Materials and Methods

The material we examined consists of standard skin-and-skull preparations or fluid-preserved specimens deposited in the following institutional collections: AMNH, American Museum of Natural History (New York); BMNH, Natural History Museum (formerly the British Museum of Natural History, London); FMNH, Field Museum of Natural History (Chicago); ICN, Museo de Historia Natural del Instituto de Ciencias Naturales de la Universidad Nacional de Colombia (Bogotá); IND-M, Instituto Alexander von Humboldt (formerly INDERENA, Villa de Leiva); MBUCV, Museo de Biología, Universidad Central de Venezuela (Caracas); MLS, Museo del Instituto La Salle (Bogotá); UMMZ, University of Michigan Museum of Zoology (Ann Arbor); USNM, National Museum of Natural History (Washington, D.C.).

We transcribed total length (TL) and length of tail (LT) from specimen tags and computed head-and-body length (HBL) by subtracting LT from TL. Length of the hind foot (HF, including claws) and of the external ear (Ear, from notch to distalmost margin of the pinna) were also transcribed from specimen tags; however, we usually remeasured hindfoot length on dried skins to check the accuracy of values recorded by the collector, and we used our values whenever large discrepancies were found. All external measurements are reported to the nearest millimeter (mm).

Craniodental measurements were taken with digital calipers and recorded to the nearest 0.01 mm, but values reported herein are rounded to the nearest 0.1 mm. The following were measured as illustrated and defined by Voss (1988, 1991): CIL, condylo-incisive length; LD, length of diastema; LM, occlusal length of the maxillary molar row; BM1, breadth of the first maxillary molar; LIF, length of one incisive foramen; BIF, breadth across both incisive foramina; BPB, breadth of the palatal bridge (between the M1 protocones); BZP, breadth of the zygomatic plate; LIB, least interorbital breadth; ZB, zygomatic breadth. In addition, we measured nasal length (NL) as the greatest longitudinal dimension of either the right or left nasal bone, and interparietal breadth (IPB) as the greatest transverse dimension of the interparietal bone.

Except as noted below, all analyzed character data were obtained from nonsenescent adults as determined by dental and pelage criteria. A specimen was judged to be adult if the molar dentition was fully erupted (with M3 exhibiting at least some wear) and if the pelage was mature; old adults (with molars worn below the widest part of the crown) were not measured or scored for qualitative characters. In effect, the specimens used for our statistical analyses approximately correspond to toothwear classes 2–4 as defined by Voss (1991).

Qualitative character variation is described herein using anatomical terminology that is explained or referenced by Reig (1977), Voss (1988), Carleton and Musser (1989), Voss (1993), Voss and Carleton (1993), and Steppan (1995).

Taxonomic History

The first species treated in this report was originally described as Aepeomys fuscatus by Allen (1912) on the basis of 12 specimens collected by W. B. Richardson near the village of San Antonio in the western Andes of Colombia. An ornithologist by training, Allen focused his description on pelage texture, pigmentation, and other features of the preserved skins he examined, while ignoring all qualitative craniodental characters (p. 89):

The genus Aepeomys was originally named by Thomas (1898) to contain two taxa, of which the designated type species was A. lugens (Thomas, 1896) from Venezuela. Because the second referred taxon, A. vulcani Thomas (1898), was based on a specimen from Ecuador, the known range of Aepeomys (sensu Thomas) bracketed the type locality of fuscatus, and Allen's (1912) generic assignment at least made biogeographic sense. However, with all of the known material of fuscatus in New York, and the types of lugens and vulcani in London, no other taxonomic assessment was possible given the incompleteness of Allen's description.

Although Aepeomys was recognized as a valid genus by Gyldenstolpe (1932) and Tate (1932b), Osgood (1933) argued that the type species (lugens) was insufficiently divergent from the many forms then referred to Thomasomys Coues, 1884, to maintain this usage and concluded that Aepeomys was just a junior synonym. Osgood's judgment was subsequently endorsed by Ellerman (1941), who recognized Thomasomys fuscatus, T. lugens, and T. vulcani as valid species in his influential monograph on rodent classification. Cabrera (1961) likewise followed Osgood's inclusive sense of Thomasomys, but treated all of the taxa formerly referred to Aepeomys as conspecific, listing fuscatus and vulcani as subspecies of T. lugens.

Apparently, neither Osgood, Ellerman, nor Cabrera actually examined specimens of fuscatus, and no new character information about this taxon was subsequently published until Gardner and Patton's (1976) review of karyotypic variation among Neotropical murids. In that report, the authors analyzed chromosomal preparations from fuscatus and six other species of Thomasomys (sensu Osgood) and remarked that Allen's form was highly distinctive. They summarized their conclusions in the context of suprageneric (tribal-level) assemblages as then understood by taxonomists (p. 32):

The second species treated in this report was originally described as Oryzomys intectus by Thomas (1921) on the basis of three specimens collected by Nicéforo María near Santa Elena in the central Andes of Colombia. Unlike Allen's description of fuscatus, Thomas's description of intectus included several relevant details of craniodental morphology (p. 356):

As the heterogeneous contents of Oryzomys were gradually sorted into species groups and subgenera by later generations of mammalogists, intectus remained an enigma, left unallocated by some authors (Ellerman, 1941) or referred to the nominate subgenus by default (Tate, 1932a; Cabrera, 1961). Despite recent advances in oryzomyine systematics that have resulted in several former subgenera of Oryzomys being elevated to generic rank (e.g., Microryzomys, Oligoryzomys, Oecomys; see Carleton and Musser [1989], Musser and Carleton [1993]), the classification of intectus has not been addressed by any published analysis or discussion of character data.

The chronology of the present study dates from the 1970s, when the late Charles O. Handley, Jr. (then curator of mammals at the National Museum of Natural History) and RSV (then a graduate student at the University of Michigan) independently examined the type series of Aepeomys fuscatus and concluded that it had been misclassified by Allen (1912) and by all subsequent compilers of Neotropical murid checklists. Whereas A. lugens (the type species of Aepeomys) was clearly allied with Thomasomys and certain other “thomasomyines”, fuscatus exhibited equally unambiguous similarities to oryzomyines. Rather than belonging to Oryzomys or other allied genera, however, fuscatus seemed to be a highly divergent form. Handley and RSV compared notes and agreed to collaborate on a joint report, but other projects intervened and only a few colleagues were informed about their results. In the meantime, Guy G. Musser examined the type series of Oryzomys intectus in London, recognized that it represented the same clade discovered by Handley and RSV, and generously provided them with copies of his notes; RSV followed this up by examining intectus himself on a subsequent visit to the Natural History Museum.

