8 Abstract

8 Abstract

8 Abstract

1 1 Title 2 Catch me if you can: Species interactions and moon illumination effect on mammals of tropical 3 semi-evergreen forest of Manas National Park, Assam, India 4 Bhatt U.M1, Habib B1, Sarma H.K2 & Lyngdoh S.L*1 5 Corresponding author – salvador@wii.gov.in 6 Department of Animal Ecology & Conservation Biology, Wildlife Institute of India, Dehradun 248001 7 Field Director, Govt. of Assam, Barpeta Road 781315 8 Abstract 9 Species interaction plays a vital role in structuring communities by stimulating behavioral 10 responses in temporal niche affecting the sympatric associations and prey-predator 11 relationships.

We studied relative abundance indices (RAI) and activity patterns of each 12 species, temporal overlap between sympatric species, and effects of moon cycle on predator- 13 prey relationships, through camera-trapping in tropical semi-evergreen forests of Manas 14 National Park. A total of 35 species were photo-captured with 16214 independent records over 15 7337 trap nights. Overall, relatively high number of photographs was obtained for large 16 herbivores (11 species, n=13669), and low number of photographs were recorded for large 17 carnivores (five species, n=657). Activity periods were classified into four categories: diurnal 18 (day-time), nocturnal (night-time), crepuscular (twilight), and cathemeral (day and night time) 19 of which 52% records were found in diurnal period followed by 37% in nocturnal phase 20 whereas only 11% photographs during twilight.

Small carnivores were strictly nocturnal 21 (leopard cat and civets) or diurnal (yellow-throated marten and mongooses); whereas large 22 carnivores were cathemeral (tiger, leopard, clouded leopard and Asiatic black bear). Analysis .

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

8 Abstract

2 23 of activity patterns throughout the 24-h cycle revealed a high degree of temporal overlap 24 (>60%) among most of the sympatric species; however, differences in the activity peaks were 25 found between most of the species pairs. Moon phase was classified according to the 26 percentage of visible moon surface as new (0-25%), waxing (25-50%), waning (50-75%) and 27 full moon (75-100%).

Moon phase did not have any correlation with activity of large carnivore 28 and large prey. The large carnivore followed the feed and starve pattern of cyclic activity. The 29 activity of small carnivore was influenced negatively by moonlight (partial correlation r = -0.221, 30 p

8 Abstract

3 46 or mesopredators, due to top predator removal [5] which can cause substantial changes in the 47 dynamics of interaction among sympatric species [6], with adverse effects on subordinate 48 species. Thus, to minimize risks, subordinate species tend to avoid encounters with dominant 49 species [7], by modifying their activity patterns according to that of the dominant species [8]. 50 Often, the prey tries to avoid the times when predators are more active [9] or segregate in other 51 niche dimensions [10].

52 Moon cycle is reported to play a significant role in activity changes, and several nocturnal 53 animals can alter their activity in response to moonlight variation [11].

Animals may adapt 54 their schedules throughout the circadian cycle to increase their fitness and allow their mutual 55 co-existence [12]. The dynamics between predators and prey depend on these adaptions too, 56 leading to a balance between their activity patterns [13]. For instance, some mammals, such as 57 rodents [11,14] and bats [15] are known to reduce their activity in brighter nights, allocate it to 58 darker periods of the night [16,17] or seek for covered areas [18,14]. This behavior is thought 59 to be due to an increment of predators hunting success during these nights [15,11,14,19,20].

60 On the other hand, other species, such as some primates [21] and some nocturnal birds [22] are 61 also known to be more active in brighter nights. This increment of activity may be related to 62 both higher predator awareness and increased food uptake success [21,22,]. Amongst abiotic 63 factors, moon cycle is reported to play an essential role in niche adaptions [11,10,23]. Several 64 nocturnal animals change their activity patterns [17,24] and habitat use [18,14,25] due to 65 moonlight and the level of that response classifies species as lunarphobic [15] or lunarphilic 66 [21]. Many species’ interaction with the lunar cycles remain still unknown, however, and a 67 better perception of the responses of other small and large sized mammals to moonlight is 68 therefore required for a full understanding of its effects.

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

8 Abstract

4 69 The tropical forest contains some of the highest levels of species diversity and abundance, but 70 many tropical species are cryptic, shy, and secretive, which makes them notoriously difficult 71 to study and their interactions with one another, remain poorly understood [26].

Recently 72 however with camera-traps, monitoring terrestrial rare, cryptic and secretive species in tropical 73 forests has become effective [27,28]. The technique has improved our ability to study terrestrial 74 movements of Asian tropical forest fauna [29], species diversity [30], the associations among 75 species [31], and their habitats [32]. In addition to recording the presence and abundance data 76 of such taxa, date and time of the captures can help in understanding the activity patterns of 77 carnivores and other mammals [29,33,34].

78 In the current study, we examine activity rhythms and effect of the moon cycle on mammals in 79 the semi-evergreen forest of Manas National Park, India, using camera traps. Objectives of the 80 study were to: 1) determine relative abundance indices (RAI) and species assemblage of MNP; 81 2) determine activity periods of each species; 3) quantify temporal overlap patterns between 82 species; and 4) investigate moonlight effect on the activity of sympatric species. Such data can 83 be used to study processes shaping ecological communities, especially whether potentially 84 competing species overlap or avoid each other temporally, and how larger species might 85 influence activity of their smaller cohorts in the same habitat.

