Environmental enrichment techniques for black-eared-opossuns (Didelphis aurita Wied-Neuwied, 1826) in captivity

Abstract This study aimed to analyze the efects of environmental dietary enrichment on the behavior of black-eared opossums that were in captivity at Centro de Triagem de Animais Silvestres (CETAS-ES), in the municipality of Serra-ES, from August to October 2022. Two food models were constructed for the study: the “Surprise Tubes” and the “Food Puzzle”. A total of 24 juvenile black-eared opossums were selected, regardless of sex, and divided into eight groups, with three animals each. Then, four groups were exposed to one model and four to the other. Each group was submitted to two conditions: experimental, with the presence of the models, and control, without the presence of the models. Each condition lasted 24 hours and occurred on two consecutive days. They were flmed with a camera trap, resulting in 3,233 videos of 25 seconds. For the elaboration of the ethogram, 24 videos of each group were selected, recorded from 6:00 PM to 7:30 PM, when the animals were more active. The opossums interacted with both models, accessing, and eating the hidden food, preferring meat over fruit. No signifcant diferences were found in the behavior of opossums in relation to the models and it was observed that the aggressive behavior signifcantly decreased when the models were present. It is concluded that the models can be used as an environmental enrichment for black-eared-opossums, bringing benefts to the reduction of aggressive behavior.

threatened, black-eared opossums do not usually attack, but exhibit aggressive behavior that includes opening their mouths to show their teeth and/or emit characteristic vocalizations (3)(4) .Furthermore, they feature deathfeigning (thanatosis) and release a foul-smelling odor, which simulates unhealthy conditions for consumption, causing the predator to lose interest in preying on them (5) .
Black-eared opossums are able to live both on land and high on trees, grasping and climbing branches with their long prehensile tail and short five-toed hands and feet, the first toe being clawless (6) .Due to their ecological plasticity, they show great adaptive efficiency to the most varied habitats, easily adapting to the environment modified by humans, including rural and urban areas (7) .
Despite the existence of laws to protect Brazilian fauna, such as laws no.5197/1967 and 9605/1998, the abuse and mistreatment, hunting, or catching of wild animals are still widely practiced.Regarding black-eared opossums, these are often killed by attacks from domestic animals and humans or road kills (8)(9)(10)(11) .
Several factors compromise animal welfare in a captive setting.The welfare status of captive animals is an attribute that is linked to their quality of life and the way they interact with their surrounding environment.Animals face potential challenges in these interactions that can lead to frustration, scarcity, or overstimulation (12) .In this sense, the incorporation of environmental enrichment techniques has emerged in an effort to improve the welfare of animals in captivity by providing appropriate environmental stimuli that mimic natural situations.It aims to create a more enriching and interactive environment (13) , which allows the expression of characteristic behaviors of each species, positively influencing the physical and psychological development of the animal (14) .Moreover, the use of entertainment strategies, based on the use of games or cognitive proposals of greater complexity, also favors the adequacy of the captivity environment, improving the welfare of animals (15) .
Environmental enrichment, a field of animal behavior recognized by Yerkes and Hedinger in the first half of the 19 th century, studies the importance of the physical and social environment for the welfare of captive animals by providing elements that allow them to express the natural behavior of the species (16)(17)(18) .This technique consists of incorporating elements into captivity that can reproduce the natural habitat of the species, allowing the animals to exercise common activities of the wild, such as locomotion, foraging, and sheltering, among others.According to Mcphee et al. (19) , there are five types of environmental enrichment: Physical: It concerns the physical structure of the enclosure, the place where the animals are kept.For this type of enrichment, elements are introduced so that the enclosure resembles the animal's original habitat as much as possible.Examples include the use of branches, substrates, vegetation, and platforms.
Sensory: This type of enrichment is one of the most used and consists of stimulating the animals' five senses: visual, auditory, olfactory, tactile, and gustatory.Examples include aromatic herbs and sounds with vocalizations.
Cognitive: It is aimed at stimulating the intellectual capacities of animals, being done by means of toys or devices, with which animals are encouraged to interact to get a reward.
Social: It is linked to intraspecific or interspecific interaction that can happen within the enclosure.In this enrichment, animals have the opportunity to interact with other species that they would naturally live with in the wild or with individuals of the same species.
Dietary: It concerns the way the animals are fed in captivity.In general, attempts are made to offer a diet closer to that found in the wild, by providing carcasses, hidden food, and/or changes to routine feeding times.
Considering the scarce reports of environmental enrichment aimed at marsupials in captivity, especially for South American species, the objective of this work was to analyze the effects of dietary enrichment on the behavior of black-eared opossums, aiming to contribute to improvements in the treatment of this species kept in captivity.

