Leptomys Thomas, 1897: 610

Musser, Guy G., Helgen, Kristofer M. & Lunde, Darrin P., 2008, Systematic Review of New Guinea Leptomys (Muridae, Murinae) with Descriptions of Two New Species, American Museum Novitates 3624 (1), pp. 1-60 : 9-23

Leptomys Thomas, 1897: 610

TYPE SPECIES: Leptomys elegans Thomas, 1897:610 .

EMENDED DIAGNOSIS: A genus of hydromyin in the subfamily Murinae , family Muridae (as Murinae is delimited by Carleton and Musser, 1984; as the contents of Murinae and Muridae are presented by Musser and Carleton, 2005) that is distinguished from all other described murine genera by the following combination of traits: (1) dorsal pelage covering head and body moderately short, dense, and silky or velvety, upperparts tawny brown, ventral coat short and dense, ranging from white everywhere to grayish white (that is washed with buff in some specimens); (2) muzzle long and narrow; (3) narrow and short bald middorsal strip between eyes and shoulders; (4) tail slender and shorter, coequal to, or longer than head and body (LT/LHB ranges from 66% to 128% in adults), tail scales very small and slightly swollen, their annuli abutting each other (not overlapping), three short hairs associated with each scale, proximal portion of tail brownish gray above, white or mottled below, distal one-fifth to one-half of tail white; (5) dorsal surfaces of front and hind feet white, hallux with claw, hind foot long and narrow with full complement of plantar tubercles in some samples, hypothenar pad absent or reduced in size in other samples; (6) two pairs of inguinal teats; (7) rostrum long and slender, interorbital and postorbital margins smooth, zygomatic arches flare moderately from sides of skull, braincase smooth and globular, interparietal about as long as wide (diamond-shaped), occiput deep, slight to no cranial flexion; (8) zygomatic plate narrow, its anterior margin concave and not projecting beyond dorsal maxillary root of zygomatic arch, its posterior edge set just behind front of first molar, superficial masseter inserting on pronounced tubercle at base of ventral zygomatic root; (9) squamosal intact except for large subsquamosal foramen, parietal does not project ventrad beyond temporal beading to form part of lateral braincase wall; (10) alisphenoid struts wide; (11) incisive foramina wide and moderately short, their posterior margins ending far anterior to front faces of first molars; (12) molar rows parallel, bony palate long and wide, its posterior margin projecting beyond back faces of third molars, palatal surface smooth with shallow palatine grooves, posterior palatine foramina level with contact between first and second molars; (13) roof of mesopterygoid fossa generally intact, penetrated by small irregular openings in some specimens; (14) pterygoid plates narrow and complete (no sphenopterygoid openings), with shallow pterygoid fossae; (15) ectotympanic (auditory) bulla small relative to skull size, bullar capsule incompletely covering periotic, which is exposed as an expansive medial wedge, posterior wall of carotid canal formed by periotic and posterior margin of eustachian tube; (16) derived carotid circulatory pattern (character state 3 of Carleton, 1980; pattern 3 described by Voss, 1988), as indicated by absent stapedial foramen, sphenofrontal foramen, squamosal-alisphenoid groove, and groove on pterygoid plate between middle lacerate foramen and foramen ovale; (17) dentary slim, ramus long and cylindrical between incisor and molar row, low ascending ramus, delicate coronoid process, prominent condyloid and angular processes, end of alveolar capsule hidden in body of ramus at level of condyloid processes; (18) upper and lower incisor enamel smooth (unmarked by grooves), uppers orange or pale orange, lowers paler, uppers opisthodont; (19) first and second upper molars (maxillary) each with four roots, third upper molar with three roots, lower molars (mandibular) each with two roots; (20) molars brachydont and narrow, occlusal surfaces cuspidate in juveniles but transform into shallow basins in adults, (21) labial cusps (t3, t6, and t9) absent or undetectable on upper molars, cusp t7 present on first upper molar, cusp t7 and anterolabial cingulum present on second upper molar in different frequencies depending upon geographic sample and species, posterior cingulum absent from first and second upper molars but present on second (seen in young animals), third upper molar very small relative to others in toothrow; (22) anteroconid formed of large anterolingual and anterolabial cusps, anterocentral cusp absent, anterolabial cusp absent from second and third lower molars, no anterior labial cusplet on first and second lower molars, but posterior labial cusplet present on first, posterior cingulum wide and narrow on first and second molars; (23) phallus has blood sinus in wall of glans extending into the urethral lappets ( Lidicker, 1968); (24) spermatozoan with sperm head bearing three elongate processes (one apical and two ventral), spermatozoan tail long ( Breed and Aplin, 1994); (25) karyotype with 2n 5 48 ( Donnellan, 1989).