This information contributes to 86 an understanding of species interactions in tropical forests and should assist in developing more 87 suitable management and conservation strategies for forest communities in the Himalayan 88 foothills.

89 Materials and Methods 90 Study Area . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

8 Abstract

5 91 The study was carried out within the 500 km2 of Manas National Park (MNP) (26°35' - 26°50' 92 N, 90°45' - 91°15' E), a UNESCO World Heritage Site, in the state of Assam, India. Manas lies 93 on the borders of the Indo-Gangetic and Indo-Malayan biogeographical realms on a gentle 94 alluvial slope in the foothills of the Himalayas, where wooded hills give way to grasslands and 95 tropical forest.

The elevation ranges between 40-170 m moll with an average of 85 m [35]; the 96 monsoon brings extremely heavy rainfall to this region, reaching up to 3,300 mm annually and 97 temperature ranges between 6-37 0C [35]. The park is home to a variety of important mammal 98 species, including the tiger, pygmy hog, hispid hare and Asian elephant [36] and also it supports 99 22 of India’s most threatened mammal species, as listed in Schedule-I of the Wildlife 100 (Protection) Act of India [37]. Together with the Royal Manas National Park in Bhutan, the 101 park forms one of the largest areas for conservation significance in South Asia, representing 102 the full range of habitats from the subtropical plains to the alpine zone [38].

MNP acquires a 103 special place from conservation aspect owing to its tropical forests, endemism and a long 104 history of social and political conflict [39]. The national park experienced a fifteen-year-long 105 ethnic and political battle starting in the mid-1980s until fledgling peace was restored in 2003 106 [40]. The violence during the conflict that followed caused large-scale damages to Manas and 107 left the park vulnerable to logging, local hunting, and poaching of important fauna, causing 108 habitat degradation and rapid loss of wildlife [41,42].

109 Methods 110 1. Field sampling design 111 Data on RAI and species assemblage was collected by deploying camera-traps (n=241) during 112 two sample periods: April 2017 to June 2017 (n=112) and December 2017 to May 2018 113 (n=129), with the whole area divided into a grid system of size 1×1 sq. km (Fig 1). The camera- . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint .

http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct.

22, 2018;

6 114 trap locations were selected based on the presence of carnivore sign, accessibility, terrain 115 features, animal trails and nallahs (seasonal drainages). At each location, a single Cuddeback- 116 color™ digital camera was set by affixing it to trees at the height of approximately 30-45 cm 117 to above the ground. The cameras were triggered by motion sensor within a range of a conical 118 infrared beam and time lag of approximately 1s between the animal detection. Relative 119 abundance index (RAI) was calculated as the sum of all detections for each species for all 120 camera traps over all days, divided by the total number of camera trap nights, and then 121 multiplied by 100 [43].

To maintain statistical independence and to reduce bias caused by 122 repeated detections of the same species, one record of each species per half an hour per hours 123 per camera-trap site was considered as an independent detection and subsequent records were 124 removed [44]. No bait was used, to avoid disproportionate increases in the frequency of some 125 species [45].

126 Fig 1. Map of the study area (MNP) showing locations of camera traps (n = 241), grids, 127 drainage and forest cover. 128 2. Activity periods 129 The date and time printed on the photographs were used to describe diel activity periods of 130 each species. The assumption was made that the number of camera trap records taken at various 131 times is correlated with the daily activity patterns of mammals. The date and time printed on 132 the photographs were used to describe the daily activity patterns of the species. As some species 133 may be partly arboreal, and the camera-traps only recorded activity at ground-level, it is not 134 possible to assess arboreal activity.

The observations were classified as diurnal, nocturnal, 135 cathemeral or crepuscular. Photos that captured an hour before and after sunrise and sunset 136 were defined as crepuscular [46]. Sunset and sunrise hours were determined using geographical 137 coordinates of the study area and the Moonphase SH software (version 3.3; Henrik Tingstrom, .

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

7 138 Kalmar, Sweden). Species were classified as diurnal (

8 161 quarter - Waxing Moon (Wx) & last quarter - Waning Moon (Wn)] and 75-100% [Full Moon 162 (Full)]. Then, the records from each moon phase were selected to assess the effect of the 163 moonlight and positioning on the time schedules of large – small carnivores and their potential 164 prey, during lunar cycle.

One-way analysis of variance (ANOVA), Tukey HSD (honestly 165 significance difference) for Post-Hoc, and partial correlation tests were conducted to measure 166 the degree of association and pairwise comparisons among records of predator-prey in each 167 moon phase.

168 Results 169 1. Relative abundance indices & species assemblage 170 A total of 35 species were recorded with 16,214 independent records over the whole sampling 171 period of 7337 trap nights. The independent records (n) and relative abundance index (RAI) 172 for the photo-captured species varied from species-wise ranging from Neofelis nebulosa (n=7, 173 RAI=0.0011) to Panthera pardus (n=298, RAI=0.0417) for large – medium carnivores, from 174 Melogale moschata (n=1, RAI=0.0001) to Prionailurus bengalensis (n=221, RAI=0.0366) for 175 small carnivore, from Axis axis (n=1, RAI=0.0003) to Elephas maximus (n=4675, 176 RAI=0.5696) for large herbivores, and from Caprolagus hispidus (n=1, RAI=0.0002) to Gallus 177 gallus (n=574, RAI=0.0853) for small herbivores.

The summarised photo captures with RAI 178 of all the species are given in table 1.

