Rhipidomys Tschudi, 1845

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 : 162-169

Rhipidomys Tschudi, 1845

Climbing rats

Relatively few specimens of Rhipidomys have been collected from Amazonia. The genus appears to be either difficult to trap or generally uncommon, although exceptionally high numbers have been collected at some localities (e.g., the PDBFF reserves north of Manaus; Malcolm, 1991b). The scattered number of localities, and the general lack of adequate series even from those sites where specimens are known, has stifled interest in the systematics of this group. Christopher Tribe’s (1996) recent revision represents a quantum leap beyond our existing understanding of the systematics of Rhipidomys , but clearly there remains much to do, and to learn, as he himself concludes (Tribe, 1996: 254–257). Hopefully Tribe’s work will en­ ergize further interest in this enigmatic group of sigmodontine rodents.

Rhipidomys is a broadly ranging genus and species exhibit considerable differentiation in body size and both external and craniodental characters. However, this large group of species can be distinguished from other sympatric murids, and specifically those of the Rio Juruá basin, by a tail that is much longer than head and body, and typically terminated by a relatively long tuft of hairs; short and very broad feet with proportionately long toes and with pale sides but darker dorsal patch over the metatarsals; and six mammae in females. Cranially (fig. 107) these taxa are readily recognizable by a combination of very shallow zygomatic notches a narrow zygomatic plate with a vertical anterior edge and no anterior projecting spine moderate supraorbital ledges that extend divergently onto the sides of the cranium as parietal ridges; a short and broad palate with the incisive foramina extending to the level of the anterior root of first molar, the mesopterygoid fossa extending to the posterior margin of the third molar, and no large posterolateral palatal pits; no or only minute sphenopalatine vacuities; a robust alisphenoid strut; small bullae with the tegmen tympani overlapping the posterior suspensory process of the squamosal; nonsulcate upper incisors; and molars cuspidate, brachydont and pentalophodont with cusps (­ids) opposite rather than alternate. The cephalic arterial pattern may either be derived (pattern 3 of Voss, 1988; Carleton and Musser, 1989) with a minute stapedial foramen, no squamosal­alisphenoid groove, and no sphenofrontal foramen, or primitive (pattern 1).

Tribe (1996) records only a single species of Rhipidomys from lowland localities of western Amazonia. This is R. leucodactylus one of the largest members of the genus. He characterizes, and maps, other similar­sized species from extreme western Amazonia, but each of these is distributed along the lower Andean slopes at elevations above 700 m Included in this group are three taxa with the derived carotid arterial system, including: R modicus , from north through central Perú (Amazonas to Junín departments); an undescribed species known only from a single locality in the Departamento de La Paz, Boliv­

ia; and R. austrinus , distributed through the yungas from northern Bolivia (Departamento de La Paz) to northwestern Argentina (Departamento de Jujuy). Also present in southern Perú is R. ochrogaster , a species restrict­ ed to the Río Inambarí in northern Departamento de Puno that is characterized by the primitive cephalic arterial pattern. Woodman et al. (1991) suggested that two species of Rhipidomys occur at the lowland Amazonian site of Cuzco Amazónico, on the Río Madre de Dios (Departamento de Madre de Dios) in southeastern Perú, describing one as large and the other smaller. Tribe (1996), who apparently examined the relevant specimens, allocates all to R. leucodactylus , and Voss and Emmons (1996), in their compilation of the fauna from this site, also list only a single species.

We obtained only nine specimens of Rhipidomys in the more than 45,000 trap­nights of effort along the Rio Jurua´, including over 19,000 trap­nights on canopy platforms (table 2), and one of these was captured by hand from a tree hole. All are large­bodied, with head­and­body lengths of 150 mm or greater. However, two forms are distinguishable among these specimens based on a number of morphological characters as well as karyotype. A single specimen of one type was collected within the Headwaters Region in Estado do Acre ; the other form was found throughout the remaining length of the river in Estado do Amazonas. Tribe (1996) examined all of the specimens we obtained from the Upper and Lower Central sections of the Rio Jurua´, and allocated these to R. leucodactylus . He also assigned our single specimen from the Headwaters Region (for which he gave the locality as ‘‘ Acre: Rio Jurua´ , above Cruzeiro do Sul’ ’ [p. 289]) to this species, based on chromosomal information that one of us (J. L. Patton) supplied to him rather than from personal examination of the specimen. Unfortunately, as we document below, the information was misleading, and emendation of Tribe’s conclusions regarding our specimens is necessary .

