General Protocols

We conducted quantitative analyses to characterize external and cranial morphology and examine nongeographic, geographic, and interspecific variation in Venezuelan Heteromys. Statistical analyses were performed with Minitab (2003; version 14.1). To test for any departure from normality in our quantitative data, we conducted Ryan-Joiner tests on the relevant residuals (see below). To correct for multiple unplanned comparisons, we determined significance within families of tests (e.g., within a series of t-tests, each comparing the mean between two samples for one of several variables) using Holm's (1979) modification of the Bonferroni procedure. In all other cases, we compared probabilities to an α  =  0.05. Past studies indicated that sexual dimorphism was insubstantial in the few cases where it was detected in species of Heteromys (Engstrom et al., 1987; Best, 1993). Nevertheless, we test for sexual dimorphism (see Nongeographic and Interspecific Variation); based on the results (see below), males and females were pooled in all other analyses. Species assignments for quantitative analyses were based on our qualitative morphological examinations.

In addition to univariate descriptive and inferential statistics (see below), we conducted principal components analysis (PCA) to examine geographic and interspecific variation from a multivariate perspective. Each PCA was performed on the covariance matrix of log-transformed cranial measurements following Jolicoeur (1963); natural logarithms (log_e or ln) were used here. PCA is a multivariate-ordination technique that extracts axes of maximum dispersion among observations (here, specimens) without regard to the group (e.g., geographic sample) to which they belong. Each resultant axis (principal component) is characterized by an eigenvector with coefficients that indicate the relationship of each variable (here, cranial measurements) to that multivariate axis. We interpreted the principal components by examination of both coefficients (elements) of the unit eigenvector and of loadings, which are Pearson product–moment correlation coefficients between specimen scores on each axis and the log-transformed variables. We only interpreted principal components with clearly distinct eigenvalues (by inspection).

Geographic and Interspecific Variation (Primary Geographic Samples)

We conducted a series of quantitative analyses using eight primary geographic samples of adult Heteromys in age class 4 (the most abundant adult age class) from selected sites across northern Venezuela (fig. 1; appendices 1 and 2). The six samples of H. anomalus spanned northern Venezuela (Tukuko, Estado Zulia, locality 97 of Anderson, 2003b; Yacambú, Estado Lara, locality 58 of Anderson, 2003b; Naiguatá, Estado Vargas, locality 120; Neverí, Estado Sucre, locality 111; Latal, Estado Sucre, locality 113; and Manacal, Estado Sucre, locality 100). We also included two highland samples from the Cordillera de la Costa representing the putative new highland species (Rancho Grande, Estado Aragua, locality 137; and Humboldt, Estado Miranda/Estado Vargas, locality 156). These geographic samples were chosen for their relatively large sample sizes and geographic dispersion. Each contains specimens from a single latitude–longitude combination (i.e., specimens with different textual locality descriptions were pooled if they held the same geographic coordinates; appendices 1 and 2; see also Anderson, 2003b). Although specimens of both H. anomalus and the new species are known from Naiguatá (localities 120 and 160), only individuals matching the morphological diagnosis of H. anomalus are included here in this primary geographic sample (locality 120).

Using specimens from these primary geographic samples, we conducted both univariate and multivariate analyses. First, we calculated descriptive statistics for specimens in age class 4 from each of the eight primary geographic samples, as well as for the two putative species (based on the same individuals but grouped by species rather than by geographic sample). We calculated these statistics for external measurements, mass, cranial measurements, and one derived variable (head-and-body length; table 1, appendix 3). Using these same individuals, we then conducted a series of two-tailed t-tests comparing means of the two species for the same measurements and derived variable (these constituted a family for correction for multiple unplanned comparisons). For each variable, we used residuals from these two samples (i.e., after taking species into account) for tests of normality, again correcting for multiple unplanned comparisons. In addition, to test for proportional differences between the species, we examined three ratios of external and cranial variables (tail length/head-and-body length; IOC/ONL; and IW/ONL). Because ratios seldom are distributed normally (Sokal and Rohlf, 1995), we conducted nonparametric tests here (two-tailed Mann-Whitney U-test). Finally, to examine the multivariate structure of the data and determine how much of the variation among specimens of the same age class (age class 4) corresponds to geographic and interspecific variation, the 11 cranial measurements for specimens from these samples were submitted to PCA.