179 180 181 182 183 . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

9 184 Table 1. Activity periods of photo-captured species through camera-trapping in Manas 185 National Park, Assam, India. SR No Species RAI N Photographic events (%) Classification Diurnal Nocturnal Crepuscular Large Carnivores 1 Panthera tigris 0.0367 269 39 50 12 Cathemeral 2 Panthera pardus 0.0417 298 58 35 7 Cathemeral 3 Cuon alpinus 0.0099 56 66 9 25 Diurnal 4 Ursus thibetanus 0.0029 27 44 41 15 Cathemeral Medium Carnivore 5 Neofelis nebulosa 0.0011 7 14 86 0 Mostly Nocturnal Small Carnivores 6 Felis chaus 0.0004 3 0 100 0 Nocturnal 7 Prionailurus bengalensis 0.0366 221 10 78 12 Mostly Nocturnal 8 Viverra zibetha 0.0303 209 5 88 7 Mostly Nocturnal 9 Viverricula indica 0.0212 114 7 82 11 Mostly Nocturnal 10 Paradoxurus hermaphroditus 0.0164 104 14 69 16 Cathemeral 11 Herpestes auropunctatus 0.0022 13 92 0 8 Diurnal 12 Herpestes edwardsii 0.0017 10 100 0 0 Diurnal 13 Herpestes urva 0.0088 62 89 3 8 Diurnal 14 Lutrogale perspicillata 0.0018 10 70 30 0 Mostly Diurnal 15 Martes flavigula 0.0023 16 75 0 25 Diurnal 16 Melogale moschata 0.0001 1 0 100 0 Nocturnal 17 Manis pentadactyla 0.0002 1 0 100 0 Nocturnal Large Prey 18 Elephas maximus 0.5696 4675 61 28 11 Mostly Diurnal 19 Rhinoceros unicornis 0.0029 21 14 86 0 Mostly Nocturnal 20 Bos gaurus 0.3457 2949 48 39 13 Cathemeral 21 Bubalus arnee 0.0236 160 26 63 12 Cathemeral 22 Axis axis 0.0003 1 0 0 100 Crepuscular 23 Muntiacus muntjak 0.1492 1068 58 31 11 Cathemeral 24 Hyelaphus porcinus 0.0050 35 40 37 23 Cathemeral 25 Rusa unicolor 0.3395 2414 22 67 11 Cathemeral 26 Sus scrofa 0.2517 1685 80 11 9 Mostly Diurnal 27 Hystrix brachyura 0.0564 319 4 83 13 Mostly Nocturnal 28 Pavo cristatus 0.0447 342 92 3 5 Diurnal Small Prey 29 Lepus nigricolis 0.0082 52 12 81 8 Mostly Nocturnal 30 Caprolagus hispidus 0.0002 1 0 100 0 Nocturnal 31 Macaca mulatta 0.0451 317 87 1 12 Diurnal 32 Macaca assamensis 0.0010 8 100 0 0 Diurnal 33 Trachypithecus pileatus 0.0014 10 100 0 0 Diurnal 34 Gallus gallus 0.0853 574 75 10 15 Mostly Diurnal 35 Lophura leucomelanos 0.0232 164 54 18 28 Mostly Diurnal Total (N) 16214 52 37 11 .

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

10 186 2. Activity periods of photo-captured species 187 Activity periods for 35 species depicts that the boundaries between the categories (diurnal, 188 nocturnal, cathemeral or crepuscular) of activity periods are not sharp (Table 1). However, 189 some animals restrict their routine activities to either the light or the dark phase, and rare 190 observations in the other period may represent activity elicited by unusual circumstances.

191 (i). Carnivores: Small carnivores were either mainly nocturnal such as Prionailurus 192 bengalensis, Viverra zibetha, and Viverricula indica, or diurnal such as Martes flavigula, 193 Lutrogale perspicillata, Herpestes urva, Herpestes auropunctatus and Herpestes edwardsii; 194 whereas Paradoxurus hermaphroditus with 69% photographs in the dark phase fell under the 195 cathemeral category (Table 1). Out of the five large – medium carnivore species; three large 196 body-sized mammals (Panthera tigris, Panthera pardus, and Ursus thibetanus) were 197 cathemeral, Cuon alpinus was diurnal, and Neofelis nebulosa was found with nocturnal nature 198 (Table 1).

199 (ii). Herbivores: Large herbivores were either mainly cathemeral such as Bos gaurus, 200 Bubalus arnee, Muntiacus muntjak, Hyelaphus porcinus, and Rusa unicolor, or diurnal such as 201 Elephas maximus, and Sus scrofa (Table 1). The only large herbivore tending toward 202 nocturnality was the Rhinoceros unicornis (Table 1). Terrestrial birds (Gallus gallus, Lophura 203 leucomelanos, Pavo cristatus) and primates (Macaca mulatta, Macaca assamensis, 204 Trachypithecus pileatus) were diurnal; whereas the routine activity of hares (Lepus nirgicolis, 205 Caprolagus hispidus) and Himalayan crestless porcupine (Hystrix brachyura) suggested 206 nocturnal nature of the species (Table 1).

207 3. Activity pattern and temporal overlap between sympatric 208 species . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

11 209 (i). Small carnivores: Eight small carnivores were sufficiently common to evaluate their 210 diel activity pattern (Fig 2). Civets and leopard cats were active during night hours; whereas 211 mongooses and yellow-throated martens were active during the daytime (Fig 2).

