Behavioural sampling techniques and activity pattern of Indian Pangolin Manis crassicaudata ( Mammalia : Manidae ) in captivity

Author Details: R.K. Mohapatra is working as Junior Research Fellow; has registered for his PhD in Utkal University, Odisha with an objective to understand behavioural patterns, space utilization, reproductive biology, haematology, effect of environmental enrichment and identification of ectoand endo-parasites of captive Indian pangolins which will be ultimately helpful in captive management and breeding of the species. Dr. S. Panda is currently working as Director, Nandankanan Biological Park, Odisha. He is the project investigator of the research project. He has co-authored 7 books and 30 research papers in the field of wildlife and biodiversity. His research interest includes wildlife conservation, in-situ and ex-situ management of wildlife and biodiversity studies.


INTRODUCTION
Pangolins are toothless small mammals with overlapping horny scales on their body, protrusible long tongue and prehensile tail, rolling up into a ball when threatened.They belong to the family Manidae of order Pholidota.Indian Pangolin Manis crassicaudata distributed throughout peninsular India, Sri Lanka, Bangladesh and Pakistan (Heath 1995;Molur 2008).The Indian Pangolin is included under Appendix II of CITES, as 'Near Threatened' under the IUCN Red List (Molur 2008), and as a Schedule I animal of the Indian Wildlife (Protection) Act, 1972.For better management, welfare and breeding in captivity there is need for a better understanding of their behaviour.
Behaviour is the exhibition of a phenotypic trait within the environmental context for which primary selective forces have shaped it, the outcome of which is inclusive fitness (Eisenberg 1981).Behavioural monitoring is a demonstrated technique that can improve the welfare of animals in zoos (Kleiman 1992), and meet conservation goals (Watters et al. 2009).Behavioural sampling method is crucial to understanding and interpreting behaviour of an animal, because sampling bias does not change the behaviour of the animal but can affect the efficiency and reliability of data collection (Fragaszy et al. 1992).
There are different methods of behavioural sampling (Altmann 1974;Martin & Bateson 1993;Lehner 1996).'Focal sampling' is simple and easy to use in a wide variety of environments (Tyler 1979;Losito et al. 1989).'Focal sampling' requires the use of a stopwatch for precise recording of the transition times between activity states.As described by Altmann (1974), 'instantaneous sampling' records the state at fixed time points.Choosing an appropriate interval between sample points is essential, particularly for 'bout lengths' for meaningful and unambiguous 'instantaneous sampling' (Engel 1996).
The present study aimed at finding out the timebudget for various above-ground activities of captive Indian Pangolins.The study also determined the optimal sampling interval for behavioural data collection with instantaneous sampling technique.

I. Animals and housing conditions
The study was conducted between February 2012 and April 2012 on six (three males and three females) Manis crassicaudata at the "Pangolin Conservation Breeding Centre" of Nandankanan Zoological Park, Odisha, India (Table 1).Each pangolin was individually housed in an enclosure 4.8x4.2x3.0 m in dimension, where 0.5m deep red laterite soil was provided as substrate over a reinforced concrete base to prevent digging and escape of the animal.On all four sides above the concrete base there was a reinforced concrete wall up to a height of 1.0m, above which chain-link mesh was provided on all four sides upto the concrete roof.An earthen mound, 2x2 m, and some hollow wooden logs and wooden poles provided enrichment material in each enclosure.The pangolins were microchipped with a passive integrated transponder (PIT) for their individual identity.The microchip number and a short reference name for each pangolin are given in Table 1.The daily husbandry routine consisted of enclosure cleaning, water maintenance, feeding and health monitoring.Health monitoring comprised of (i) observation through video recordings to detect inactivity or change in behaviour, (ii) intermittent direct observation for any sickness, and (iii) faecal sample examination to detect the presence of gastrointestinal parasites.Every month physical examinations were carried out with each pangolin for measuring the length and weight of the body, for detection of any injury or external parasites and for secretion from their natural orifices like nose,