DESCRIPTION: Thomas’s description of external traits is accurate. In all samples of Leptomys the adult dorsal pelage is soft, dense, and short (6–12 mm long, depending on the species), and either velvety or silky to the touch. The black-tipped guard hairs either match the overhairs in length or barely extend beyond them, which partly gives the fur its uniform velvety appearance. Upperparts are tawny brown with dark brown highlights over back and rump and buffy along sides of body, thighs, and upper arms (figs. 4–6, 15, 19). Overhairs in the dorsal pelage are mostly gray with brown or buff tips. This dorsal coloration applies to all species except L. signatus , which has a pale buffy gray dorsal coat without the rich buff and dark brown tones seen in the other species. The head is the same color as the back in all species except L. signatus , which displays a large and conspicuous white blaze (fig. 4). Some specimens of L. elegans have a distinct but much less pronounced pale mark on the head between the eyes (evident in AM, AMNH, and PNGNM samples). Between the eyes and shoulders the fur is permanently parted to reveal a middorsal strip of balding skin 10–15 mm long and 1–2 mm wide. The pattern is conspicuous in species with short coats but less detectable in the longer pelage of L. paulus , n. sp. The moderately large and dark brown pinnae appear naked but are actually scantily covered with minute dark hairs. Long mystacial and superciliary vibrissae extend well beyond the pinnae when laid against the head, the genal vibrissae reach the base of the pinnae, and the usual murine array of facial submental and interramal sensory hairs and ulnar and tarsal vibrissae are present. A dark brown mask around the eyes extends onto the muzzle in some species but not others. The muzzle at base of the mystacial vibrissae appears unpigmented in one species (because only short scattered pale hairs cover it) but is brown or grayish brown in all the other species (due to the dense covering of pigmented hairs), and the cheeks are white.

The ventral coat is also soft, dense, and shorter (5–6 mm long) than the dorsal pelage, and the underparts vary in coloration (figs. 4– 6). Many specimens of Leptomys ernstmayri and L. paulus , n. sp., are grayish white, some with a pale or dark buffy wash, or whitish gray over most of the venter; a white patch or narrow strip extends from chin to chest in some individuals and many have a white inguinal region. The grayer and buffy tones result from hairs with gray bases and short unpigmented or buffy tips; the paler tones (whitish gray) reflect hairs with gray bases and longer white tips. Underparts are white everywhere (hairs are unpigmented from base to tip) in other specimens, particularly those in samples of L. elegans , as Thomas described the holotype (‘‘creamy white’’), L. signatus , and L. arfakensis , n. sp.

The tail averages much longer than head and body in the species of Leptomys with smaller body size, L. ernstmayri and L. paulus , n. sp. Among those physically larger Leptomys , the tail is about coequal to length of head and body in L. signatus and L. arfakensis , n. sp., and shorter than the head and body in L. elegans (tables 2, 4). The tail is covered in annuli of very small, slightly swollen, and inconspicuous scales (15–20 per cm). The scale hairs (three emerge from beneath each scale) are fine and laid flat against the tail so it appears naked. From about one-fifth to one-half of the distal portion of the tail is all white (depending on the species and geographic sample), the dorsal proximal region is brownish gray, and the ventral surface below this brownish gray segment ranges from white to mottled.

Dorsal surfaces of front and hind feet, including digits and claws, are all white and densely covered with short silvery hairs that form sparse tufts at the base of each claw. The area proximal to ankle and wrist is brown. The second digit of the front foot is half the length of the third, which is the longest; the fourth is slightly shorter, and the fifth is the shortest, extending just beyond the base of the adjacent digit. The hind foot is long and narrow, 25%–28%, on average, of head and body length, depending upon the species ( Tate, 1951: 222, considered Leptomys to be ‘‘specialized through the considerable elongation of the hind foot, which may function as an incipiently hopping organ’’). The first digit is very short, its claw barely or not quite reaching base of the adjacent digit; the three middle digits are longest and subequal in length; the fifth digit is short but longer than the first, its claw reaching to proximal third of second phalanx. Palmar surfaces are naked and brown, and possess the usual complement of five pads (three metacarpals, a thenar, and smaller hypothenar). A naked plantar surface, which is partially brown and adorned with four conspicuous interdigital pads, is common to all species of Leptomys , but presence or absence along with size of the hypothenar is variable among samples. In all species, the thenar is elongate and conspicuous. By contrast, the hypothenar is either absent or ranges from a tiny nubbin (not always evident on dry skins and best detected in fluid-preserved material) to a moderately large pad (but always smaller than the thenar), depending upon the species and geographic samples.

