The study was conducted year-round, during wet and dry seasons (December 2016 to January 2019) at the Raia Olimpica, a large man-made pool (Figure 2b) on the campus of University City, Universidade de São Paulo, Brazil. 23º 33’ 21 S 46º 43’ 14 W, altitude 722 m (WGS84 EGM96 Geoid. Google Earth Satellite Imagery, 2019). The pool area is approximately 247 500 m2, surrounded by trees and extended grass areas, in the midst of one of the world’s largest megalopolises (Figure 2). The Pinheiros river, the principal corridor for capybara migration, one of two rivers partially surrounding the City, is separated by an intercity highway. The city of São Paulo lies within the Atlantic Forest domains of the São Paulo State. Peak temperatures in the summer months (December - March) reach 30ºC, with an average temperature of 25ºC and median rainfall of 170 mm. In the winter (June - September), temperatures can fall to below 12ºC, with an average temperature of 19ºC and median rainfall of ≤ 50 mm (World Meteorological Organization, 2019).

In 2013, free-living capybaras invaded the pool area from the nearby Pinheiros river through a constructional breach in the water canalization system. A total number of two adults and five pups were reported to have entered, which grew to 40 animals within a four-year period then divided into two major groups and several satellite males (status at end of 2016).

Group I: 1 dominant male (original alpha male), 14 adults (including the original adult female), and 15 juveniles.

Group II: 1 dominant male (a direct descendant of the original alpha male and female). After growing into a sexually mature male, he was expulsed by the alpha male, together with a second mature male (likely from the same litter). Over time, five adult females from group I joined the two males, forming group II.

Satellite males: distributed throughout the pool area; three sexually mature males that were driven out by the group I alpha male.

The polygynous society of capybaras represents a ridged linear hierarchy, where the alpha male agonistically protects his breeding rights. Nevertheless, subordinate females have been observed to leave the main group to reunite with satellite males, either temporarily or to form new groups. Likewise, subordinate males opportunistically attempt to mate with females from established groups (Herrera & Macdonald, 1993; Moreira et al., 2013)

For group confinement, we constructed a budget-friendly flexible corral (Figure 3a) using a scaffold frame (ea. element L: 150 x H: 100 cm). The leg extensions were forced into the ground and connected to one another using a metal wire (gauge 14). A continuous metal fence was then installed from the inside (1.5 x 45 m in circumference, ± 127 m2), fixed to each scaffold frame using the same metal wire. The top portion of the fence was curved inward and maintained in position by tightening the upper portion of the fence every three to four m with metal wire. The aim of the fence’s flexibility was to impede any push-throughs or possible jump or scaling escapes. Capture strategies for individual capybaras, namely satellite males, were twofold: (a) employ classical conditioning methods to lure the male into the corral, or (b) alternatively, use specially-constructed single-animal traps.

Figure 3
Corral & cage-traps. a) flexible corral design, b) single-trap, with removed doors for habilitation, c), depicting trapped capybara male (satellite).

A single piece of 2 x 3 m (8 mm) steel net, with a 15 x 15 cm mesh, was folded to create a continuous U-shaped cage. The cage was placed over a pre-mounted steel frame and welded to all contact points. The structure was reinforced with steel rods. The U-shaped frame served as a track for the removable wood-plates, making the cage light enough to facilitate transport. The wooden front and back doors measured 100 x 115 x 1.2 cm, and the wooden floor measured 100 x 195 x 1.2 cm. The wood was made from a waterproof medium-density fiberboard (MDF) for external use. In addition, a wheel support was mounted onto each corner so that it could be rolled away. On location, these wheels were buried into the ground, assisting in keeping the cage in place. The trapping mechanism was based on two possible guillotine-door trigger concepts: by weight, pushing the floor down when entering, or by bait contact.

During habituation of the capybara satellite males to the traps, both sides of the doors remained open. Holes were placed on only one side of the wooden floor to allow the sugarcane to be placed upright (Figure 3b). When it was determined, via observation, that the animal was comfortable enough to enter the cage and feed, the cage was set for capture. The back-side was closed, the front door remained in the open position, the bait was set, and a simple nylon string was attached directly to the bait or to the floor to trigger the guillotine-door to slide shut (Figure 3c). The track also contained foam-cushion to minimize the noise of the door closing.