Other independent discoveries of essentially the same facts about fuscatus and intectus were made by MGL and VP; their results were based on new specimens and/or new characters (not previously considered by Handley and RSV), resulting in this collaboration. With the recent death of Charles Handley, it seems appropriate to name the new genus for him in recognition of his past participation in our work and in tribute to his many important contributions over the course of a long and productive career in Neotropical mammalogy (fig. 1).

Handleyomys, new genus Figures 2–11

Type Species: Aepeomys fuscatus Allen, 1912.

Geographic Distribution: Known from the western and central Andean cordilleras of Colombia at elevations from ca. 1500 to 2800 m above sea level (fig. 12).

Contents: Two species as redescribed below.

Morphological Diagnosis and Description: Adult body pelage fine and soft, uniformly dull brownish-gray dorsally, usually darker (sometimes almost blackish) middorsally than on flanks; ventral pelage dark gray frosted with paler gray or buff, not sharply countershaded. Mystacial, superciliary, genal, submental, interramal, and carpal vibrissae present; mystacial hairs neither very short nor very long, extending posteriorly to (but not beyond) caudal margins of pinnae when laid back against cheeks of properly made-up skins. Pinnae not large but clearly visible above fur of head, sparsely covered with short dark hairs. Manus sparsely covered dorsally with short pale (whitish or silvery) hairs; ventral surface naked and unpigmented, with five separate plantar tubercles (two carpal and three interdigital pads); claws short, neither conspicuously elongated nor unusually recurved. Pes long and narrow, with outer digits (I and V) much shorter than middle three (claw of dI extending to middle of first phalange of dII, claw of dV extending just beyond first interphalangeal joint of dIV); conspicuous ungual tufts of long silvery hairs rooted at bases of claws on dII–dV, but pedal dorsum otherwise only sparsely covered with short pale or dark-banded hairs; plantar surface (including heel) naked, weakly pigmented (grayish in life), with two metatarsal and four interdigital pads; indistinct squamae (scale-like tubercles) sparsely distributed along outer distal plantar surfaces, but not in center of sole. Tail about as long as combined length of head and body, apparently naked (a sparse caudal pelage is only visible under magnification), and unicolored (dark above and below). Mammae six in inguinal, abdominal, and postaxial pairs.

Skull with long, tapering rostrum flanked by shallow but distinct zygomatic notches; interorbital region hourglass-shaped, neither greatly inflated nor unusually constricted, with rounded supraorbital margins; braincase moderately inflated and rounded, without prominent temporal crests, ridges, or beads. Zygomatic plate (in lateral view) moderately broad, its anterior edge vertical or nearly so (not sloping backward), with a rounded (never angular or spinous) anterodorsal contour. Premaxillae short (not produced anteriorly beyond incisors to form a rostral tube with nasals). Incisive foramina neither very short nor greatly elongated, averaging about 60% of diastemal length (not extending posteriorly between molar rows) and widest near premaxillary/maxillary sutures. Palatal bridge wide and long, without median ridge or deep lateral gutters; posterolateral pits usually large and often complex, typically consisting of two or more foramina recessed in a common fossa on each side. Mesopterygoid fossa not penetrating anteriorly between molar rows; bony roof of fossa complete or perforated only by narrow slits, never conspicuously fenestrated. Alisphenoid strut absent (buccinator-masticatory foramen and foramen ovale confluent). Carotid arterial morphology primitive, with orbitofacial circulation supplied by separate supraorbital and infraorbital branches of large stapedial artery (= pattern 1 of Voss, 1988); course of supraorbital stapedial ramus marked by prominent squamosal-alisphenoid groove and sphenofrontal foramen. Postglenoid foramen separated from large subsquamosal fenestra by slender hamular process of squamosal. Tegmen tympani not overlapping squamosal, or tegmen tympani-squamosal overlap weak (not involving a distinct posterior suspensory process of the latter bone). Bullae small; pars flaccida of tympanic membrane present, large; orbicular apophysis of malleus well developed.

Mandible with well-developed falciform coronoid process; lower incisor alveolus without distinct capsular process on lateral mandibular surface. Basihyal morphology unknown.

Incisors ungrooved, with yellow-orange enamel bands; upper teeth small, narrow, deeper than wide, opisthodont; lower teeth unremarkable in coloration or morphology.

Maxillary molar rows parallel; principal cusps arranged in opposite labial/lingual pairs, bunodont when unworn but quickly eroded with age to same level as other enamelled structures (occlusal surface of most adult teeth more-or-less planar); labial and lingual reentrant folds long and interpenetrating (incipiently lophodont sensu Voss, 1993); M1 anterocone entire, not divided into labial and lingual conules (anteromedian flexus absent); anteroflexus very deep, extending lingually beyond dental midline on M1 and M2; anterolophs and mesolophs large, fused with corresponding (antero- and meso-) styles on labial margins of M1 and M2; posterolophs distinct on M1 and M2, persisting with moderate to heavy wear; M3 subtriangular, smaller than more anterior teeth, with most of the same occlusal elements but usually without distinct hypocone or posteroloph. First maxillary molar with one accessory labial root (four roots total); M2 and M3 three-rooted.

Anteroconid of first mandibular molar (m1) undivided by median flexid, fused with protolophid and/or anterolophid to enclose a persistent internal fold of uncertain homology (anteroflexid and/or protoflexid); anterolophid absent on m2 and m3; anterolabial cingulum absent or indistinct on m2, consistently absent on m3; mesolophids and posterolophids large and well developed on all mandibular teeth; ectolophids consistently absent. First mandibular molar with an accessory labial root, occasionally also with an accessory lingual root (three to four roots total); m2 and m3 each with two small anterior roots and one large posterior root (three roots total).

Tuberculum of first rib articulates with transverse processes of seventh cervical and first thoracic vertebrae; second thoracic vertebra with greatly elongated neural spine; entepicondylar foramen of humerus absent. Thoracicolumbar vertebrae 19; sacrals 4; caudals 28–30, with or without hemal arches4; ribs 12.

Stomach (in two dissected specimens of H. intectus) unilocular and hemiglandular, without any extension of glandular epithelium into corpus; bordering fold crosses lesser curvature slightly to right of incisura angularis (between that flexure and the pylorus); bordering fold crosses greater curvature opposite incisura angularis. Gall bladder absent (in two dissected specimens of H. intectus). Phallic and other male reproductive characters undetermined.