In all the material we have examined, females have two pairs of teats, both inguinal in position, as previously noted by Thomas (1897: 610) and others ( Tate, 1951, and Flannery, 1995, for example).

The skull in all species of Leptomys is elongate (figs. 7–10). The rostrum is long and slender as viewed from a dorsal perspective, its lateral margins broken only by the slight bulge of each nasolacrimal canal. Dorsolateral outlines of the interorbital and postorbital regions are smooth, without ridging. The cranium is smooth and oval in outline: the temporal ridges are indicated only by indistinct roughened places along the squamosal-parietal sutures, each lamboidal ridge forms an inconspicuous linear beading, and each mastoid is only sightly inflated. The interparietal is roughly diamond-shaped, slightly wider than long. Thin zygomatic arches gently bow outward from sides of the skull, the jugal component of each one is short.

From a lateral view, the convex dorsal outline of the cranium arches above the dorsal plane of the rostrum, and continues back to define the arcuate posterior outline of a deep occiput, which overhangs the occipital condyles. The rostrum is deep and somewhat rectangular in outline, the nasals and premaxillaries project as a short tube well beyond the incisor faces, and the nasolacrimal capsules are only slightly inflated. The concave anterior margin of the narrow zygomatic plate does not project anterior to the dorsal maxillary root of the zygomatic arch (no zygomatic notch). Most of the ventral maxillary root of the zygomatic plate originates anterior to the first molar, but the posterior third of the root is situated above the front third of the molar. The squamosal root of the zygomatic arch is located low on the side of the braincase. Dorsal to the squamosal root the wall of the braincase is formed entirely by the squamosal (without the contribution of a ventrally projecting extension of the parietal). Posterior to the squamosal zygomatic root and dorsad of the auditory bulla, the squamosal is intact and forms a seamless union with the exoccipital in most specimens, but in a few others the squamosal-exoccipital suture is perforated by a small or minute subsquamosal foramen. The junction of the orbitosphenoid, alisphenoid, and frontal bones forms a solid section of the braincase wall, unbroken by a sphenofrontal foramen. The inner walls of the braincase are smooth, without the texture associated with squamosal-alisphenoid grooves. A wide bony alisphenoid strut (present in all skulls surveyed) separates the foramen ovale accessorius from the combined buccinator-masticatory foramen. A dorsal postglenoid foramen and ventral postalar fissure, each ranging from small to spacious, separates the dorsal and anterior margins of the auditory (ectotympanic) bulla from the adjacent squamosal. The periotic exposed along the dorsal margin of the bullar capsule may or may not extend anteriorly to touch the posterior margin of the squamosal, but in no specimen does it project anteriorly as a tegmen tympani to overlap the squamosal. Projecting from the anterodorsal margin of the bullar capsule into the postglenoid space is a bony triangular process.