Nongeographic and Interspecific Variation (Site of Sympatry)

To assess nongeographic variation (including tests for sexual dimorphism) and examine interspecific differences between the species further, we took advantage of one site of sympatry where large samples of both species are available. For Los Canales de Naiguatá (localities 120 and 160), many specimens of each species exist in museum collections; these specimens correspond to all age classes, but age classes 1–4 are especially well represented. Hence, this site provides an opportunity to test for sexual dimorphism (and age-related variation) and examine interspecific variation (across age classes) without the complication of geographic variation in either species. To assess and characterize any differences between the sexes in each species, we performed a two-way General Linear Model (GLM; unbalanced ANOVA) for each cranial measurement, with age class and sex as the two fixed factors (and testing for an interaction term of age class by sex). Here, we used only age classes 1–4, because very few specimens were available in age classes 5 and 6. These GLMs for each species test for differences among the age classes (controlling for sex), differences between the sexes (controlling for age class), and any possible interactions between age class and sex. This series of GLMs for age class and sex made up a family for correction for multiple unplanned comparisons. Similar to the analyses for the principal geographic samples, we used residuals from these samples (after taking into account both age class and sex) for each variable to test for departures from normality (again correcting for multiple unplanned comparisons).

We assessed interspecific differences at Los Canales de Naiguatá in several ways. First, we calculated descriptive statistics of each species for specimens from this site in each of the age classes 1–4, respectively. We calculated these statistics for all 11 cranial measurements (appendix 4). Second, we performed a second series of two-way GLMs for each cranial measurement, with age class and species as fixed factors, and testing for an interaction term of age class by species. As in the earlier GLMs, we used only age classes 1–4. These analyses test for differences among the age classes (controlling for species), differences between the species (controlling for age class), and any possible interactions between age class and species. This series of GLMs for age class and species formed a family for correction for multiple unplanned comparisons. As above, residuals from these samples (here, after taking into account both age class and species) for each variable were used to test for departures from normality (once again correcting for multiple unplanned comparisons). Third, we performed a PCA using cranial measurements for all available specimens from Los Canales de Naiguatá with intact skulls (age classes 1–6). Because of high levels of variation in interparietal width and interparietal length, those measurements were excluded from this analysis, which was conducted with the remaining nine cranial measurements.

Geographic Variation in the Highland Species

We also performed analyses aimed at assessing geographic variation within the highland species. First, we conducted a PCA on cranial measurements of all specimens of the putative new species in age class 4 with intact skulls (i.e., from all known localities of the species; appendix 1). As above, interparietal width and interparietal length were excluded from this analysis, which was conducted with the remaining nine cranial measurements. We also conducted two-tailed t-tests between the primary geographic samples of Rancho Grande and Humboldt (see above) to detect any difference in mean for each external and cranial measurement (and the derived variable head-and-body length) for specimens in age class 4 (these composed a family for correction for multiple unplanned comparisons). As before, we conducted a Mann-Whitney U-test to assess differences between the populations for the tail length/head-and-body length ratio.

Holotype

USNM 517536, adult male; skin and skull in excellent condition (fig. 3). Collected on 20 March 1960 from Venezuela: Aragua: Rancho Grande, near Biological Station, 13 km NW Maracay [10°21′N, 67°40′W] at 3576 ft [1090 m] elevation by Charles O. Handley, Jr., and Daniel I. Rhymer; original number COH, Jr. 7778.

Paratypes

We designate as paratypes the following 31 specimens (adults in age class 4; skins and skulls in good condition) from the type locality, and housed in eight museum collections (appendix 1): EBRG 190, 4317; KU 120294, 120295; MCNUSB I-423, 424, 581, 1357; MIZA 20-024, 20-028, 20-049, 20-130; UF 13709; UMMZ 110768, 110771; USNM 517533, 517534, 517538–517541, 517543, 517545, 517546, 517558, 517563, 517564, 517566, 517567, 517569; UWZM S.31418.