Karyotype: Gardner and Patton (1976: table 2) reported a diploid number (2n) of 54 chromosomes and a fundamental number (FN) of 62 from three karyotyped specimens of Handleyomys fuscatus (USNM 507267–507269); both the X and Y chromosomes are submetacentrics (op. cit.).

Comparisons with Aepeomys: As recognized prior to this report (e.g., by Musser and Carleton, 1993; Ochoa et al., 2001), the genus Aepeomys contained four other nominal taxa in addition to fuscatus: lugens Thomas (1896), vulcani Thomas (1898), ottleyi Anthony (1932), and reigi Ochoa et al. (2001). Of these, ottleyi has long been considered a subjective junior synonym of A. lugens (see Osgood, 1933), an assessment with which we completely agree. Aepeomys reigi, however, is a valid species distinguished from A. lugens by subtle but consistent morphological characters and by its highly distinctive karyotype (Ochoa et al., 2001). The fourth nominal taxon, vulcani, is based on a partially crushed holotype (BMNH 98.5.1.10) that exhibits none of the distinctive attributes shared by A. lugens and A. reigi; as independently determined by C.O. Handley, Jr. (personal commun.) and VP (in prep.), vulcani is referable to the genus Thomasomys (sensu stricto), wherein its possible synonymy with other nominal taxa remains to be determined.

Thus restricted (excluding fuscatus and vulcani), Aepeomys is a morphologically diagnosable taxon that can be unambiguously distinguished from Handleyomys despite a superficial resemblance in external features. (Both genera include drab-colored, soft-furred mice with naked tails about as long as heads-and-bodies; narrow hindfeet with six plantar pads, short outer digits, and long ungual tufts; and six mammae.) Because field identifications of these allopatric5 taxa are not problematic, taxonomic comparisons are more usefully focused on their many salient points of difference in skeletal and visceral comparisons.

In dorsal and lateral cranial views (figs. 4, 5, 8) the rostrum of Aepeomys appears long and tapering, with a prominent bony tube formed by the nasals and premaxillae that projects anterodorsally well beyond the incisors; the rostrum of Handleyomys appears short and blunt by comparison, and lacks any trace of a nasal-premaxillary tube. Flanking the base of the rostrum on each side of the skull in Handleyomys is a shallow but well-defined zygomatic notch formed by the free anterodorsal margin of a broad, more-or-less vertically oriented zygomatic plate. By contrast, the zygomatic notches are indistinct in Aepeomys because the zygomatic plate is slender and slopes posterodosally without a free anterodorsal projection. The interorbital region is hourglass-shaped (concave-sided in dorsal view) with rounded supraorbital margins in both taxa (fig. 9A, B), but this part of the skull is relatively narrower in Handleyomys, a consequence of its less inflated frontal sinuses and olfactory bulb. The braincase is likewise more compact and globular in Handleyomys by comparison with the more inflated and ovoid calvarium of Aepeomys.

In ventral cranial view, the two genera are most conspicuously distinguished by palatal morphology. Whereas the palate of Aepeomys is relatively short (not extending posteriorly much behind the molar rows) and lacks well-developed posterolateral pits (fig. 10A), the palate of Handleyomys is long (extending posteriorly well behind the molar rows) and is perforated posterolaterally by one or more large pits between M3 and the mesopterygoid fossa (fig. 10B). In addition, the wide mesopterygoid fossa of Aepeomys is flanked by long, narrow parapterygoid fossae, but the mesopterygoid fossa of Handleyomys is relatively narrower and flanked by shorter, broader parapterygoid fossae; in both genera, the mesopterygoid roof is entirely bony or is perforated only by small foramina.

In Aepeomys, the bulla is firmly anchored to the rest of the skull anterodorsally by broad overlap between the tegmen tympani and a posterior suspensory process of the squamosal, and the postglenoid foramen (through which the eponymous vein exits the cranial lumen) is smaller than the subsquamosal fenestra (fig. 11A). By contrast, the bulla of Handleyomys is not firmly anchored to the skull anterodorsally because the squamosal lacks a posterior suspensory process; in consequence, the postglenoid foramen is much wider than the subsquamosal fenestra (fig. 11B).

The upper molar dentitions of both genera are highly distinctive (fig. 6). The first maxillary molar of Aepeomys is relatively longer and narrower than the shorter, broader M1 of Handleyomys, and the unworn labial cusps (paracone and metacone) of M1–M2 in Aepeomys are much taller and sharper than their homologs in Handleyomys. With increasing toothwear, the principal molar cusps of Aepeomys persist as distinct tubercles, whereas the molar cusps of Handleyomys are quickly worn down to form a common planar surface with other occlusal features. Differences in the length and orientation of the principal labial and lingual flexi are also striking. In Aepeomys, the internal segments of the major labial folds (paraflexus and metaflexus) are oriented anteroposteriorly down the midline of M1–M2, and the lingual flexi (protoflexus and hypoflexus) are shallow; longitudinal (rather than transverse) enamelled crests are therefore prominent in these teeth. By contrast, the principal labial flexi of Handleyomys slant transversely across the midline of the tooth to interpenetrate with much longer lingual flexi, resulting in the morphology that Voss (1993: 20) termed “incipient lophodonty”. The unworn anterocone of M1 is deeply divided by an anteromedian flexus in Aepeomys, whereas the M1 anterocone is undivided in Handleyomys. Corresponding differences are apparent in the lower molars as well (fig. 7).

Whereas 13 ribs have been reported for Aepeomys lugens (see Steppan, 1995: table 5), all examined postcranial skeletons of Handleyomys fuscatus (ICN 12786) and H. intectus (ICN 12160, 12164) have 12 ribs.

Visceral differences that we observed between dissected examples of Aepeomys lugens (MBUCV I-2793, I-2794) and Handleyomys intectus (ICN 16092, 16093) include gastric morphology and occurrence of a gall bladder. The stomachs of both genera are unilocular (single-chambered), but gastric glandular epithelium in Aepeomys is restricted to a small, pouch-like structure on the greater curvature (closely resembling the condition illustrated by Carleton [1973: fig. 5C] for Oxymycterus rutilans). By contrast, the entire right half (antrum) of the stomach is lined with glandular epithelium in Handleyomys (closely resembling the hemiglandular condition illustrated by Carleton [1973: fig. 2D] for Zygodontomys brevicauda). A large and distinct gall bladder is present in a cleft between the left and right halves of the cystic lobe of the liver in Aepeomys (as previously reported by Voss, 1991: table 4), but both dissected specimens of Handleyomys lack a gall bladder.