In his discussion of R. leucodactylus , Tribe (1996: 204–205) states that ‘‘Considerable geographic variation is present in this species. Karyotypic or molecular data may in the future demonstrate that it is composite, but on the basis of currently available material it cannot be split unquestionably into separate taxa even at the subspecies level The most divergent populations occur in the Peruvian region of Inka (i.e., the former departments of Cuzco and Madre de Dios) and the Ucayali valley, where pelage is often shorter without gray bases to ventral hairs tails are less hairy with shorter pencils, and hind feet have a narrower dark patch; skulls are generally rather smaller, with less pronounced supraorbital ridges, narrow mesopterygoid fossae and smaller molars.’’ This combination of features characterizes the single specimen we have from the headwaters of the Rio Jurua´, and differentiates it from those downriver. The two also differ in karyotype (see below) and in sequence data from the cytochrome­b gene. Figure 108 illustrates the relationship between cytochrome­b haplotypes, based on 500 bp, for four specimens of Rhipidomys . These include two specimens from the central and mouth regions of the Rio Juruá that share close similarity (Kimura two­parameter distance of 1.22%) and the single specimen from the Headwaters Region, which is more than 12% different. This specimen groups with one

from Cusco Amazónico in southeastern Perú that shares similar morphological features, although the two are still quite different in sequence (7.31%).

We allocate the larger series of our specimens to the species R. leucodactylus , primarily on the revised diagnosis provided by Tribe (1996: 203), who describes the tail as ‘‘... long, abundantly furred distally, with long pencil’’ and the hind feet with a ‘‘... dark patch covering most of foot and extending onto digits.’’ The second species, apparently without an available name (Tribe, 1996), we describe here.

Rhipidomys gardneri , new species

HOLOTYPE: MVZ 168938 (Museum of Vertebrate Zoology, University of California, Berkeley ), adult male, collected June 7, 1984, by Richard M. Warner (original number 711); skin with skull and mandibles, in good condition; and liver and kidney tissue, originally preserved in liquid nitrogen, maintained at —76° C in the frozen collection of the Museum of Vertebrate Zoology.

TYPE LOCALITY: Reserva Cusco Amazónico, left (= north) bank of the Río Madre de Dios, 14 km east of Puerto Maldonado, De­ partamento de Madre de Dios, Peru´, elevation about 200 m (12°33̍S, 69°03̍W). A detailed site description, including climate vegetation, soils, history, and maps can be found in Duellman and Koechlin (1991) Lists of mammals present are given in Woodman et al. (1991, 1995) and Voss and Emmons (1996).

DIAGNOSIS: A large­bodied species with short fur without gray bases to the ventral hairs; a long but sparsely haired tail with a short terminal pencil of hairs (<6 mm; fig 109); hind feet with a narrow dark dorsal patch limited to the metatarsals; cranium with only moderately developed supraorbital ridges; deep zygomatic notch when viewed from above (figs. 107 and 110); distinctly narrow mesopterygoid fossa with a scalloped anterior border (fig. 111); short toothrow (<6 mm, on average); and karyotype with 2n = 44, FN = 50 (fig. 112).

PARATYPES: MVZ 168999, an adult male and MVZ 168960, an adult female, both from the type locality, and MNFS 1409, from the left bank of the Rio Juruá opposite Igarapé Porongaba, Acre, Brazil. All three specimens are skins with skull and mandible, in good condition, and with tissues preserved

either frozen at —76° C or in 95% ethyl alcohol.

MEASUREMENTS OF HOLOTYPE: TOL, 365; TAL, 215; HF, 33; E, 24; CIL, 36.16; ZB, 21.31; MB, 14.97; IOC, 6.40; RL, 13.41; NL, 13.12; RW­1 , 7.55 ; RW­2 , 5.94 ; OL, 13.49; D, 10.10; MTRL, 6.30; IFL, 7.74; PBL, 16.19; AW, 7.21; OCB, 8.45; BOL, 6.25 MPFL, 7.57; MPFW, 2.85; ZPL, 3.47; CD, 13.66.

