation (by themselves) do not show any influence on leaf initiation but cumulatively show influence on leaf initiation. Availability of seasonal water had a strong impact on fruiting indicating that there is a significant relationship between one-month lag rainfall and fruiting. It seems that changes in temperature and rainfall pattern have a pronounced effect on the phenology of A. malaccensis. This information may be used as a baseline for further evaluation of phenological variations for this vulnerable tree with reference to climate change. The study suggests that there is a need to develop a long-term monitoring strategy on phenological aspects of A. malaccensis in order to understand the impact of climate change on phenology.

INTRODUCTION
Human-wildlife conflict (HWC) traditionally arises from a rivalry or antagonism between humans and wildlife , or between people over wildlife and/or its management . The former typically emerge from territorial proximity between humans and wildlife, conflict over the same resource or even a direct threat to human wellbeing. People-people conflicts on the other hand, characteristically emerge when disparate values clash in the face of management decisions (Nyhus 2016).
While humans and wildlife have a long history of interaction, the frequency and complexity of conflicts has grown in recent decades, mainly because of the exponential increase in human populations and concomitant human footprint, expansion of some wildlife distributions (Chapron et al. 2014), as well as a frequent inability of institutions that are meant to mediate such conflicts to respond effectively (Anthony et al. 2010). HWC often pits disparate values against one another (Tajfel 1981;Kellert 1993;Young et al. 2010) and demands attention from economic, legal, social and environmental policy makers (Knight 2000;White et al. 2009;Nyhus 2016). Moreover, these values influence people's behaviour towards wildlife and institutions responsible for conservation (Manfredo & Dayer 2004;Manfredo 2008;Dickman et al. 2013). Therefore, HWCs are best managed through a shared understanding of the broader context of the situation, necessitating both natural and social science approaches (Dickman 2010;, and often utilizing workshops Reed et al. 2009;WWF 2015). This shared understanding is of key importance to finding long-lasting solutions to such conflicts, and to avoid potential escalation Anthony et al. 2010).
The identification, differentiation and meaningful involvement of all affected stakeholders and the mapping of their goals and opinions on the resource(s) in question and potential mitigation strategies are crucial before crafting or implementing management decisions (Reed 2008;Reed et al. 2009;White et al. 2009;. Recent cases where stakeholder analysis and participatory strategies have been applied with the aim of conflict resolution range from conflicts concerning Hen Harriers Circus cyaneus in Scotland ), Eastern Imperial Eagles Aquila heliaca in Hungary (Kovács et al. 2016), to livestock depredation by large carnivores in South Africa (Anthony & Swemmer 2015). Before engaging with wider actors, however, it has been suggested that organizations first develop a coherent understanding of the issue within their own institution and/or with institutions that share common values, serving to enhance channels of communication and catering to a unified backing of wider stakeholder engagement (FAO 2002), particularly in contexts where complex multi-actor governance models exist (Funtowicz et al. 1999). Thus, there has been greater realization by management authorities that focusing on both wildlife and human dimensions together is critical, as opposed to treating them separately, even within organizations (Clark et al. 1996;Baruch-Mordo et al. 2009;Treves et al. 2009).

Mauritius Fruit Bats
Bats are the only mammals native to the Mascarene Islands, consisting of Mauritius, Réunion and Rodrigues (Fig. 1). Historically, three fruit bat species occupied these islands: one is now extinct (Pteropus subniger), leaving one species each on Mauritius (P. niger, Kerr 1792) and Rodrigues (P. rodricensis). Once widespread over Mauritius, the Mauritius Fruit Bat population decreased considerably from its original population due to habitat loss and degradation, cyclones, invasive alien species, climate change and illegal hunting (Hutson & Racey 2013;Vincenot et al. 2017). Due to lack of major cyclones for well over a decade, however, the population has increased, thus shifting its IUCN Red List status from Endangered (2008) to Vulnerable (in 2013), which was also based on an assurance that culling would not be considered (Hutson & Racey 2013). Assessing the status of this bat species has been complicated by discrepancies in population estimates yielded by different census techniques, ranging in 2015 from ~50,000 by the Mauritian Wildlife Foundation (MWF), to ~90,000 by the National Parks and Conservation Service (NPCS) (Hansard 2016). In October 2016 a population estimate was undertaken by the NPCS in collaboration with the Forestry Service and MWF, using both evening dispersal counts and direct counts, which are believed to be more accurate (Kunz 2003), yielding an estimate of ~62,000 individuals.
Mauritius Fruit Bats are considered keystone species as they provide critical pollinating and disseminating services (Vincenot et al. 2017). They are mainly nocturnal or crepuscular, and roost chiefly in primary forests or areas containing a mixture of native and introduced plant species. Bats may travel long distances to visit orchards and garden fruit trees for exotic fruits when their natural food supplies are limited (Aziz et al. 2016). The reported level of fruit damage by bats has ranged from 9.3%

12075
and 11.4% on Lychee Litchi chinensis and Large Mango Mangifera indica trees, respectively (Oleksy 2015), to over 50% of Lychee trees (Hansard 2016). Despite a subsidized tree netting scheme, and due in part to alleged significant increases in fruit damage by bats and the lobbying of fruit growers for its lethal control, the government passed the Native Terrestrial Biodiversity and National Parks Act in November 2015, legalizing the culling of any wildlife that has attained 'pest' status. Consequently, a highly controversial government sanctioned cull was conducted in November-December 2015, with a reported 30,938 bats culled (Hansard 2016). A second official cull was conducted in December 2016 in which 7,380 bats were killed (Hansard 2017). This culling largely contributed to a subsequent uplisting of the species from Vulnerable to Endangered by the IUCN in 2018 (Kingston et al. 2018). The Mauritian Fruit Bat cull has pitted a number of stakeholders and their values against one another (MWF 2016). This sensitive situation, involving disputed bat population and fruit damage estimates, and the role of culling to alleviate fruit damage, requires joint actions from fruit growers, local organizations and governmental bodies, and also calls for a deeper understanding of the conflict by conservation organizations to provide a basis for developing effective management strategies. In order to improve this understanding, we utilized a workshop targeted specifically to conservation organizations to map how they perceive the conflict landscape by identifying the scope and scale of human-bat interaction issues associated with relevant actors in Mauritius, and to propose strategies to navigate forward. It specifically aimed to explore intra-stakeholder complexities involved in preventing and resolving conflicts and fostering coexistence between people and bats, acknowledging data deficiencies along the way.

METHODS
As an overarching framework, but restricted to organizations with similar values, we utilized Lasswell's (1971) general strategy for problem solving that undertakes five 'intellectual tasks': (1) clarify the goals of people involved or affected by the problem and its solution.
(2) describe the history and trends of the problem (including empirical data on the biophysical and cultural context of the problem and relevant processes such as decision making).
(3) understand the relationships of all factors that have influenced, affected, or caused the problem.

12076
(4) project the trajectory, severity, and consequences of future developments.
(5) invent, appraise, and select alternatives. In addition, we incorporated a number of relevant sub-frameworks drawing from examples from the literature on the targeted theme.
To implement this framework, we convened a oneday workshop for MWF and NPCS staff in May 2017. All staff who were directly or indirectly engaged with the fruit bat conflict were invited, and included organization directors, project managers, and field-level officers. Participants were provided with a pre-workshop package consisting of a schedule, and group member allocation along with assigned readings and tasks. The workshop consisted of introductory sessions on the background of human-wildlife conflict and its mitigation, the Mauritius Fruit Bat, and an outline for group exercises (see Appendix 1). These were followed by three parallel group sessions, the composition of which was based on maintaining equally sized groups and personnel expertise and awareness. Each group had a number of iterative tasks to complete including an ongoing assessment of knowledge gaps and/or research needs (Table 1). A group-appointed rapporteur recorded notes on both a flip chart and notebook, then communicated findings back to all workshop participants at the end of the day. Notes for each group were subsequently compiled and categorized according to pre-defined conceptual codes according to the sub-frameworks used, and were largely descriptive in nature.
Secondly, in June 2017, we administered a follow-up questionnaire to all workshop participants consisting of two parts. First, we captured information on length of involvement in their organization, and perceived knowledge of the fruit bat conflict prior to the workshop. Second, we requested their opinion as to (i) whether the workshop met their expectations, (ii) assisted them to see and appreciate the wider conflict landscape, (iii) what was particularly useful with the workshop, and (iv) how it could be improved. Univariate statistics were computed using SPSS ver. 22 (IBM Corp 2013). Qualitative responses to questionnaire items were analysed using emergent content coding (Stemler 2001).

RESULTS
A total of 20 participants representing staff from MWF (18) and NPCS (2) attended the workshop, and contributed to its results. Below, we present findings from the group exercises, including coded indications of knowledge level of the respective concept/stakeholder by workshop participants (bold = well known; normal font = somewhat known; italics = unknown).

Group A
Group A participants identified 18 stakeholders involved in human-bat interaction, ranging from highly influential and important fruit-growers, to leisure parks holding relatively little influence and power in the

12077
conflict. In addition, there were a number of 'unknown' actors of varied influence and importance, including the role of religious organizations (Table 2). Group A also explicated a number of current interactions between stakeholder groups, outlining the perceived level of hostility, stakeholder activity, and current expressions of the conflict. These interactions represented public, government, and NGO sectors (Appendix 2), ranging from varied responses to media campaigns, frustration with current mitigation strategies (tree netting), and conflicting government mandates across ministries.

