Construction and Validation of a Systematic Ethogram of Macaca fascicularis in a Free Enclosure

Behavioral studies in non-human primates have become ideal models for further investigations into advanced cognitive function in humans. To date, there is no systematic ethogram of the cynomolgus monkey (Macaca fascicularis) in a free enclosure. In a field observation of 6012 subjects, 107 distinct behaviors of M. fascicularis were preliminarily described. 83 of these behaviors were then independently validated through a randomized cohort and classified into 12 behavioral categories. 53 of these behaviors were then selected to accurately reflect the daily mundane activity of the species in a free enclosure. These findings systematically document the behavior of M. fascicularis in a free enclosure for use in further investigations.


Introduction
Behavioral science has garnered considerable attention in contemporary neurological and psychological scientific circles as a platform for studying advanced cognitive function in humans. The United States declared 2000-2010 as the ''Decade of Behavior'' [1], in line with the increasing volume of research into psychopathology and human behavior. This phenomenon has been accompanied by a rising number of behavioral studies on human subjects [2].
However, due to ethical considerations, certain advanced functional and structural investigations cannot be conducted in humans [3,4,5,6,7,8,9,10,11,12,13,14,15]. Non-human primate models offer a viable alternative, as humans share neuroanatomical and psychological homology with these species. But, behavioral descriptions and classifications on non-human primates are often incomplete and scattered across disparate studies. Therefore, to support future behavioral investigations and concomitant quantitative studies, it is imperative to establish systematic ethogram of non-human primates on a species-specific basis [16,17].
Several macaque species have demonstrated promise as nonhuman primate models for behavioral studies. In the 1930's, Skinner et al. [18] constructed a systematic ethogram of a captive lion-tailed macaque (Macaca silenu) through behavioral descriptions, definitions and classifications, as well as comparisons with other monkey species. Adams et al. [19] studied the sexual behavior of the adult cynomolgus monkey (Macaca fascicularis), uncovering the value of this macaque species as an animal model for human behavioral studies. Jolly et al. [20] reported this species is capable of acquiring feeding and social behaviors by example, indicating its behavioral homology to humans. William et al. [21] cataloged the behavior of 8 male rhesus monkeys (Macaca mulatta) for approximately 3 years under conditions of separation and confinement. Furthermore, Zhang, J. [22] investigated the post conflict settlement of group-living primate, among adult females Sichuan snub-nosed monkeys Rhinopithecus roxellana, which is crucial understanding for primate's competition and cooperation, and indicated that the pattern of post conflict affiliation demonstrates that the R. roxellana belongs to a tolerant species. In addition, Mear and Harlow [23] investigated the development of play behaviors in 8 rhesus infants over a 12-week period. Ultimately, Yamada, Kazunori [24] reviewed the studies concerning the social behavior of Old World monkeys and concluded that the social environments plays an important role in their daily routine on how to manage themselves.
However, to date, no systematic ethogram of the cynomolgus monkey (M. fascicularis) has been established. In this study, a comprehensive systematic ethogram (including behavioral descriptions, definitions and classifications) for M. fascicularis in a free enclosure was constructed and validated which satisfies the technical requirements for future hypothesis-testing and quantitative studies on the species.