Comparisons with Oryzomys: Even excluding intectus, the genus Oryzomys remains a morphologically heterogeneous collection of species that defies meaningful diagnosis. Because several systematic studies have recently suggested that Oryzomys—even in its strict modern sense (Musser and Carleton, 1993)—is not monophyletic (see Myers et al., 1995; Steppan, 1995; Weksler, 1996; Bonvincino and Moreira, 2001), we base our comparisons on a small group of taxa that appear to be closely related to the type species, O. palustris (Harlan). This core concept of Oryzomys essentially corresponds to Goldman's (1918) palustris group and additionally includes O. couesi (Alston), O. dimidiatus (Thomas), O. gorgasi Hershkovitz, and other possibly valid species currently regarded as synonyms of O. couesi (see Sánchez et al., 2001). In the following paragraphs we use Oryzomys in this narrowly defined sense, acknowledging that additional genera must eventually be named to contain at least some of the many other species currently referred to this long-abused taxon. Carleton and Musser (1989) provided detailed illustrations and descriptions of many morphological traits of O. palustris that should be consulted for additional details of characters mentioned below.

Species of Oryzomys are much more boldly marked externally—with abruptly countershaded heads and bodies and sharply bicolored tails—than species of Handleyomys, which lack distinct countershading and have unicolored-dark tails. Whereas the plantar surface of the hindfoot is almost entirely covered with squamae and the hypothenar (lateral tarsal) pad is indistinct or absent in Oryzomys (see Carleton and Musser, 1989: fig. 9A), squamae are only present on the outer distal margins of the sole and the hypothenar pad is distinct in Handleyomys (fig. 3). Although well-developed fringes of silvery hairs are present along the plantar margins of the hindfoot, the claws are naked in Oryzomys; by contrast, no conspicuous hairy fringes occur along the plantar margins, but the claws on digits II–V are concealed by long tufts of ungual hairs in Handleyomys. Additionally, Oryzomys has eight mammae in four pairs (inguinal, abdominal, postaxial, and pectoral), but Handleyomys has only six teats in three pairs (the pectoral pair is absent; see Voss and Carleton [1993: fig. 8] for a map of muroid mammary loci).

Oryzomys is cranially distinctive by virtue of its deep zygomatic notches (the zygomatic notches are distinct but much shallower in Handleyomys; fig. 8), convergent interorbital region with beaded supraorbital margins (the interorbit is hourglass-shaped with rounded margins in Handleyomys; fig. 9), widely fenestrated mesopterygoid roof (the mesopterygoid roof is entirely bony or only narrowly perforated by sphenopalatine openings in Handleyomys; fig. 10), and highly derived (pattern 3) carotid circulation (versus pattern 1 in Handleyomys; see Voss [1988: fig. 18] for illustrated examples and explanations of these morphologies). The lower incisor root is contained in a prominent bony capsule on the lateral surface of the mandible below the base of the coronoid process in Oryzomys, but a distinct capsular process is absent in Handleyomys. Whereas the molar dentition of Oryzomys is persistently tubercular and nonlophodont, the principal cusps of Handleyomys molars are quickly worn down to a flat (planar) occlusal surface dominated by interpenetrating labial and lingual flexi (conforming to the incipiently lophodont morphotype defined by Voss [1993: 20]).

Despite the many points of external and craniodental difference noted above, Oryzomys and Handleyomys resemble one another in all postcranial-skeletal and visceral characters scored for this report. In particular, both taxa have 12 ribs and unilocular-hemiglandular stomachs; the gall bladder is likewise absent in each genus.

Phylogenetic Relationships and Other Comparisons: Pentalophodont sigmodontines—those with well-developed mesoloph(id)s—have traditionally been referred to one or the other of two suprageneric groups: one cluster including Thomasomys and its putative allies (“thomasomyines”), the other Oryzomys and its supposed relatives (oryzomyines; see Voss [1993: 21–25] for a historical review of this dichotomy). Although any discussion of the relationships of Handleyomys in this report will soon be superseded by taxon-dense phylogenetic analyses of “thomasomyines” (by VP) and oryzomyines (by M. Weksler) that are currently nearing completion, some preliminary remarks about the data at hand are appropriate.

Among those morphological characters with well-defined states for which hypotheses of polarity have been proposed in the literature (table 1), it is noteworthy that none supports a close relationship between Aepeomys (a “thomasomyine”) and Handleyomys. Instead, all of the tabulated character states shared by these taxa appear to be plesiomorphic on the basis of current assessments of evolutionary transformations in sigmodontine morphology. Likewise, we found no compelling evidence of recent common ancestry between Aepeomys and Handleyomys among characters that we do not tabulate, most of which have less well-defined states and/or more ambiguous polarity.

In contrast, Handleyomys shares an impressive number of shared-derived resemblances with Oryzomys, including (1) a long palate with large posterolateral pits; (2) absence of overlap between the tegmen tympani and a posterior suspensory process of squamosal; (3) absence of an anteromedian flexus on M1; (4) 12 ribs; and (5) absence of a gall bladder. Because traits 1, 2, 4, and 5 are currently thought to be oryzomyine synapomorphies (Voss and Carleton, 1993; Steppan, 1995), we hypothesize that Handleyomys is a member of that clade, within which other taxonomic comparisons are appropriate. Three oryzomyine genera deserve particular attention.

In his original description of Oryzomys intectus, Thomas (1921) compared it with Melanomys, remarking general similarities in cranial shape and molar morphology. However, species of Melanomys differ from Handleyomys by their coarser, glossier, more richly colored (usually chestnut- or reddish-brown) pelage; relatively shorter tail (LT < HBL); shorter hindfeet with blackish soles and large, almost-naked claws; mammary complement of eight teats; relatively shorter rostrum; relatively broader, convergent-beaded interorbit; relatively much narrower parapterygoid fossae; highly derived (pattern 3) carotid circulation; lack of subsquamosal fenestrae; and narrower, nonlophodont molars (see Goldman [1918, 1920] for descriptions and illustrations). In fact, we have not found a single trait in which Melanomys and Handleyomys resemble one another that is not also shared with many other small oryzomyines. Prima facie, they do not appear to be closely related.