Group B
Group B participants identified 13 environmental and 17 social risk factors associated with the human-bat interaction, along with knowledge gaps (Appendix 4), which would necessitate targeted investigation before and during extended dialogue with other stakeholders. Environmental risk factors included the influence that climatic conditions (e.g., cyclones), forest health and composition, fruiting season, fruit tree pruning and protection, and bat behaviour have on the conflict. Social risk factors were also varied, ranging from market disparities, powerful lobbying interests, media influence, distrust, and folklore.
Further, Group B participants assessed both the perceived and real costs of conflict, with an indication of level of knowledge concerning these factors (Appendix 3). Most discrepancies between perceived and real costs of the conflict were economical in nature, including those relating to fruit tree maintenance, the price of fruit, and the potential impact on tourism if Mauritius' world renowned reputation in conservation is seen as eroding.

Group C
Group C was assigned to outline what policy and management measures are, and potentially could be, leveraged to mitigate conflict between fruit bats and the various stakeholders. Results are outlined in Appendix 5, conforming to the same scheme of level of knowledge about the effectiveness of policy and management options. Measures identified by workshop participants included extended tree netting and pruning service to fruit growers (both backyard and larger orchards), initiating decoy crops, increased bat awareness campaigns, stricter control on fruit prices, and expanded research on bat ecology.

Workshop Assessment
Fifteen (75%) workshop participants completed and returned the questionnaire, representing both the MWF (13), and the NPCS (combined response from 2 participants). Length of time employed in their respective organizations ranged from 0.5-20 years (x ̅ = 7.9, sd = 5.63). On a 10 point scale (1=very low to 10=very high), prior knowledge regarding the fruit bat conflict ranged from 5 to 9 (x ̅ = 7.4, sd=1.39), and was greater among those who held higher positions within their organization and/or those who worked directly with the bat issue.
On a scale of 1 to 10 (1=not at all to 10=completely), participants rated whether the workshop met their expectations, and opportunity was granted to explain their response. Scores ranged from 3 to 8 (x ̅ = 6.0, sd =1.65). Those with higher scores noted that the workshop helped to (i) increase appreciation of the wider legal, social, and institutional aspects of the conflict, (ii) provide intra-agency exposure and awareness of the conflict complexity, and (iii) provide a much-needed platform to hear other agency views (and challenges) associated with the conflict.
Workshop participants were asked more specifically to rate how well the workshop helped them to see the wider social and management aspects of the issue both within their own organization and with another

12078
conservation organization. Scores ranged widely (x ̅ = 6.1, sd = 2.53), with those with higher scores noting how well the workshop helped them to understand the breadth of stakeholders directly or indirectly involved in the conflict, to see underlying issues, and recognize political dimensions of conservation conflicts (including public and political resistance). Others commented on how well the workshop disclosed how even two proconservation organizations can have disparate opinions on how to manage such conflicts. For those who perceived themselves to have moderate experience in conflict management and resolution, the workshop did not add much to their understanding of the breadth of social and management facets of this particular conflict. Participants believed the workshop was particularly useful in that, before extending dialogue with other stakeholders, it: • involved group sessions within conservationoriented stakeholders in which issues could be openly discussed and debated; • encouraged wider understanding of models by which conservation conflicts can be framed; and • provided pre-workshop readings and introductory sessions which facilitated improved framing of workshop tasks.
Finally, ideas on improving such workshops included eventually expanding stakeholder representation, extending its duration to 3-4 days, developing a common strategy to move forward, providing a broader array of theories, case studies, and bat research, and allowing for prolonged inter-group discussions on findings.

DISCUSSION
Our initial findings demonstrate that inter-and intraorganizational workshops designed to map conservation conflict landscapes, before extending dialogue with a wider spectrum of stakeholders, can be of immense value in a number of ways. First, a broader array of stakeholders can be acknowledged at the onset, each with varying degrees of influence and importance which, in turn, allows for more strategic and prioritized engagement (IFC 2007). Second, conflict nodes between stakeholders and their intensity can be identified, facilitating more nuanced strategies for addressing particular conflict dimensions, and allowing for a more appreciative inquiry of the conflict typology that currently exists, or may develop in the future. Third, delineating environmental and social risk factors including both perceived and real conflict costs can assist the designing of more complex mitigation strategies including more focused awareness raising campaigns, as well as leveraging existing and potential policy and management options (Dickman 2010).
Finally, by recognizing where knowledge gaps exist, conservation organizations can channel appropriate resources towards research needs and/or solicit support from other stakeholders for both research and appropriate monitoring.
We believe initial conflict mapping workshops of this nature can elevate pan-organizational understanding of conservation conflicts and build consensus by identifying, appreciating, and eventually communicating the positions and values of stakeholders, and their justification. Of course, this is only the first step in realizing true resolution, as other stakeholders may have vastly different or contrasting opinions, attitudes and values concerning the conflict ). Moreover, we recognize that in-house workshops represent only one of many options for participatory and non-participatory processes which can be used to address conservation conflicts (Reed et al. 2009). Nevertheless, our assessment demonstrates that organizations would benefit from in-house workshops in order to develop an inclusive and coherent approach to engage other stakeholders before taking that next step.
Our findings also suggest that such workshops should extend to a minimum of three days, eventually involve more stakeholders, and generate more tangible outcomes in terms of mitigation strategies. We recommend, however, that such preliminary workshops be restricted to a limited number of stakeholders sharing similar values, involving relevant personnel who interact both directly or indirectly with other stakeholders (including the general public) in HWC issues. Doing so prompts a more collective and nuanced strategy for navigating forward as an organization, and for reducing the risk of conflict escalation. In our case, the fate of an entire species, and the services it provides, may depend on it.

Mapping of human-wildlife interaction of Mauritius Fruit Bat
Anthony et al.

MWF x Public High
MWF educates and raises public awareness on bat conservation to improve attitudes towards bats and their conservation · little/no change in attitude · people still unaware of importance of bats · some public participate in saving injured bats, but majority do not.
Fruit growers x MWF High MWF: Provide info to farmers (netting / pruning) but with mixed results · Fruit growers believe MWF 'do not understand their problems' · because MWF has low influence, the opinion of growers is strongly influencing government decision · damage level not based on scientific results

Government x Civil society High
Government mandated to both protect wildlife and farmer interests, and wants to appease voters through approving bat culls · Ministry of Agro-Industry and Food Security has conflicting mandates (wildlife protection and food production) · Conservation and animal welfare NGOs lobby for bat protection Press x Public Minimal Press reporting on bat issue to public · Press provide media coverage (good info) · Press communicate wrong or distorted info, leading to negative public opinion · Press has been ambivalent: strongly encourage culling before cull, and after cull was more nuanced · Public: blame MWF for high population of bats · Encourages illegal culling

Appendix 3. Perceived and real cost factors identified by workshop participants
Note: bold = well known; normal font = somewhat known; italics = unknown

Perceived Costs of Conflict Real Costs of Conflict
· cleaning under fruit trees · removal of fruit trees leads to less fruit · ▲ price of fruit · availability of fruits: less fruit on backyard tree to eat or give to neighbours; market fruit usually available, just expensive · bat extinction will deprive future generations of wildlife · less fruit leads to ▼ revenue · sleep disturbance · affects tourist industry negatively if bats culled · Mauritius international reputation as biodiversity champion tarnished · ▼ bats leads to ▼ forest regeneration · psychological impact of culling · physical injury from installation/ removal of netting · boycott in export of Mauritian fruits if bats culled · subsidy of netting · cost of culling · cost of surveys (bat population and questionnaire) · cost of nets and installation and removal · cost of pruning · cleaning under trees · other methods to keep bats away (guarding, lights, fire crackers, shooting)

12081
Appendix 5. Existing and potential policy options identified by participants, and level of knowledge regarding these options Note: bold = well known; normal font = somewhat known; italics = unknown

Existing Policy/Management Proposed Policy/Management Relevant Considerations
Netting subsidy (75%) scheme Full canopy netting · extend netting scheme to more than (current) half of all trees in orchards <2 acres, and 5 for backyard growers. · service provider to train (i) teams in community allowing free net installation for backyard growers, and (ii) orchard staff which would increase uptake and effectiveness in orchards

Appendix 4. Environmental and social risk factors identified by workshop participants
Note: bold = well known; normal font = somewhat known; italics = unknown; arrows indicate effect between variables

Environmental Risk Factors Social Risk Factors
Environmental characteristics/land use and management • ▼native forest extent leads to▲ bats' reliance on exotic fruits • ▼forest quality leads to ▲ bats' reliance on exotic fruits • ▲ urbanisation leads to ▼tree abundance/density • ▼cyclones leads to ▲ bat population • lychee season leads to less native food source for bats • ▲ commercial fruit growers leads to ▲ fruit which, in turn, leads to ▲ bats Inequality and power • lobbying by influential groups (fruit exporters, NGOs) • political decision based on popularity (backyard growers as large voting base) • press influence (affects public perception) • lack of education leads to less informed judgement • control of fruit price (for economic gain) and/or unfair trade practices (limiting supply) can lead to and maintain inflated fruit prices Human Behaviour • ▲ pruning and netting effectiveness leads to ▼bat damage to fruits • ▲ capacity/willingness to utilize netting leads to ▼ bat damage to fruits • ▲ bat culling leads to ▲ illegal killing of bats by public • orchard owners: ▲ resources for tree protection leads to ▲ tree protection • backyard growers: ▼ resources for tree protection leads to ▲ bats feeding in backyards Vulnerability and Wealth • Backyard growers and small scale planters cannot afford netting nor installation which leads to inability to reduce damage by bats and birds • Returns from harvest significant percentage of annual revenue for orchard owners • Physical incapacity to install nets may deter use