All four 212 nocturnal small carnivores had shown high temporal overlap between them, with the highest 213 overlap was found between leopard cat and small Indian civet with an overlap coefficient, 214 Δ4=0.93 (±0.15), followed by leopard cat and large Indian civet (Δ4=0.90); whereas least 215 overlap (Δ4=0.74) was found between large Indian civet and palm civet (Fig 2). Leopard cat 216 had a strong bimodal pattern, with a stronger peak at around 23:00 hr and a less pronounced 217 peak from about 01:00 to 04:00 hr. Large Indian civet also showed the bimodal pattern as it 218 increases post-sunset and reaches its peak at around 22:00 hr and then starts to decline; again, 219 it starts rising post-midnight and reaches its peak at about 01:00 hr and then begins to fall (Fig 220 2).

Other two civets (small Indian civet and Asian palm civet) had also shown a bimodal 221 pattern, but with differences in peaks; less pronounced peaks for both the species were between 222 19:00 to 23:00 hr and 03:00 hr, whereas stronger peaks were about at 04:00 hr and 17:00 to 223 22:00 hr respectively (Fig 2). High temporal overlap was found between all the four diurnal 224 small carnivores, with the highest coefficient value of Δ1=0.84 (±0.09) between crab-eating 225 mongoose and yellow-throated marten, followed by crab-eating mongoose and grey mongoose 226 (Δ1=0.77); whereas least coefficient value (Δ1=0.54) was found between small Indian 227 mongoose and yellow-throated marten (Fig 2).

Small Indian mongoose showed a unimodal 228 pattern, and it increases post 05:00 hr and reaches its peak at around 11:00 hr and then starts to 229 decline gradually until 18:00 hr (Fig 2). Grey mongoose, crab-eating mongoose, and yellow- 230 throated marten were active throughout the light phase, had a bimodal activity pattern, with a 231 stronger peak at 06:00, 15:00 and 16:00 hr, whereas less pronounce peak at 15:00, 08:00 and 232 06:00 hr respectively (Fig 2). Melogale moschata (n=1), Felis chaus (n=3), and Lutrogale .

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

12 233 perspicillata (n=10) had the fewest detections and therefore, were not considered for activity 234 analysis (Table 1). 235 Fig 2. Temporal overlap among small carnivores in Manas National Park, Assam, India. 236 Individual photograph times are indicated by the short vertical lines above the x-axis.

The 237 overlap coefficient (Δ1 / Δ4) is the area under the minimum of the two density estimates, as 238 indicated by the shaded area in each plot. The abbreviations of species’ names are CEM-Crab- 239 eating Mongoose, SIM-Small Indian Mongoose, GM-Grey Mongoose, YTM-Yellow Throated 240 Marten, LC-Leopard Cat, LIC-Large Indian Civet, SIC-Small Indian Civet, and PCPalm 241 Civet.

242 (ii). Large carnivores: Among the activity patterns of large carnivores, tigers and leopards 243 showed the highest daily activity overlap Δ4 = 0.82 (± 0.03) for any 2 species of top carnivores 244 in the study area, followed by leopards and Asiatic black bears (Δ1 = 0.82); whereas lowest 245 overlap Δ1 = 0.10 (± 0.07) was found between clouded leopards and dholes (Fig 3). Leopard 246 was active throughout the day and night but was more active during daylight, with peaks in the 247 early morning and late afternoon; tiger had also shown cathemeral activity pattern but was least 248 active from about 10:00 to 15:00 hr (Fig 3).

Two activity peaks (between 21:00 and 23:00 hr 249 and between 2:00 and 4:00 hr) were observed for clouded leopards, suggesting a bimodal 250 activity pattern of the species (Fig 3). Dholes, showed a unimodal pattern of activity, with 251 peaks between 06:00 to 09:00 hr (Fig 3). However, Asiatic black bears were relatively 252 cathemeral, with 44% of their photo-captures obtained during daytime and 41% records 253 obtained during night-time; whereas least activity was observed between 20:00 to 02:00 hr 254 (Table 1, Fig 3). A synchronized least-active pattern was noted between midnight and 03:00 hr 255 for all the detected large carnivores (Fig 3).

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

13 256 Fig 3. Temporal overlap among large carnivores in Manas National Park, Assam, India. 257 Individual photograph times are indicated by the short vertical lines above the x-axis. The 258 overlap coefficient (Δ1 / Δ4) is the area under the minimum of the two density estimates, as 259 indicated by the shaded area in each plot.

The abbreviations of species’ names are T-Tiger, L- 260 Leopard, CL-Clouded Leopard, WD-Wild Dog, and ABB-Asiatic Black Bear. 261 (iii). Carnivores and their prey: High temporal overlap was found among nocturnal 262 prey species such as Indian hare with leopard cats and civets whereas red junglefowl, and kalij 263 pheasant was active during the daytime, hence had shown large overlaps with mongooses and 264 yellow-throated marten (Fig 4a). Activity patterns of large carnivores and its prey showed 265 variable temporal overlap with the highest overlap between tiger and wild buffalo (84%) 266 followed by sambar (84%), hog deer (84%), and gaur (79%) (Fig 4b).

In case of leopard highest 267 overlap was found with hog deer (76%) followed by gaur (74%), barking deer (72%) and wild 268 buffalo (72%) (Fig 4b). Clouded leopard had shown highest overlap with Himalayan crestless 269 porcupine (66%), followed by sambar (65%) whereas dhole had maximum overlap with wild 270 boar (52%) and barking deer (50%) (Fig 4b). Chital (n=1) and hispid hare (n=1) had only one 271 detection and therefore, were not considered for the analysis (Table 1). 272 Fig 4. Pairwise temporal overlap (Δ1 / Δ4) between (a) small carnivore vs. prey and (b) 273 large carnivore vs. prey in Manas National Park, Assam, India.

