Breeding and reproductive behaviour of the neotropical opossum , Didelphis marsupialis insularis , Allen 1902 under captive conditions

Given the increase in hunting pressure on wildlife in developing neo-tropical island nations, captive wildlife production systems have been experiencing increasing popularity in these regions (Brooks et al. 2010; Bulte and Damania, 2005; Hoffman and Cawthorn, 2012; Nogueira and Nogueira-Filho, 2011). Non-domesticated species like the agouti/cutia (Dasyprocta leporina, Dasyprocta spp.) and the lappe (Cuniculus paca/Agouti paca) (Brown-Uddenberg et al., 2004), have all been successfully studied and bred under captive conditions using production systems. These captive wildlife systems have proven to be effective at meeting their required objectives that include, serving as a conservation and population management tool and further functioning as a form of sustainable utilization of neo-tropical species in the region. Abstract Developing a captive breeding system for the widely hunted Caribbean species of opossum Didelphis marsupialis insularis can greatly aide in the management and conservation of this species in the neo-tropics. Although this species possesses ideal traits for captive breeding in this region (tolerance to high heat and humidity, high reproductive rate, and resistance to disease), challenges due to its aggressive behaviour and limited information on its breeding behaviour have prevented a system from being developed for this species. The present study describes a breeding system, and the reproductive behaviour of this species under captive conditions. Six (1 male; 5 females) adult opossums were maintained and managed for breeding over a ten (10) month period. Pouch litter sizes averaged 5±2.5 with a range of 2 to 8. Gestation length was found to be 13.25±0.96 days and 4 litters (n=23) were successfully weaned at 11-13 weeks. It was found that the male D. m. insularis exhibits behaviours of interest that can serve as indicators for receptivity of the female, and overall, that this species can be successfully reared and bred under captive conditions in the neo-tropics.


Introduction
Given the increase in hunting pressure on wildlife in developing neo-tropical island nations, captive wildlife production systems have been experiencing increasing popularity in these regions (Brooks et al. 2010;Bulte and Damania, 2005;Hoffman and Cawthorn, 2012;Nogueira and Nogueira-Filho, 2011). Non-domesticated species like the agouti/cutia (Dasyprocta leporina, Dasyprocta spp.) and the lappe (Cuniculus paca/Agouti paca) (Brown-Uddenberg et al., 2004), have all been successfully studied and bred under captive conditions using production systems. These captive wildlife systems have proven to be effective at meeting their required objectives that include, serving as a conservation and population management tool and further functioning as a form of sustainable utilization of neo-tropical species in the region.
Didelphis marsupialis insularis is the popularly hunted opossum species that inhabits the smaller Caribbean islands of Trinidad and Tobago and the Lesser Antilles . The neo-tropical opossums' high reproductive rate, ability to tolerate tropical climates and a wide, omnivorous diet has led it to being identified as an ideal candidate for captive animal breeding in the region (Tardieu et al., 2017). However, this species also poses special challenges that hinder the development of a captive breeding system, namely its aggressive nature and the lack of information available on its breeding behaviours. To date, there have been no effective captive breeding nor captive management systems recorded for this species in the Caribbean.
As many studies have highlighted differences even among closely related species, in Didelphis (Caceres, 2005;Tardieu et al., 2017Tardieu et al., , 2019b, the utilization of captive breeding systems conducted for related Didelphis species may not be widely applicable. Therefore, studies are required on this particular subspecies to determine captive reproduction management in the neo-tropical region.
The overall aim of this research was to trial a captive reproduction system for D. marsupialis insularis on the neo-tropical island of Trinidad and Tobago, and further to note the reproductive behaviours of this species under these conditions. If successful, it could eventually serve as a useful conservation tool and further be used for animal production in the Caribbean region.
The specific objectives of this study were (1) to develop and test a captive housing and management system for the breeding of D. m. insularis in the neo-tropics and (2) to document the reproductive parameters and behaviours of D. m. insularis in captivity.