Behaviour and management of conflict-causing species • bats non-territorial, thus damage by bats widespread • ▲ protection by law leads to ▲ bat population
Distrust and animosity • people upset because of fruit predated by bats, noise and faeces/ residue • annoyance over legal protection of bats leads many Mauritians to consider bats as 'pests'

• Distrust towards conservationists ('do not understand farmers losing fruits'; 'all they want to do is to protect bats')
Beliefs and Values • as bats are believed to be nocturnal, their habits are unknown • hunting is considered normal (acceptable) killing • perceptions of bats due to folklore ('evil creatures') • bats are considered by some to be edible, thus it is more acceptable to kill them (cultural for some sections of the population)

• religions do not promote killing • superstitions (bats 'dark and evil', 'vampires', 'get entangled in people's hair')
Humans have had an uneasy relationship with wild animals since the dawn of human evolution as they preyed upon, were prey, and competed with wild animals (Knight 2013). As human populations dispersed across the globe during the Late Quaternary they were the main driver of the extinction of large mammalian fauna (Bartlett et al. 2016). For example, human dispersal into North America during the Pleistocene probably caused the extinction of 35 genera of mammals (Faith & Surovell 2009). Sites with mass killing of megafauna by Palaeolithic hunters have been documented across continents (Barnosky et al. 2004).
The remnants of this rivalry can be perceived in cultural practices of traditional societies, and cultural beliefs involving dangerous animals such as werewolves, vampires and others which are a metaphor of the pervasive human belief of the 'beast within', 'bestial', etc. Violent killing by humans is denounced in the idiom of natural predation where criminals and enemies are termed 'jackals', 'wolves', etc. (Knight 2013).
This rivalry is closely inter-twined with human expansion into wilderness habitats (Knight 2013). Even today, this continues, and tends to be the most intense in settlements at the forest edge, in many cases due to colonization of forests by frontier populations (Rudel & Roper 1997).
The modern depiction of this rivalry is termed human-wildlife conflict (HWC), defined by the IUCN World Parks Congress  as "…when the needs and behavior of wildlife impact negatively on the goals of humans or when the goals of humans negatively impact the needs of wildlife. These conflicts may result when wildlife damage crops, injure or kill domestic animals, threaten or kill people".
In terms of usage, a conflict is typically defined as 'an active disagreement between people with opposing opinions or principles; or fighting between two or more groups of people or countries' (https://dictionary. cambridge.org/dictionary/english/conflict, viewed 08-04-2018).
Therefore, HWC suggests 'conscious antagonism between wildlife and humans' and implies that wildlife act consciously and often places wildlife entities on an equal footing with people in the role of combatants, even though they cannot represent themselves in the political sphere against people (Raik et al. 2008;Peterson et al. 2010). The use of this term in which wildlife are blamed for every encounter or incident places culpability entirely on the wildlife side of the equation, suggesting that wildlife assert their interests to undermine human goals . This promotes antagonism towards wildlife that can exacerbate the problem, hinder resolution and can result in people directing their anger, frustration on wildlife with potentially adverse conservation outcomes for endangered species (Peterson et al. 2002;Brook et al. 2003;Redpath et al. 2015). Besides the HWC approach is often ineffective because it has led to purely technical solutions being proposed that may have worked in particular circumstances but have not addressed the underlying issues (Redpath et al. 2015). For example translocating wildlife to resolve 'conflicts' has often failed to achieve its objectives due to lack of understanding of the species' behaviour and/ or the underlying issues (Athreya et al. 2011). Often the increasing human population densities and expansion into forest areas that result in such incidents (Newmark et al. 1994) are not addressed.
To the best of my knowledge, the earliest reference to 'conflict' between wild animals and humans was in the early 1990s (Sukumar 1991;Newmark et al. 1994). Before the term 'conflict' became popular, more precise terms such as crop raiding and livestock depredation were used to describe incidents involving wildlife (Jhala 1993;Oli et al. 1994). The use of this term has increased over time: Treves (2009) carried out a Google search based on the keywords "human AND wildlife AND conflict OR depredation OR damage", and Google Scholar returned 3140 hits between 1992-1999, and 8060 between 2000 and 2007.
Its popularity stems from its simplicity and ease of usage to describe a diversity of situations involving wildlife. Thereby it has become a buzz word used to amplify conservation initiatives, create funding opportunities, increase research productivity and create a sense of urgency that limits the array of potential solutions that may arise when the situation is more accurately described (Peterson et al. 2010). In many cases the damage or threat is exaggerated for gains, for example, in Japan the scale of concern over bears greatly exceeds the actual damage done by the animals (Knight 2013).
To understand in what context this term humanwildlife conflict is used in conservation literature, Peterson et al. (2010) carried out a meta-analysis of 422 case studies of HWC and found that over 95% of the 422 cases referred to animal damage in some form to (i) resources such as food, (ii) property, or (iii) attacks on people. Only one case represented a typical example of 'conflict' where there was human retaliation against Magpies (Cracticus tibicen) that repeatedly attacked specific humans that they considered threats (Warne & Jones 2003). Less than 4% related to human-human conflict such as those between conservationists and other parties on how wildlife should be managed (Peterson et al. 2010).
Thus human conflicts are often projected onto wildlife (Knight 2013), and may in fact be a symbolic vehicle for expression of social conflict between people at the local, national and international levels, such as between conservation movements and developers or between people and protected area management termed 'human-state conflict' (Knight 2013). In Japan, widespread concern about the bear is balanced by local support for the bear, based on the premise that given the extent of human colonisation into bear territory, it is humans that are problematic with regard to the bear and not vice versa (Knight 2013). These human-human conflicts need to be distinguished from human-wildlife impacts.
Therefore, more precise description of the issue at hand may lead to better solutions. For instance crop raiding is a widespread problem in forest fringe areas where the cultivation of edible crops attracts wild herbivores. When crop raiding is described as crop raiding instead of as 'conflict' then better solutions may emerge depending on the location, the crops cultivated and the herbivores in question. Whereas the conflict terminology is provocative and emotional which could create more problem than it solves, particularly if sensationalised by the media (Bhatia et al. 2013;. In many cases rodents and monkeys cause more economic loss to people than large mammals such as bears, elephants and the great apes which take a disproportionate amount of the blame (Knight 2013).
Therefore in the Anthropocene, where the rate of species extinction is accelerating (Sanderson et al. 2002;Barnosky et al. 2011) there is a growing realization that humans need to move beyond their past history which has framed the narrative regarding wildlife. Finding ways to increase tolerance and coexistence with wildlife ) is needed to slow down the population declines of iconic megafauna. If not, future generations will no longer have the privilege of sharing their world with large charismatic animals. There are many examples of human tolerance to wildlife and acceptance of certain levels of loss of crops and livestock (Knight 2013). To protect human interests, however, innovative solutions need to be explored in forest fringe areas to ameliorate the situation.
This term which is problematic, semantically incorrect and which masks the underlying complexities of particular situations, needs to be avoided. It is well The Snow Leopard Panthera uncia is distributed throughout northern Nepal along the boundary with China over an area of 22,625.34km 2 (Aryal et al. 2016). Snow Leopards research and conservation activities have been focused mostly on protected areas (PA) of Nepal (Jackson & Alhborn 1989;Oli 1991;Kyes & Chalise 2005;Khatiwada et al. 2007;Ale 2007;Ale et al. 2007;Devkota 2010;Karmacharya et al. 2011;Wegge et al. 2012;Aryal et al. 2014Aryal et al. , 2016. To date, research institutions, conservationists and students have given very low priority to surveys and monitoring of Snow Leopard outside Nepal's PAs, in spite of the existence of suitable habitat. A few Snow Leopard studies based on sign surveys were conducted outside PAs in Humla District (R. Jackson in litt. 2003; Khatiwada & Ghimirey 2009;FoN 2014) and Bhajang District (FoN 2014). Of the total potential habitats preferred by Snow Leopards in Nepal, 65% were located outside PAs (Jackson & Ahlborn 1990). More recent species distribution modeling, however, shows that only 34% of the total Snow Leopard habitat in Nepal lies outside the PA network, comprising a significant portion of potential unprotected habitats located in western Nepal, including Humla, Bajhang, and Bajura (Aryal et al. 2016).
The paucity of surveys and the resulting lack of reliable data for unprotected lands presumably increases leopard vulnerability to local extinction from poaching and retaliatory killing (R. Jackson pers. comm. 2017), and lack of effective government actions against poaching and wildlife trade aggravates the situation. Therefore, urgent collection of baseline data on Snow Leopard presence, distribution and status outside of PAs is warranted. To date, information on distribution and status of Snow Leopard is mostly based on anecdotal evidence and few sign surveys (Jackson & Hunter 1995). Due to challenging habitat structure, site accessibility and the cryptic nature of this felid species, direct sighting such as detection is virtually impossible (Jackson et al. 2006;Ale & Brown 2009). Camera trapping is a preferred method for detecting such rare and elusive species. In addition, sign surveys and local residents' interviews generally lack scientific rigor for reliable status assessment (Jackson et al. 2006). Considering these facts, we conducted the first ever camera-trapping survey of Snow Leopards in the Limi valley of Humla District of Nepal. The findings of this study will make an important baseline for future monitoring and help government and conservation partners in conservation planning.