Etymology

The adjective catopterius derives from the Greek katoptêrios (from kata, “down” and opter, “eyewitness, scout”; Brown, 1956: 279, 693). Katoptêrios was applied to a “height that commands a view” (Liddell and Scott, 1961) in classical texts (e.g., by the geographer Strabo) and is used here as an adjective to mean “commanding a view” or “having a spectacular or wide/panoramic view” (M. Knight, in litt.). We apply this name to the new species in reference to its presence on the rugged and steep-sided Cordillera de la Costa along the littoral of northern Venezuela, from which magnificent views derive, especially of the Caribbean coast. Suitable common names include Overlook Spiny Pocket Mouse (English); and Ratón de Abazones del Mirador, Ratón Bolsero del Mirador, and Ratón Mochilero del Mirador (Spanish).

Distribution

Known only from the Cordillera de la Costa of north-central and northeastern Venezuela (figs. 1, 4Figure 4 (Continued), 5). The species has been documented from all four major subranges of the Cordillera de la Costa: the main Aragua-Carabobo chain and the El Ávila massif (which together form the Serranía del Litoral), the Serranía del Interior, and the Macizo Oriental. The known elevational range of the species is 350–2425 m, but the vast majority of known records of the species correspond to highland sites above ca. 700 m. See also Sympatry and Zones of Contact with Heteromys anomalus, below.

Figure 4

Maps showing collection localities for Heteromys in central and eastern Venezuela and in Trinidad and Tobago. Locality numbers correspond to data given in the Gazetteer (appendix 1). Heteromys anomalus continues its distribution to the west (not shown) in northwestern Venezuela, northern Colombia, and the Río Magdalena Valley in central Colombia (see Anderson, 2003b). Only georeferenced localities are plotted here with symbols. Gray shading denotes regions higher than 1000 m in elevation.

Figure 4 (Continued)

Figure 5

Landscape on the southern slopes of the Serranía del Litoral of the Cordillera de la Costa in Estado Carabobo and Estado Aragua. Here, Heteromys catopterius inhabits areas above ca. 1000 m (principally cloud forests), and H. anomalus occurs in other forested areas at lower elevations. Rancho Grande (locality 137), the type locality of H. catopterius, lies near the western peak. Photographed by Eliécer E. Gutiérrez in January 2008, from the northern outskirts of the city of Maracay.

Figure 6

Dorsal external morphology of adult male specimens in age class 4 of two species of Heteromys from the Rancho Grande/El Limón region in Estado Aragua, Venezuela, illustrating interspecific variation based on samples in close geographic proximity: Heteromys catopterius (top), Rancho Grande, near Biological Station, 13 km NW Maracay, 3576 ft [1090 m] (USNM 517536; holotype of Heteromys catopterius; locality 137) and H. anomalus (bottom), Rancho Grande [region], El Limón, 4 km NW Maracay, 524 m (USNM 517544; locality 50).

Diagnosis

A species of spiny pocket mouse with adults showing the following combination of characters (figs. 3, 6, 7, 8, 9): p4 (lower permanent premolar) with three lophids; P4 (permanent upper premolar) with straight, moderately long fold in anterior margin of posterior loph; tubercle or swelling at posteroventral border of infraorbital foramen absent; mesopterygoid fossa formed by long, thin hamular processes of pterygoids; optic foramen especially small, with posterior margin formed by strong bar of bone; parietomastoid suture parallel to parietal crest throughout its length; rostrum long and wide (only slightly to moderately tapered anteriorly), without anterodorsal flare; anterior portion of premaxillary convex (inflated), forming a smooth (not stepped) lateral border of rostrum; interorbital constriction wide; braincase wide and moderately inflated; interparietal wide; skull average for genus (ONL 34.74–38.66 mm in adult specimens of age class 4; table 1, appendix 3); dorsal coloration dark slate gray or black and only moderately grizzled with thin ochraceous hairs intermixed among spines; distinctive patch of dark coloration present on dorsal and external surfaces of forearms, continuous with dark coloration of flanks; ears dark gray to black and medium in size; orange band on flanks absent; plantar surface of hind feet naked.