The bars indicate 274 percentage temporal overlap among carnivore species with potential prey and the whiskers 275 above the bars indicate standard errors.

276 4. Moon phase effect on prey-predator relationship 277 Photographs of large carnivores, small carnivores, and potential prey species were analysed 278 during four moon cycles (New, Wx, Full, and Wn) (Fig 5). Differences were observed in 279 carnivore community with respect to the moon cycle, with highest records of large carnivores . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint .

http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct.

22, 2018;

14 280 in full moons except for dhole while small carnivores had more photographs at new moon 281 phase, except for Asian palm civet (Fig 5). Dhole activity was found mainly diurnal with only 282 9% photographs in nocturnal periods, out of which around 58% records were recorded in darker 283 nights (Fig 5). On the other hand, Asian palm civet had more photographs in full moon (33%); 284 and 21% photographs were recorded in new moon phase (Fig 5). All three photo-captured small 285 prey such as red junglefowl, kalij pheasant, and Indian hare had more photographs in the new 286 moon phase (Fig 5).

Larger prey showed almost uniform activity in all moon phases, with 287 highest records in new moon (wild buffalo, hog deer, wild boar, Himalayan crestless porcupine 288 and Indian peafowl); whereas the remaining three species (gaur, barking deer and sambar) had 289 more photographs in a full moon (Fig 5).

290 Fig. 5. The proportion of nocturnal records of (a) small carnivore vs. prey, (b) large 291 carnivore vs. prey, and (c) large carnivore vs. small carnivore in different moon phases 292 in Manas National Park, Assam, India. The bars indicate species records in different moon 293 phases. The dashed line is to separate small carnivore and their prey, large carnivore and their 294 prey, large carnivore and small carnivore respectively. 295 One-way ANOVA result pointed significant difference only for small carnivore (F= 5.007, 296 p

15 304 the zero-order correlation showed statistically significant, negative correlation between small 305 carnivore and moon visible surface (r= -0.205, p

16 315 Discussion 316 The current study provides baseline information on activity patterns and temporal overlaps of 317 mammals of Manas National Park as well as it is also the first of its kind of research on moon 318 illumination and effect of moon phases on prey-predator interactions in tropical forests of India. 319 Results from the present study are mainly concordant with basic accounts of natural history 320 (i.e., whether a species is most active during the day or night) [50]. We also compared our 321 results with those of previous studies on species body size, activity pattern and temporal 322 overlaps of mammalian fauna.

The lunar cycle results largely showed that the moonlight has a 323 stronger effect on the activity of the prey than on the behavior of the predator. 324 Evaluation of the camera trapping data revealed that the study area had a healthy habitat for 325 the mammalian fauna. All the major fauna from MNP was photo-captured during the survey 326 confirming 35 species. Out of the 35 recorded species, 1 (Chinese pangolin) is classified as 327 Critically Endangered, 6 (tiger, dhole, Asiatic elephant, wild water buffalo, hog deer and hispid 328 hare) are classified as Endangered, 8 (leopard, clouded leopard, Himalayan black bear, one- 329 horned rhinoceros, gaur, sambar and capped langur) are classified as Vulnerable, 1 (Assamese 330 macaque) is classified as Near Threatened while the remaining 19 species are classified as least 331 concern [51].

The previous camera-trapping studies in Manas National Park provides 332 information on relative abundances of tigers and their prey [52], carnivore diversity [53], and 333 density estimation of carnivores and herbivores [40]. Recently, Borah et al. [54] provide info 334 on density estimation of common leopard and clouded leopard; whereas Lahkar et al. [55] 335 explained about diversity, distribution and photo-capture rate of mammals of MNP. In the 336 present study, we examine activity rhythms and the lunar cycle effect on the mammalian fauna 337 in the semi-evergreen forest of Manas National Park.

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

17 338 1. Diel activity patterns and temporal overlap 339 Van Schaik and Griffiths [29] explained variation in activity periods for Indonesian rainforest 340 mammals using species body size as the primary factor influencing activity patterns.

The theory 341 suggests that smaller mammals (10 kg) are more cathemeral because of energy 343 requirements and associated feeding commitments. The intensive camera-trap survey provided 344 one of the most detailed studies of activity periods in mammals of MNP under natural 345 conditions and classified activity patterns into four categories [29]. In the present study, all the 346 photo-captured small mammals are found to be mainly nocturnal (jungle cats, leopard cats, and 347 civets) or diurnal (smooth-coated otter, yellow-throated marten, and mongooses), as predicted 348 by the van Schaik and Griffiths model.

The results showed that the medium-sized mammals 349 are cathemeral (barking deer, hog deer, and wild boar) and diurnal (wild dog), and the larger- 350 sized mammals such as tiger, leopard, Asiatic black bear, gaur, wild buffalo, and sambar are 351 active during both day and night hours which is also in accordance with the Schaik and Griffiths 352 model.