Ellerman (1941: 341) astutely noted the “highly aberrant dentition” of Oryzomys intectus with respect to other congeners, and remarked that “… this species should, I think, be transferred to the genus Nectomys.” Although the incipiently lophodont molars of Handleyomys (figs. 6, 7) and Nectomys (see Hershkovitz, 1944: fig. 4; Gómez-Laverde et al., 1999: fig. 4) are a noteworthy point of shared-derived resemblance, these genera are dissimilar in most other respects. Species of Nectomys are much larger rats (adult HBL > 160 mm) that differ from Handleyomys in many external and craniodental traits, including their coarser, glossier, water-repellent fur; relatively longer tail (LT > HBL); very large hind feet with prominent plantar squamae, well-developed interdigital webbing, and naked claws (Hershkovitz, 1944: fig. 2c, d); mammary complement of eight teats; relatively shorter rostrum; much deeper zygomatic notches and correspondingly broader zygomatic plates; convergent-beaded interorbit; highly derived (pattern 3) carotid circulation; and lack of subsquamosal fenestrae. That Ellerman did not, in fact, transfer intectus from Oryzomys to Nectomys probably resulted from the lack of supporting evidence from other (nonmolar) characters.

Although Voss and Carleton (1993) stated that all oryzomyines possess eight mammae in four pairs (inguinal, abdominal, postaxial, and pectoral), subsequent research has shown that the oryzomyine genus Scolomys is characterized by having only six teats because the pectoral mammary pair is absent (Patton and da Silva, 1995). Handleyomys likewise has only six teats (lacking pectoral mammae), but the phylogenetic interpretation of this similarity is ambiguous. Six mammae are thought to represent the plesiomorphic condition among Neotropical sigmodontines (according to the ingroup and outgroup assumptions explained by Voss, 1993: 12), so the shared possession of this primitive feature by Scolomys and Handleyomys suggests that these genera might be basal to an oryzomyine radiation of eight-mammate taxa. Alternatively, if Scolomys and Handleyomys are descended from an oryzomyine ancestor with eight teats, their shared loss of pectoral mammae provides evidence of a sister-group relationship. Obviously, distinguishing between these (and other) alternatives is impossible in the absence of a comprehensive analysis of oryzomyine phylogeny, but it is relevant that no other special similarities suggest a close relationship between Scolomys and Handleyomys.

Indeed, no oryzomyine genus appears to exhibit any compelling pattern of synapomorphies with Handleyomys apart from those that support tribal monophyly. Likewise, none of the morphologically diverse species currently referred to Oryzomys (sensu Musser and Carleton, 1993) seems to show any special morphological similarity with H. fuscatus or H. intectus. Given the Linnaean convention of binomial nomenclature and the absence of evidence for close relationship to other oryzomyines, we treat Handleyomys as a genus, but we acknowledge that assigning ranks to supraspecific clades is a biologically arbitrary exercise. Determining the hierarchical position of this taxon with respect to other sigmodontine lineages is the key issue, and one that remains to be determined by future phylogenetic studies.

Geographic Variation and Species Limits

All of the material of Handleyomys that we examined comes from elevations above 1500 m in the western Andes (Cordillera Occidental) and central Andes (Cordillera Central) of Colombia. Because the cool, humid habitats that prevail above 1500 m in these mountain ranges are separated by the hot, semi-arid lowlands of the upper Río Cauca (fig. 12), we tested the hypothesis that populations of Handleyomys in the western Andes (including the type locality of fuscatus) are morphologically distinct from allopatric populations in the central Andes (including the type locality of intectus).

Close visual inspection of skins and skulls from both regions revealed three qualitative characters that exhibit significant geographic variation in trait frequencies. (1) The short hairs that cover the dorsal surface of the hind feet are dark-banded (with dense concentrations of melanin that are distinctly visible under low magnification) in all examined specimens from the western Andes, but most specimens from the central Andes have hind feet that are covered dorsally with pale (pure-white or indistinctly pigmented) hairs. (2) The nasal bones are long (extending posteriorly well beyond the premaxillary-maxillary suture; fig. 13B) in most examined specimens from the central Andes, whereas most specimens from the western Andes have short nasals (truncated posteriorly at or near the premaxillary-maxillary suture; fig. 13A). (3) The incisive foramina are anteriorly constricted (with lateral margins that are abruptly narrowed at or near the premaxillary-maxillary suture; fig. 14A) in most specimens from the western Andes, but these diastemal openings are smoothly tapering (with evenly rounded lateral margins like parentheses; fig. 14B) in most specimens from the central Andes. Although none of these characters exhibits fixed differences between western and central Andean samples (table 2), it is noteworthy that no examined specimen is geographically atypical in all three scored traits. Thus, no specimen from the western Andes has pale hind feet, long nasal bones, and smoothly tapering incisive foramina; nor does any specimen from the central Andes have dark hind feet, short nasal bones, and anteriorly constricted incisive foramina. Therefore, these aggregate data suggest that the two cordilleras are inhabited by morphologically distinctive populations of Handleyomys.

We also calculated summary statistics for external and craniodental measurements of adult Handleyomys to assess univariate patterns of morphometric divergence between populations from the western and central Andes (table 3). External measurement means appear to be similar from the two cordilleras, but we did not apply statistical tests because specimens were measured in the field by many different persons (including inexperienced undergraduates) using unknown protocols. We used one-way Analyses of Variance, however, to test for sample differences in craniodental measurements, all of which were taken by us using identical methods. No significant differences were found between cordilleran samples in Condylo-incisive Length, Length of Diastema, Breadth of Incisive Foramina, Breadth of Palatal Bridge, and Zygomatic Breadth. Other craniodental measurements showed statistically significant differences, but some of these are not empirically compelling (e.g., BM1 and LIB, with mean differences < 0.01 mm). The most visually obvious differences indicated by these tests are in Length of Nasals (already scored as a qualitative trait; table 2, fig. 12), and Interparietal Breadth. Although the latter measurement exhibits slight overlap between western and central Andean specimens (8.2–10.2 mm versus 6.2–8.8 mm, respectively), the mean morphological difference is obvious when skulls are viewed side-by-side (fig. 15). Contrasts between western and central Andean skulls in Molar Length, Length of the Incisive Foramina, and Breadth of the Zygomatic Plate are visually subtle in some pairwise comparisons of local population samples but visually conspicuous in others (see below).