ADDITIONAL MEASUREMENTS: See table 46.

DESCRIPTION AND COMPARISONS: This also is a large­bodied species, but slightly smaller than R. leucodactylus in nearly every measurement (table 46). The single specimen ob­

tained from the Rio Juruá is still in juvenile pelage, although all three molars are in place and exhibit some wear (e.g., age class 3). As a result, this specimen is difficult to compare to specimens of R. leucodactylus from the Rio Juruá in external characters. In the four specimens examined, including both this young individual and three adults from southeastern Peru´, the darkened surface of the hind feet is narrower and does not extend onto the toes (i.e., the toes are pale from base to tip), and the venter is paler, with graybased hairs limited to the midline of the throat and chest; the chin and inside of both fore and hind legs are white. The tail is long

15704, locality 7) and R. gardneri , (right ­MNFS 1409, locality 2).

but with only a short pencil of hairs extending beyond the tip (6 mm; fig. 109). The tail is also less hairy with individual hairs barely extending over two scales. As a result, the scales are clearly visible, although small, averaging 16 rows per cm. In direct comparison to specimens of R. leucodactylus of similar toothwear (e.g., JLP 15704), the skull has deeper zygomatic notches (figs. 107 and 110) and a narrower and longer mesopterygoid fossa that possesses a scalloped anterior bor­ der with a distinct, posteriorly projecting median spine (fig. 111).

COMMENTS: Tribe (1996) regarded the holotype and paratype of R. gardneri from southeastern Perú as R. leucodactylus yet noted their distinction from other specimens he diagnosed separately as that species. While we have not seen specimens reported by Woodman et al. (1991, 1995; also listed in Voss and Emmons, 1996) from the type locality, we assume these to represent R. gardneri .

DISTRIBUTION AND HABITAT: The single specimen we caught came from a canopy platform trap on the left bank of the Rio Juruá opposite Igarapé Porongaba (locality 2) in Estado do Acre. The habitat here is várzea forest that is flooded only occasionally rather than annually. Specimens at the type locality were taken in traps placed in trees in undisturbed lowland evergreen forest that did not flood during the wet season (see Woodman et al., 1995). This species is distributed from the headwaters of the Rio Juruá at least to the lowlands of southeastern Peru´, and, following Tribe (1996), probably into the Ucayali valley.

ETYMOLOGY: Named in honor of Alfred L. Gardner, who introduced one of us (J. L. P. both to mammalogy (in 1963) and to Amazonia (in 1968), for his many seminal contributions to the systematics and distribution of South American mammals.

REPRODUCTION: The single individual from the Rio Juruá was a nulliparous female, in juvenile pelage with the minimal toothwear for assignment to age class 3. The three adults from the type locality are adults ; the holotype had abdominal testes (12 mm Χ 7 mm, length to width) and small vesicular glands (7 mm long) suggesting that it was nonreproductive.

KARYOTYPE: This species has a diploid number of 44 and a fundamental number of 50 (fig. 112B). As only a female was examined, the sex chromosomes are unknown However, the autosomal complement is likely to have 17 pairs of large to small acrocentric, three pairs of small metacentric, and one pair of large subtelocentric autosomes and the X­chromosome is considered to be a

medium­sized subtelocentric, one with a very small short arm. This is the chromosome complement listed by Tribe (1996: 111, 113), which Patton erroneously told him by letter characterized all specimens that we karyotyped.

Figures

Fig. 108. The single most­parsimonious tree based on an exhaustive search, for haplotypes of the mitochondrial cytochrome­b gene (500 bp) for specimens of climbing rats, Rhipidomys, from southwestern Amazonia. Length = 183 steps, CI = 0.885, RI = 0.741. The sequence for MVZ 168938 (Cusco Amazónico, Madre de Dios, Perú is from Smith and Patton (1993). Tree is rooted by comparison to sequences of Thomasomys aureus (also from Smith and Patton, 1993). Bold numbers at internal nodes are bootstrap values based on 1000 replicates; percentages are average Kimura two­parameter distances. Haplotypes are identified by field number and locality.