Didelphis marsupialis Linnaeus, 1758

PATTON, JAMES L., DA SILVA, MARIA NAZARETH F. & MALCOLM, JAY R., 2000, Mammals Of The Rio Juruá And The Evolutionary And Ecological Diversification Of Amazonia, Bulletin of the American Museum of Natural History 2000 (244), pp. 1-306 : 45-49

Didelphis marsupialis Linnaeus, 1758

TYPE LOCALITY: ‘‘America’’; restricted to Surinam by Thomas (1911).

DESCRIPTION: This is the largest species of didelphid marsupial in the region, with weights often over one kg. Individuals of black­color phase were present at all localities and were numerically dominant over those of the white phase at the four localities where both color morphs were recorded (the terra firme sites of Sobral, Penedo, Altamira, and Ilhazinha—localities 4, 7, 9, and 16, respectively). As a genus, Didelphis needs no comparison to any other marsupial wherever it occurs in tropical or subtropical forests, as it is readily separable from all other taxa. This is the largest marsupial in the community, lacking spots above the eyes, with long, coarse guard hairs covering dense underfur, and a very characteristic musky odor that marks the animal’s presence even when not seen. However, species identifications can sometimes be problematic, especially in areas of near or actual sympatry between D. marsupialis and D. aurita in southeastern Brazil (Cerqueira, 1985) or between D. marsupialis and D. albiventris in the Guianan region (Catzeflis et al., 1997; Lavergne et al., 1997). Selected measurements (mean and range) for 14 adult specimens, separated by sex, are given in table 5. Since only a single adult male was obtained, we are unable to assess the degree of sexual dimorphism present, nor are we able to assess variation due to age. Both sexual dimorphism and age contribute signficiantly to variation among individuals in mensural characters in Didelphis (Gardner, 1973) .

MOLECULAR PHYLOGEOGRAPHY: We have examined variation in the mtDNA cytochrome­b gene (the initial 660 bp) for 21 individuals from 13 localities within the Rio Juruá basin as well as from other localities throughout Amazonia (fig. 38; table 6). Sequence variation is relatively slight, as D. aurita from the Mata Atlântica of coastal Brazil and D. marsupialis from Central and South America differ by only 2.9%. These species of Didelphis are, however, reciprocally monophyletic with regard to their mtDNA haplotypes, with bootstrap values of 81 and 79 supporting each, respectively (fig. 39) Moreover, despite the low level of overall sequence divergence within D. marsupialis , some phylogeographic structure is present within this species across its sampled range All haplotypes from the Rio Juruá form a monophyletic assemblage, with bootstrap support of 81%; these differ among themselves by an average of only 0.304%. Similarly, haplotypes from central Amazonia north of the Rio Solimões (Rio Jaú and left [= east] bank of Rio Negro) are monophyletic, with 82% bootstrap support. These are more different among themselves, with an average Kimura two­parameter distance of 1.112%, but the variation does not separate into differences between samples on opposite sides of the Rio Negro. Finally, there is a series of samples from the upper Rio Negro eastern Ecuador, and Guyana, each of which are monophyletic unto themselves, but for which there is no clear geographic set of hierarchical relationships apparent, even in

comparison to haplotypes from southern Central America ( Costa Rica and Panama´) or from the southeastern Brazilian Amazon (Carajás, Estado do Para´). Consequently, while the level of sequence divergence within D. marsupialis is quite minimal, especially in comparison to that in other marsupials (see other accounts, below), there is some phylogeographic structure across its sampled range. It is unclear whether the low level of sequence divergence within Didelphis reflects a relative recency in the evolution of its member species, or simply slower rates of molecular evolution. In either case, the relatively high levels of similarity among haplotypes of specific phylogeographic regions cannot be assumed to result from differences in population structure (e.g., higher gene flow rates), despite the fact that Didelphis is often characterized as a nearly nomadic small mammal with large home ranges and great individual movement distances (Gardner 1982).

DISTRIBUTION AND HABITAT: This species was taken at 10 of the 16 primary sample sites and in each of the four regional areas from the headwaters to the mouth of the Rio Juruá (appendix A). At mid and lower river localities, we found D. marsupialis only in mature or second growth terra firme forest in disturbed garden plots, or in traps placed in the ecotone between terra firme and igapó forest. At the headwaters localities, specimens were taken in both nonflooded and locally inundated river­edge forest. This appears to be primarily an upland, nonflooded forest taxon, that inhabits a wide range of usable microhabitats, including secondary growth indicative of heavy disturbance by humans. Nearly all specimens were taken in traps placed on the ground, although several individuals were seen and shot as they

climbed in trees or vine tangles less than 2 m off the ground and five specimens were taken in canopy traps at heights above 10 m (Penedo, locality 7; Vira­Volta, locality 14; and Ilhazinha, locality 16).

REPRODUCTION: Harder (1992) presents a brief summary of the literature on reproduction in Didelphis species. Among our specimens, all females of age class 3 or greater (from Gardner, 1982; age approximately 7– 11 months, see also Tyndale­Biscoe and MacKenzie, 1976) were either parous (as judged by the orange color of the pouch re­ gion) or with pouch young; one of two females younger than age 3 showed signs of reproductive activity; and all females of age class 2 or less were nulliparous. For males maximal testis size (length> 15 mm) was found in all specimens of age class 2 and older. Females with pouch young were taken in all months of our trapping effort, including August, September, October, November, February, March, and June. These data suggest that reproduction can occur in all months of the year, during both wet and dry seasons Year­round reproduction also characterizes

populations in primary rainforest in French Guiana (Catzeflis et al., 1997), but contrasts with the llanos (‘‘plains’’) of eastern Colombia that are highly affected by seasonal changes and where reproduction was limited to the dry season (January through August; Tyndale­Biscoe and MacKenzie, 1976). Litter size averaged 5.1 (range 4–7, n = 9). This value is within the range of litter sizes reported from other localities in South America ( Colombia, 4.5 [Tyndale­Biscoe and Mac­ Kenzie, 1976]; and French Guiana, 4.7 to 6.5 [Catzeflis et al., 1997; Charles­Dominique, 1983; Julien­Laferrière and Atramentowicz, 1990]).

KARYOTYPE: 2n = 22, FN = 20 (fig. 40A). The autosomal complement as well as both X­ and Y­chromosomes are acrocentric and graded by size (see Gardner, 1973; Reig et al., 1977; Palma and Yates, 1996). We karyotyped four specimens from two separate localities ( MNFS 372 , 396 , 437 , 1035 ) .

SPECIMENS EXAMINED (n = 39): (1) 1m, 1f — MNFS 1160, 1190; (2) 2f — MNFS 1182–1183; (b) 1m — MNFS 1003; (c) 1m — MNFS 1035; (3) 1m, 4f — MNFS 1515– 1517, 1639, JUR 203; (4) 1f — MNFS 1452; (6) 1f — JLP 15654; (7) 2m, 5f — MNFS 372, 396, 413, 429, 437–438, JLP 15354; (9a) 1m — MNFS 931; (9) 2m, 7f — JLP 15939, 15990, 15999–16000, 16020, 16044, 16070, 16075–16076; (11) 1m — JLP 15751; (12) 1m, 1f — MNFS 683, JLP 15798; (13) 1m — JUR 467; (15) 1m, 4f — JUR 468, 508, 521, 549, 560);.