Study area
The study area was the Brazilian Wild Animal Screening Center (CETAS) owned by IBAMA-ES, the Brazilian Institute of the Environment and Renewable Natural Resources, located in the Environmental Protection Area (APA) of Jacuném Lagoon, in Barcelona, in the municipality of Serra-ES, Brazil.The research was authorized under order No. 12950928/2022 and case No. 02009.000902/2022-13,published in the IBAMA system.
CETAS-IBAMA-ES is a unit that aims to receive, identify, tag, screen, evaluate, recover, rehabilitate, and allocate healthy wild animals back to nature (20) .In Espírito Santo, these animals are brought in from inspection seizures, rescues, or voluntary deliveries from individuals.They come from the Greater Vitória region, which comprises the municipalities of Vitória, Vila Velha, Cariacica, Viana, Guarapari, Serra, and Fundão.
Black-eared opossums are brought to CETAS-ES through various means including inspection seizure, rescue, and private individuals.These opossums are commonly victims of domestic animal attacks, cars collisions, or electrocutions.Thenat CETAS-ES, they are weighed, sexed, inspected, and evaluated by the caretakers and/or volunteers of the Marsupial Project.Subsequently, they are allocated and housed in enclosures (74 x 70 x 76 cm) until they are considered ready for release.The Marsupials Project is part of Últimos Refúgios Institute, a non-profit socioenvironmental and cultural organization.Since 2017, the institute has been actively involved in the preservation of Brazilian marsupials, collaborating with CETAS-ES for their rescue and rehabilitation (21) .
Since they are omnivores, black-eared opossums are fed fruits such as bananas, apples, oranges, papaya, and grapes daily.They are also fed meat, which can be quail or rat, three times weekly.The enclosures are cleaned every morning, at the same time as the water in the drinkers is changed and the leftover food in the feeders is removed.Young opossums are provided with special care and diet, including a blend of goat's milk and various fruits, supplied in 1-ml syringes.Moreover, if they are in litters, these are always kept together.

Animals
For the experiment, a total of 24 black-eared opossum juveniles were selected, including males and females, with weights ranging from 58 to 142g, under evaluation at CETAS-ES.The opossums were divided into eight groups, with each group consisting of three animals (Table 1), which were allocated to a cage with dimensions of 74 x 70 x 76 cm (Figure 3).The groups were formed according to the animals' arrival at CETAS-ES; therefore, some groups consisted of animals from the same litter.For identification purposes, each opossum in the group was individually tagged: the first animal with adhesive tape on the left paw, the second animal with adhesive tape on the right paw, and the last animal was not tagged.

Dietary enrichment
Dietary enrichment was achieved by the introduction of food hidden in containers constructed especially for the study, known as models, which were adapted to meet the specific needs of the black-eared opossums (D. aurita).

Model 1 -"Surprise Tubes"
Created based on the model developed by animal caregiver Aisa Coco (22) for Procyon lotor (raccoons), the "Surprise Tubes" model consists of a wooden rectangle measuring 20.5 x 25.5 cm, to which small PVC pipes of various shapes were attached (Figure 1).The food (animal feed, meat, and fruits) was concealed inside the pipes, encouraging the animal to reach for the food.

Model 2 -"Food Puzzle"
The "Food Puzzle" model was proposed by Banton-Jones (23) to be used with striped skunks (Mephitis mephitis).It consists of a wooden box with dimensions of 23.5 x 12 x 8.5 cm, which contains three smaller boxes measuring 5 x 5 cm, each with a lid and handle (Figure 2).The food items were stored separately in the smaller boxes, encouraging the opossum to attempt to open the boxes and obtain the food.

-Collection procedure
The eight groups of three opossums were randomly distributed among the enrichment models, with four groups subjected to Model 1 and the other four to Model 2. Each group underwent two conditions: experimental and control, over two consecutive days.The order of the conditions was randomized among the groups (Table 1).
In the experimental condition, or dietary enrichment condition, the behavior of the animals was recorded when they were subjected to the model.The food was placed inside the experimental model, which was then introduced into the cage in the late afternoon, around 16h:30, remaining there for 24 consecutive hours.In the control condition, the behavior of the animals was recorded in the absence of the model, also for 24 consecutive hours.In this condition, the food was provided in standard CETAS-ES containers, which are rounded, with a diameter of 13 cm and an opening at the top.
The behavior of the animals in each condition was recorded using a KOVOSCJ camera trap, model H982, with 1080P full HD image quality and night vision.The camera was positioned outside the cage, supported by a tripod, facing towards the dietary enrichment model (Figure 3).Considering that opossums have crepuscular and nocturnal habits (24) , the recordings were made from August 27 to October 22, 2022, starting at 16h:30 and ending at 04h.The camera was programmed to be triggered by animal movements and each recording lasted 25s.