Results
From April to December 2010, 107 behavioral items were created based on the Posture-Action-Environment principle [25], and the scanning method was employed to verify these behaviors among 6012 M. fascicularis subjects in a free enclosure. 83 of these behaviors were then validated on 40 randomly selected female subjects and classified into 12 behavioral categories .From this set of 83 behaviors, 53 behaviors were then selected from these 12 behavioral categories: i) ingestion, ii) thermo-regulatory, iii) rutting and estrous, iv) mating, v) resting, vi) parental, vii) amicable, iix) conflict, ix) vigilance, x) communication, xi) locomotive and xii) miscellaneous behaviors, more detail data, please see File S3. Across all 12 behavioral categories per day and 12 behavior categories within all observational phases per day, resting and vigilance behaviors occurred at a significantly greater average daily duration per subject than other behaviors (11.566 and 10.396the median value, respectively; Fig. 1, Table 1 and table 2). Across all 53 behaviors and all observational phases, watching company (a vigilance behavior), quadrupedal walking on floor (a locomotive behavior), and sitting on floor (a resting behavior) were the most frequent behaviors observed (77.106, 47.886, and 46.276 the median value, respectively; Table 3). The average durations per subject for all 12 behavioral categories on both a daily and observational phase basis are listed in Tables 1 and 2, respectively. The average counts per subject for all 53 behaviors on both a daily and observational phase basis are detailed in Tables 3 and 4, respectively.
Ingestion behaviors occurred at a moderately higher average daily duration per subject with respect to other behavioral categories (1.366 the median value; Table 1). Chewing (5.576 the median value), picking remaining food (3.946 the median value), and feeding while sitting (3.736the median value) were the most frequent ingestion behaviors observed across all phases with the notable exception of phase P2 (Fig. 2, Table 3) ingestion. In this phase, drinking (1.326 the median value) was the most frequently observed ingestion behavior. In addition, the average daily count per subject for ingestion behaviors was significantly higher in phase P5.
Thermo-regulatory behavior occurred at a higher average daily duration per subject as compared to other behavioral categories (2.656 the median value; Table 1). Rutting, estrous and mating behaviors occurred at a significantly lower average daily duration per subject relative to other behavioral categories (0.036 and 0.016 the median value, respectively; Table 1). Average daily counts per subject and relative frequencies for thermo-regulatory, rutting and estrous, and mating behaviors are shown in Figure 3. Embracing, a thermo-regulatory behavior, accounted for the highest proportion of these behaviors across all phases (5.846 the median value). Presenting buttocks was the most frequent rutting and estrous behavior (0.616 the median value), and copulating (0.466 the median value) was the most frequent mating behavior across all phases. (Fig. 3, Table 3).
Resting behavior, the behavioral category of longest duration, occurred at a significantly higher average daily duration per subject as compared to other categories (11.566the median value; Table 1). Across all phases, sitting on the floor (46.276the median value) was the most frequent resting behavior, distantly followed by perching on shelf (6.976 the median value)shelf, and hanging on window or door (6.846 the median value; Fig. 4, Table 3).
Parental and amicable behaviors occurred at a moderately higher average daily duration per subject as compared to other categories (1.566 and 1.186 the median value, respectively; Table 1). Among parental and amicable behaviors, four (4)  Table 3).
Conflict behaviors occurred at a significantly lower average daily duration per subject as compared to other categories; vigilance behaviors, in contrast, occurred at a significantly higher average daily duration per subject as compared to other categories (0.016 and 10.396 the median value, respectively; Table 1). Among conflict and vigilance behaviors across all phases, watching company was the most frequent behavior (77.106 the median value), distantly seconded by shifting position (10.616 the median value; Fig. 6, Table 3).
Locomotive behaviors occurred at a moderately lower average daily duration per subject as compared to other categories (0.826 the median value; Table 1). Among locomotive behaviors across all phases, quadrupedal walking on floor (47.886 the median value), climbing (13.446 the median value), walking on shelf (5.106 the median value) , and standing (4.306 the median value) were the most frequent behaviors (Fig. 7, Table 3).
Communication behaviors occurred at a significantly lower average daily duration per subject as compared to other categories; miscellaneous behaviors occurred at a moderately lower average daily duration per subject (0.026 and 0.666 the median value, respectively; Table 1). Among communication and miscellaneous behaviors, scratching body by foreleg (21.906 the median value) was the most frequent behavior, distantly followed by scratching by hind leg (3.366 the median value), shaking body (2.796 the median value), and yawning (2.786 the median value; Fig. 8, Table 3).