353 We found differences in the activity peaks of tiger and leopard, but there was no active temporal 354 separation between predators probably owing to their similar morphology and hunting 355 strategies [2]; however, significant time overlap between them was evident. Tigers are 356 opportunistic predators [56] and had considerably higher activity overlap (>75%) with gaur, 357 wild buffalo, sambar and hog deer [57,58]. However, their diet includes birds, fish, rodents, 358 insects, amphibians, reptiles in addition to other mammals such as primates and porcupines 359 [56]. The present study also found higher overlap (>65%) with Himalayan crestless porcupine 360 as compared to leopard’s overlap.

Leopard’s activity overlapped (>70%) with all the prey 361 species ranging from medium to large sized prey [59,60]. Leopard showed higher temporal .

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

18 362 overlaps with medium-sized prey [60] such as barking deer in comparison with tiger’s overlap. 363 Asiatic black bear tends to be diurnally active [61] or crepuscular [62]. The current study found 364 cathemeral nature of the species as it has to spend most of its time in search of food for energy 365 requirements under a very high competition with conspecifics (for vegetal food/and animal 366 matter) as well as other carnivores (for the animal matter) [63].

367 The study showed that clouded leopard activity was predominantly nocturnal, similar to the 368 studies of Gumal et al. [64], and Azlan & Sharma [65] from Peninsular Malaysia, and also 369 Kanchanasaka [66], and Grassman et al. [67] from Southern Thailand. Austin et al. [68] 370 recorded activity peaks at crepuscular hours in two radio-collared clouded leopards. However, 371 the overall activity pattern from radio-telemetry studies (n=4) indicated two peaks at 18:00- 372 02:00 hr and 08:00-12:00 hr [69]. The present study also found a bimodal pattern but with 373 different peaks at 21:00-23:00 hr and 2:00-4:00 hr.

In case of its prey, the study found the high 374 temporal overlap with sambar and Himalayan crestless porcupine [57] as compared to the other 375 prey species. It is possible that clouded leopard terrestrial activity is higher at night-time due 376 to the avoidance of leopards in the study area being more active on trees during daytime (A. 377 Wilting, pers. comm). However, studies suggest clouded leopards be more terrestrial [70,71,72] 378 with the use of trees primarily for resting [70,73]. The low capture rate of 7 photos in 7337 trap 379 nights in our study, does not necessarily reflect low numbers of the felid, but rather a decreased 380 probability to capture it along wildlife trails and roads that are frequented by high numbers of 381 leopards and tigers, the top predator of the area.

The species is known to use a dimension that 382 was not covered in our sampling, namely trees higher up than 60 cm above ground [74]. 383 The only canid species recorded during the study was wild dog (dhole). Dholes showed less 384 temporal overlap with their dominant competitors (tiger, leopard and clouded leopard) because 385 they were more active during the daytime and crepuscular hours and less active in full darkness, .

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

19 386 similar to most other studies of India and Southeast Asia (67,75,76,77]. Dholes’ diet includes 387 a wide variety of prey species, ranging from small rodents and hares to gaur [2,78,76,79]. In 388 tropical semi-evergreen forests of Southeast Asia, the species appear to persist in smaller packs 389 and consume medium-sized prey [75], as smaller packs are more energetically advantageous 390 in the rainforest where large prey species are scarce, thick vegetation favors stalk and ambush 391 hunting techniques over cursorial hunting, and competition with tiger and leopards [80].

The 392 present study also found the high temporal overlap of dhole with medium-sized prey such as 393 barking deer and wild boar as compared to other large-sized prey. 394 The present study recorded the two small cats (leopard cat and jungle cat) to be strictly 395 nocturnal which is consistent with other reported studies only in case of leopard cat [81,82], 396 yet there are studies contradicting nocturnality of leopard cats (83,65,84]. According to Prater 397 [81], jungle cat is diurnal as well as crepuscular and can even kill porcupine species which are 398 nocturnal. In this study, jungle cat is found to be strictly nocturnal as reported by Majumder et 399 al.

[85], though our result is insignificant because of only three captures of a jungle cat. The 400 small Indian civets and large Indian civets are found active in nocturnal hours which is 401 consistent with other reported studies [86,87]; whereas with 69% photographs in darker hours 402 of Asian palm civet also supports other study of Duckworth [88] and Azlan [89] which suggests 403 crepuscular or nocturnal nature of the species. Mongooses are predominantly diurnal or 404 cathemeral [90] and the present study found a strictly diurnal pattern for all the three photo- 405 captured mongooses. The yellow-throated marten was completely a diurnal species in the study 406 area [91] but can hunt both by day and night and also known to attack young deer species [81].

407 On the other hand, results of the current study showed a Himalayan crestless porcupine as a 408 nocturnal species which is consistent with the previous studies by Menon [92]. Indian hare 409 which is mainly active during crepuscular and nocturnal hours [93] are found to be active only 410 in nocturnal phase in this study. According to several studies [13,94], the diel activity of many .

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

20 411 felids is associated with the activity pattern of their prey. The main reason of these small cats 412 being nocturnal in the study area could be that rodents (Himalayan crestless porcupine) and 413 hares (Indian hare and hispid hare), their primary preys, are generally nocturnal [81,94] 414 although we could not quantify this point through camera trapping for small-sized prey.

415 2. Do species respond differently to moonlight? 416 Moonlight has usually been thought to increase predation risk by enhancing the ability of 417 predators to detect prey [14,95], therefore leading to decreased activity or shifts in prey 418 foraging efficiency in the presence of bright moonlight [16,96,97]. The results of the present 419 study demonstrate that the effects of moon illumination on activity across nocturnal mammal 420 species. The response of nocturnal mammals to the moonlight differs among taxa and may vary 421 according to several determinants, such as phylogeny, trophic level, sensory systems and 422 habitat type [13,97].