Analysis procedure
For the analysis of images, the Boris software program (Behavioral Observation Research Interactive Software) was used to record the duration and frequency of occurrence of observed behaviors in the videos.
Comparisons between different enrichment models were conducted using the Mann-Whitney test, which is suitable for comparing independent nonparametric measures.This analysis allowed for determining if there were significant differences in behaviors concerning the different models used.All analyses were conducted using the JASP 0.16.1 software program, which provides statistical tools for data analysis.The choice of non-parametric tests was due to the absence of normal distribution in the observed data.
Thus, using only the videos recorded from 18h:00 to 19h:30, when the animals were most active, behaviors were identified and categorized to create an ethogram (Table 2).The observed behaviors were categorized as an event or a state.For events (short behaviors), the frequency of occurrence was recorded, and for states (prolonged behaviors), the duration of the behaviors was recorded.

Table 2. Behaviors identified in the recordings of Didelphis aurita and their respective definitions
Note: Behaviors with the superscript number 1 represent a state, whereas those marked with the superscript number 2 represent an event.* Regarding the eating behavior, food was provided in bowls during the control condition and within the model during the experimental condition.In all conditions, meat, fruits, and animal feed were provided.

Behavioral category Behavior Description
Exploratory 1-Explore environment (stationary) 1 The opossum remains stationary in the enclosure, making regular head movements from side to side, as if sniffing the air.
2-Explore environment (walking) 1 The opossum moves around the enclosure, making regular head movements from side to side, as if sniffing the air.
3-Explore the model 1 The opossum moves over or around the model, looking, sniffing, biting, or manipulating the model with its hands.The opossum brings the food to its mouth using its hands or takes the food with its mouth and ingests it.

5-Drink 1
The opossum drinks the bowl water.
Body maintenance

7-Yawn 2
The opossum performs involuntary and long inhales with its mouth open.

Movement 8-Climb 1
The opossum climbs the branches present in the enclosure or climbs the bars.

9-Manipulate cage bars 1
The opossum scratches and bites the cage bars.
10-Enter the burrow 2 The opossum moves into the burrow.

11-Leave the burrow 2
The opossum moves out of the burrow.
12-Carry food with tail 1 The opossum uses its tail to carry food from one place to another.
13-Carry object with tail 1 The opossum uses its tail to carry the object from one place to another.
14-Carry food with hand/mouth 1 The opossum uses its hands or mouth to pick up food and carry it from one place to another.
15-Carry object with hand/mouth 1 The opossum uses its hands or mouth to carry an object from one place to another.Social 16-Aggressive social interaction 1 Biting, scratching, stealing food from another animal in the group.17-Peaceful social interaction 1 Smell and look at another animal within the group.
18-Look at another animal in the cage next door 1 Direct eye contact with another animal from the adjacent enclosure.
Rest 19-Stay in the burrow 1 The opossum remains stationary inside the burrow.

Qualitative results
It was noted that among the four groups exposed to the "Food Puzzle" model, two of them managed to open all three boxes, and the opossums consumed all the provided food.In the remaining two groups, only one of the boxes was opened; one contained feed, the other contained meat, and the rest remained closed.Regarding the groups exposed to the "Surprise Tubes" model, it was observed that the opossums inserted their snouts into the Noronha C E et al.