Thirty days prior to the first conditioning attempt (May 2017), behavioral observations were conducted, following descriptions by Altmann (1974), for three-hour periods during the day (either 07:00-10:00 or 14:00-17:00) or during the night (19:00-22:00). Observations were carried out at distances 1-100 m by direct observation or using visual magnifying aids and recording devices. The observations were conducted by a single investigator, studying the overall group dynamics; individual conduct; sleeping, resting, and eating periods; social interactions; and breeding behavior. The collected data served as a behavioral reference for observations after the contraceptive treatment.

To evaluate the success of our strategies in reducing capture and handling associated stress, we focused on direct behavioral observations (Skinner, 1938; Aguiar & Moro-Rios, 2009), such as conduct, body language, and vocalization. Capybaras have a broad spectrum of vocal communication (Sánchez, 2003; Barros, Tokumaru, Pedroza, & Nogueira, 2011; personal communication, 2017/2018). Interpretations were done by one observer and were based on vocal (V) and body language (BL) expressions for an individual’s, and group’s emotional state, discriminating between three principal states: (1) general well-being, curiosity, and positive excitement (V: high-pitched purring-clicking, often contagious to the rest of the group. BL: inquiring, probing, sniffing closing-in); (2) distress, fear, and irritation (V: squeaking, whistle, bark, teeth-chatter. BL: neck-hair stands up, muscle tremors, urinating, fleeing); (3) agonistic behavior (V: teeth chattering, bark. BL: neck-hair stands up, sham-attacks, real attacks, biting). We created several benchmarks for the purpose of measuring the progress of classical conditioning and behavioral training.

Initially, we used the concept of classical conditioning (or Pavlovian conditioning), taking advantage of unconditional stimuli (food) that was natural to capybaras, specifically sugarcane. Once attracted by the smell and voluntarily approaching the food source, referred to as the unconditional response (not based on behavior), we accompanied each food offering with click-like oral sound. This subsequently created a classical conditioning scenario by combining the unconditional response with a conditioned stimulus, which over a short period of time, turned into a conditioned response, providing the fundamental principles to expand the strategy to operant conditioning (or, Skinner Principle).

Conditioning goals: habituation to humans; voluntary approximation, hand-feeding; habituation to corrals/traps; entering corrals/traps voluntarily (bait present); herding/entering corrals; and remaining calm during recovery. For each goal, the number/time of sessions needed for the first-time achievement was recorded.

Click-sound: starting with the first training session, a click-sound was used to achieve a conditional stimulus.

Positive experience: independent of scheduled training events, food was also placed in the corrals/traps to reinforce the association of the location with a pleasant experience.

Conditioning Sessions: fourteen days prior to the first capture event (June 2016), we initiated the conditioning process. Sessions lasted 30-60 min., three times a week, with continued behavioral observations conducted ad libitum.

Capybaras are herbivorous, and although specific grass and aquatic plants are their staple food, they do have a great desire for sugary plants, especially sugarcane (lat. Saccharum officinarum) (Borges & Colares, 2007; Ferraz, 2007; Felix et al., 2014). Sugarcane was cut into approximately 2-foot-long pieces, and either hand-fed to the capybaras (Figure 4) or placed upright into the ground to attract the animals, which works much more effectively than leaving stalks uncut and flat on the floor (personal communication, 2017).

Figure 4
Conditioning/Training/Reinforcement w/ sugarcane; a) herding b) corral entry c) corral positive experience/reinforcement d) darted animal, while other members close by w/o agitation or agonistic behavior e) return of an treated animal, with members relocating towards the end of the corral, expressing no irritation or panic, and members outside the corral calmly grassing. (red arrows – sugar cane; yellow arrow – tranquilizer dart)

Starting with the first contact, one observer emitted a click-sound to accompany each event, when approaching the animals or animals approached the observer, during food presentation/hand-feeding during herding (Figure 4a), and entering the corral (Figure 4b), with emphasis before and during food offerings. The click-sound was not used if food was not immediately associable, which was to avoid the negative effect of ‘conditioning extinction’, thus, possibly undoing the training attempts.

Aside from direct training attempts, to eventually achieve an operant conditioning effect, efforts to complement the positive experiences with the event location were made. Food was placed periodically and independently from events, into the corral and traps, reinforcing their voluntarily entrance and pleasant experience during procedures and recovery (Figure 4c–e).