Materials and Methods Study Area
Limi Valley is located in the trans-Himalayan steppe environment in Humla District of Nepal (Fig. 1) Wolf Canis Lupus himalayensis (Werhahn et al. 2017).

Site selection and camera trapping
We conducted a camera-trapping survey in the Limi Valley during July and August, 2015 to conduct a rapid assessment of Snow Leopard status. We divided the total study area into 5x5 km 2 grid cells considering 24km 2 (11-37 km 2 ) as the minimum home range of the Snow Leopard in western Nepal (Jackson 1996) and placed one or two cameras in each grid. Due to time limitations we could not cover an entire study area and placed camera-traps only in the least disturbed locations where we frequently observed Snow Leopard signs. We set up 17 Bushnell Trophy HD cameras in 17 locations (Fig. 1) within 10 grids following preferred habitat features like ridgelines, travel corridors and marking sites (Jackson et al. 2006). A minimum distance between camera-traps was kept at 2.5km, and they were placed at elevations from 3,092-4,608 m. Since the purpose of this preliminary survey was to assess the presence of Snow Leopard, we did not account for spatial autocorrelation of camera-trap sites. One camera-trap was lost thus only data from the remaining 16 camera-traps were used in the analysis.

Data management and analysis
We managed camera-trap images using the program described by Sanderson & Harris (2013). We defined capture events as independent images of a species at a location captured at least 60 minutes apart. We computed the relative abundance index (RAI) for all the mammals, birds and human recorded during the surveys. RAI was expressed as the number of independent images per 100 trap-nights (Sanderson & Harris 2013;Jenks et al. 2011).

Relative Abundance Index (RAI)
A total survey effort of 195 trap-nights recorded 39,839 images from 16 camera traps. We sorted out and analyzed images of mammals, birds and human (n = 1,110) and discarded the ghosts, i.e., false images and also the unknown images (n = 38,729). The survey recorded 95 independent images of four species of mammals: Snow Leopard (n = 6 images), Blue Sheep (n = 6 images), Beech Marten (n = 14 images), and Pikas (n = 69 images). In addition to mammals, images of different species birds (n = 711 images) and human (n = 99 images) were also recorded. The birds had the highest RAI of 45.00 captures/100 trap-nights followed by Pika (30.45), human (13.64), Beech Marten (5.45), Snow Leopard (2.73), and Blue Sheep (2.73) ( Table 1).

Photographic evidence of Snow Leopard
Snow Leopard was captured in two different locations (Til Gomba and Chhongerche) and the time to first detection was 16 days. We used pelage patterns (Jackson et al. 2006;Ale et al. 2014) to identify different individuals from camera-trap pictures. The dorsal side of the tail, forelegs, right hind legs and right flanks were particularly informative. We recorded total of three individuals, including two adults and one sub-adult (Image 1). In Til Gomba, individual 1 was recorded on 28 July 2015 from 00:06:05 h to 00:36:04 h (1a in Image 1) and individual 2 was captured on 30 July 2015 from 23:16:40-23:22:42 hr (2a in Image 1). The adjoining location Chhongerche also produced two events. Individual 3 was captured on 28 July 2015 from 17:14:36-17:14:37 hr (3b in Image 1) and individual 2 was captured on 31 July 2015 from 03:37:45-03:38:45 hr (2b in Image 1). In Til Gomba, individuals 1 and 2 were captured while scraping on the surface. In Chhongerche, individual 3 was captured while scentmarking a rock (3b in Image 1). Both these locations are dominated by broken terrain and cliffs providing good habitats for Snow Leopard and the middle elevations contain open areas with good grazing grounds for Blue Sheep.

Discussion
This is the first camera-trapping survey in the Limi Valley that successfully recorded Snow Leopards during a short survey period in two locations. An opportunistic survey carried out in early 2000 suggested the presence of Snow Leopards in the Limi Valley (R. Jackson pers. comm. 2017). A study conducted in 2007 in the adjoining area which included some parts of the Limi Valley also showed a good abundance of Snow Leopard signs (Khatiwada & Ghimirey 2009    The need to undertake biodiversity studies is accelerated by the rapid destruction of forests, particularly in the tropics including the Western Ghats. The number of small carnivore species reported from different protected areas of Kerala vary, e.g., 11 species from Parmbikulam Tiger Reserve (Sreehari & Nameer 2016), nine species from Eravikulam National Park (Nikhil & Nameer 2017), and Wayanad Wildlife Abstract: A study on the small carnivores in Silent Valley National Park (SVNP), southern Western Ghats, Kerala, India was conducted from September 2015 to April 2016, using the camera trap technique. Seven species of small carnivores were recorded during the study. The most common species of small carnivore of SVNP was Viverricula indica (44%) followed by Paradoxurus jerdoni (20%) and Herpestes vitticollis (17%). The other small carnivores found at SVNP were Herpestes fuscus (7%), Prionailurus bengalensis (6%), Aonyx cinereus (5%) and Martes gwatkinsii (1%). P. jerdoni and M. gwatkinsii are endemic to the Western Ghats. We discuss the niche partitioning among small carnivores in SVNP.
Sanctuary (Sreekumar & Nameer 2018). The first record of Martes gwatkinsii from Parambikulam Tiger Reserve was reported by , and the social behavior, feeding habits and activity pattern of Martes gwatkinsii were reported from Pampadum Shola National Park (PSNP) (Anil et al. 2018).  reported the presence of Herpestes smithii in Parambikulam Tiger Reserve and Chinnar Wildlife Sanctuary, and Herpestes fuscus in Parambikulam Tiger Reserve and Eravikulam National Park. The lack of details on small carnivores from the Silent Valley National Park (SVNP), except on the sighting records of M. gwatkinsii (Christopher & Jayson 1996) and habitat characterization of M. gwatkinsii (Balakrishnan 2005), prompted the present study. We report the status and distribution of small carnivores in SVNP.

Materials and Methods Study Area
Silent Valley National Park is part of the Nilgiri Biosphere Reserve and has an extent of 237.52km². The

Camera Trap Survey
Digital scout cameras having passive infra red sensors for heat and motion detection (Cuddeback Attack model C1) were used for the current study. Camera trap stations were placed in the west coast tropical evergreen forest (1A/C4) and southern montane wet grasslands (11A/C1/DS2). Overall a 100 trapping stations (Fig. 1) were identified based on the presence of the indirect evidence of the small carnivores (Mudappa 1998). The camera traps were set at a height of 30cm above the ground and at least 250m apart from each other (Sreehari & Nameer 2016;Nikhil & Nameer 2017;Sreekumar & Nameer 2018). The cameras were set up in default mode with the time-delay between pictures as fast as possible in daytime and the time-delay of five seconds between pictures during night time. The camera trap locations were marked using Garmin GPS eTrex 30. The cameras were kept open for 24 hours a day. The date and time of exposure were automatically recorded by the camera on the images, as and when the images were taken. At each trapping stations, each camera was opened for 15 days. Thus, a total of 1,500 camera-trap days, monitoring 36,000 hours were carried out in the Silent Valley National Park. The data analysis was done using the statistical packages such as the XL STAT (Version 2016.03.30846), andPAST (Hammer et al. 2001).
Microhabitat parameters were documented at each of the camera trapping sites. Microhabitat parameters that are crucial for the survival of the small carnivores, such as, canopy height (clinometer), canopy cover (visual estimation), height of shrubs (stems <10cm girth at breast height) and ground vegetation (herbaceous plants <50cm in height, measured with tape), litter depth (average of four measurements taken around the trap using a calibrated probe), and basal area of trees >30cm girth, densities of shrubs (within 2m radius), trees, climbers, buttresses and canes, and distance to the nearest large tree (measured with a tape to a tree >60cm girth), frequency of natural hollow in the trees etc., were taken in the camera trapping sites. At each camera trap site, a circular plot having a dimension of 5m radius was taken and 100 such plots were enumerated for the microhabitat parameters listed above. Thus, a total of 7,850m 2 area was sampled. The relationship between these microhabitat variables on the distribution of small carnivores in the study area was analysed using discriminant analysis.

Results and Discussion
We recorded seven species of small carnivores in SVNP representing four families such as Viverridae, Herpestidae, Mustelidae and Felidae. This comprises two herpestid, mustelid, and viverrid species each, and one felid species ( Fig. 2; Table 1).
Of the total 607 photographs of all the mammals (20 species) obtained, 165 images (seven species) were of small carnivores. The most common species recorded was Viverricula indica (72, 44%) followed by Paradoxurus jerdoni (33, 20%) ( Table 1). The camera trap success rate of small carnivore was 10.90%.

