Roost-Colony protection strategy from raiding by enemies in the Indian Flying Fox Pteropus mediu s

Pteropus medius roost in the open in trees. They are susceptible to attack by different predators and competing species. Roosting bats produce alarm calls in response to perceived danger. To assess anti-predator behaviors to watch the arrival of enemies, in these bats, we collected data on the directions roosting bats faced in two roost trees in West Bengal, India. It is revealed that the bats, irrespective of roost spots, are accustomed to maintaining the hanging posture facing all direction though out the day. These differences in hanging poses were statistically significant (P<0.05). We hypothesized that such orientation in hanging stance is the induced impact of anti-predator behavior to ensure vigil in all directions.


INTRODUCTION
Roosting bats are vulnerable to attack by predators and competing species (Table 1).The Indian flying fox Pteropus medius roosts in trees.They are attacked by various predators and competing species off and on.We noted the hanging postures of these bats in two roost trees.It is revealed that irrespective of the hours of the daytime some of the colony-members may be seen hanging, watching out for attacks from inimical groups.Such a strategy enables bats to emit alarm calls to alert the colony members to escape the danger.
We hypothesized that in such directions hanging, bats keep a vigil.Since there exists no report in this respect, we took the liberty to bring the said fact, at least in tree roosting bats, into, the notice of the research workers.

Roost Trees and Study Regions
We selected two big roost trees.Of the two, the silk flower tree Albizia lebbeck is located in the village Joteghanashyam (22°31′10.0″

Data Counting Date and Time Slots
We noted hanging postures of P. medius bats at frequent intervals during any one of the times scheduled 05:22-11:59h; 12:00-14:59h, 15:00-16:59h and, 17:00h to onwards until they started to leave the roost site for foraging.We recorded their front view hanging direction from August 2, 2008, to December 31, 2014, at Joteghanashym and February 9, 2015, to May 3, 2019, at Simla.The purpose of noting the front view hanging direction was to confirm the visual strategy as an anti-predator behavior of these bats.To note if roost trees have any impact in such type of vigilance, one less bushy and deciduous tree (A.lebbeck) and the other bushier as well as evergreen were taken into the study program sequentially one after other.

Observation Strategy
Two persons took part in recording the hanging postures simultaneously but individually to avoid any sort of ambiguity in observations and data records.We used field binoculars (Nikon Aculon A211 12*50) for observations during data collection.

Statistical Analysis
We analyzed the data with R-studio 3.6.3and PAST 3.14, and one-way ANOVA was applied to ascertain the impact of the roost tree.To assess the fixed and random effects of hanging directions, we applied the Generalized Linear mixed effect Model (GLMM).
We used GLMM (Tables S2 & S4), Eigen, and S4 package and conducted Shapiro-Wilk tests for normality followed by the Poisson family and a Kruskal-Wallis.

RESULTS AND DISCUSSION
In the course of studies, irrespective of time schedules and study sites, we encountered 552834 bats facing in any one direction, as seen in (Figure 1, 2).From a statistical view-point such, mouth orientation with respect to the four scheduled study hours was typical pattern of distribution.We recorded directions 6-950 and 3-295 P. medius faced in two roosts (Table 2).The results of GLMM (Tables S2 & S4) and ANOVA (P=0) tests suggest that there exists no significant difference with respect to the direction of the bats face in either of the roost trees.The results of the Kruskal Wallis test indicated a standard order of statistics median and distribution of each median value in the four frontal-facing directions (Figure 3).From the results, it is apparent that the bats P. medius habituated to hang in all directions, viz.east, south, west, and north-facing states throughout the roosting hours (Tables S1 & S3).It is also evident that less than 20% of individuals were in north-facing hanging posture.In contrast, more than 80% were east, south, or west-facing with respect to different hours of the day time Paschim Medinipur district and the tamarind tree Tamarindus indica is situated 75 km away of A. lebbeck in the village Simla (23022′44.20″N, 86038′ 47.02 ″ E) in Purulia district of West Bengal, India.There were big ponds in the vicinity of both the study sites.These trees are umbrella-shaped with a large number of branches and sub-branches.The A. lebbeck (16.15 m tall with a 1.73 m diameter trunk) is deciduous while, T. indica (17.67 m tall, a trunk 3.33 m in diameter) is evergreen.

Figure 1
Figure 1 3D-Pie chart showing P. medius projecting the mouth towards east, south, north, and west in the A. lebbeck roost tree at Joteghanashyam, Paschim Medinipur, West Bengal, India.

Figure 2
Figure 2 3D-Pie chart showing P. medius projecting the mouth towards east, south, north, and west in the T. indica roost tree at Simla, Purulia, West Bengal, India.

Figure 3
Figure 3 Normal order of statistic medians and distribution of P. medius bats' mouth facing directions irrespective of roost trees, based on the results of Kruskal Wallis test.

Table 2
Data [range, (mean±SE)] on the front facing hanging directions of P. medius bats in the roost trees, A. lebbeck and T. indica.

Table S3
Total number (%) of bats P. medius encountered at different hanging directions during the four scheduled study period at