To summarize geographic patterns of multivariate morphometric variation, we calculated Mahalanobis distances (D) among five geographic samples, two from the western Andes and three from the central Andes (table 4). In general, morphometric distances between geographic samples from the same cordillera are smaller than distances between samples from different cordilleras, a pattern that can be heuristically summarized by cluster analysis (fig. 16). Because some sample sizes are small, clustering sequences determined by small differences in computed distances (e.g., within the central Andean cluster) are probably not biologically significant, and some large values (e.g., between Valle del Cauca and Western Risaralda) might be artifactually inflated. Nevertheless, the substantial estimated mean distance between western and Central Andean samples (ca. 6.5 within-sample multivariate standard deviations) reinforces the conclusion previously based on qualitative character data that these mountain ranges are occupied by morphologically differentiated populations of Handleyomys.

A valid criticism of cluster-analytic summaries of geographic variation is that hierarchical patterns are forced on data that may not, in fact, be hierarchically structured (de Queiroz and Good, 1997). In particular, discontinuous variation (as resulting from speciation) cannot be distinguished by cluster analysis from intraspecific clinal variation without taking the geographic distribution of the analyzed samples into consideration (op. cit.). Intraspecific clinal variation is not, however, a plausible explanation for our results, because geographic and morphometric distances are not correlated in Handleyomys. The Quindío/eastern Risaralda sample, for example, clusters with geographically distant samples from Antioquia rather than with the geographically adjacent sample from western Risaralda (just across the Cauca valley). Nevertheless, it is possible that such geographically proximate but morphometrically divergent populations of Handleyomys might show some degree of intergradation.

To further assess the distinctness of adjacent western Andean versus central Andean populations, we used principal components analysis to summarize patterns of multivariate craniometric variation in our samples from El Campamento (in western Risaralda; locality 14) and La Suiza (in eastern Risaralda; locality 12). The first two principal components account for about 75% of the total variance (table 5). Projected specimen scores (fig. 17) indicate complete sample separation on the first component, the coefficients of which suggest that La Suiza specimens differ from El Campamento specimens by their longer incisive foramina, narrower zygomatic plates, longer nasals, and narrower interparietals (essentially the same pattern of morphometric divergence previously identified by our univariate comparisons of central versus western Andean series). The second principal component appears to represent a general size factor, with coefficients that resemble the usual muroid pattern of multivariate craniodental growth allometry (Voss et al., 1990; Voss and Marcus, 1992).6

We interpret these results as indicating the presence of two species of Handleyomys, one in the western Andes and the other in the central Andes, despite the apparent absence of any completely fixed qualitative character difference. Fortunately, names are already available for both taxa: H. fuscatus for the populations inhabiting the Cordillera Occidental, and H. intectus for populations in the Cordillera Central. We briefly summarize the distinguishing attributes of each taxon in the following accounts.

Handleyomys fuscatus (J.A. Allen, 1912)

Type Material and Type Locality: Allen's (1912) original description was based on the holotype (AMNH 32230, preserved as a skin and skull) and 11 paratypes, all of which were collected between 7000 and 8000 ft near San Antonio, Valle del Cauca, Colombia, by W.B. Richardson from 5 January to 31 March 1911.

Geographic Distribution: All referred specimens of Handleyomys fuscatus are from the western Andes (Cordillera Occidental) of Colombia between 1700 and 2580 m above sea level in the departments of Valle del Cauca and Risaralda.

Emended Diagnosis: A species of Handleyomys distinguished by hindfeet that are covered dorsally with dark-banded (never pure-white) hairs; nasals that are usually truncated posteriorly at or near the premaxillary-maxillary suture; incisive foramina with lateral margins that are usually constricted abruptly anteriorly; and wide-shallow interparietals.

Variation: All of the material at hand of Handleyomys fuscatus comes from two clusters of adjacent localities, one of which is near Cali in the department of Valle del Cauca and the other in western Risaralda department. Although univariate mean differences between samples from these regions are not impressive (table 6), it is noteworthy that the pattern of divergence is not attributable to a simple size factor. Whereas specimens from Valle del Cauca average larger than specimens from western Risaralda in several dimensions (e.g., CIL, LD, LIB, ZB, IPB), the same specimens have absolutely and relatively smaller incisive foramina (LIF, BIF), more slender zygomatic plates (BZP), and shorter nasals (NL). Because only a few measurable adults are available from Valle del Cauca, it is possible that some of these differences are artifactual, but the estimated multivariate distance between Valle del Cauca and western Risaralda (>5 within-sample standard deviations; fig. 16) is larger than that observed between any other conspecific pair of samples analyzed in this report and suggests that substantial genetic variation might exist among Handleyomys populations from the Cordillera Occidental. In the current absence of other supporting evidence, it does not seem necessary or useful to recognize more than a single taxon in the western Andes, but future karyotypic and molecular datasets should be carefully assessed for geographic variation among populations here referred to H. fuscatus.

Specimens Examined: Risaralda, El Campamento (ICN 12799–12827), El Empalado (ICN 12700–12703, 12705–12710), La Jalea (ICN 12704), Los Planes (ICN 12783–12793, 12795–12798), Mampay (ICN 15277), Siató (ICN 12208, 12725–12727); Valle del Cauca, Campamento Corea (IND-M 3657, 3659), El Queremal (ICN 6903), Finca la Playa (ICN 4376, 4381, 4383, 4388, 4389), Peñas Blancas (USNM 507267–507269), San Antonio (AMNH 32227–32233, 32236–32239; FMNH 20109).

Handleyomys intectus (Thomas, 1921)

Type Material and Type Locality: Thomas's (1921) original description was based on the holotype (BMNH 21.7.1.17, preserved as a skin and skull) and two paratypes, all of which were collected at Santa Elena, near Medellín, Colombia, by Nicéforo María in December 1919 and January 1920.

Distribution: All referred specimens of Handleyomys intectus are from the central Andes (Cordillera Central) of Colombia between 1500 and 2800 m above sea level in the departments of Antioquia, Quindío, and Risaralda.

Emended Diagnosis: A species of Handleyomys distinguished by hindfeet that are usually covered dorsally by pure-white (or indistinctly pigmented) hairs; nasals that usually extend posteriorly well beyond the premaxillary-maxillary suture; incisive foramina that usually have smoothly tapering (never abruptly constricted) lateral margins; and narrow-deep interparietals.

Variation: Available samples of Handleyomys intectus are remarkably similar throughout the known range of the species, with measurement means that seldom differ geographically by more than about 0.1 mm (table 7). The only noteworthy exception to this tendency are the interparietals of specimens from Quindío and eastern Risaralda, which are markedly narrower, on average, than those of specimens collected in Antioquia. Two specimens collected near El Retiro (ICN 16091, 16093) in the department of Antioquia are unusually small but do not appear to be exceptional in other respects.