Discussion
The systematic ethogram of M. fascicularis in a free enclosure precisely describes, defines and classifies the behaviors observed in the daily mundane routine of this species. The most important characteristic of this ethogram is the coding and transfer of raw video-based observations into digital data for further statistical analysis. By employing this methodology, the ethogram satisfies the technical requirements for future hypothesis-testing and quantitative studies [16,17] by enabling NOLDUS software [26] to perform advanced statistical analysis on behavioral data. Another technical merit of this ethogram is that it can be updated in real-time through direct field observation and data importation into the software for later analysis.
In addition to these technical advantages, underlying psychological elements can be assessed through an analysis of the quantitative data on observable behaviors. For example, appetite can be measured by the frequency and duration of ingestion behavior; the social relationships between subjects can be analyzed by the frequency and duration of amicable, conflict and vigilance behaviors; sexual drive can be measured by the frequency and duration of rutting, estrous and mating behaviors; activity levels can be ascertained by the frequency and duration of rest and locomotive behaviors; and social interaction and affect can be studied through the frequency and duration of communication and miscellaneous behaviors.
In the current study, resting and vigilance behaviors were the most prominently occurring behavioral categories, with sitting on the floor and watching company dominating their respective categories; this finding did not vary significantly by observational phase. It can be reasonably surmised that this species dedicates substantial time to conserving energy through rest and promoting social ties/avoiding conflict through vigilance. As to ingestion behaviors, it should be noted that there were three (3) daily feedings at 7:30, 10:30 (phase A3) and 16:00 (phase P5). The more frequent drinking behavior, absence of chewing behavior, and significantly lower average daily count per subject value in phase P2 is likely attributable to this fixed feeding schedule. Moreover, during the 10:30 feeding, primarily fresh fruit was supplied; this dietary difference may explain the fact that the average daily count per subject value in phase A3 was less than half the corresponding value in phase P5. As to rutting, estrous, and mating behaviors, it appears that these behaviors peak in frequency in phases A3-A4 and P5; whether this phenomenon is correlated or just coincident with the feeding schedule remains an open question. As to parental and amicable behaviors, the average daily count per subject value peaks and grooming behaviors are more frequent in phase A3-A4; as a speculation, the late morning may mark a time period for social bonding in this species. As to conflict, vigilance and locomotive behaviors, the average daily count per subject value peaks in phases A3-A4 and P5. This finding may be correlated with the feeding schedule, as conflict, vigilance and locomotive behaviors may be heightened in order to protect feeding position and/or food from other monkeys. As to communication and miscellaneous behaviors, the average daily count per subject value peaks in phases P2-P3; this category is primarily concerned with self-directed behaviors such as self-scratching and shaking. The early afternoon hours may mark a time period for solitary selfcaring activities in this species. There are several limitations to this study. First, while observation of M. fascicularis in the wild would have been preferable, issues of subject identification and repeat observation required the use of a free enclosure environment. Second, it should be noted that observations spanned daytime hours only (9:30-11:30 and 14:30-16:30); thus, nocturnal behaviors were not investigated nor recorded in this study. Third, parturition behavior was not observed nor recorded. Fourth, behaviors were classified according to specific ecological functions. As a result, the edges of some behavioral sets may overlap, which may lead to inaccuracies in behavioral classifications. Fifth, some less commonly occurring behaviors may have been overlooked through human error. Further behavioral studies on the species can address these limitations.
In conclusion, this systematic ethogram of the cynomolgus monkey (M. fascicularis) in a free enclosure provides a platform for future behavioral investigations and concomitant quantitative studies on this species. These studies, combined with others on non-human primates, should provide further insight into advanced cognitive function in humans.

Location
The M. fascicularis Feeding Base of Zhongke Experimental Animal Co., Ltd. (hereinafter ''the company'') is located in Suzhou, P.R.C. at E 31u079030 to 31u079060, N 120u199080to 120u199150. The company imported the M. fascicularis subjects from Guangdong and Vietnam in 1990 and established a domestication and breeding base for these monkeys.

Observation
First, the initial observations of 6012 M. fascicularis subjects followed by a detailed verification were performed by the scanning method to construct the preliminary ethogram. Then, an independent validation by focal observation was performed on a randomized cohort of 40 young adult female M. fascicularis subjects (each weighing 3-6 kg and aged 8-16 years)who were selected by means of simple random sample, seed 20101207, from the original population (n = 6012). The subjects were housed in 25 free enclosures measuring 8 m63 m63 m (L6W6H), given water and libitum, and fed daily with fresh fruit, vegetables and compound high-nutrition monkey food. Two (2) male and 20 female subjects were placed into each free enclosure to match the wild male: female ratio range of 1:7-11. There was no statistically significant   (3) high-pixel videos of the subjects were recorded using a SONY 1100 megapixel camcorder. Then, a Lenovo PC was used to transfer and convert the videos into a viewable format for analysis through NOLDUS Observer XT software (version 10.0, Noldus Information Technology, Leesburg, PA) [26]. The list of behaviors was encoded to satisfy the operational requirements of the software [27]. Three (3) qualified observers blindly watched the videos and used the software to record data on the frequency and duration of each behavior. Interobserver reliability between the three (3) observers was determined to be greater than 85% for each behavior.
A six (6)-month preliminary observation was performed from April-December 2010. More than 1000 hours of video footage was collected to obtain a comprehensive view of the subjects' behavior. A systematic ethogram was then constructed using behavioral observations collected from qualitative sampling ad libitum, focal animal sampling and instantaneous scans. Sampling was limited to daytime hours after feeders completed their daily cleaning procedure (9:30-11:30 and 14:30-16:30 daily).