2023, Cienc. Anim. Bras., V24, e-76165E
holes of the pipes to reach the food, thus gaining easier access.No food was found in any of the groups at the time of the removal of the "Surprise Tubes" model.
Another behavior observed in the eight groups of black-eared opossums was the relationship between the size of the animals and the ease of feeding in the enrichment models presented.The D. aurita specimens with larger size and weight were able to open the boxes of the "Food Puzzle" model more easily, using mostly their mouths, whereas the smaller opossums presented to the same model, besides taking longer to open the boxes, managed to open them unintentionally by bumping into the boxes.In contrast, in the "Surprise Tubes" model, the smaller opossums had more agility and ease in reaching the food that was inside the tubes by introducing their snouts into the holes of the tube and entering the spaces between the tubes, whereas the larger opossums had more difficulty in reaching the food.
During the filming of the experimental condition of one of the groups subjected to the "Food Puzzle" model, two opossums died.They were found dead in the cage, near the entrance of the burrow, early in the morning, by the CETAS-ES caretaker who collected them, so it was not possible to see if there were apparent injuries, leaving the cause of death inconclusive.Therefore, it was necessary to start a new filming with new subjects.The data from this group were not used for the analysis.In the footage, it was also observed that the surviving opossum moved rapidly throughout the cage and did not interact with the model.No document on the cause of death of the two opossums was found in CETAS-ES.
Despite being present in the ethogram, since they were observed in the first analysis of all videos, the behaviors of carrying food/object with the tail were not observed in the videos selected for quantitative analysis and, therefore, are not included in Table 3.Moreover, the behavior of carrying objects with the mouth was witnessed, with the lids of the boxes of the "Food Puzzle" model, as well as twigs and leaves, being taken into the burrow.

Quantitative results
The frequency and duration of opossum behaviors in the presence (experimental condition) and absence (control condition) of the model were compared using Friedman's tests (Table 3).A significant difference was observed between the experimental and control conditions in only two behavioral categories: social interaction and eating behavior.In the social interaction category, it was found that there were more aggressive than passive interactions (Table 3).Aggressive interactions were significantly more frequent than passive interactions both in the presence (t = 4.32, p < .001)and in the absence (t = 3.24, p = .011) of the model, shown by post-hoc comparisons with the Bonferroni correction.However, when comparing the experimental and control conditions, aggressive interactions were significantly more frequent in the absence of the model (t = 3.51 p = .005).Effect size was moderate (Kendall's W = 0.3).
Differences were seen in the duration of eating behavior, and the opossums ate more meat than fruit.(Table 3).The post hoc comparisons pointed to a significantly higher consumption of meat than fruit before the model (t = 3.01 p = .02).However, the effect size was small (Kendall's W = 0.16).Statistically significant differences were not found when comparing the frequency and duration of behaviors in the presence of the Tubes model and the Puzzle model using Mann-Whitney tests.Table 3 shows that the standard deviation values were relatively high compared to the means, indicating individual variation in the frequency and duration of behaviors.The absence of significance in most of the performed comparisons (between experimental and control conditions and between models) may be related to this large behavioral variability observed between individuals.

Discussion of qualitative results
Despite the fact that the size of D. aurita influenced the interaction with the presented enrichment models, the objectives of interaction with the model and access to food were achieved.These results were also achieved by Murray et al. (25) and Hogan et al. (26) , in their research using food enrichment for Petaurus australis and Lasiorhinus latifrons, respectively.Murray et al. (25) had positive results in their experiment with Petaurus australis, an Australian arboreal marsupial with a predominant diet of eucalyptus nectar, sap, and pollen.The authors report that the experiment presented expected results since P. australis interacted with and investigated the gum tree, which was used for the food enrichment technique.Hogan et al. (26) studied the natural foraging behaviors of Lasiorhinus latifrons, a herbivorous Australian terrestrial marsupial, by exposing them to a circular lawn with roots, a eucalyptus branch, and buried food.
Interactions also occurred with parts of the "Food Puzzle" model outside the feeding situation.One of the animals was observed carrying the lid of a box into the burrow.The observation occurred in a video that was not selected for analysis since it was filmed outside the time range of greater activity of the animals.The observation of the behavior suggested, in a first analysis, the possibility of nest building (27) .However, since it was a behavior exhibited by only one opossum, the largest specimen of all participating animals, in a group with only juvenile females, excluding the possibility of pregnancy or nursing of joeys, it was later found that the animal had used the box for foraging in the burrow.A female was seen carrying the box into the burrow, in addition to the lid, twigs, and leaves, sleeping on the objects taken.
The death of the two opossums was considered an unexpected fact, as it is not common at CETAS-ES for juvenile specimens to die compared to neonates.The specimens that died had been at CETAS-ES for at least a week, were from different litters, and were healthy at the beginning of the recordings, before being exposed to the "Food Puzzle" experimental condition.According to the study by Baggio (28) , agonistic and antisocial behavior, common in representatives of the genus Didelphis, is already present in the juvenile development stage.The author studied opossums in captivity and observed that young animals should be separated in different environments since disputes, fights, and even cannibalism can occur.The author reports that she observed some juvenile opossums escaping from the enclosure and entering another, with younger opossums, where a specimen was found feeding on one of the animals.In another study, by Kajin et al. (29) , who examined a population of black-eared opossums in an Atlantic Forest area in Rio de Janeiro for nine years, a higher mortality rate was observed in lactating and young opossums, while the values found for adults were lower.The authors report that the high mortality of neonates may be linked to the reproductive strategy of marsupials, which show little investment in gestation and great investment in lactation, and the death of lactating opossums becomes common due to this stage of life being more critical for survival.In the case of weaned joeys, the authors explain that the high mortality rate may occur due to the vulnerability still present at this stage of life.