The short and wide incisive foramina are conspicuous when the skull is viewed from a ventral perspective, and are located about midway between incisors and anterior margins of the first molars, occupying about 55% of the diastemal length. The wide bony palate is much longer than each molar row, a reflection of the short incisive foramina, and projection of the palate beyond posterior margins of the third molars. Its surface is mostly smooth, the pair of palatal grooves shallow and inconspicuous; posterior palatine foramina are even with the contact between first and second molars. The roof of the mesopterygoid fossa is intact in some specimens, pierced by small irregular openings or larger but still modest sphenopalatine vacuities in others. The adjacent narrow pterygoid plates (5 parapterygoid plates) are also intact (no sphenopterygoid vacuities) except where each is pierced by the ventral opening of the foramen ovale, and no groove scores the posterolateral area of the ventral surface. The posterolateral margin of the pterygoid, which forms a ridge lateral to the foramen ovale in many murines (see the illustrations in Musser and Newcomb, 1983, for examples), is reduced to a delicate threadlike bone, which is often lost when skulls are cleaned. Long and narrow bony eustachian tubes project from the ectotympanic bullae, which are very small relative to size of the skull (length of bulla is about 12% of occipitonasal length). The bullar capsule does not cover the entire surface of the periotic, leaving exposed a broad wedge of that element separating the capsule from the basioccipital; the anteriomedial margin of the periotic and anteodorsal surface of the base of the eustachian tube form the posterior wall of the carotid canal (a reflection of the small ectotympanic capsule and broadly exposed periotic). The middle lacerate foramen separates the bullar capsule from the posterior margin of the pterygoid plate. None of the specimens we surveyed exhibited any trace of a stapedial foramen. When present, this opening penetrates the fissure between the bullar capsule and periotic (the petromastoid fissure), but these two elements join seamlessly in our examples of Leptomys .

All species of Leptomys possess a derived carotid circulatory pattern (character state 3 of Carleton, 1980; pattern 3 described by Voss, 1988; and the plan diagrammed for Oryzomys palustris by Carleton and Musser, 1989). Although this assertion is not based upon dissection of soft tissue, the disposition of certain cranial foramina and osseous landmarks evident on cleaned skulls clearly point to such a specialized arterial circulation. A stapedial foramen is absent, as is a sphenofrontal foramen at the junction of orbitosphenoid, alisphenoid, and frontal bones. There is no squamosal-alisphenoid groove scoring the inner surface of each wall of the braincase, and no groove extending from the middle lacerate foramen to the foraemen ovale on the ventral posterolateral surface of each pterygoid plate. There is a shallow trough running diagonally over the dorsal (inner) surface of each pterygoid plate. These landmarks and the circulatory pattern they reflect are similar to the derived arterial configurations described for Sulawesian Crunomys and Sommeromys ( Musser and Durden, 2002) , and a suite of Philippine ( Musser and Heaney, 1992; Musser et al., 1998) and Sundaic ( Musser and Newcomb, 1983) murines, as well as North and South American cricetids ( Carleton, 1980; Steppan, 1995; Voss, 1988; Weksler, 2006). In this configuration, the stapedial artery is either absent or minute and serves only the periotic region. The supraorbital and infraorbital arteries, normally branches of the stapedial, are absent with the result that the ophthalmic and internal maxillary circulation is supplied by a secondary arterial connection branching from the internal carotid artery and coursing obliquely across the dorsal surface of the pterygoid plate. This derived carotid arterial circulatory pattern is contrasted with the other two found within muroid rodents particularly clearly through the descriptions and diagrams provided by Bugge (1970), Carleton (1980), Carleton and Musser (1989), and Voss (1988).

The dentary is slim, especially that portion of the ramus anterior to the molar row, and its labial and lingual surfaces are moderately smooth (figs. 8–10). A deeply concave posterior margin separates large and projecting condyloid and angular processes, and the smaller and more delicate coronoid process is isolated from the condyloid by a wide and deep sigmoid notch. The alveolus for the lower incisor terminates at the base of the condyloid process but is unmarked by any external swelling or lateral projection.

Enamel layers of upper incisors are orange (diluted to pale orange in many specimens), those of the lowers are paler, even cream in many examples. The faces are smooth, lacking either grooves or shallow sulci, and the face of each upper is gently convex. The enamel face of each upper incisor does not wrap around the side, as is the pattern in the Philippine Chrotomys , for example, but resembles the enamel coverage seen in Rattus ( Musser and Heaney, 1992:79) . The incisors emerge from the rostrum at slightly less than a right angle to the occlusal plane of the molars (slightly opisthodont). The sharp-tipped lower incisors project appreciably beyond the dentaries (figs. 8–10).

Alveolar patterns for upper and lower molar roots are similar among samples of Leptomys . Each first and second upper molar is anchored by four roots (one large anterior, one large posterior, and two somewhat smaller lingual roots); the third upper is anchored by three large roots coequal in size. Each of the lower molars has two roots (anterior and posterior). In their original description of Paraleptomys and comparisons with Leptomys, Tate and Archbold (1941: 2) correctly noted that each upper and lower third molar in Leptomys was anchored by three roots; later, Tate (1951: 222) incorrectly stated that each upper and lower third molar had but a single root.

