Marmosops Matschie, 1916

VOSS, ROBERT S., TARIFA, TERESA & YENSEN, ERIC, 2004, An Introduction to Marmosops (Marsupialia: Didelphidae), with the Description of a New Species from Bolivia and Notes on the Taxonomy and Distribution of Other Bolivian Forms, American Museum Novitates 3466, pp. 1-40 : 5-11

Marmosops Matschie, 1916

TYPE SPECIES: Marmosops incanus ( Lund, 1840) by original designation.

DIAGNOSIS: Species of Marmosops can be distinguished from other small didelphid marsupials by the following combination of character states: eye surrounded by mask of blackish fur contrasting in color with paler fur of crown and cheeks; dorsal pelage uniformly colored, unpatterned; manus mesaxonic (digit III longer than adjacent digits II and IV); manual claws small, not exceeding fleshy apical pads in length; central palmar surface of manus smooth (not densely tuberculate); digit IV of pes longer than adjacent digit III; caudal scales in spiral series; middle hair of each caudal­scale triplet petiolate (narrower basally than at midlength), thicker, and usually more heavily pigmented than lateral hairs; tail longer than the combined length of head and body, not incrassate; ventral surface of tail­tip modified for prehension; premaxillary rostral process usually well developed; postorbital processes usually absent; petrosal usually exposed on lateral surface of braincase through fenestra between squamosal and parietal bones; maxillary fenestrae absent; posterolateral palatal foramina small, not extending lingual to M4 protocones; maxillary and alisphenoid not in contact on floor of orbit; extracranial course of mandibular nerve (V 3) enclosed by anteromedial strut of alisphenoid bulla (secondary foramen ovale present); fenestra cochleae exposed laterally (not enclosed in a bony sinus); I2–I5 with approximately symmetrical rhomboidal crowns increasing in breadth from front to back; P2 and P3 subequal in height; molar dentition highly carnassialized; centrocrista strongly V­shaped, its apex high above trigon basin; lower canine (c1) procumbent and premolariform (always with a flattened, bladelike apex and usually with a small posterior accessory cusp); p2 taller than p3; dp3 fully molariform; m3 hypoconid labially salient; entoconid large and well developed on m1–m3.

COMPARISONS WITH MARMOSA : Species of Marmosops have long been confused with Marmosa , which they superficially resemble in size and external appearance, but these taxa are only distantly related ( Jansa and Voss, 2000; Voss and Jansa, 2003) and can readily be distinguished by qualitative integumental and craniodental characters ( table 2). In the field, two external characters are useful for generic identification. Whereas the manus of Marmosops is mesaxonic, with digit (d) III distinctly longer than the adjacent digits (dII and dIV; fig. 3A), the manus of Marmosa is paraxonic, with dIII and dIV subequal in length (fig. 3B). Easily determined from live specimens (or properly fluid­preserved material) with elastic phalangeal joints that can be manipulated to align and straighten adjacent digits for relative length comparisons, this character is often hard to score from dried skins with hard, bent, twisted, or otherwise distorted fingers (see Voss and Jansa, 2003: character 10). A second distinguishing external trait is the morphology of the caudal pelage, of which the central hair in each caudal­scale triplet is petiolate (narrower basally than at mid­length), thicker, and usually more darkly pigmented than the lateral hairs in Marmosops ; by contrast, the three hairs of each caudal­scale triplet are subequal in thickness and similarly shaped and pigmented in Marmosa ( Voss and Jansa, 2003: character 27).

In dorsal cranial comparisons, these taxa are distinguished by the usual absence of distinct postorbital processes of the frontal bones in Marmosops (fig. 4B) versus the usual presence of distinct postorbital processes in adults of most species of Marmosa (fig. 4A). In lateral cranial view, the petrosal capsule that encloses the paraflocculus and semicircular canals (= pars canalicularis or pars mastoideus) is usually exposed through a fenestra between the squamosal and parietal bones in Marmosops (fig. 5A), but there is no such lateral petrosal exposure in Marmosa (fig. 5B). In ventral cranial view, Marmosops is distinguished by the invariant presence of a secondary foramen ovale (fig. 6B), which is just as consistently absent in Marmosa (fig. 6A).

Marmosops and Marmosa are similar in most dental traits, but the morphology of the deciduous lower third premolar (dp3) appears to provide a consistent difference. In all examined juvenile specimens of Marmosops (fig. 7, upper panel) this tooth is fully molariform because the trigonid includes a distinct paraconid, protoconid, and metaconid. By contrast, the dp3 of juvenile Marmosa (fig. 7, lower panel) is not fully molariform because the bicuspid trigonid lacks a distinct metaconid.