Family Viverridae
Out of the three species of viverrids (Nameer 2015) of Kerala, V. indica (Image 1) and P. jerdoni (Image 2) are found in SVNP. V. indica was the most common species of small carnivores, photo-captured 72 times (Fig. 3), between an altitudinal range of 900-1,200 m, and from the rainforests as well as from the grasslands. In the previous studies done in the Kerala part of the Western Ghats in Parambikulam Tiger Reserve (Sreehari & Nameer 2016) and in Wayanad WS (Sreekumar & Nameer 2018), V. indica was the most abundant species of small carnivore. Mudappa (2002), however, had reported that the V. indica is the most common small carnivore in the drier forests of the southern Western Ghats and rare in the tropical wet evergreen forests.
Paradoxurus jerdoni (Image 2) is an endemic small carnivore restricted to the rainforests of the Western Ghats (Rajamani et al. 2002). P. jerdoni was the most common small carnivore in Kalakkad-Mundanthurai Tiger reserve followed by V. indica (Kumar et al. 2002). A total of 33 captures of P. jerdoni were obtained during the study period from SVNP, and there was a single direct sighting in the night (06 October 2015) from Sairandri (Fig. 3). All the captures of the P. jerdoni were from the tropical evergreen forest and between the altitudes of 900-1,200 m.
During the present study, 27 captures were obtained, and there were also two independent sightings of the species from Sairandri (07 October 2015) and another from Panthanthode (24 February 2016) (Fig. 4).
H. fuscus (Image 4) is found in the forests of the southern Indian hill ranges at 900-1,850 m (Mudappa 1998) and is also seen in Sri Lanka (Phillips 1984). The previous records of this species from the Western Ghats are from Parambikulam Tiger Reserve Sreehari & Nameer 2016), and Eravikulam National Park Nikhil & Nameer 2017). During the present study, 10 captures were obtained between an altitude range of 900 and 1,200 m (Fig. 4). In southern India, H. fuscus is found from an altitude range of 492 and 2,032 m and is reported from different hill ranges of the Western Ghats such as Coorg, Nilgiri Hills, Palni Hills, Anamalai Hills, High Wavy Mountains and Agasthyamalai Hills Mudappa & Jathanna 2015). Of the two species of the otters seen in the Western Ghats, only the Aonyx cinereus could be found in SVNP that was captured five times during the current study (Fig. 5,Image 6), and all the captures were above 1,000m. The only previous records of the Aonyx cinereus from the Western Ghats were from Eravikulam National Park (Perinchery et al. 2011;Nikhil & Nameer 2017), Anamalai Tiger Reserve (Prakash et al. 2012) and Wayanad WS (Sreekumar & Nameer 2018). There is, however, a record of this species from the northern Western Ghats in Maharashtra (Punjabi et al. 2014).

Family Felidae
Prionailurus bengalensis (Image 7) is the only small cat recorded during the present study and 10 camera trap images were obtained from the SVNP between an altitude range of 900 and 1,200 m in evergreen forest Park (Nikhil & Nameer 2017), and Wayanad Wildlife Sanctuary (Sreekumar & Nameer 2018).

The microhabitat preference of the selected small carnivores in Silent Valley NP
The differential preferences for microhabitat variables in the study area by small carnivores were examined using discriminant analysis (Table 2). This helps to understand whether there is any niche partitioning between and among the species concerning the habitat variables studied. The pair-wise Fisher's distances (blue cells) and associated P values (red cells) clearly show that there is no significant difference in the clusters, thus indicating that the selected small carnivores show no significant niche partitioning (Table 3, Fig. 6).

Conclusion
The Silent Valley National Park which constitutes one of the few pristine rainforests of the Western Ghats is a home for many endemic and threatened species including small carnivore fauna. Even though some of the high-altitude areas of the SVNP could not be surveyed due to logistical reasons, it supports seven species of small carnivores. The disturbed habitats are vulnerable to incursions by more widespread species at the cost of restricted range species. The absence of widespread species of small carnivores in the national park indicates the intact habitats of SVNP.    (Daniels 2008;Rajashekar & Venkatesha 2008; Bhattacharya et al. 2010; Dandapat et al. 2010Ghosh et al. 2010;Khera et al. 2010; Dhanya 2011Sethi & Vashisth 2013). Urbanization and industrialization, leading to the loss of suitable foraging locations and nesting spaces in urban and rural areas has contributed much to the declining sparrow populations (Cramp et al. 1985;Rao 2000;Summer & Smith 2003;Robinson et al. 2005;Pineda et al. 2013). Besides, many reasons have been suggested for the decline of sparrow populations such as lack of old fashioned buildings and weedy gardens (Monika 2005), changes in agricultural practices, predators (Summers & Smith 2003;Vincent 2005;Shaw et al. 2011), competitions (Vincent 2005Khera et al. 2010, Mason 2006, disease (Vincent 2005), environmental pollution (Chamberlain et al. 2005;Vincent 2005;Balmori & Hallberg 2007;Dhanya 2011), electromagnetic radiation (Balmori & Hallberg 2007), lack of insect availability, nest sites, substratum, nesting materials, food items and roosting sites (Chamberlain et al. 2005;Vincent 2005;Mason 2006;Bohner & Witt 2007;Klok et al. 2008;Dhanya 2011). Use of unleaded fuel results in methyl nitrite during combustion, which is harmful for soft-bodied insects, as they form the major diet for sparrow chicks. This was also suggested to be a threat for sparrows (Summer-Smith 2007).
Hopping near grocery shops, picking up fallen grains and clearing out insect pests, they were once common sights in our markets and urban areas. The birdwatchers and nature enthusiasts in the district were concerned about the decline in one of the commonest urban birds. Also, no studies in the population of sparrows were conducted in Kannur District. Hence this study was taken up by an NGO, the Malabar Awareness and Rescue Centre for Wildlife-Kannur, aimed at documenting sparrow populations in the district, the threats faced, the perception of the public towards sparrow conservation and possible conservation action.

Materials and Methods
A press release was published in all leading newspapers in the district (with the details of the project and contact numbers of the volunteers) to identify the potential sparrow inhabiting areas in Kannur District (Fig.  1). The areas communicated by the respondents were visited during 08:00-11:00 hr and 15:00-18:00 hr and the numbers of sparrows sighted were recorded using point count method (Bibby et al. 1998) from March to July 2015.
An open-ended type questionnaire survey was done in markets, rural and urban towns in the district, targeting shop keepers, workers, and the local people. All the respondents were between 35-55 years old. The questionnaire survey was conducted in the local language (Malayalam).

Result and Discussion Status of sparrows in Kannur District
A total of 35 sites were surveyed and 553 sparrows were recorded in Kannur District (Fig. 1). Compared to

12100
urban towns, more number of sparrows were found in smaller towns and in rural areas as reported earlier.
Among the 140 individuals interviewed 89% stated to have seen the sparrows earlier (in the past), but only 56% of the people stated to have seen them in the present. This indicated that 33% of the respondents had seen the sparrows earlier but not at present. Hence, this data suggested that there has been a decline in the populations of sparrows in the district.

Perspective of public regarding sparrows
Sparrows are generally believed to be useful to the public (Fig. 2). As major pest control agents, they pick up insects and worms (39%) from food grains; they clean the surroundings by pecking on thrown out food materials (21%) and maintain an ecological balance (15%). Hence, they were believed to play a vital role in maintaining the health of the ecosystem. Twenty percent of the respondents felt that sparrows have an aesthetic value due to their cheerfulness and tweeting sounds to make the surroundings lively, and some respondents (5%) believed that sparrows were a good omen when they

12101
for the birds), water scarcity, the destruction of nests and the loss of habitat and nesting spaces due to urbanization (Bokotey & Gorban 2005), had decreased the endurance of sparrows in towns. These were the other reasons suggested by the public in the questionnaire survey. Various conservational plans were suggested by the respondents (Fig. 4) of which the provision of nest boxes (47%) were the most recommended suggestion to enhance sparrow population. Also, spaces for sparrow nesting have been thought of during modernisation of buildings; especially in towns (5%). Planting roosting trees (14%) of small heights of less than 5m are found to support and host good numbers of roosting sparrows (Dhanya & Azeez 2010). Provision for feeders with grains (11%) and water bath (5%) could also help to regain sparrow population. Other suggestions (13%) were to maintain eco-friendly and clean environments by minimizing pollution and by reduced use of pesticides. A section of respondents who believed mobile towers to be the major cause of the decline in the population, had suggested to minimize construction of mobile towers (5%) as well.

Efforts for conservation of sparrows
During the study, students and the public were involved in population assessment of sparrows in the district. Awareness programs were conducted in local colleges and schools to educate students about the importance of sparrows. Mass participation of public was assured by conducting sparrow photography competitions, during which the public had spent time to watch and observe sparrows in the city.
Sparrows are expected to build nests in any available places including nest-boxes (Shaw et al. 2008) and studies showed that artificial nest boxes can enhance the population of sparrows in urban and sub-urban areas (Chethan 2012). Hence, with the preliminary knowledge of the status of sparrows in the district and with the suggestion from the public, we decided to create and fix nest boxes for sparrow conservation. A total of 100 wooden nest boxes were fixed in identified sparrow inhabiting sites in the district, which later was found to be effective. Hence with this preliminary study, we were able to map out some of the sparrow population in the district and understand its status and potential threats. Moreover, the project had created a network of students and members of the public who stood for conservation of sparrows in the district. Furthermore, with the continuous support from stake holders, we could create a bigger network of public, students, shopkeepers, vegetable sellers, etc. who can be utilized to monitor the sparrow population. Along with provision of more nest boxes and by planting short roosting trees and maintaining urban gardens, sparrow population in the district could be enhanced.
Ghats of Kerala and Wayanad in particular (Shaji & Easa 1995Easa & Basha 1995). None of them could locate the specimens from Day's locality, which may either be due to the rarity of the species or unavailability of the exact location of Day's collection. Dario neela was described in 2018 based on the collections from Wayanad and the study also suggested that Day's collections were in fact Dario neela and not Dario urops (Britz et al. 2018).
A recent survey in the Kabini watershed of the Cauvery basin in Wayanad helped in documenting the further distribution of this species in the Kerala part of the Western Ghats (Fig. 1). We provide new information on the habitat and ecology of the species. Specimens were collected and a few were reared in captivity to study the social and reproductive behavior of the species.