Figure 7

Dorsal views of the crania of adult male specimens in age class 4 of two species of Heteromys from the Rancho Grande/El Limón region in Estado Aragua, Venezuela, illustrating interspecific variation based on samples in close geographic proximity: Heteromys anomalus, Rancho Grande [region], El Limón, 4 km NW Maracay, 524 m (left, USNM 517544; locality 50) and H. catopterius, Rancho Grande, near Biological Station, 13 km NW Maracay, 3576 ft [1090 m] (right, USNM 517536; holotype of Heteromys catopterius; locality 137).

Figure 8

Dorsal views of the crania of specimens of Heteromys catopterius from the Rancho Grande region (Estado Aragua) in several age classes, showing growth-related differences in size and shape. All specimens derive from localities 131 and 137, only ca. 2 km distant from each other at elevations of 3576–4900 ft [1090–1494 m] (USNM 517550, age class 1; USNM 517549, age class 2; USNM 517536, age class 4; USNM 562956, age class 5). Such marked differences among age classes are typical of species of the genus (see Anderson and Timm, 2006: 14) and complicate identification and characterization of nongeographic, geographic, and interspecific variation. See appendix 1 for full provenience.

Figure 9

Dorsal views of the crania of adult female specimens in age class 4 of Heteromys anomalus (two crania on left) and H. catopterius (two crania on right) from Venezuela, showing interspecific and geographic variation: H. anomalus from 10 km WNW Valera, Isnotú, 930 m (Estado Trujillo; locality 82 in Anderson, 2003b) and 5 km S and 25 km E Carúpano, Manacal, 185 m (Estado Sucre; locality 100) and H. catopterius from Rancho Grande, near Biological Station, 13 km NW Maracay, 3576 ft [1090 m] (Estado Aragua; locality 137) and Hotel Humboldt, 9.4 km N Caracas, 2181 m (Estado Miranda/Estado Vargas; locality 156). See appendices 1 and 2 for full provenience.

Figure 10

Plot of specimen scores on the first two axes of the principal components analysis of 11 cranial measurements of adult specimens of Heteromys in age class 4 from the eight primary geographic samples across northern Venezuela. Specimens of H. anomalus are marked by circles, and individuals of H. catopterius are denoted by triangles. PC1 is a size axis (with larger specimens appearing toward the left side of the plot), and PC2 represents a shape difference (see text and table 2).

Table 1

Descriptive Statistics and Results of Comparisons of Adults in Age Class 4 of Heteromys anomalus and H. catopterius in Venezuela for External Variables, Mass, and Cranial Measurements

Mean ± standard error, sample size, and observed range (minimum–maximum) are given for each species. Probabilities associated with differences in mean between the two species in two-tailed t-tests (or in median for the Mann-Whitney U-test, for the tail length/head-and-body length ratio) are indicated as follows: *  =  P≤ 0.05; **  =  P ≤ 0.01; ***  =  P ≤ 0.001; n.s.  =  P > 0.05. To correct for multiple unplanned comparisons, we determined significance within this family of tests using Holm's (1979) modification of the Bonferroni procedure; here, this altered the statistical conclusion only for hind foot length, which was no longer significant after correction for multiple comparisons. Note that for all significant comparisons, H. catopterius averages larger than H. anomalus. Linear measurements are in millimeters, mass is in grams, and the ratio is multiplied by 100 to yield a percent.