Ethics Statement
All procedures described were observational under normal rearing circumstance and did not involve physical manipulation of the subjects or changes to their environment or diet. Animal care and housing procedures were in compliance with Chinese regulatory requirements (see addendum entitled ''Ethics Statement on Non-human Primate Research'') and AAALAC statements.
In brief, complete animal husbandry and veterinary care was provided daily. Animals were fed a nutritious standardized diet, supplemented daily with fresh fruits and vegetables. Animals had unrestricted access to potable water. Animal enclosures were cleaned daily. Animals were observed daily by trained care-takers. Any observed abnormality, disease or injury was reported to the veterinary staff for diagnosis and treatment; this veterinary support was documented in both hardcopy and electronic formats.
In addition, this study was performed in strict accordance with the recommendations in the Guide for the Care and Use of

Definitions
Behavior is generally defined as a ''combination of posture and action, with obvious environmental adaptation functions.'' Each discreet behavior occurs at a location within the environment. Therefore, the behavior of M. fascicularis in a free enclosure environment can be segregated into three discreet elements: posture, action and location. [25] These elements are defined as follows: Posture: a particular bearing of the subject's body. Action: a coordinated series of contractions, extensions, rotations and/or translations of the subject's body part(s) occurring in a relatively short period of time. Location: a specific position within the free enclosure environment.

Ethogram Construction
Posture Coding. Through observation, nine (9) postures were identified and encoded. These postures can be separated into two (2) groups: static postures (standing, sitting, lying, mounting, hanging) and dynamic postures (walking, galloping, jumping, climbing).These postures are defined are defined as follows: Standing: supporting the body erect on hind legs or all four limbs for at least 5 seconds Sitting: resting on the buttocks. Lying: resting recumbent on the floor or shelf. Mounting: rising upon the rear of another subject for copulation. Hanging: suspending on four limbs from the skylight. Walking: translating by steps on the floor. Galloping: translating swiftly such that all limbs leave the floor for an instant. Jumping: springing up from the floor by the muscular contraction of the limbs. Climbing: ascending by all limbs along the window or the door.
Action coding. All actions are encoded and defined in the attached addendum entitled ''File S1 and File S2''.
Posture-Action-Location System. From April to December 2010, 107 behavioral items were created based on the Posture-Action-Environment principle, and the scanning method was employed to verify these behaviors among 6012 M. fascicularis subjects in a free enclosure. 83 of these behaviors were then independently validated and classified into 12 behavioral catego-ries through focal observation of a randomized cohort of 40 young-adult female M. fascicularis subjects (each weighing 3-6 kg and aged 8-16 years) selected from the original population (n = 6012). 53 of these behaviors were then selected belonging to 12 categories based on ecological function: ingestion, thermoregulatory, rutting and estrous, mating, resting, parental, amicable, conflict, vigilance, locomotive, communication and miscellaneous behaviors. These behavioral categories are defined below: Ingestion behavior: taking food material into the subject's digestive system. Thermo-regulatory behavior: regulating body heat through movement. Rutting and estrous behavior: arousing sexual interest in another subject. Mating behavior: sexual intercourse between oppositesex subjects. Resting behavior: refraining from activity. Parental behavior: providing supervision and care to the subject's offspring. Amicable behavior: providing friendly actions towards another subject, such as grooming. Conflict behavior: struggling between subjects resulting from incompatible or opposing demands. Vigilance behavior: responding to threatening external stimuli. Locomotive behavior: engaging in activity (as opposed to resting behavior). Communication behavior: transmitting information to other subjects via gestures, emitting sounds, etc. Miscellaneous behavior: a set of defined miscellaneous actions.

Supporting Information
File S1 Behaviorial Patterns of M.F.