Molars are brachydont and narrow. In the maxillary and mandibular rows, the elongate first molar forms about half of the row, and the second molar about a third; the round, simple third molar is very small compared with the other two, comprising about 15% of the row (figs. 11, 12). Simple occlusal surfaces are clearly cuspidate on juvenile molars but transform with age and wear into surfaces dominated by shallow basins. If our interpretation of cusp homologies is correct, the first and second upper molars in all species of Leptomys appear to lack the series of labial cusps (t3, t6, and t9) that are present in most other murines; either they have been completely lost, or during embryonic development have coalesced so completely with the adjacent row of central cusps as to be undetectable. Occlusal surface of each first upper molar appears to consist of a diagonal anterior pair of large round cusps t1 and t2; some specimens have a slight bulge or cusplet where cusp t3 would normally occur, but that identity is impossible to verify in our material. Cusps t4 and t5 form the second row. Between them and the robust, round posterior cusp t8 is a lingual cusp or ridge that we identify as cusp t7. The second upper molar shares a similar configuration except only cusp t1 remains of the anterior row and cusp t7 is present or absent, depending upon species and geographic sample (table 6). In the large sample of L. ernstmayri from the Huon Peninsula (localities 19 and 20 in the gazetteer), a distinct shelflike cingulum forms the anterolabial border of each second upper molar (fig. 11); this structure occurs, in variable frequency, only in other geographic samples we identify as L. ernstmayri and in one example of L. elegans (table 6). In one specimen the anterolabial border of the second molar bears a cusplet that may represent cusp t3. A discrete posterior cingulum forms the posterior border of each second upper molar in young animals, but coalesces with cusp t8 after wear and becomes undetectable; we could not assess its frequency among our samples because most specimens are too old. We did not see a posterior cingulum on the first upper molar in any example, even juveniles. The simple round and peglike third upper molar is without significant occlusal topography in most adults. Judged from juveniles, cusp t1 and a coalesced cusp t4–t5 appear to form most of the molar.

Coronal surfaces of lower (mandibular) molars are simple but not as evidently cuspidate as are the uppers, and any cusp topography present in juveniles wears quickly down to a series of round, transverse, or oblong basins (figs. 11, 12). The anterior half of each first lower molar is basically a deep basin bounded anteriorly by anterolingual and anterolabial cusps (the anteroconid) and posteriorly by the protoconid and metaconid. A diagonal basin consisting of coalesced hypoconid and entoconid defines the posterior half of the molar. A thin posterior cingulum forms the posterolabial border. Two transverse basins, each bounded by dentine rims, and a posterior cingulum, comprise the occlusal surface of the second lower molar. Some specimens exhibit small labial cusplets. An oblong basin forms the chewing surface of the very small, round and peglike third molar. In one species, L. arfakensis , n. sp., represent- ed by the holotype and paratype, the third lower molar is excessively reduced in the latter and absent in the former (fig. 22).

Many additional anatomical systems, along with other aspects of Leptomys , remain to be investigated in detail, but some information on phallic anatomy, karyology, and sperm morphology is available. Phallic morphology was described by Lidicker (1968, 1973; Lidicker and Brylski, 1987; the specimens dissected, AMNH 191406 and 194940, were identified as L. elegans , but are actually examples of L. ernstmayri ). Donnellan (1989) sampled the karyotype (2n 5 48) for L. elegans , which differs from the postulated ancestral Australo-Papuan murine karyotype (see Baverstock et al., 1983, and references cited there), excluding Rattus , by the presence of a pair of submetacentrics. Spermatozoan morphology and its significance were discussed by Breed (1997, 2004) and Breed and Aplin (1994), who sampled both L. elegans and L. ernstmayri . In addition, Yarao (in litt., 2005) completed a study of gastrointestinal tract morphology in L. ernstmayri that is not yet published. Genetic studies that included Leptomys have been limited. Watts and Baverstock (1994, 1996) sampled a specimen of L. ernstmayri (identified as L. elegans ) in their inquiry covering intergeneric clustering in Australo-Papuan murids based on immunological comparisons (microcomplement fixation of albumin), and Leptomys has been included in a phylogenetic study based on molecular data derived from a wide sampling of murines ( Rowe et al., 2008).