Specimens Examined: Antioquia, Finca Cañaveral (ICN 16091–16095), La Ceja (AMNH 61575; MLS 187vii, 188vii), La Forzosa (ICN 16073–16079), Río Negrito (FMNH 70296–70306), Santa Elena (BMNH 21.7.1.17–21.7.1.19 [the type series], AMNH 37734), Ventanas (FMNH 70332–70338); Quindío, El Roble (AMNH 32928, 32931–32933, 32937, 32940, 33021), Salento (AMNH 32939); Risaralda, La Suiza (ICN 12104, 12158–12179, 12891).

Habitat Descriptions

Departamento Antioquia, Municipio Anorí (9 km S Anorí), Vereda Roble Arriba, bosque La Forzosa, 1775 m: The protected area known as La Forzosa (fig. 12: locality 1) is a ca. 450 ha fragment of primary premontane wet forest (“bosque muy húmedo Subtropical” of Espinal and Montenegro, 1963) on the northeastern flanks of the central Andes east of the upper Río Nechí (Cuervo et al., 2001). Frequent mists, high annual precipitation (2600–3500 mm), and high relative humidity (77–95%) characterize the local climate, which includes a short dry season from December to February and a longer wet season from May to October (op. cit.). The local topography is complex,

Carlos A. Delgado V. collected seven specimens of Handleyomys intectus at this site between 7 and 11 January 2001. The animals were taken in Sherman live traps set in mature forest festooned with bromeliads, ferns, moss, and orchids; no specimens were taken in traps set along the forest edge or in clearings. Traps that took H. intectus at La Forzosa were placed on the ground among the elevated roots of palms and other trees, beneath the large leaves of low-growing understory plants, and inside rotting logs. All captured individuals apparently entered the traps at night. Sympatric rodents trapped at the same locality included Heteromys australis (voucher: ICN 16072), Rhipidomys latimanus (ICN 16087, 16088), Oryzomys albigularis (ICN 16084–16086), and Melanomys caliginosus (ICN 16080–16083).

Departamento Antioquia, Municipio El Retiro (4 km S El Retiro), Vereda Puente Peláez, Finca Cañaveral, 2100 m: Perched on the western flanks of the central Andes in the valley of the Río La Miel, the landscape of Finca Cañaveral (fig. 12: locality 2) is dominated by anthropogenic habitats, principally pastures (for milk cattle) and agricultural fields, but relictual forest (“bosque húmedo Subtropical” according to Espinal and Montenegro's [1963] map) survives along the margins of streams that descend from continuous forest on steeper slopes above the valley floor. Carlos A. Delgado V. collected three specimens of Handleyomys intectus on two brief visits to this site, once in August 2000 and again in March 2001. The animals were trapped at night in dense forest (including Cecropia sp., Quercus humboldti, Inga sp., Schefflera sp., Guatteria goudotiana, Vismia sp., Piperaceae, Rubiaceae, ferns, orchids, and bromeliads) along the banks of a small unnamed tributary of the Río La Miel. Other rodents taken in the same traplines as Handleyomys at this site were Akodon affinis (vouchers: ICN 16089, 16090), Microryzomys minutus (ICN 16096), and Reithrodontomys mexicanus (ICN 16097–16100).

Departamento Risaralda, Municipio Pereira, Corregimiento La Florida, Vereda La Suiza, 1900–1950 m: The area known as La Suiza (fig. 12: locality 12) is a protected area in the valley of the Río Otún on the western slope of the Cordillera Central. Although the biological station at La Suiza is surrounded by pastures, agricultural fields, and alder (Alnus acuminata) plantations, some primary cloud forest (“bosque muy húmedo Montano Bajo” of Espinal and Montenego, 1963; including large specimens of Quercus humboldti) persists along the nearby Quebrada La Hacienda and on surrounding hillsides. Local rainfall is bimodally distributed, with average monthly maxima in June (237 mm) and November (290 mm); January is the driest month on average (121 mm), and the mean annual total is 2038 mm (Aguilar and Rangel, 1994). The median annual temperature is about 24°C, with average daily maxima and minima of 33° and 16°, respectively (op. cit.).

The small mammal fauna at La Suiza was sampled by ICN researchers on two separate occasions in 1992. On the first visit to this site, MGL and Angélica Peñuela trapped for five consecutive nights (22–27 February) in the dry season; on the second, MGL and Rocío Polanco trapped for another five consecutive nights (8–13 June) in the rainy season. The equipment used on both visits consisted of medium-size (ca. 80 ;ts 90 ;ts 230 mm) Sherman live traps, large (ca. 100 ;ts 115 ;ts 380 mm) Sherman live traps, and small snap traps (Museum Specials), all of which were placed on the ground and similarly baited. Traplines were located along the margins of forest clearings, along narrow trails through the forest, and along the banks of Quebrada La Hacienda. Altogether, these faunal-sampling efforts amounted to 1163 trap-nights, of which 603 were in February and 560 in June. Handleyomys intectus was the most commonly trapped species at La Suiza on both visits, accounting for 50% of total captures (table 8). All trapped animals were recovered at dawn and appeared to have been captured during the previous night.

Departamento Risaralda, Municipio Pueblo Rico, Vereda Siató, 1520–1620 m: The steep-sided western foothills of the Cordillera Occidental near Pueblo Rico (fig. 12: locality 13) have been largely cleared for agriculture and animal husbandry, but fragments of cloud-forest vegetation (“bosque muy húmedo Montano Bajo” of Espinal and Montenegro, 1963) persist along the margins of local streams. Quantitative climatic data are unavailable, but local precipitation is bimodally distributed with maxima in April–June and October–November and minima in January and July. The local fauna of small mammals was sampled in the valley of the Río Siató (several kilometers SE Pueblo Rico), where MGL and 33 undergraduate students from the Universidad Nacional set medium-size Sherman live traps, National wire live traps (ca. 145 ;ts 145 ;ts 410 mm), and Museum Specials along the forested banks of the Quebrada La Cristalina for eight nights (20–29 September) in the late-dry/early-wet season of 1991. All traps were set on the ground. A total of 21 captures were recorded in 405 trap-nights, including 4 specimens of Handleyomys fuscatus, 10 Akodon affinis (vouchers: ICN 12765–12767), 4 Oryzomys albigularis (ICN 12715, 12716, 12719), 2 Melanomys caliginosus (ICN 12717), and 1 Oligoryzomys fulvescens (ICN 12718). All trapped animals were encountered at dawn and were assumed to have been captured during the previous night.

