First evidence of gregarious denning in opossums (Didelphimorphia, Didelphidae), with notes on their social behaviour

Abstract

The Didelphidae are considered solitary opossums with few social interactions, usually limited to mating-related or mother–pouch young interactions. Anecdotal reports suggest that additional interactions occur, including den sharing by a few individuals, usually siblings. Here, we report novel observations that indicate opossums are more social than previously thought. These include nest sharing by males and females of Marmosa paraguayana, Gracilinanus microtarsus and Marmosops incanus prior to the onset of the breeding season and without signs of sexual activity; this is taken to indicate early pair-bonding matching and cooperative nest building. We also recorded den sharing among recently weaned siblings of Didelphis aurita and Caluromys philander. In addition, we observed 13 individuals of Didelphis albiventris representing three age classes resting without agonistic interactions in a communal den. These are the first reports of gregarious behaviour involving so many individuals, which are either unrelated or represent siblings from at least two litters, already weaned, sharing the same den with three adults. Sociality in opossums is probably more complex than previously established, and field experimental designs combining the use of artificial nests with camera traps or telemetry may help to gauge the frequency and extent of these phenomena.

1. Background

New World opossums of the family Didelphidae are usually considered solitary, barely tolerating the presence of conspecifics. In fact, most social interactions are limited to those involving mating (e.g. pre-copulatory or copulatory behaviour), agonistic and defensive behaviour or to those between a lactating female and her pouch young [1]. For example, Caluromys philander only tolerates the presence of other adults in a feeding spot 2–3 m away. Trespassing that threshold distance usually triggers hissing and other aggressive behaviour, ultimately resulting in increasing distance between individuals [2].

Anecdotal reports, however, suggest that some level of structured interactions exist, including females of Caluromys and Didelphis species responding to distress calls from young other than hers [2]. Captive specimens apparently tolerate or ignore other individuals, as seen in Marmosa murina and C. philander [3], and up to four unrelated Didelphis virginiana were seen sharing a den in captivity [4,5], but den sharing is only known in free-living D. virginiana [6] and Chironectes minimus [7].

In the wild, den sharing by several individuals (other than a mother and her unweaned litter) is unknown, and when dens or nests are shared, it is either by a pair in reproductive activity or by siblings, and when a mother shares a nest with her offspring, it usually lasts only shortly beyond weaning. Gregarious behaviour in American metatherians, however, has been suggested from the findings of over 30 individuals of the fossil stem-metatherian Pucadelphys andinus [8]. Although gregarious behaviour is known for several more derived Australian marsupials [9–13], no similar phenomenon has been reported in living American opossums.

Here, we report observations of social and gregarious behaviour in several Didelphidae species, including the first record of a large group of Didelphis albiventris sharing the same den, and evidence of premating pair-bonding for nest construction in several species.

2. Material and methods

Observations were made at two different sites. The first is an Atlantic forest site in Rio de Janeiro state, southeastern Brazil, in the course of an 8-year population dynamics study. Up to 312 artificial nests were set at five heights (0, 2.5, 5, 7.5 and 14–21 m), distributed across three 1.44 ha grids checked monthly from June 2003 to October 2010 (details on experimental design in Loretto & Vieira [14]). All animals recorded in artificial nests were marked and released after data collection. Nests were used throughout the year by both sexes of all species, and the presence of external morphological reproductive signs indicated a reproductive season between June and March. Five didelphid species regularly used artificial nests as shelter or den sites, yielding several records on nest use.

The second site was the Universidade Federal de Pernambuco (UFPE) campus, in Recife, northeastern Brazil, where we made an isolated observation. Animals were observed, photographed and immediately released on campus grounds.

3. Results

After 17 127 nest checks, we recorded 10 nests with multiple individuals for five species in the Rio de Janeiro site, as well as 28 nests with females carrying pouch youngs or sharing the nest with unweaned juveniles, and other 308 records of solitary individuals (see the electronic supplementary material). Most pair interactions showed no visible evidence of reproductive activity (table 1).

Table 1.

Interactions observed in artificial nests in the Rio de Janeiro study site.

period of breeding season	date	species	group	interaction	reproductive activity	observations
onset	Aug 2008	Gracilinanus microtarsus	male–female pair (one adult ♂, one subadult ♀)	nest sharing	no	nest at 2.5 m height
onset	June 2010	Marmosa paraguayana	male–female pair (one adult ♂, one subadult ♀)	nest sharing	no
end	Feb 2010	Caluromys philander	confirmed siblings (one juvenile ♂, two juvenile ♀)	nest sharing	no	marked while on mother's pouch, recently weaned. Canopy nest
end	Mar 2007	Didelphis aurita	possible siblings (two juvenile ♀)	nest sharing	no	recently weaned and same age (only first upper and second lower molars erupted). Nest at 5 m height
end	Jan 2006	Marmosops incanus	same-sex pair (two subadult ♂)	den or nest sharing	no
end	Feb 2005	M. incanus	male–female pair (one subadult ♂, one adult ♀)	nest sharing and building	no	two-chambered nest indicating association for nest building
end	Mar 2007	M. incanus	male–female pair (one adult ♂, one adult ♀)	nest sharing and building	no	two-chambered nest indicating association for nest building
end	Jan 2010	C. philander	unknown (one adult ♂, one adult ♀ with four newborn pouch young)	nests in proximity	yes	male nest set 1.5 m away from female nest, on the same tree; ca 15 m height
end	Feb 2007	C. philander	unknown (two adults, sex undetermined)	nest sharing	unknown	possible adults based on size, fled before identification and sexing
middle	Nov 2007	M. paraguayana and C. philander	unknown (two adult ♂)	nests in proximity	no	nests on the same tree, at 2.5 m (M. paraguayana) and 5 m (C. philander)