COMPARISONS WITH PARALEPTOMYS: Named and described by Tate and Archbold in 1941, their diagnosis of Paraleptomys consisted of the short statement that it required ‘‘comparison only with Leptomys , from which it is separated by its normal hind foot (elongated in Leptomys ), and by the absence of m3/3’’ ( Tate and Archbold, 1941: 1). The genotype, P. wilhelmina , was based on two lots of specimens obtained from the northern ramparts of the Snow Mountains (Pegunungan Maoke), which form part of the massive Central Cordillera in western New Guinea (Province of Papua, Indonesian New Guinea [5 Irian Jaya]). Of the 48 specimens in the first lot, 43 were collected 9 km northeast of Lake Habbema at 2800 m and five came from the Bele River Camp, 18 km northeast of Lake Habbema at 2200 m; the second lot of 30 specimens was obtained at 2150 m and 1800 m on ridges 18 km and 15 km, respectively, southwest of Bernhard Camp, near the Idenburg River ( Tate, 1951: 223; the bulk of this material resides at AMNH).

True Paraleptomys wilhelmina comprises the samples from 2200 and 2800 m, and these remain the only record of the species. The series from 2150 m and 1800 m represents a morphologically distinct but unnamed separate biological species (‘‘may be separable as a faintly differentiated race’’, observed Tate and Archbold [1941: 2]; also see Helgen in Musser and Carleton, 2005: 1431). A single specimen identified as P. wilhelmina has been reported from the Tifalmin Valley at 1800 m at the eastern margin of the Star Mountains in western Papua New Guinea ( Flannery, 1995), but this example, along with an unreported series collected in the Star Mountains (in BBM-NG), is a sample of a third species, yet undescribed, that is endemic to the Star Mountains of Western and Sandaun (West Sepik) provinces of Papua New Guinea (K. M. Helgen, in litt., 2007).

The only other currently recognized species of Paraleptomys is P. rufilatus , which was named and described by Osgood in 1945 and is currently recorded from montane forest formations in the North Coast Ranges—the Cyclops, Torricelli, and Bewani Mountains ( Musser and Carleton, 2005: 1431). This distinctive species is larger than P. wilhelmina in body size and differs in pelage coloration ( Flannery, 1995).

All samples of Paraleptomys are from high mountain forests, whether from the Central Cordillera or from the North Coastal Ranges.

In the following comparisons between Paraleptomys and Leptomys , we generally focus on P. wilhelmina , the type species of the genus, as our example of Paraleptomys , and our observations are drawn primarily from skins and skulls stored at AMNH.

External features: Body form and lengths of appendages relative to body size resemble closely these features in Leptomys (compare the photograph of a live P. rufilatus with those of several Leptomys that are reproduced in Flannery, 1995). Paraleptomys wilhelmina is smaller in body size than any species of Leptomys , but the larger P. rufilatus is about the same size as the smallest Leptomys , L. paulus , n. sp. (compare measurements in Tate, 1951, and Flannery, 1995, with those listed in tables 2 and 3). The tail is longer than length of head and body (LTV/LHB 5 106%–112%) in Paraleptomys , which falls within the range of many Leptomys (tables 2, 3). Even conformation of the hind foot in Paraleptomys , which Tate and Archbold (1941) and Tate himself (1951) thought to be different from that of Leptomys , is similar in the two genera. Although the narrow hind foot of P. wilhelmina is shorter than that in any species of Leptomys , it is about 25% of head and body length (26% for P. rufilatus ), a proportion falling within the average proportional range for the five species of Leptomys (25%–28%). Lengths of digits relative to overall length of hind foot is close: about 43% in both genera (as determined from dry study skins), as are lengths of individual digits relative to one another.

The most telling external distinction between Paraleptomys and Leptomys is pelage quality and color. The entire dorsal coat of P. wilhelmina is dark brown, dense (10–12 mm thick), and soft and woolly in texture ( P. rufilatus has buffy brown sides), contrasting sharply with the bright tawny brown, short, and velvety coat common to Leptomys . Dark, grayish white underparts characterize P. wilhelmina ( P. rufilatus has a white throat); three species of Leptomys have pure white ventral coats, in samples of the other two species underparts range from white to whitish gray, some with a pale buffy wash.

Skull: Paraleptomys and Leptomys share the same general cranial shape as well as many proportional aspects: ‘‘in many respects very like Leptomys ,’’ wrote Tate and Archbold (1941: 2) (see Flannery, 1995, for informative illustrations of skulls representing the two genera). Paraleptomys , however, diverges from that basic conformation in the following diagnostic features:

(1) The interparietal in Paraleptomys is much wider than long (about as wide as long in Leptomys ).