Materials and Methods
Fishes were collected mostly by sieving with clothes and mosquito nets. A few specimens were fixed in 10% formalin and transferred into 70% ethanol for permanent storage. Ten individuals with two males and eight females were selected, grouped and reared under captivity to study the social and breeding behavior of the species. Physical parameters of the stream habitat, viz, substrate, canopy cover, stream temperature, stream type, and stream width were recorded at each collection site. Substrate was classified as bedrock, boulder, cobble, pebble, gravel, sand, and mud. Canopy cover was measured using a spherical densiometer.
Temperature was measured using a submersible digital thermometer (Mextex-Multi Thermometer).
Elevation, altitude and latitude of sampling locations were recorded using global positioning system meter (Garmin-GPS 72H) (±10m). Morphometric measurements were taken point-to-point to the nearest 0.1mm using digital calipers.

Results
The morphological data of the species is given in Table 1.
There is an appreciable difference in the morphological attributes of male and female (Images 1 & 2). Males were larger than females with well-developed fin rays. The pectoral fin length and caudal fin length were found to be higher in males when compared with females. Caudal peduncle length and pre-anal length were higher in females.

Distribution in Wayanad
The specimens were collected from seven different  (Table 2). All these are lower order streams draining into Kabini, an east flowing river (Fig. 1). All these collection localities were sections of the streams flowing through evergreen forest patches and the species could be recorded from an elevation range between 700-1,050 m.

Habitat and Ecology
The flow rate was minimal and the collection localities were characterized mostly by pools and runs. The substrates were constituted by sand, mud, gravel, pebbles, boulders and bed rocks. Boulders and sand dominated in the study cites altogether (80%) and the average constitution of the pebbles and gravels were of 10% in the sites of collection. All the collection sites were characterized by heavy litter fall over various substrates. The water temperature varied from 17.9-22.6 0 C. The canopy cover varied from 70.38-97.43 %. The physical parameters of various sites at the time of collection are given in Table 3. Shallow regions of the streams were heavily occupied by vegetation like Lagenandra sp., Colocasia sp. etc. which acted as fish cover. D. neela specimens were mostly collected from submerged leaf litter, tree roots, submerged vegetation and Ochlandra clumps. The co-occurring species recorded from various collection localities were Devario cf. malabaricus, Barilius gatensis, Haludaria fasciata, Schistura cf. nilgiriensis, Schistura semiarmata and Neolissochilus wynaadensis.
Although not a shoaling species, D. neela tend to live in small groups of 5 -10 individuals in the wild, with a welldefined territory for each individual. The groups are typically formed by two males and several females defending territories near each other. In such groups, one male always dominated the other invariably. Larger dominant males defended larger territory when compared with the inferior male and females. The dominant males were darker colored (black with a bluish tinge) and the colour disappeared when the animal was stressed (Image 3). The distal margin of the fins were an iridescent blue-green. The inferior males were greyish brown in colour with faded brown vertical bands which fade away soon after preservation. The distal margin of fins were similar in colour and appearance to that of dominant males. Females were smaller than males and had a beige-brown colour on the body with irregular vertical bands and the fins were devoid of the iridescent colour. The  caudal spot was present in all specimens (Image 1).

Breeding behavior
The spawning behavior of Dario neela was studied in a well-planted glass tank, which mimicked the natural ecosystem from where the fish were collected. A large tank (190L) was used to study the social and breeding behavior of the species. Tank water properties like pH and temperature were almost similar to the stream condition. The tank was planted with locally available aquatic vegetation like Lagenandra toxicaria and Cabomba sp. Hiding spaces were provided in the form of large rocks and artificial caves. The substrate was set with sand and leaf litter collected from the streams.
Dominant males mostly occupied the heavily planted regions of the tank with enough hiding spaces and aggressively defended the territory (typically 40cm 2 ) from the inferior male and females. Females were found to hide under leaf litter and rocks, while the inferior male defended smaller territory when compared to the dominant male.
Male Dario neela chased the females and displayed his bright colors to attract the females, shivering and shaking his body. Gravid females when ready to spawn followed the male to his territory. Mating is usually attained by a spawning embrace, where the mates wrap around each other (Fig. 2). The female released eggs after two to three fake matings and the males fertilized it. Eggs were usually scattered in dense vegetation, in caves or under leaf litter. After spawning, the male chased away the female preventing her from eating the eggs. Eggs were found to hatch within 48 hours in normal conditions, with the larvae hiding in vegetation immediately after hatching. Males were found to protect their territories aggressively after spawning.

Discussion
Dario neela is the only species of badid fish known from the Western Ghats of Kerala. The present study described the distribution of this rare species in the Wayanad part of Nilgiri Biosphere Reserve. Dario urops was described based on the specimens collected from a stream draining into Barapole tributary of Valapattanam River in Karnataka and the specimens collected by Francis Day (1875-1878 from Wayanad were originally assigned to this species (Britz et al. 2012). A recent study (Britz et al. 2018), however, revealed that the specimens collected from Wayanad is in fact a separate species and named it as Dario neela, referring to its blue coloration.
Dario neela could not be recorded from any other locations outside Wayanad and in Wayanad the species could be recorded only from the east flowing streams draining into the Kabini River, suggesting that the species is endemic to the Kabini River System in the Western Ghats of Kerala.
The population trends of this species is not well studied and the present study indicates that the species is a habitat specialist, recorded only from less disturbed clear water mountain streams which might have accounted for its rarity.
The antiparallel spawning embrace has been reported in Badidae, Anabantoides, Nandidae and Channidae by Barlow et al. (1968). A similar type of spawning embrace was observed in Dario neela. The present success in captive breeding helps in developing future ex situ conservation plans for this species.  Aquarium fish keeping is one of the oldest hobbies in the world and next only to photography in popularity (Das et al. 2005;Singh & Ahmed 2005). The high demand for ornamental fishes has made them an important component of the world fish trade (Andrews 1990;Singh & Ahmed 2005;Tlusty et al. 2013); however, the aquarium industry is sighted as both positively (socio-economic and livelihood Abstract: Fifty-eight ornamental fish species belonging to five orders, 13 families and 36 genera occur in Himachal Pradesh. The dominant family is Cyprinidae (46.55%) followed by Nemacheilidae (15.51%); Sisoridae, Poeciliidae, Osphronemidae (9.89%); Cobitidae (5.17%); Amblycipitidae, Ambassidae, Badidae, Gobiidae, Helostomatidae, Cichlidae and Characidae (1.72%). Of the 58 species, 27.58% are exotic and have been mainly imported for aquarium keeping. The exotic species are being introduced in the region without any regulation, subsequently turning invasive and threatening the indigenous fauna. Thus, there is a need for developing scientific guidelines and regulatory mechanisms for importing exotic aquarium fishes. On the other hand, the breeding and culture of indigenous fishes can be a profitable venture, provided there is an availability of a standardized breeding technology. Such an enterprise will go a long way in conservation of native fishes, improving livelihoods as well as raising the socioeconomic status of local communities.
Himachal Pradesh is located in the western Himalaya between 30.36667-30.2 0 N and 75.78333-79.06667 0 E and altitudes ranging from 320-7,000 m. It has four physiographic zones (i) Shiwalik, (ii) Lower Himalayan, (iii) Higher Himalayan, and (iv) Trans Himalayan zone. The state has enormous potential for fishery in terms of aquatic resources with approximately 300km of perennial rivers, 775km of seasonal rivers (Satluj, Beas, Ravi, Chenab and Yamuna), 60,000ha reservoirs and 2,000ha, lakes and ponds including two Ramsar Sites, Pong Dam and Renuka Wetland.
A review of literature reveals that although much work has been undertaken on the general fish resources of Himachal Pradesh (Day 1875(Day -1878Hora 1937;Menon 1962Menon , 1987Menon , 1999Bhatnagar 1973;Seghal 1974;Tilak & Hussain 1977;Sharma & Tandon 1990;Johal et al. 2002Johal et al. , 2003Dhanze & Dhanze 2004;Mehta & Uniyal 2005;Sharma 2014), no information is available on the potential aquarium fishes. For the first time, an attempt has been made to produce a comprehensive list of ornamental fishes recorded from the waters of Himachal Pradesh.

Material and Methods
Fishes were collected from the Beas, Yamuna, Satluj, Ravi and Chandra Bhaga rivers in Himachal Pradesh and their tributaries using a combination of gears including cast net, scoop net and hand net. Fish specimens were preserved in 4% formalin solution and deposited in the High Altitude Regional Centre, Zoological Survey of India (ZSI), Solan, and identified using standard literature (Talwar & Jhingran 1991;Jayaram 2010). Conservation status of the fish species is based on the IUCN Red List of Threatened Species (2017) and nomenclature is as per Eschmeyer et al. (2016). Six fish species viz. Barilius modestus Day, 1872, B. sacra Hamilton, 1822, Raiamas bola (Hamilton, 1822, Schistura himachalensis Menon, 1987, Paraschistura punjabensis (Hora, 1923) and Triplophysa microps (Steindachner, 1866) which were not collected in the present study have been included based on records in published literature (Tilak & Hussain, 1977;Dhanze & Dhanze, 2004;Mehta & Uniyal 2005;Sharma 2014).