Description

Dorsal pelage (fig. 6) dark slate gray or black (sharply contrasting with soft, white pelage of venter) and moderately grizzled with thin ochraceous hairs intermixed among spines, typically moderately spiny, but soft in some populations; distinctive patch of dark coloration present on dorsal and external surfaces of forearms, continuous with dark coloration of flanks; ventral and internal surfaces of forearms white; orange band on flanks absent; ears dark gray to blackish and medium in size; tail strongly bicolored, and slightly to moderately longer than head-and-body length (table 1, appendix 3); hind feet average for genus (table 1, appendix 3), with plantar surface naked; skull (figs. 3, 7, 8, 9) average for genus (ONL 34.74–38.66 mm in adult specimens of age class 4; table 1, appendix 3); rostrum long and wide (only slightly to moderately tapered anteriorly), without anterodorsal flare; anterior portion of premaxillary convex (inflated), forming a smooth (not stepped) lateral border of rostrum; interorbital region moderately inflated; interorbital constriction wide; braincase wide and moderately inflated; parietal and temporal crests weakly to moderately developed; interparietal variable (moderately wide to very wide), sometimes with distinctive lateral and anterior points; incisive foramina usually short, wide, and anteriorly tapering; tubercle or swelling at posteroventral border of infraorbital foramen absent; mesopterygoid fossa formed by long, thin hamular processes of pterygoids; parapterygoid fossa deep; postalar fissure large and distinctly rounded anteriorly; optic foramen especially small, with posterior margin formed by strong bar of bone; parietomastoid suture parallel to parietal crest throughout its length; anterodorsal lobe of periotic capsule variable in size, sometimes causing notable undulation in posterior margin of squamosal (see Geographic Variation; see schematic drawing in Anderson, 1999: 619); jugular foramen (anterolateral to occipital condyle) medium in size; dental formula: (incisors 1/1, canines 0/0, premolars 1/1; molars 3/3) × 2  =  total 20; molars variable in size (notably large and robust in some populations but typical for genus in others); lateral borders of lophs of molars smooth, not pointed; lophs of M2 subequal in width; P4 (permanent upper premolar) with straight, moderately long fold in anterior margin of posterior loph, and with smoothly curved posterior border; p4 (lower permanent premolar) with 3 lophids; m3 not reduced, approximately same width as p4.

Measurements of the Holotype (mm)

total length 321, tail length 186, hind foot length 37, ear length 21, ONL 36.98, ZB 17.52, RL 16.40, NL 15.18, IOC 9.27, SB 16.14, MTR 5.88, IW 10.47, IL 5.08, PB 13.48, SD 11.78. Mass not available.

Karyology

Schmid et al. (1992) reported karyotypic information for three individuals of the species now known as Heteromys catopterius from the northern slope of the Cordillera de la Costa along the Maracay–Ocumare de la Costa highway at 940 m in Estado Aragua (locality 130), very near the type locality of Rancho Grande (MIZA 20-391, 20-392, and 20-394; appendix 1). These individuals possessed a standard karyotype with a diploid number of 2n  =  60 and a fundamental number of FN  =  72. Note, however, that those authors included the four arms of the sex chromosomes in the count for the fundamental number; under standard counting protocols, this karyotype would have a FN  =  68 (M. Aguilera, in litt.). Various banding and staining techniques were conducted as well; the Y chromosome shows exceptional quinacrine-positive heterochromatin, indicative of a preponderance of AT-rich sequences (Schmid et al., 1992). Although the standard karyotype of H. catopterius is the same as that of H. anomalus (see account of that species below), some species of the genus with identical standard karyotypes have been shown to be very distinct based on DNA sequences and morphology (i.e., H. nubicolens and nearby populations of H. desmarestianus; Anderson et al., 2006; Anderson and Timm, 2006; Anderson and Jansa, 2007); future banding studies are necessary to compare the cytogenetics of H. catopterius and H. anomalus in more detail.