II. Behavioural observation
The following terms are used to describe present behavioural study: (i) 'Ethogram'-a complete inventory of the behaviour patterns of species (Tinbergen 1951); (ii) 'Focal Sampling'-continuous behavioural observation that records each sample period for each focal individual (Altmann 1974); (iii) 'Instantaneous behavioural sampling'-the documentation of individual's current activity at preselected moments in time (Altmann 1974); (iv) 'Activity pattern'-the time spent in different aboveground behaviour when pangolins could be observed through CCTV (close circuit television) cameras.
Data for this study was obtained from six numbers of pangolins and comprised 1377 hours of video recordings for each pangolin between 1500 and 0800 hr on 81 consecutive observational nights.Since the Indian pangolins are nocturnal, secretive, solitary and intermittently active, behavioural observation was made through digital video recording system assisted by infrared enabled CCTV cameras installed at each enclosure using focal animal sampling.Though the enclosures of each pangolin were regularly inspected during the remaining time, i.e., from 0800 to 1500 hr, of the observational day, no above-ground activity was observed and was not subjected to analysis.An ethogram of commonly observed behaviour of pangolins was prepared from the preliminary observations (Table 2).The observed behaviours with their respective time intervals were documented in Excel worksheets.
The 'focal samples' comprised of the durations of respective behavioural events extracted by continuous behavioural scan from recorded video files.The focal samples were used to create instantaneous sample data set.For example, behavioural scans at every 60 th second were used to create the instantaneous sample for a oneminute interval.Similar behavioural scans were carried out to prepare instantaneous samples at 2, 3, 4, 5, 10, 15, 20 and 30 minute intervals.When behaviour was measured with instantaneous sampling at intervals, the sampled behavioural data stands for all the behavioural events that occurred during that particular interval.Therefore, 'instantaneous behavioural samples' at different intervals, extracted from the 'focal behavioural samples' were multiplied with the sampling interval to represent 'behavioural score'.

III. Data analysis
Estimates of sample size included mean, standard deviation and standard error calculated for focal behavioural samples as well as instantaneous behavioural samples.The mean 'bout duration' of any behaviour was calculated by extracting at least 50 samples from focal sampling data during initial observations.Similarly, the 'mean duration of repeatability' of any behaviour was calculated from more than 30 samples extracted from the focal sampling data.Spearman rank-order correlations (r s ) were calculated to correlate mean bout duration, mean duration of repeatability and mean time spent in displaying the behaviour.To find out optimal sampling interval for Instantaneous behavioural sampling, the behavioural data from focal samples and instantaneous samples at different intervals were subjected to the following statistical tests.Pearson product correlations (r) were used to correlate average time spent in different behaviour.Analysis of variance (Zar 1984) was used to test the null hypothesis that the expected means of all samples taken by different sampling methods are equal.Rejection of null hypothesis only implies that the group means were not the same.To find out the

Behaviour Definition
Bathing (BA) Individual lying submerged or moving inside water pool.

Bipedal Stand (BS)
Standing on its legs with head elevated to survey the surroundings.

Climbing (CL)
Climbing movement on wire mesh or on dry tree branches.

Coiling (CO)
Rolling itself in to a sphere.

Digging (DI)
Scooping of soil by forelimbs and throwing the dugout soil using hind limbs.

Drinking (DR)
Head lowered below level of back and consuming water by lapping movement of tongue.

Exploration (EX)
Short head movements directed towards pangolin's proximate environment usually followed by sniffing or touching the substrate by snout or forelimb.

Feeding (FI)
Head lowered below the level of back and consuming ants and termites using frequent in and out movement of tongue.

Pacing (PC)
Repetitive quadrupedal waking movements in same path.

Secretive (SC)
Looking out from the mound holes while the entire body is inside the burrow, head visible.

Social Interaction (SI)
Interaction between the individuals of neighbouring enclosures along chain-linked mesh.(Mother-young and male-female interactions have not been included in this study.) Walking (WK) Forward progression in a typical quadrupedal movement and tail usually held up the ground.
Other Urination, defecation or any other behaviour not included in the ethogram.

RESULTS
The 12 behaviours compared between six Indian Pangolins (Table 1) were presented in the ethogram (Table 2).Mean time spent (± standard deviation) in different behaviours (in minutes) by individual Indian Pangolins is given in Table 3.This study used analysis of focal behavioural sampling data as the standard to compare instantaneous sampling data at different interval.Means and standard errors for all behaviours The correlation values gradually decreased with the increase in sampling intervals (Table 5).Taking individual behaviours into consideration, walking (r=0.904,P<0.05), pacing (r=0.915,P<0.05) and climbing (r=0.876,P<0.05) behaviours are highly correlated in all the intervals of instantaneous sampling.We also found that the repeatability of the behaviours have a significant negative correlation with the mean time spent in that behaviour (r s =-0.721;P=0.008).By comparing the difference of means between focal samples and instantaneous samples with LSD value, it was found that the mean values of the behaviours, except social interaction samples, with interval ≤ 5 minutes were not significantly different from focal samples (Table 6).A peak in activity was observed between 2000-2100 hr in the present study on Indian Pangolins, except for the pangolin P395, which was found most active between 2300-2400 hr, followed by decreasing activity thereafter (Fig. 1).Pangolins were active, i.e., exhibited different behaviour over ground for 129.02±46.45minutes (N=81days) on average in a day during the study period.Out of the total active period, time spent on different behaviour were feeding (14±4.32%,N=6), drinking (0.72±0.56%,N=6), bipedal stand (2.3±1.73%,N=6), coiling (0.61±0.43%,N=6), digging (3.67±3.65%,N=6), exploration (6.59±3.91%,N=6), climbing (0.68±0.64%,N=6), bathing (0.18±0.06%,N=6), walking (59.34±22.33%,N=6), and secretive (1.84±0.83%,N=6).For two pangolins social interaction scored 1.35±0.63%,and pacing was 19.18±15.15%.The beginning of activity was usually directed towards feeding behaviour that started 3.5±1.8minutes (N=81days) after emergence of pangolins from burrows, except the pangolin P395 (30.8±24.6 minutes).Mean time allocated in individual pangolins is given in Fig. 2. The differences in equitability of behavioural diversity exhibited were tested for individual pangolins.No significant difference was observed in behavioural diversity (F=0.577,N=6, P=0.719); with Shannon-Weiner diversity index (H) value, 0.791>H>1.586.It was found that pangolins spent most of the time on the ground (73.25±58.23%)exhibiting, walking, pacing, digging and coiling behaviour (Fig. 3).The time spent near the feeding platform is the second highest (9.66±4.28%)that spend for feeding and associated behaviours.Pangolins used the enrichment items (2.99±2.35%)for climbing behaviour.They used given water pools (0.82±0.78%) for drinking, bathing, water inspecting behaviour.Time spent in mound utilisation (2.76±2.41%)includes walking over mound and mound inspecting behaviours.