Results and Discussion
A systematic list of 58 ornamental fish species belonging to five orders, 13 families and 36 genera from various ecosystem of the state is summarized in Tables 1 and 2 The exotic Cyprinus spp. has commercial value but due to its hardy nature, beautiful colour and disease resistance are used as aquarium fishes till they reach their fingerling stage. These exotic fishes have also entered the various natural water bodies (streams of Beas and Satluj River) of the region and are well established in the Pong dam, Govind Sagar Reservoir and Pandoh Dam.
Native fishes recorded as ornamental (Table 1) are hillstream species that are threatened by various anthropogenic stresses, viz., over exploitation, illegal fishing, invasive species, habitat loss and destruction due to channelization of water, and upcoming hydroelectric projects. Breeding and farming of these ornamental fish species can help in the restoration and conservation of indigenous fish fauna. Further, it will be a promising alternate livelihood for the farmers of the region. Thus the ornamental fish trade will go a long way to provide employment in the region.
The conservation status following the IUCN Red List of Threatened Species (2017) has revealed that among the 42 native fish species, 30 species (71.4%) come under the 'Least Concern' (LC) category; two species (4.8%) under 'Data Deficient' (DD) category and 10 species (23.8%) under 'Not Evaluated' (NE) category.
About 90% of the freshwater ornamental fish exported from India are wild caught indigenous species (Silas et al. 2011). Raghavan et al. (2013) stated that more than 1. 5 million freshwater fish belonging to 30 threatened species were exported from India to Europe, US and other Asian countries from 2005 to 2012. Without any focus on conservation and sustainable use, freshwater fishes are collected from nature as an open access resource for the Most of the Ramsar sites in India are recognized either because they are representative natural wetlands (Group A, Criterion 1) (Islam & Rahmani 2008) or based on information available on avian diversity and their abundance (Group B, Criterion 5 to 9) (Islam & Rahmani 2008). Besides birds (Ramakrishna et al. 2006;Kumar 2008Kumar , 2009, faunal studies of the Ramsar sites of India are largely restricted to fishes Dua & Chander 2009;Saikia & Saikia 2011). In India aquatic insect diversity is never used as criterion for recognition of Ramsar site. Among the 26 Ramsar sites Abstract: Odonate diversity of Nalsarovar Bird Sanctuary, a Ramsar site in Gujarat, was studied between January 2015 and July 2017. A total of 46 species belonging to two suborders, six families, and 27 genera were recorded, which included 14 species of Zygoptera (damselfly) and 32 species of Anisoptera (dragonfly). Out of the 46 species, 40 species are new records for the Nalsarovar Bird Sanctuary. The record of Enallagma cyathigerum Charpentier, 1840 in Gujarat needs verification. Need to monitor changes taking place in Odonata species composition after influx from Narmada canal at Nalsarovar is emphasized.
Keywords: Damselflies, diversity, dragonflies, odonates, protected area, wetland in India, odonate diversity of only three wetland sites are known (Palot & Soniya 2000;Kirti & Singh 2000;Singh et al. 2017). The present study on the Odonata was aimed to generate a basic database of the aquatic fauna of Nalsarovar Bird Sanctuary.
Biodiversity status and list of important species dependent on candidate wetland is a prerequisite for declaring any wetland as 'Ramsar site' (Ramsar Regional Center -East Asia 2017). Published literature on the fauna of NBS is either incomplete (GEER Report 1998) or erroneous (Kumar 2009 published list of odonates is yet available (GEER Report 1998;Kumar 2009), this study was taken up to fill up the lacunae in our knowledge.

Materials And Methods
Study area NBS consists primarily of a 120.82km 2 area of the much larger natural low-lying area, and is situated about 64km to the west of Ahmedabad City, in the central part of Gujarat State, India (Fig. 1). It is considered a freshwater ecosystem, which gets inundated during south-west monsoon (June to September); its water becomes saline during summer (March-May) due to evaporation (GEER Report 1998). NBS is located at 22.81790 0 N and 72.04530 0 E, and it receives water from two rivers: Brahmini and Bhogavo (Singh 2001). It is a natural lake, originated by the elevation of the land between present-day Gulf of Khambhat and Gulf of Kachchh during the late quaternary period, thereby breaking the connection between the two gulfs. The area of Nalsarovar has remained a shallow depression with water depth ranging between 1.5-2.0 m as the land did not rise up to the height of mainland Gujarat or Saurashtra (Prasad et al. 1997).
The lake has around 300 small and big islands. The basin of the lake is elongated and nearly elliptical with gentle sloping margins. All around the basin, there is sandy to clayey shoreline. The water temperature rises up to 35 0 C during the month of May and falls below 15 0 C in January. The average rainfall is about 580mm (Kumar et al. 2006).
It is the largest wetland bird sanctuary in Gujarat, and one of the largest in India. About 48 species of phytoplanktons and 71 species of flowering plants, including 30 species of aquatic macrophytes, are recorded in this natural lake (Kumar et al. 2006). For establishing Narmada Canal network in Saurashtra, a site known as Bhaskarpura was created as storage reservoir of Narmada Canal water (GEER Report 1998). Narmada canal water started flowing in 2003, and since then the canal water is percolating to the Nalsarovar via Vadala depression (Fig. 1).

Sampling methods
This study was carried out between January 2015 and July 2017 in post monsoon period. Odonates were closely observed at the shallow edge of the wetland with naked eyes and occasionally using 7X35 binoculars. We also surveyed the marshy area of the adjoining villages and the small island areas within the lake (Images 10 & 11). Voucher specimens of some species were collected using insect collecting net. The specimens were either preserved in 70% alcohol or kept in envelopes, labeled with details of the collection. Odonates were counted using point count method (Small shire & Beynon 2010; Rohmare et al. 2016) on the peripheral area of the sanctuary. Occurrence status was worked out on the basis of the frequency of occurrence as follows: >50%common, 25-50%-Uncommon, 5-25%-Rare and <5%very rare.
The species were identified with the help of photographic guides (Subramanian 2009;Nair 2011;Kiran & Raju 2013) and a suitable taxonomic book (Fraser 1933;1934;1936). The scientific names are adopted from the revised nomenclature by Subramanian & Babu (2017).

Results
A total of 46 species belonging to two suborders and 27 genera under six families were recorded in and around NBS. Fourteen species of Zygoptera (damselfly) and 32 species of Anisoptera (dragonfly) were recorded. In this study, both Zygoptera and Anisoptera were represented by three families each (Table 1; Images 1-9) .
At Nalsarovar, the most dominant families were Libellulidae with 26 species and Coenagrionidae with 10 species, respectively. Remaining families had two or three members each (Table 1). On 27 September   (Table 1).