(2) Paraleptomys has a shorter bony palate relative to length of skull (about 20% of occipitonasal length), which reflects the short- er molar rows due to the missing third molars (relatively longer in Leptomys , about 25% of occipitonasal length).

(3) The ectotympanic bulla is large in Paraleptomys (about 15% of occipitonasal length), and has a short bony eustachian tube relative to bullar length. The ectotympanic capsule covers much of the periotic, with its anterodorsal margin abutting the basioccipital; the back of the carotid canal is formed by capsule and basioccipital only. ( Leptomys has a much smaller bulla, its length about 12% that of the occipitonasal dimension, and a relatively longer bony eustachian tube. The ectotympanic capsule rests on top of the periotic leaving exposed a wide and long wedge of that element that completely separates capsule from basioccipital; the rear wall of the carotid canal comprises periotic and capsule only.)

(4) Paraleptomys possesses a carotid circulatory pattern that is derived for muroid rodents in general but primitive for members of subfamily Murinae (character state 2 of Carleton, 1980; pattern 2 described by Voss, 1988; conformation diagrammed for Oligoryzomys by Carleton and Musser, 1989). We saw this pattern in certain cranial foramina and osseous landmarks in cleaned skulls, as well as dried blood vessels left on incompletely cleaned skulls. No sphenofrontal foramen exists at the junction of orbitosphenoid, alisphenoid, and frontal bones; no squamosal-alisphenoid groove scores the inner surface of each wall of the braincase, and no shallow trough extends diagonally over the dorsal (inner) surface of each pterygoid plate. There is a large stapedial foramen in the petromastoid fissure, and a deep groove extending from the middle lacerate foramen to the foramen ovale on the ventral posterolateral surface of each pterygoid plate. This disposition of foramina and grooves indicates that the stapedial artery branches from the common carotid, enters the periotic region through a large stapedial foramen, and as the infraorbital artery exits the periotic through the middle lacerate foramen, then courses in a short groove on the outside of the pterygoid plate to disappear into the braincase through the alisphenoid canal from which it emerges to run through the anterior alar fissure into the orbit. The supraorbital branch of the stapedial is absent. The circulatory pattern is widespread among murines ( Musser and Newcomb, 1983; Musser and Heaney, 1992), and is also found in some North and South American cricetids ( Carleton, 1980; Steppan, 1995; Voss, 1988; Weksler, 2006). Leptomys exhibits a carotid circulation that is derived for murines (character state 3 of Carleton, 1980; pattern 3 described by Voss, 1988; and the plan diagrammed for Oryzomys palustris by Carleton and Musser, 1989), as indicated by no stapedial foramen or sphenofrontal foramen, no squamosal-alisphenoid groove, no groove extending from the middle lacerate foramen to the foramen ovale on the ventral posterolateral surface of each pterygoid plate, and a shallow trough running diagonally over the dorsal (inner) surface of each pterygoid plate (see the previous section covering general description of Leptomys ).

(5) The dentary in Paraleptomys has a large coronoid process relative to area of the dentary surface (relatively smaller and more delicate in Leptomys ).

Dentition: Shapes of the upper and lower incisors, coloration of enamel, and the rightangle projection of the uppers from the rostrum (opisthodont configuration) are similar in Paraleptomys and Leptomys . They also share brachydont molars, with the first and second each anchored by four roots (large anterior and posterior roots, two smaller lingual holdfasts). Outside of these similarities the dentitions are quite different. Basically, Paraleptomys has fewer molars and those present have less complex occlusal surfaces, which take the form of a series of enamel basins very early in the animal’s life (comparable formation of basined laminae occurs much later in the life of Leptomys ). The morphology is similar to other hydromyins with no third molars and basined occlusal surfaces, such as the species of Hydromys , Baiyankamys , Parahydromys , and Crossomys , for example.

(1) Paraleptomys lacks upper and lower third molars (present in Leptomys ; one of two known examples of L. arfakensis , n. sp., is without lower third molars).

(2) No discernable cusp t7 exists on either the first or second upper molar in Paraleptomys (present on the first molars in Leptomys and also on the second molars in some samples; see table 6).