After the individual or group confinement, the target animal was anesthetized with a combination of ketamine and dexmedetomidine (9 mg kg^-1, Syntec, Brazil, and 5 µg kg^-1, Zoetis, Brazil, respectively) which was administered intramuscularly into the hind leg musculature, using specialized darts and a CO₂ projector (X-Caliber, Pneu-Dart, Inc., Williamsport, Pa.).

After sedation, the animals were brought to our field clinic for biometric examination (weight, size, gender), ID marking, and health status evaluation. During different events, the procedures performed included administration of an immunocontraceptive, collection of semen, and in the final stage of the project, hemigonadectomy of the treated and control animals. After each procedure, the animal was returned to the corral, until fully recovered.

This study was approved by the Brazilian Ministry of Environment, SISBio: 54634-2 and Universidade de São Paulo’s Ethics Committee for Use of Animals in Research, Ceuavet: 9553260816.

The large flex-corral withstood any escape attempt by the animals, usually achieved by probing the fence for weak spots (pushing the head into the fence at different locations). No scaling or jumping was observed. The fence’s flexibility and the use of metal wire circumvented any breakage of the structure and prevented injuries to the animals. The corral’s large oval shape allowed the trapped animals to relocate to different parts when humans entered to retrieve or return a sedated animal. We believe that this key design feature contributed to avoiding panic by providing enough ‘safety-distance’. The prior habituation efforts, in conjunction with the ‘pleasure’ experience of finding reinforcing food in the corral and traps when entered voluntarily, contributed toward the calm capture, which was maintained throughout the duration of the animals’ captivity.

After several cage-building trials, the final concept produced a cage/trap with a light-structure-design, allowing for easy relocation, as site adjustments were frequent. Although lightweight (33 kg), its structure was resistant enough to contain the trapped animal for prolonged periods of time. The trapping mechanism used was efficient in automatically trapping the animal.

The first crucial task was to habituate the groups to our presence and for the animals to associate us and the click-sound with food. Surprisingly, very little time was needed to achieve this. The first session started with an initial distance of approximately 50 m, slowly moving toward the group while using the click-sound for the first time. Then the investigator stopped and remained in a position approximately 20 m away. The mood status of the animals was observed to be: (V) some animals barked; (BL) some animals got up, some were cautious or perhaps fearful, most of the group members turned and faced the direction of the water, and some juveniles dived into the water. Time invested at the location was approximately 20 min.

The second session, carried out on the following day, started from 50 m, the mood status of the animals was observed: (V), an initial two or three barks. The investigator approximation was done in intervals of 5 m, using a click-sound, until they remained approximately 10 m from the first members of the group. Sugarcane was forced upright into the ground, while also holding two pieces in the hand. The food was positioned up-wind to aid in the olfactory detection of the sugarcane. (BL), the animals closest to the observer got up into a sitting position, while the other animals remained as they were. Within five min., two individuals started to move toward the bait and carefully started to feed. Within a further three min., most members had gotten up and walked toward the sugarcane. The original members turned toward the investigator, hesitantly and cautiously starting to accept the food by hand. (V), animals made the singing-like high-pitched purring-clicking sound that is associated with positive excitement (personal communication, 2017). Throughout the entire session, the investigator emitted the click-sound intermittently. The goal of accepting human presence and hand-feeding was achieved in only two sessions and a total time of ≤ 50 min. was invested.

During the second session, the corral and trap were prepared with several pieces of sugarcane (Figure 4c). Some was placed at the entrance and some was distributed throughout the corral/trap. During the next visit on the following day, the sugarcane had been consumed, and there was evidence of capybara having visited, in the form of feces.

Session three occurred two days after session two at approximately 09:00 and was the first capture event for procedures. The same investigator and two additional staff members, with sugarcane in hand, got close to the group. The distance between the group and the corral was approximately 220 m. While closing in, using the click-sound, the investigators stopped at a distance of approximately 10 m from the group. Within 2 min., the first members (apparently the same individuals from session two) came straight toward the food offerings. (V), on their way, they were observed to be emitting the purring-sound, and shortly after, three more members followed. The investigators used alternating ‘stop’ and ‘go’ to lure the capybaras toward the corral, using the bait as a tease and periodically allowing the animals to chew for a few seconds on the sugarcane, until the corral was reached (Figure 4a). Half-way through, some members started to stay behind, but they eventually reunited with the group. As the capybaras dictated the progress rhythm, the walk was slow; although, the goal of herding the animals over a lengthy distance was achieved.