Discussion
Prasad (2004) reported seven species from the Nalsarovar during a general faunal survey of Gujarat State by the Zoological Survey of India. In the present study, 46 species of odonates were recorded. Hence, 40 species are additions to the list of Odonata of NBS. Enallagma cyathigerum Charpentier, 1840 recorded by Prasad (2004) was not encountered in the present study. In Gujarat, Enallagma cyathigerum Charpentier, 1840 was reported from Anandpura Village (Mandal Tahsil) and Nalsarovar of Ahmedabad District by Prasad (2004). This species is widely distributed in Europe and Northern Asia with only two records from India. In India, this species has been recorded from Kashmir and West Bengal (Fraser 1933;Srivastava & Sinha 1993). Rohmare et al. (2015Rohmare et al. ( , 2016 had reported the species from central Gujarat, however, it was considered a misidentification (Rathod 2017). Rathod (2017) had not encountered this species anywhere in Gujarat State.
The Odonata diversity of Thol Bird Sanctuary (Mokaria 2015) and Pariej wetland  of Gujarat have been reported recently. Both the wetlands are located in central Gujarat within the direct distance of 55km (Thol Bird Sanctuary) and 65km (Pariej Wetland) from Nalsarovar and is fed by Narmada canal waters. Their reported species diversity was only 15 (Thol) and 29 (Pariej). The differences in reported numbers of species might be attributed to the structural differences and area of the wetland, and the relative efforts made by the researchers. In India, the faunal study of Odonata is done only on three Ramsar sites. The present study on Nalsarovar is fourth one. Comparison of the diversity of Odonata with other Ramsar sites across the country may not be meaningful as several factors such as biogeographic zones, the intensity of study, and period influence the reported diversity. However, a comparison with Keoladeo National Park would be worth as it is located in the adjoining state (Rajasthan) and in the same biogeographic zone (semi-arid zone).
Only 37 Subramanian et al. 2008;Dholu 2015). Long-term ecological studies of such Indicators of the ecosystem should be undertaken for wetland monitoring and conservation. This study was done after the implementation of Narmada Canal, and it is unfortunate that no information about the odonate fauna of Nalsarovar before implementation of Narmada canal project is available. Nalsarovar is rich in macrophytes with at least 30 species (GEER Report 1998), and when the water permanency of Nalsarovar increases with the percolation from Narmada canal, the aquatic vegetation is likely to flourish. In such a situation, the impact on Odonata species at NBS needs to be monitored. range of this species.
As all Indian species under Storena are misplaced in the genus as per WCS (2017), we looked out for characters of S. gujaratensis matching with other genera. We noticed presence of hook shaped dorsal cymbial flange, large tegulum, moderately long and thick embolus and a few of these characters typically found in Suffasia Jacque, 1991. The members of Suffasia can be differentiated from other genera by presence of dorsal cymbial flange overlapping palpal tibia, presence of cymbial lateral pit and swollen venter of the abdomen in male, whereas females can be distinguished by the epigyne structure with frontal entrance openings and the course of the copulatory ducts (Jocqué 1991;Jocqué 1992). Although all characters of male did not match with the Suffasia especially absence of cymbial lateral pit and AME not being small but Suffisia also shows high variation in the palp structure within the genus. Further, we found the S. gujaratensis male palp structure closely resembled Suffasia attidiya Benjamin, 2007 from Sri Lanka by having large embolus and tegulum and RTA short. Therefore, here we transfer Storena gujaratensis to Suffasia. So far, only three species of the Suffasia, viz., S. ala Sen et al., 2015, S. keralaensis Sudhikumar et al., 2009 andS. tigrina (Simon, 1893) are reported from India (WSC 2017). In this paper, we provide additional morphological characters for Suffasia gujaratensis comb. nov. along with illustrations, description of male and natural history notes. High variation is noticed amongst Suffasia spp. and therefore a revision of this genus is urgently needed.
Carapace covered with grey hairs, dense in anterior half. Chelicerae with hairs, presence of sclerotized chillum with hairs. Sternum having triangular extensions which correspond with slight concavities in coxae, uniformly covered with bristles and hairs, bristles with warty appearance, integument rough having net like pattern. Maxillae wider at base gradually narrowing posteriorly, bordered ridge on prolateral surface. Labium longer than wide, arrow shaped. Abdomen dorsally covered with brown hairs, thin scutum ventrally covering book lungs and epigynal area. Ventral abdomen uniformly covered with brown and black color hairs and bristles intermixed. Tracheae small and broad covered with brown hairs, situated just in front of spinnerets, colulus with two hairs. Anterior spinnerets long with two segments, Posterior spinnerets two segmented with apical segment dome shaped.
Palp: Tibia digitiform with two short blunt apophyses and elevated retrolateral margin with distinct process. Cymbium with lateral fold, distally truncated with a notch in the middle; dorsal cymbial flange hook-shaped overlaying palpal tibia. Conductor large flap-like seen distally; tegulum large, sclerotized with very short and blunt tegular extension; embolus slender long and originates at 6 o'clock position of tegulum on short embolic base. Comments: Suffasia gujaratensis comb. nov. possess some unique characters which have not been previously reported in Suffasia like presence of distal cymbial notch, dorsal cymbial hook, large tegulum and absence of cymbial lateral pit. Though, S. attidiya possess large tegulum, moderately long embolus and short RTA, diagnosis for the genus is weak as variations within Suffasia spp. is high (like in the structure of cymbial flange, presence-absence of cymbial lateral pit; tibial apophysis structure, tegular shape and size; embolic length, etc.). Therefore, we consider Suffasia as a species complex and multiple specimens of both the sexes for all the species will help in assigning robust diagnosis for the genus.
Distribution: Jambughoda Wildlife Sanctuary, Gujarat, India. Tea, Camellia sinensis L. (O. Kuntze) plantation provides habitats for thousands of insect species including pests and their natural enemies like parasitoids and predators. The immense value of predators in pest suppression has been well understood by entomologists and there is a renewed interest in biological pest suppression. Classical biological control or periodic inundative release of natural enemies has been most effective in cropping systems where large-scale use of insecticides or their ecologically disruptive practices are minimal (David & Easwaramoorthy 1988). Green lacewings are known to have tolerance to commonly used pesticides (Bigler 1984), and they are relatively easy to rear in captivity (Tulisalo et al. 1984). Laboratory culture and augmentation of Mallada desjardinsi (= boninensis) is feasible through Corcyra cephalonica larvae and artificial diet (Vasanthkumar et al. 2012).
Mallada desjardinsi is an important predator of pests such as mealy bugs and aphids. In tea they are important predator of Red Spider Mite Oligonychus coffeae (Vasanthkumar et al. 2012).
Distribution study in India showed Bengaluru, Karnataka, to have the highest density of M. desjardinsi population (26.6% and 5.05 ± 0.108 per plant) in the areas sampled (Boopathi et al. 2016).
The larvae of green lacewings are important predators largely used as biological agents. They feed on pest thrips, aphids, scales, caterpillars, and spider mites infesting a variety of plants (McEwen et al. 2001). Adults of green lacewing generally are not predatory and feed on nectar, pollen or honeydew while a few of them are predatory (Coppel & Mertins 1977).
Generalist predators prefer to take prey of whatever size they can handle (Dong & Polis 1992;Finke 1994). If these prey include younger conspecifics or other predators, then control of the herbivore population is not guaranteed. Intraguild predation (IGP) is a combination of the killing and eating species that use similar, often limiting, resources and are thus potential competitors (Polis & Myers 1989); however, there has been little research on the IGP of chrysopids, especially on the M.
the increase in number of species of lacewings correlates with increasing winter temperature, while they decrease with increasing summer precipitation (McEwen et al. 2001).
Preference was mostly for abandoned egg sacs and spiderlings. In the field it was observed that the fully grown larvae of M. desjardinsi, roamed near and consumed 22% eggs of C. nigra, 20% of N. mukerjei, 11% of C. argyrodoformis and 10% of Cyrtarachne sp. and below 10% of the other spider species (Fig. 2).
In the laboratory studies Vanitha et al. (2009) showed that when egg sacs were offered to fully grown larvae of Chrysoperla, they consumed eggs of Oxyopes javanus and Clubiona drassodes, whereas no consumption was observed when the mother was present. The popula-   (Fig. 3). Noppe et al. (2012) reported that green lacewing, Chrysoperla carnea was the superior intraguild predator, winning 88.9% when the experiment was repeated in petri dishes without plant material, regardless of whether green bugs or eggs of Ephestia kuehniella Zeller were offered as focal prey.
Intraguild predation by M. desjardinsi can be regarded as a mechanism for enabling survival when the red spider mite prey is scarce. Nevertheless, the intraguild predation of M. desjardinsi may reduce pest suppression in tea plantations. remnants of tepal bases.

Phenology
Flowering and fruiting are probably from March to September.

Conservation status
The species is presently known only from Uttara Kannada, Karnataka and Idukki, Kerala (Fig. 2). The population size of this species at Thodupuzha, Idukki is small with very few reproductively mature individuals. Apart from this, based on herbarium information a population is present at Uttara Kannada District, Karnataka. Even though there are no recent collections from this locality, we cannot rule out the probability of finding it there. There are less than 25 mature individuals in the subpopulation at Thodupuzha, but we are not sure about the number of mature individuals or the quality of the habitat at Uttara Kannada. Due to its limited distribution knowledge from several botanical studies in the Western Ghats, we assume it is highly restricted and found in a few severely fragmented locations with a plausible threat that could impact the status of the species. We recommend immediate surveys in the type locality to determine its conservation status.
explains the research methodology adopted and the constraints faced during the study. The ban on shahtoosh shawl production is good news for the conservationist but not for locals. Tourism and handicrafts are the main sources of income in the state. In this context, the author answers the question: 'Can the sustenance of people be sacrificed for nature conservation?' Though it depends purely on the situation, the author strongly argues that the decision should not be taken in a hasty manner. Given due considerations of the demographics of the state, the author concludes that some do reap the benefit of the ban: illegal traders. The trade in shahtoosh continues even after 2002.
Chapter Three presents a figure which might really indicate the size of the trade network (p.44). The chapter concludes that the failure of conservationists to understand regional politics and socio-economic relations is a hurdle to their success. The author fears that the very purpose of the ban might be defeated by illegal trade.
Chapter Five, aptly titled "The Micropolitics of the Ban on Shahtoosh: Costs and Reparations", sums it all up and it's my favourite chapter. Some of the bitter truths of the shahtoosh trade are pointed out here: (i) the public is not aware of the source of wool and they are mostly carried away by false propaganda (ii) the weavers are miserably exploited by a few influential traders (iii) the trade for chiru wool is done by barter system and the list goes on. I spoke to some Kashmiri students on campus and they were also not aware of the facts. This was really intriguing and this book will come as an eye-opener for them. The book provides other useful insights, as well.
The second half of the book focuses on joint forest management. After reading the previous chapters, one can really understand ground realities and attitudes. As with the shahtoosh ban, the implementation of the Joint Forest Management was not very successful. The author, again, points a finger at caste politics and the disparity in economic standards. The style adopted is similar to the one she used for the shahtoosh issue. She begins with an introduction to forest management in the state (Chapter Six) and follows it up with details of the implementation of the JFM in Chapter Seven. Finally, Chapter Eight deals with the micropolitics of implementing joint forest management.
The second part of the book might not be that interesting to wildlife biologists. However, in the last chapter, the author points out the similarity in the two conservation actions taken in the state. Though the shahtoosh ban hampered the economy, the JFM attempted to aid and revamp livelihoods. Both attempts failed due to various reasons, including militancy. However, the take-home message for readers, especially conservationists, is that conservation policies should be based on local realities for long term stability. Otherwise, with the State as a dictator, one can only hope for partial success. There is a need to balance conservation hegemony and community needs.
This book is the outcome of the author's doctoral research work at the School of Oriental and African Studies, University of London. Regrettably, each chapter's abstract is written carelessly. However, the author is to be commended for the delicacy with which she highlights various aspects of the problems and conflicts in Jammu and Kashmir (p.32). Overall, the book enlightens readers and stresses the need to focus on socio-economic backgrounds in nature conservation efforts.