Discussion of quantitative results
We expected to obtain more significant differences between the conditions with the presence and absence of the models, especially in behaviors such as exploring the environment since they occurred frequently.Black-eared opossums are known to have intense scansorial activity, i.e., great climbing ability, as reported by Vieira e Camargo. (30).The relatively high standard deviation values indicate that behavior varied widely around the mean, suggesting large individual variation and implying that studies with larger sample sizes could assess differences between models and between experimental and control conditions.The significant occurrence of more aggressive behavior in the absence of the model indicates that enrichment may have an effect on decreasing the aggressive behavior of opossums.This result provides support for the hypothesis that environmental enrichment can promote increased welfare in captive animals.
Black-eared opossums are considered solitary and antisocial animals and usually tend to avoid contact with other individuals of the same species by acting aggressively when in contact with another, except during breeding periods (31) .In this study, no sociability behavior was observed among black-eared opossums, even in groups with juveniles from the same litter, confirming the common solitary habit of the species.Some opossums reacted defensively when another opossum approached to catch food, threatening to bite, or advancing to push it away.Thus, the decrease in aggressive behavior in the presence of the models is indicative of an improvement in the quality of life of captive animals.
Furthermore, the results also indicated a significantly higher consumption of meat compared to fruit by black-eared opossums.According to Santori et al. (32) , the preference of D. aurita for a certain type of food is not yet clearly established, as the available information on the feeding habits of these animals comes from analysis of fecal and stomach contents, as well as direct observation.Carvalho et al. (33) reported that although protein consumption from arthropod ingestion was one of the main diet items of the black-eared opossums they analyzed, there were no significant differences in relation to a diet composed of fruits and/or seeds when comparing animals from different climatic seasons, sexes, age classes, and habitat fragments.In turn, Hsu et al. (34) and Hume (35) found that the Virginia opossum (Didelphis virginiana), which is closely related to and has a similar diet as the black-eared opossum (D. aurita), is an opportunistic omnivore that, when kept in captivity, can survive on dog food with some natural supplements, such as fruits and crickets.
Since generalist food consumption by D. aurita is common, they easily adapt to different types of diets (36)(37) .The feeding of black-eared opossums is directly related to food availability, so a hypothesis for meat preference, as observed in the present study, meets the four factors that influence an animal's food choice, according to Owen (38) : availability, palatability, accessibility, and the energy return obtained from the food.

Conclusion
The research demonstrated that black-eared opossums (D. aurita) successfully interacted with the "Surprise Tubes" and "Food Puzzle" models, adapted and built for the experiment, managing to access and eat the hidden food.Furthermore, it is important to note that significant effects were found when comparing the experimental and control conditions, such as the greater aggressive interaction in the absence of the enrichment models compared to when they were present, indicating that the models had an influence on decreasing this behavior.Another significant result was the higher consumption of meat than fruit in the presence of the models, which may be indicative of the dietary preference of these animals in captivity.We conclude that it is possible to apply environmental enrichment models for captive opossums using simple and low-cost materials.The results suggest that the use of enrichment can stimulate behaviors characteristic of the species, such as foraging and climbing, for example, and decrease aggressive behavior, providing an improvement in the quality of life for these animals in the captive environment.We highlight that this study used a small sample size, with four groups of three animals each, to test each enrichment model.Thus, we indicate that more studies with larger samples of captive black-eared opossums are needed to confirm the results found in this study.

Figure 3 .
Figure 3. Positioning of the trap camera for the experiment recording

Table 1 .
Distribution of environmental enrichment models and group composition

Model Order of presentation of the experimental condition** Individuals Sex Weight at the beginning of the experiment
Note: * The individuals selected to form these groups come from different litters.** 1 st experimental condition (presence of the model) before the control condition; 2 nd control condition before the experimental condition (see text for detailed explanation).

Table 3 .
Frequency and duration of opossum behaviors in the presence (experimental condition) and absence of the model (control condition)