Departamento Risaralda, Municipio Santuario, Vereda El Campamento, 2400–2500 m: This locality on the eastern slope of the Cordillera Occidental (fig. 12: locality 14) is characterized by steep terrain with patches of disturbed (selectively logged) primary cloud forest (“bosque muy húmedo Subtropical” of Espinal and Montenegro, 1963), agricultural fields, and pastures (fig. 2). The local climate is presumably similar to that described by Sánchez-Páez et al. (1991) from the nearby Parque Nacional Natural Tatamá (see Los Planes, below). ICN researchers Pedro Sánchez Palomino and Marcela Morales trapped small mammals at El Campamento on six nights (19–24 and 26 November) during the rainy season of 1991. All traps (medium-size Shermans and Museum Specials) were set on the ground inside the forest fragments. Seventy traps were set each night, for a total of 420 trap-nights. Handleyomys fuscatus was the most abundant of seven species taken at this locality, accounting for about 47% of all recorded captures. In order of decreasing capture frequency, the other taxa trapped at El Campamento were Oryzomys albigularis (vouchers: ICN 12001–12023, 12616, 12617), Microryzomys minutus (ICN 12720–12722), Akodon affinis (ICN 12614, 12615), Reithrodontomys mexicanus (ICN 12723), Thomasomys cinereiventer (ICN 12623), and T. aureus (unvouchered).

Departamento Risaralda, Municipio Santuario, Vereda Los Planes, 2530 m: The eastern slopes of the Cordillera Occidental from the municipal capital of Santuario (ca. 1500 m) to the vicinity of Los Planes (fig. 12: locality 15) have been almost completely deforested for agriculture and animal husbandry, with only fragments of natural vegetation remaining along watercourses. Above 2500 m, however, is an extensive landscape of primary cloud forest (“bosque muy húmedo Montano Bajo” of Espinal and Montenegro, 1963) that is protected by the Parque Nacional Natural Tatamá. The ICN base camp at Los Planes was an abandoned shelter about an hour and a half by mule from finca Las Delicias (ca. 2100 m) via a dirt road that parallels the Río San Rafael. Collections were made just inside the park boundary, about 500 m above the south bank of the river, in a zone of transition between the denuded lower slopes and the primary forest. The eastern slopes of Tatamá are said to have a median annual temperature of about 17°C and an annual precipitation of 2000–2800 mm; precipitation is bimodal, with maxima in May and October–November and minima in February and July (Sánchez-Páez et al., 1991).

Small mammals were trapped at Los Planes on eight consecutive nights (27 October to 4 November) in the rainy season of 1991 by an ICN team consisting of Pedro Sánchez Palomino, MGL, and Yaneth Muñoz-Saba. All of the traps (medium-size Shermans and Museum Specials) were set on the ground in low, shrubby secondary vegetation that was said to be about five years old by local inhabitants; primary forest, however, occurs in close proximity, and some traps were set along the forest edge. Handleyomys fuscatus was the third most-abundant species taken in 460 trap-nights at Los Planes, where it comprised about 22% of all recorded captures (table 9). All trapped animals were recovered at dawn and were assumed to have been captured during the previous night.

Summary

Available habitat descriptions and trapping records indicate that species of Handleyomys occur in both primary cloud forests and adjacent anthropogenic vegetation (secondary growth). Although climatic data are rarely available from the exact sites where specimens have been collected, plotting collection localities on published ecological maps suggests that Handleyomys is most frequently encountered in relatively cool and very humid environments (12–18°C mean ambient temperature; >2000 mm annual precipitation), corresponding to the vegetation zone that Espinal and Montenegro (1963) designate as “bosque muy húmedo Montano Bajo” (table 10). In the classification of tropical montane forest vegetation suggested by Grubb (1977), the natural habitat at most Handleyomys collection localities would be called Lower Montane Rain Forest, but specimens have also been taken in habitats corresponding to Upper Montane Rain Forest in his system. In Chapman's (1917: 84–93) classification of Colombian life zones, Handleyomys is a member of the humid-Subtropical and humid-Temperate faunas.

All of the other sympatric small mammals that have been collected with Handleyomys are likewise known primarily from wet forests. Whereas some are commonly encountered in foothill landscapes (e.g., Heteromys australis, Melanomys caliginosus, Sigmodontomys alfari), others are typically found at middle elevations (e.g., Oryzomys albigularis), and a few are genuinely upper-montane taxa (e.g., Caenolestes fuliginosus, Thomasomys spp.). Conspicuously absent from all known sites where Handleyomys have been collected are any small mammals primarily associated with lowland forests (e.g., Proechimys spp.) or with semiarid habitats (e.g., Sigmodon hispidus, Zygodontomys brevicauda).

Available trapping records suggest that Handleyomys is predominantly terrestrial and nocturnal, but these data are less than conclusive because traps were not simultaneously set in trees, and because trapped animals encountered at dawn might have been caught before dark the previous day. However, terrestrial habits are consistent with inferences based on morphology: the long-narrow hindfoot of Handleyomys (with short, nonopposable outer digits; fig. 3) resembles the condition seen in many other murids with better-documented terrestrial behavior (e.g., Amazonian species of Oryzomys; see Malcolm [1991], Patton et al. [2000], and Voss et al. [2001] for relevant trapping data). By contrast, semiarboreal and arboreal muroids typically have shorter, broader hindfeet with longer outer digits, of which dV is often semiopposable (Voss et al., 2001: fig. 53B, C). The relatively short, naked-appearing tail of Handleyomys is another indication of terrestrial locomotion because arboreal murids usually have relatively longer tails with longer, coarser hairs that sometimes form a terminal tuft (Voss et al., 2001: fig. 61). Lastly, the dull, drab pelage of Handleyomys is similar to that of many other nocturnal montane murids and contrasts with the more richly pigmented fur of some taxa that are known to be active in the daytime (e.g., Scotinomys, Melanomys; Hooper and Carleton [1976], Gardner [1983]).

Other aspects of the natural history of Handleyomys are obscure. No noteworthy dietary information is available; only a few ectoparasite records have been published (Brennan, 1968); nesting sites, litter sizes, and gestation lengths are unknown; and data from year-long trapping programs are not available to assess reproductive seasonality. Obviously, the absence of concrete information about so many biological details provides considerable scope for productive future fieldwork on these animals.