Housing and Caging design
The study site was located in Chaguaramas, Trinidad and Tobago (10.6825° N, 61.6341° W) and the breeding system was run for a 10 month period from January 2018 to October 2018. The animal caging system was housed within an indoor compound with natural lighting, using a 12 hour light and 12 hour dark cycle, supplemented by fluorescent bulbs, electricity and water. Cages were constructed to hold adult opossums individually and the custom caging design was based on the 'Single Corridor housing system' described by Tardieu and Garcia (2018) and illustrated in Figures 1, 2 (Figures 1and 2).
Fine wire mesh (approximately 0.03 m x 0.015 m) was utilized for the entire enclosure to achieve the following 1)   prevent the damage and slippage of opossum feet through the caging unit, 2) prevent any aggression and harm to the animals that might occur from tails or limbs slipping though the wire holes to the neighbouring cages and finally 3) to reduce access to pest species like rodents.
In order to observe the behaviour of the adults in the evening and night-time periods, an outdoor camera system with three units was utilized (Zmodo TM 720HD Outdoor WiFi IP Camera). Camera units were placed in fixed locations and positioned to offer the following views; Camera 1: Bird's eye view of the entire unit; Camera 2: Inside view of the male unit; Camera 3: (Roving camera) view of any one female that was being bred. Behavioural observations were conducted using an ad libitum sampling method, using an interval type of data recording. Each animal enclosure unit was furnished with a food and water bowl and a 'bucket-type' den as described by Tardieu and Garcia (2018

Animal collection and management
Didelphis marsupialis insularis adults were wild-caught from various locations in the Republic of Trinidad and Tobago and were captive held separately for at least three days before being introduced into the breeding system ( Figure 4). Opossums were weighed and possessed an average body mass of 2.7 ± 0.8kg and were between one to three years old, based on body condition and teeth wear (Ramírez and Osorio, 2014). A total of six (n=7) animals were utilized for this trial study; females (n=6) and males (n=1). The experimental system held during breeding trials, a total of 6 adults (5F:1M). All females were identified as being in reproductive condition (Table 1). Water was offered ad libitum and the diet consisted of commercial based dog chow, seasonal fruits and vegetables and meat/protein items (e.g. fish, chicken parts). The diet was based on reported diets fed to captive specimens in rehabilitation facilities in the neo-tropics (W. Rollock, personal communication, 2017). Food was offered once daily, between 17:00-18:00h.

Reproduction/breeding management
The male behaviour was closely observed for behaviours of interest. Using the camera system, observations were conducted following daily feeding and cleaning procedures, to determine if the female was in oestrus, or receptive for breeding based on the findings of Tardieu et al. (2019a). Once interest was detected by the male, the male would be allowed access to the female via the removal of the access panel ( Figure 2). The breeding pair would be closely monitored and the male freely allowed access to the female, until either mating had been observed or a given period of time (4-5 days) with no mating had elapsed. The access door to the female would then be closed and the male and female separated into their individual cages.
After mating the female would be monitored for a period of 11-14 days for signs of parturition. For this study observation of unsuccessful newborns (newborns that did not make it into the pouch to attach to a nipple) on the cage flooring was utilized as signs of parturition. Didelphis are reported as producing more young than they can support (Krause and Krause 2006) resulting in many newborns not surviving to the next phase of development. A week following the 11-14 day parturition period, the female would then be restrained and her pouch checked for suckling young. After a period of 12-13 weeks, (time period based on by Tardieu et al. (2019a) young opossums would be weaned (removed) from the female to allow the female to begin cycling again. Descriptive statistics were used and means were provided through the manuscript when necessary.

Reproductive behaviours
Male behaviour provided fairly reliable indicators of female receptivity. These included three noteworthy actions of interest; (1) increased interest in the form of time spent near to the female's enclosure door, (2) marking and (3) clicking vocalizations similarly reported by Tardieu et al. (2019a) these reproductive behaviours are detailed in Table 2.

Reproductive parameters
A total of five mating episodes with four parturitions were documented over the ten month period. Four litters were successfully weaned from females over this period.
Mating was observed to occur at two separate time periods; late evening (20:00h) or early morning (5:00-6:00h) and the male took a mean of 2.8 ± 1.3 days to mate with a female after being allowed access to a female. It was found that two females produced multiple litters during the observation period (that demonstrated the repeatability of the mating protocol); however, only one female had both litters survive to weaning age. Important reproductive parameters of this species that were obtained from this study are summarised in Table 3.