After a 30 min. walk, the five individuals from the herding attempts would (BL) cautiously follow the investigator into the corral (the attraction of the bait seemed to be stronger than the uncertainty). (V), the animals would continue to emit the purring sound. The goal of entering the corral/first capture was achieved in one session, with the total time to capture being ≤ 20 min. (Figure 4b). We did recognize that the individuals initiating the approach, first to be hand-fed, and first to be herded and enter the corral, were not the dominant male or female (identified by body size and phenotypic characteristics). Another unexpected side effect, after taking the first five members into the corral, was that within the next 20 min. the entire group, including the dominant adults, had gathered around the corral, trying to get in, with multiple members emitting the purring sound. During this first procedure event, small pieces of sugarcane were given to the members periodically, inside and outside the corral, to maintain their interest and to sustain the overall tranquil atmosphere.

During the anesthetic application and animal removal (Figure 4d) and after procedures, when returning sedated animals for recovery, the other group members relocated to the far-most side of the corral but remained tranquil (Figure 4e). (V), after the initial capture, no further purring-sounds were emitted. (BL), other than some probing at the fence, sitting down, or chewing sugarcane leftovers, no apparent fearful or irritated behaviors were observed, nor were any agonistic behaviors (toward other group members or staff). The goal of animals remaining calm during manipulation and recovery was achieved. The total time of the first event time (time of intervention) was approximately six hours.

After the first event, food was offered by hand-feeding, accompanied by the click-sound biweekly, during 30 min. reinforcement sessions. In addition, the corral was prepared with bait to maintain/reinforce the positive association with the location.

At each consecutive conditioning session, the response to the click-sound was almost immediate. In less than five min. (a fast response time for capybaras), most of the group members would approach the investigator to receive their food-treats, (V) making the purring-sound for the remainder of the feeding time. (BL), no stress-behavior was observed. Periodically, members and satellite males were observed inside the corrals, even without food present.

Over a period of 18 months, a total of 5 events were executed; each time the group members were successfully called and herded into the corral. Throughout all sessions and events, no agonistic or panic-like behavior was observed, except the occasional superficial attack by dominant animals toward subordinate members for the right to feed first.

There are a myriad of techniques available to measure stress, which are mainly endocrinological and monitor stress-associated hormones or their metabolites. In addition, there are many well-defined ethological programs (DeNicola & Swihart, 1997; Romano et al., 2010; Copeland, 2015). However, for this project, because of logistics, resources available, and the irregular capture of free-ranging capybaras, the observational interpretation of vocalizations, behavior, and body language, was considered the only viable form of measurement. Although aware of potential investigator bias and the risk of subjective interpretation, driven by strong expectations (Tuyttens et al., 2014), we believe that the results are clear and objective enough to avoid any misinterpretation.

All conditioning goals were achieved in relatively short time, such as habituation to human presence; voluntary approximation, allowing to be hand-fed; habituation to, and voluntarily entering, corrals/traps; and allowing to be herded over distances and taken into the corrals. Equally important, the aims to maintain a calm atmosphere when handling the animals in the corral, removing a sedated animal, or retuning the same animals for anesthetic recovery, with other members present, were all achieved. The conditioning method allowed the investigators to execute the selective capture of target animals at any time when needed, compared with random trapping. As there is no available literature on similar methods or experience with this species, it is not possible to discuss the time investment or quality of the conditioning.

Capybaras are inquisitive animals and, when there is no threat present, they are quite calm and approachable, especially if food is involved. How quickly they respond to offered food depends on the individual animal’s character and experience, as well as the transient group dynamics. As soon as one member breaks the contact barrier, the rest will follow shortly, suggesting a learning behavior by imitation. The capybaras quickly learned to associate the tongue-click sound with food, and after only a few repetitions, the approaching vehicle and individual staff members were associated with food and greeted with excitement.

Great caution is warranted to not feed excessive amounts of sugarcane to capybaras, as this may lead to intensive fermentation, potentially provoking tympany. In addition, working with hand-fed bait, a calm situation can quickly turn into a light feeding-frenzy when the animals get too close to each other, leading to fights of feeding-dominance.

Based on our behavioral observations, the capture strategies employed in this study appeared to provide several positive effects. Using the concept of classical conditioning was successful, with animals rapidly associating the click-sound with food offerings and approaching the investigator to be hand-fed. Animals could easily be herded into the corral while maintaining a calm demeanor throughout all capture events. These results demonstrate that the applied conditioning methods for capybaras are useful in facilitating frequent capture needs.