The current study suggests that the moon phases are also likely to 423 influence how prey distribute their activities through time to face different predation risk 424 periods.

425 Large-sized prey species activity were not significantly affected by moon phases as they 426 showed uniform activity with highest record of photographs in new moon (wild buffalo, hog 427 deer, wild boar, Himalayan crestless porcupine and Indian peafowl) and full moon (gaur, 428 barking deer and sambar) (Figs 5 and 6b). Large carnivore did not get influenced by moonlight 429 as they follow the feeding and starvation pattern of cyclic activity across a lunar cycle (Fig 6a). 430 We, therefore, suggest that large carnivores switch their type of prey, they hunt in different 431 moon phases, their hunting efficiency increases in the full moon and the greater foraging 432 benefits they take in brighter nights as they are more photo-captured during a full moon (Fig 433 5).

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

21 434 Small-sized prey species were more active during brighter nights to avoid predation risk against 435 smaller carnivores (Fig 5). This anti-predator behavior is already well recognised for the 436 species such as marsupials and rodents [33].

However, statistics showed that moonlight did not 437 influence the activity of small prey (Fig 6d). Small carnivores displayed a higher level of 438 activity during the darker nights when reduced brightness hampers their visual detections by 439 large carnivores which were active more in brighter nights (Fig 5). Tukey and partial 440 correlation tests also highlighted that moonlight had negative influence on the small carnivore 441 activity; their activity decreases with an increasing moonlight intensity (Fig 6c, Table 2). 442 Fig. 6. Photo-captures of (a) large carnivore, (b) large prey, (c) small carnivore, and (d) 443 small prey in a lunar cycle in Manas National Park, Assam, India.

The bars indicate 444 percentage moon visible surface in a complete lunar cycle. The lines indicate average records 445 in each day from all 7 lunar cycles, and the whiskers above and below the lines indicate 446 standard errors.

447 Conclusion & limitation 448 Our result suggests that despite historical ethnopolitical conflict and continued threats in some 449 areas, MNP supports a diversity of mammalian fauna of conservation concern, including 450 clouded leopards, dholes, tigers and other species. Adaptations are bidirectional and take place 451 over at least two dimensions: spatial and temporal [9,98] and our study focuses primarily on 452 the temporal component and provides some interesting insights into the diel activity patterns 453 and temporal overlap among mammals of MNP. The current study also highlights the 454 significance of incorporating moon illumination into movement and activity pattern of 455 mammals as well as interactions between prey-predator in tropical forests of India.

The brighter 456 hours or full moon lights shows an inverse relation in the activity pattern of prey and predator. .

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

22 457 Camera trapping is effective in recording species interaction but with certain limitations such 458 as the inability to account for detection probability, which is bound to vary with species [76]. 459 Placement of camera traps should be done depending on size, habitat and activity pattern of 460 species.

Like in our study, some of the species are at least partially, or even predominantly, 461 arboreal such as clouded leopard, small prey and primate species. Hence, activity patterns of 462 such species would be better explained if camera-traps were deployed in species-specific 463 habitats. Moonlight effects were not only related to the trophic level and were better explained 464 by phylogenetic relatedness, visual acuity, and habitat cover.

465 Acknowledgements 466 We thank the director, dean & research co-ordinator, wildlife institute of India. We are thankful 467 to the Doyil Vengayil & Syed Asrafuzzaman, Department of Science and Technology, 468 Government of India to carry out the study on the clouded leopard (Neofelis nebulosa). 469 Paniram, Tapan, Anukul, Dipul, Dipen and Dilli are thanked for their assistance in the field. 470 We thank the Dept., Environment & Forests, Govt., of Assam, field staff of Manas National 471 Park, for permissions and field support.

472 References 473 1. Case TJ, Gilpin ME. Interference competition and niche theory.

Proceedings of the 474 National Academy of Sciences. 1974 Aug 1;71(8):3073-7. 475 2. Karanth KU, Sunquist ME. Prey selection by tiger, leopard and dhole in tropical forests. 476 Journal of Animal Ecology. 1995 Jul 1:439-50. 477 3. Steinmetz R, Garshelis DL, Chutipong W, Seuaturien N. Foraging ecology and 478 coexistence of Asiatic black bear bears and sun bears in a seasonal tropical forest in 479 Southeast Asia. Journal of Mammalogy. In press.

CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

23 480 4. Caro TM, O'doherty G. On the use of surrogate species in conservation biology. 481 Conservation biology. 1999 Aug;13(4):805-14. 482 5. Prange S, Gehrt SD. Response of skunks to a simulated increase in coyote activity. 483 Journal of Mammalogy.

2007 Aug 20;88(4):1040-9. 484 6. Kamler JF, Stenkewitz U, Macdonald DW. Lethal and sublethal effects of black-backed 485 jackals on cape foxes and bat-eared foxes. Journal of Mammalogy. 2013 Apr 486 16;94(2):295-306. 487 7. Berger KM, Gese EM. Does interference competition with wolves limit the distribution 488 and abundance of coyotes?. Journal of animal Ecology. 2007 Nov 1;76(6):1075-85. 489 8. Carothers JH, Jaksić FM. Time as a niche difference: the role of interference 490 competition. Oikos. 1984 Mar 1:403-6.