Discussion
The results from this study have shown that the neotropical subspecies, D. marsupialis insularis can be captive bred in the neo-tropics and generated reproductive parameters and breeding observations. Based on these  findings, an adult reproductive female D. marsupialis insularis opossum has the potential to produce at least two litters, with approximately five to eight young per year. This was similarly found to be the case with D. albiventris which also has been observed to be capable of producing two generations from one reproductive season (Astúa and Geise, 2006). Determination of receptivity of the female for breeding is vital for the reproduction of any species. Previous methods of oestrus detection for Didelphis opossums relied solely on laboratory tests. One laboratory method, used by Jurgelski Junior and Porter (1974), provided much detail on how to determine receptivity of the female, but involved laboratory conditions and relatively frequent handling of the female. However, outside of the laboratory this method may be impractical, too invasive and negatively affect the behaviour of the animals thus reducing productivity.
Alternatively a technique using reproductive behaviour analysis, as described by Silva et al. (2017) may be more appropriate for this application. This technique is supported by observations from local wildlife carers, who have indicated that males gave behavioural cues, such as audio and visual signs, when the female was receptive (W. Rollock, personal communication, 2017). Ryser (1992) suggested that females were able to advertise their breeding status by means of chemical signals, and that although males would be able to gather information from these signals, regarding the time to oestrus, they may not become aware of oestrus more than one or two days in advance.
Behavioural analysis as a means of oestrus detection, was supported by the observations made in this study, as it was found that given proper observation, the cues of behavioural interest shown by the male, did serve as fairly accurate detectors of oestrus.
The five mating episodes that resulted in four parturitions in this study were solely based on cues detected by observing the male and provide support that this species can be bred in captivity using the behavioural cues of the male and not relying on the cues of the female, as has been the practice with most domesticated animals (Herren, 2012).
The one mating episode between the male and female that proved to be unsuccessful may have been due to the male misreading the signals of the female or that the female may have been a first time breeder. It was observed that mating under captive conditions can take a physical toll on the male who up until the brief mating act, was found to be aggressively chased and snapped at by the female. The breeding male in this study received minor and sometimes major injuries particularly to his extremities i.e. tail. This is supported by Fritz (1971) who reported damage to the tail or tail injuries with the opossum.
Copulations were found to occur either in the late evening (8pm) or early morning (5-6am) which do not agree with those observations of Tyndale and Biscoe (Tyndale-Biscoe, 2005). These authors cited between midnight and dawn as the period when fertile copulations usually occurred. This inconsistency may have been a result of the different Didelphis species studied, as Tyndale's report was conducted on wild North American opossums (D. virginiana).

Weaning age
Young opossums have been observed to cannibalize the mother if allowed to remain with her for too long. Reports of damage to the tail, ears and skin of the female by her own young, have been made for D. virginiana (Jurgelski Junior and Porter, 1974), while cannibalization among newly weaned litter mates has also been recorded in Didelphis species housed in the same cage (McManus 1970).
Weaning age for D. virginiana young was recommended to be approximately 90-100 days (13-14weeks) (Jurgelski Junior and Porter, 1974). While for D. aurita weaning occurred at 143 days (20 weeks) (Gentile and Cerqueira, 1995) and at 3.5 months (14 weeks) for D. albiventris (Regidor and Gorostiague, 1996). Using reported weaning ages from wild caught individuals cared for in neo-tropical rehabilitation clinics (W. Rollock, personal communication, 2017), young from four litters were successfully weaned at about 77-91days (11-13 weeks) which was closer to the range quoted for D. virginiana species.
Weaning age is important for animal reproduction, as it determines the productivity of the species under captive conditions. It also can be used for the evaluation of population dynamics, the onset of reproduction, reproductive potential, and time between generations all being important parameters for ecological, conservation and evolutionary biology studies. Moreover, correct estimates of female reproductive capacity allows for a more precise definition of several parameters (Astúa and Geise, 2006). Re-breeding or time between successive litters is another important reproductive parameter for animal reproduction. Under laboratory captive conditions Jurgelski Junior and Porter (1974) found that 2-8 days after weaning the female re-enters oestrus and so can be mated, while Cutts et al. (1978) reports 20 days for one female. Reynolds (1952) and Green et al. (1996) studies have found that removal of young will usually lead the female to a return to oestrus within 7-14 days. The study herein observed a period of approximately 16 days and while only two re-matings were initiated, both resulted in parturitions. However, of the two parturitions only one female maintained young successfully to weaning age. Cutts et al. (1978) have also noted that non-virgin, wild caught females were more apt to rebreed and become pregnant under captive conditions. Further that the greatest number of failures to become pregnant occurred with wild caught, virgin females who Cutts et al. (1978) explained may just not be naturally vigorous breeders. This agrees with our findings as one of the rebred females had previously carried young and was able to successfully wean her litter. While the other female did not appear to have carried pouch young when she was caught and when rebred, though able to get pregnant, was unable to carry to weaning age, as the newborns were unsuccessful at attaching to nipples.
In conclusion, this study shows that the neo-tropical opossum D. m. insularis can be captive bred in the neotropics. Furthermore, behavioural cues can be used to determine the receptivity of the female for mating. Finally, it provides evidence and lends support to the development and use of captive breeding systems for conservation and production of 'non-domesticated' wildlife species like the neo-tropical opossum in the Caribbean.