COMPARISONS WITH GRACILINANUS : Museum specimens of Marmosops are sometimes misidentified as Gracilinanus , apparently the result of using inaccurate published keys. 5 Despite such confusion, the two genera can be distinguished unambiguously by several external morphological differences ( table 2). Unlike Marmosops , which has a mesaxonic manus with dIII longer than the other digits, Gracilinanus (like Marmosa ) has a paraxonic manus with dIII and dIV subequal in length. Also, whereas Marmosops has spirally arranged caudal scales, the caudal scales of Gracilinanus are arranged in predominantly annular series ( Voss and Jansa, 2003: character 23). Lastly, the central hair of each caudal­scale triplet is petiolate, thicker, and usually more heavily pigmented than the lateral hairs in Marmosops , whereas the central hair is not conspicuously differentiated from the lateral hairs in Gracilinanus ( Voss and Jansa, 2003: character 27).

5 Among other characters erroneously used in relevant key couplets by Hershkovitz (1992: 6) and Anderson (1997: 30), presence/absence of gular glands, plantar pad morphology, nasal proportions, and presence/absence of postorbital processes do not reliably distinguish species of Marmosops from Gracilinanus .

a

Cranially, Marmosops consistently lacks maxillary fenestrae, palatal openings that are consistently present in Gracilinanus on each side of the maxillopalatine openings near the M1/M2 commissure (fig. 8). The morphology of the deciduous lower premolar (dp3) might also distinguish these taxa, but we have examined so few juvenile Gracilinanus (two specimens of G. agilis and one G. emiliae , all with incomplete trigonids) that the diagnostic value of this trait is unclear.

COMPARISONS WITH THYLAMYS : Specimens of Marmosops have occasionally been misidentified as Thylamys despite the large number of consistent differences that distinguish these very dissimilar taxa. Whereas the dorsal body pelage of Marmosops is uniformly colored and unpatterned, the middorsum of Thylamys is distinctly darker than the flanks, with a more­or­less sharp line of transition between the two colors on either side ( Voss and Jansa, 2003: character 7). Additional external differences ( table 2) include the morphology of the central palmar surface of the manus ( Voss and Jansa, 2003: character 11), the occurrence of lateral carpal tubercles ( Voss and Jansa, 2003: character 12), occurrence of plantar pelage on the tarsus ( Voss and Jansa, 2003: character 16), caudal hair morphology ( Voss and Jansa, 2003: character 27), and caudal incrassation ( Voss and Jansa, 2003: character 28). Craniodental differences between these genera are likewise numerous, including the presence/absence of a rostral premaxillary process ( Voss and Jansa, 2003: character 29), shape of the nasal bones ( Voss and Jansa, 2003: character 33), size and extent of the posterolateral palatal foramina ( Voss and Jansa, 2003: character 41), lateral exposure of the fenestra cochleae ( Voss and Jansa, 2003: character 47), and upper premolar proportions ( Voss and Jansa, 2003: character 55).

PHYLOGENETIC STATUS AND RELATIONSHIPS: Although distinguished from certain other ‘‘marmosines’’ by a relatively small number of morphological characters, Marmosops appears to be a natural group based on recent phylogenetic analyses in which generic monophyly was tested rather than assumed ( Jansa and Voss, 2000; Voss and Jansa, 2003). The most taxonomically complete analysis to date ( Voss and Jansa, 2003) included only five species of Marmosops , but among these were members of the diminutive M. parvidens complex as well as large Amazonian and southeastern Brazilian forms; generic monophyly was supported in separate parsimony analyses of both nonmolecular (morphological + karyotypic) characters and sequences from exon 1 of the nuclear gene encoding the Interphotoreceptor Retinoid Binding Protein (IRBP), in a likelihood analysis of the IRBP data, and in a parsimony analysis of a combined­data (nonmolecular + IRBP) supermatrix. The only discrepant result concerning generic monophyly, Kirsch and Palma’s (1995) report that Gracilinanus agilis and Marmosops dorothea are almost indistinguishable by thermal elution of scnDNA heteroduplexes, is now known to be an artifact of specimen misidentification (see Voss and Jansa, 2003: 57).

Several phylogenetic analyses of molecular data (e.g., Kirsch and Palma, 1995; Jansa and Voss, 2000; Voss and Jansa, 2003) suggest that Marmosops belongs to a didelphine clade that also includes Gracilinanus , Lestodelphys , and Thylamys . The recently described genus Chacodelphys may also belong to this group ( Voss et al., 2004). Although Patton et al. (1996) recovered Marmosops as the sister taxon of Caluromys , the cytochrome­ b sequences they analyzed appear to be substitution­saturated at this level of taxonomic comparison ( Jansa and Voss, 2000: fig. 12), and the relationship in question ( Marmosops + Caluromys ) had only trivial bootstrap support in their parsimony results. Palma and Spotorno’s (1999) report of a sister­group relationship between Marmosops and Metachirus (supported by very large bootstrap values in their neighbor­joining and parsimony trees) is an unexplained anomaly that is not consistent with any other analytic results or character data of which we are aware.