I. Behavioural sampling
An observational study should show that their sampling technique is reliable, i.e., they have small errors of measurement and that the scores of individuals show stability, consistency, and dependability for behaviour being studied.If the measure is not reliable, it cannot be expected to show lawful relationships with other variables being studied (Mitchell 1979).In most behavioural studies the standard is focal sampling (Mitlohner et al. 2001) but instantaneous sampling is not uncommon.Instantaneous sampling proved to be an accurate reflection of the amount of time spent by an animal in various behaviours (Leger 1977).Several studies have used instantaneous sampling technique with one minute (Bashaw et al. 2003(Bashaw et al. , 2007;;Skibiel et al. 2007;Macri & Patterson-Kane 2011), two minutes  5).This may be due to less repeatability of the behaviour (Table 3) than other behaviours.Therefore the study suggests use of focal sampling to record social interaction behaviour.Besides, repeatability of behavioural measures also depends on the bout duration of the behaviour.Studies have shown that the shorter the interval the more accurate duration is obtained from time sampling (Powell et al. 1975;Sanson-Fisher et al. 1980).If short-duration behaviours are of interest and percentage observation time is the response dimension of interest it is wise to choose shorter intervals as far as possible.
Instantaneous sampling is often useful to avoid arbitrary definitions of behaviour's start and end time, which may be difficult to specify.It needs less time and number of efforts than focal sampling, and it can easily be carried out in conjunction with sampling other behavioural and ecological data (Doran 1992).One can study the behaviour of animals by circular scan in particular interval of time and that interval can be used for recording data from another animal.Instantaneous sampling can improve efficiency by avoiding observer fatigue and inability to process desired volume of information.If the initial observation yield a high correlation between focal behavioural sampling and instantaneous sampling, the investigator may use the later for further recording to save time and efforts.These outcomes lead many to recommend instantaneous sampling as a mode of choice (Dunbar 1976;Leger 1977;Powell et al. 1977;Shahon-Fisher et al. 1980;Ary & Suen 1983;Tagha et al. 1985).

II. Activity pattern
The result obtained in this study is similar to that of Mishra & Panda (2010) that report captive Indian Pangolins spend most of their time walking inside the enclosure under simulated conditions of surroundings.The time of sunset plays a role in initiating the activities of the pangolins and the activity pattern recorded for each individual is very specific to each (Mishra & Panda 2010).Results from the present study suggest that the pangolins were active intermittently between 1700 and 0500 hr.This is similar to the activity pattern reported for Sunda Pangolins by Challender et al. (2011).The peak period of activity was during 2000-2100 hr with one individual variation.Pangolin P395 was most active during 2300-2400 hr.It is clear from the present study that variations in behavioural diversity were observed between individual pangolins.Stereotypic pacing behaviour was observed in some of the individuals, predominantly in P5FE (55.04±30.91%,N=81 days) and P473 (2.29±2.13%,N=81 days).The forms and frequencies of pacing observed in this study were comparable to those reported from Sunda Pangolins (Challender et al. 2011).Maintaining Indian Pangolins in natural enclosures, in which they can perform a full range of natural behaviour and exert control over their environment, is likely to offer some protection against behavioural deficits.
Furthermore, environmental enrichment has had some success in reducing the development of stereotypes (Bashaw et al. 2003(Bashaw et al. , 2007;;Skibiel et al. 2007).Development of environmental enrichment for Indian Pangolin is a matter for further study.

Figure 2 .
Figure 2. Mean time allocation in different behaviours in individual Indian Pangolins at captivity.Data presented here refer toTable 3. Behaviour types refer to description in Table 2.

Figure 3 .
Figure 3. Mean time allocation in enclosure space utilisation by Indian Pangolins.Key: * Over mound activity refer to behaviour like walking over mound.