(3) The second upper molar in Paraleptomys is without an anterolabial cingular shelf or cusp (an anterolabial cingular shelf is present in some samples of Leptomys ; see table 6).

(4) Paraleptomys has a second upper molar consisting of only two basined laminae—cusp t8 and a posterior cingulum is not present. The second lamina is oriented toward the vertical, about 30 ° from the midline of the bony palate (cusp t8 is present and forms about a third of the occlusal surface of the second molar in Leptomys ; a posterior cingulum is evident in young dentitions but generally loses its identity with wear; and the second lamina is oriented toward the horizontal, about 70 ° from the midline of the palate).

(5) The first lower molar of Paraleptomys has a very small posterior cingulum relative to overall size of tooth; the posterior cingulum is either not present on the second lower molar or represented by an indistinct, minute bump (the posterior cingulum is large and conspicuous on both first and second lower molars in Leptomys ).

Disappearance of the third upper molars in all hydromyin genera except Leptomys is apparently attended by loss of the posterior lamina (the large cusp t8) and posterior cingulum on each second upper molar (‘‘Assumption of m 2 of the terminal position in the toothrow has so modified its form that the postero-external tubercle has been eliminated’’, opined Tate and Archbold [1941: 2]), and in Paraleptomys a reorientation of the second lamina from the horizontal toward a nearly vertical position. Attending the absence of third lower molars is a reduction in size of the posterior cingulum on each first molar (at least in Paraleptomys ) and diminution to a minute remnant, or actual loss, of the posterior cingulum on each second molar. This generalization does not apply to the holotype of Leptomys arfakensis , n. sp., which does not have lower third molars, yet exhibits a prominent posterior cingulum on each first and second lower molar (fig. 22).

Tate and Archbold (1941: 2) noted that the first upper molar in Paraleptomys was ‘‘slightly broader’’ than its counterpart in Leptomys , and the fissures separating the laminae deeper, differences we cannot verify and do not appreciate.

Phallus and spermatozoa: These contrasts are extracted from the literature; we have not verified them by direct observation of specimens.

(1) Paraleptomys has a simple phallus compared with that of Leptomys (comparisons based on L. ernstmayri and P. wilhelmina ; Lidicker, 1968). In both genera, the wall of the glans contains a conspicuous blood sinus extending into the bases of the urethral lappets, but in Paraleptomys , the primary blood sinus is medial to the inner crater (situated between inner and outer craters in Leptomys ). Paraleptomys has few or no tiny papillae and no denticles adorning the rim of the outer crater wall (large papillae and tiny denticles along the rim in Leptomys ), no dorsal groove extending from the dorsal notch in the rim of the outer crater wall (prominent dorsal groove in Leptomys ), a proximal baculum in which the shaft is bowed ventrally and its base strongly notched (straight shaft without a notch in the base in Leptomys ), and tips of lateral bacular processes oriented ventrally (point distally in Leptomys ).

(2) Leptomys and Paraleptomys share a similar spermatozoan morphology, differing primarily in the shorter tail of the latter ( Breed and Aplin, 1994; L. elegans , L. ernstmayri , and P. rufilatus were sampled).

Summary: Like the species of Leptomys , those of Paraleptomys are clearly terrestrial. The two genera also resemble each other in external body form and proportions. Their crania and mandibles are similar in overall shape and dimensional proportions—these elements appear much less highly derived than those in most other hydromyins. But Paraleptomys is characterized by a unique combination of traits relative to Leptomys : dark, thick, and woolly pelage; differently shaped interparietal and relatively shorter bony palate; large bullae relative to skull size, with more of the periotic cloaked by the ectotympanic capsule; a cephalic arterial circulation that is primitive for murines; larger coronoid process relative to area of dentary; no third molars; first and second upper molars with simple (compared to Leptomys ) and basined occlusal surfaces (no cusp t7 or anterolabial cingulum), second upper molar without cusp t8 and posterior cingulum; lower molars also exhibiting simple and highly basined chewing surfaces, and very small or no posterior cingula; and differences from Leptomys in phallic and spermatozoal features. This suite of traits does not characterize any of the specimens we examined representing the five species of Leptomys . Our observations, derived primarily from dry museum skins with accompanying cleaned skulls, do not support the notion that either Paraleptomys or Leptomys is polyphyletic, rather just the opposite: the species in each form separate monophyletic clusters.