491 9. Lima SL, Bednekoff PA. Temporal variation in danger drives antipredator behavior: 492 the predation risk allocation hypothesis. The American Naturalist. 1999 493 Jun;153(6):649-59. 494 10. Lucherini M, Reppucci JI, Walker RS, Villalba ML, Wurstten A, Gallardo G, Iriarte A, 495 Villalobos R, Perovic P. Activity pattern segregation of carnivores in the high Andes. 496 Journal of Mammalogy. 2009 Dec 15;90(6):1404-9. 497 11. Clarke JA. Moonlight's influence on predator/prey interactions between short-eared 498 owls (Asio flammeus) and deermice (Peromyscus maniculatus). Behavioral Ecology 499 and Sociobiology.

1983 Sep 1;13(3):205-9.

500 12. Schoener TW. Resource partitioning in ecological communities. Science. 1974 Jul 501 5;185(4145):27-39. 502 13. Harmsen BJ, Foster RJ, Silver SC, Ostro LE, Doncaster CP. Jaguar and puma activity 503 patterns in relation to their main prey. Mammalian Biology. 2011 May 1;76(3):320-4. . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint .

http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct.

22, 2018;

24 504 14. Kotler BP, Brown JS, Hasson O. Factors affecting gerbil foraging behavior and rates 505 of owl predation. Ecology. 1991 Dec 1;72(6):2249-60. 506 15. Morrison DW. Lunar phobia in a neotropical fruit bat, Artibevs jamaicensis 507 (Chiroptera: Phyllostomidae). Animal Behaviour. 1978 Aug 1;26:852-5. 508 16. Daly M, Behrends PR, Wilson MI, Jacobs LF. Behavioural modulation of predation 509 risk: moonlight avoidance and crepuscular compensation in a nocturnal desert rodent, 510 Dipodomys merriami. Animal behaviour. 1992 Jul 1;44(1):1-9. 511 17. Kotler BP, Brown JS, Dall SR, Gresser S, Ganey D, Bouskila A.

Foraging games 512 between gerbils and their predators: temporal dynamics of resource depletion and 513 apprehension in gerbils. Evolutionary Ecology Research. 2002;4(4):495-518. 514 18. Emmons LH, Sherman P, Bolster D, Goldizen A, Terborgh J. Ocelot behavior in 515 moonlight. Advances in neotropical mammalogy. 1989;1989:233-42. 516 19. Mougeot F, Bretagnolle V. Predation risk and moonlight avoidance in nocturnal 517 seabirds. Journal of Avian Biology. 2000 Sep;31(3):376-86. 518 20. Jetz W, Steffen J, Linsenmair KE. Effects of light and prey availability on nocturnal, 519 lunar and seasonal activity of tropical nightjars.

Oikos. 2003 Dec;103(3):627-39. 520 21. Gursky S. Lunar philia in a nocturnal primate. International Journal of Primatology. 521 2003 Apr 1;24(2):351-67.

522 22. Wilson MD, Watts BD. Effect of moonlight on detection of Whip‐poor‐wills: 523 implications for long‐term monitoring strategies. Journal of Field Ornithology. 2006 524 Mar;77(2):207-11. 525 23. Michalski F, Norris D. Activity pattern of Cuniculus paca (Rodentia: Cuniculidae) in 526 relation to lunar illumination and other abiotic variables in the southern Brazilian 527 Amazon. Zoologia. 2011 Dec 14;28(6). . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint .

http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct. 22, 2018;

25 528 24. Mori E, Nourisson DH, Lovari S, Romeo G, Sforzi A. Self‐defence may not be enough: 529 moonlight avoidance in a large, spiny rodent. Journal of Zoology. 2014 Sep;294(1):31- 530 40. 531 25. Penteriani V, Kuparinen A, del Mar Delgado M, Palomares F, López-Bao JV, Fedriani 532 JM, Calzada J, Moreno S, Villafuerte R, Campioni L, Lourenço R. Responses of a top 533 and a meso predator and their prey to moon phases. Oecologia. 2013 Nov 1;173(3):753- 534 66. 535 26.

Ridout MS, Linkie M. Estimating overlap of daily activity patterns from camera trap 536 data. Journal of Agricultural, Biological, and Environmental Statistics. 2009 Sep 537 1;14(3):322-37.

538 27. Karanth KU, Nichols JD. Estimation of tiger densities in India using photographic 539 captures and recaptures. Ecology. 1998 Dec 1;79(8):2852-62. 540 28. Cutler TL, Swann DE. Using remote photography in wildlife ecology: a review. 541 Wildlife Society Bulletin. 1999 Oct 1:571-81. 542 29. van Schaik CP, Griffiths M. Activity periods of Indonesian rain forest mammals. 543 Biotropica. 1996 Mar 1:105-12. 544 30. Kitamura S, Thong-Aree S, Madsri S, Poonswad P. Mammal diversity and conservation 545 in a small isolated forest of southern Thailand. Raffles Bulletin of Zoology. 2010 Feb 546 28;58(1).

547 31. Ngoprasert D, Lynam AJ, Sukmasuang R, Tantipisanuh N, Chutipong W, Steinmetz R, 548 Jenks KE, Gale GA, Grassman Jr LI, Kitamura S, Howard J. Occurrence of three felids 549 across a network of protected areas in Thailand: prey, intraguild, and habitat 550 associations. Biotropica. 2012 Nov;44(6):810-7. . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint .

http://dx.doi.org/10.1101/449918 doi: bioRxiv preprint first posted online Oct.

22, 2018;

You can also read