In addition, we made a single and unique observation in the UFPE campus, where specimens of D. albiventris are routinely seen wandering at night, coexisting with stray cats and dogs that live on campus, nesting in trash bins or any other suitable cavities. In July 2012, the University maintenance staff called us to remove individuals of D. albiventris nesting in a small (ca 30 × 30 cm) partially buried concrete box used to house electrical connections, where the presence of opossums routinely results in interrupted electric service. We found a group of 13 opossums sharing the box as a den site (figure 1); they were awake, but all resting together with no agonistic behaviour whatsoever. At least three adult individuals were present (based on body size; dentition and sexes were not recorded); one of them actively sought to protect the smaller ones (mid-sized, possibly subadults; and smaller juvenile specimens) during removal.

Figure 1.

Specimens of D. albiventris found in a communal den at UFPE campus, Recife, Brazil. At least nine specimens (out of a total of 13) of different sizes can be seen on the photo. (Online version in colour.)

4. Discussion

These are noteworthy reports of gregariousness or cooperation in didelphids. Apart from previous reports for wild D. virginiana [6] and C. minimus [7], and for captive C. philander [3], these are the first reports on den or nest sharing for three additional genera. While the reports involving juvenile C. philander and Didelphis aurita are of confirmed or putative siblings, the other three records (Marmosops incanus) at the end of the breeding season have no known cause, and may be unrelated to reproduction. Didelphid juveniles are typically recorded during December–May period in the Southern Hemisphere, when they are weaned and begin recruiting to adult and solitary life. Juvenile nest-sharing records may be evidence of sibling cooperation as a strategy to mutually increase safety, as seen in many other vertebrates. Male–female association beyond the breeding season is unknown and may relate to either extended parental care (February 2005 record) or a second oestrous period following loss of a previous litter (March 2007 record). Maternal investment in a second litter is uncommon in small didelphids, but was recorded twice in C. philander in the Rio study. Although M. incanus seem to be semelparous, a second oestrous could be triggered if the first litter is lost.

Premating behaviour is known for Monodelphis domestica [15,16], C. philander [2] and D. virginiana [17], and their courtship involves chasing and mildly aggressive behaviour. There are no reports of either captive or free-living opossum species where males and females cooperate to build nests. Our report is unique in recording premating and post-breeding nest building and sharing by four different free-living opossums.

This is also the first report of a communal, gregarious den for more than five free-living D. albiventris. Our instance could be interpreted as a single family, but it is unlikely that it was composed of parents, an additional adult and a single litter. While litter size in D. albiventris may reach 12 pups in higher latitudes [18], litters in northeast Brazil range from three to nine pups, averaging 6.5 [19,20], yet 13 individuals representing at least three cohorts were seen in the communal den. Subadult and juvenile D. albiventris recorded in July represent litters born in the previous breeding season, approximately in August and December [20,21]. If all den inhabitants were related, then both litters received extended parental care and heightened mutual benefits of safety in numbers as a result of their gregarious strategy, which has never before been reported, just as the presence of three adults caring for two litters.

5. Conclusion

Although isolated, these reports cast new light on our understanding (or lack thereof) of didelphid social behaviour, and may influence current views of social evolution in marsupials. While large Australian marsupial species (kangaroos and wallabies) may form feeding or resting groups of up to 10 individuals (and several groups may form ‘mobs’ of up to 50 individuals) [13], gregarious denning in small-bodied Australian species is also very rare and restricted to only a few species [9–12]. It is clear from our findings that social interactions in opossums are not restricted to mating and mother/offspring relation, and other gregarious interactions may not be uncommon in the wild, where males and females may pair for nest building prior to reproductive activity and oestrous. If so, behaviourally triggered oestrous in free-living didelphids, already recorded in captive C. philander and M. domestica [22,23], may be more frequent than is currently recognized. Current methods for studying didelphid ecology, mostly based on live traps, are insufficient for properly assessing their natural denning and nesting behaviour. These can be more adequately monitored using artificial nest boxes, for example. It is possible that use of larger boxes than those currently used associated with camera traps or remote monitoring may reveal the true extent of communal denning and more complex pair interactions in Didelphidae.

Ethics

All field procedures complied with current ethical standards in small mammal research [24] and Brazilian Law 11794 and Guidelines from the National Animal Experimentation Control Board (CONCEA). Field activities in Rio de Janeiro site were conducted under ICMBIO/IBAMA collecting permits 02001, 004671/98-51 and 16704-1.

Data accessibility

Raw data on nest occupancy in the Rio de Janeiro site, and additional photographs recording specimens on the communal den are available as electronic supplementary material.

Authors' contributions

D.A. conceived the manuscript, organized and analysed the data and wrote the main draft, R.A.C., P.F.M. and A.R.M. collected the field data on the communal den, D.L. collected and analysed the data on all other species. All authors drafted the manuscript and gave final approval for publication.

Competing interests

We declare we have no competing interests.

Funding

D.A. is currently supported by a CNPq fellowship (306647/2013-3) and FACEPE. R.A.C., P.F.M. and A.R.M. were supported by fellowships from FACEPE (MSc), CNPq (PIBIC) and FACEPE (PIBIC), respectively. D.L. was supported by FUJB, CNPq, PIBIC/CNPq, PROBIO (MMA—GEF), PRONEX and FAPERJ.