Comparisons

Externally, the dark slate-gray dorsal pelage (only moderately grizzled with ochraceous hairs; fig. 6) of Heteromys catopterius differentiates it from most currently recognized species of the genus. Three types of dorsal pelage are present in the genus (pale brown, strongly grizzled; dark brown, moderately grizzled; and dark slate gray, nearly uniform to moderately grizzled). In contrast to the new species, Heteromys gaumeri, H. oasicus, and most populations of H. anomalus have pale brown dorsal pelage that is strongly grizzled with thin ochraceous hairs intermixed among the spines (Anderson, 2003b). In addition, specimens of H. gaumeri possess an orange band on the flanks that is absent in H. catopterius. The dwarf species H. oasicus is much smaller than H. catopterius (total length 227–250 mm for adults in age class 4; Anderson 2003b). Heteromys gaumeri is restricted geographically to Belize, Guatemala, and the Península de Yucatán in Mexico (Williams et al., 1993). Heteromys oasicus inhabits only the Península de Paraguaná in northwestern Venezuela, to the northwest of known records of H. catopterius (Anderson, 2003b). Heteromys anomalus inhabits much of northern South America and comes into geographic contact with H. catopterius in several areas; because the pelage of H. anomalus is atypically dark in a few portions of its range (i.e., similar to that of H. catopterius), detailed comparisons are necessary to ensure proper identification (see below). The pelage of H. catopterius also differs from the dark-brown pelage of H. nubicolens and members of the H. desmarestianus species complex, which occur in Mexico and Central America far from the known distribution of H. catopterius (Anderson and Timm, 2006; H. d. crassirostris also has a marginal distribution in South America in extreme northwestern Colombia; Anderson, 1999).

Heteromys catopterius shares dark slate-gray pelage with several other species of the genus (H. australis, H. nelsoni, H. oresterus, and H. teleus). Both H. nelsoni and H. oresterus display extremely soft dorsal pelage, in contrast to the pelage of H. catopterius, which is spiny in most populations (all known populations except those from the Macizo Oriental). They are distributed in restricted highland regions of extreme southeastern Mexico and southwestern Guatemala (H. nelsoni) and the Cordillera de Talamanca in Costa Rica (H. oresterus; Rogers and Rogers, 1992; Williams et al., 1993; Reid, 1997), respectively. Heteromys teleus and H. australis are extremely similar to H. catopterius externally, although their pelage tends to be even darker than that of H. catopterius. Like the new species, they also inhabit South America but are not known to come into geographic contact with it. Heteromys teleus is endemic to western Ecuador, and the distribution of H. australis extends from northwestern Ecuador and extreme eastern Panama through much of Colombia to western Venezuela; Anderson, 1999, 2003b; Anderson and Jarrín-V., 2002). Cranial characters are necessary to distinguish them from H. catopterius.

Cranial characters serve to separate Heteromys catopterius from all other species of the genus (figs. 3, 7, 8, 9). Heteromys catopterius differs from H. gaumeri by the narrow postalar fissure and deeply concave (collapsed) anterior termination of the premaxillary of H. gaumeri. Additionally, the form of the premaxillary in H. gaumeri creates a distinct step in the lateral border of the rostrum in dorsal view, a character lacking in the new species (Anderson et al., 2006). Heteromys catopterius contrasts with the Central American and Mexican species H. desmarestianus, H. nelsoni, H. nubicolens, and H. oresterus by its especially small optic foramen and straight, only moderately long fold in the anterior margin of the posterior loph of P4. Consistent, but more subtle, differences exist in comparison with other South American species. Heteromys oasicus is much smaller than the new species and has a much less-developed rostrum (Anderson, 2003b). In comparison with H. catopterius, the braincase of H. australis is even more inflated and the rostrum less developed. Heteromys teleus is very similar to the new species cranially but has wider (bowed) zygomatic arches and an uninflated braincase. These three species inhabit areas of South America to the west of the known distribution of H. catopterius. Cranial differences also separate H. catopterius from H. anomalus (see detailed comparisons below).

External differences exist between Heteromys catopterius and H. anomalus. Externally, most populations of H. anomalus have a pale brown dorsal pelage that is strongly grizzled with thin ochraceous hairs, in stark contrast to the dark slate-gray dorsal pelage (only moderately grizzled with ochraceous hairs) of H. catopterius. In addition, H. catopterius possesses a distinctive patch of gray or black on the forearm; the homologous patch is present in H. anomalus as well, but it is typically paler and brownish. However, some populations of H. anomalus from mesic areas (e.g., some regions of Estados Lara, Monagas, Sucre, and Zulia in Venezuela, and the island of Tobago in Trinidad and Tobago) exhibit a much darker dorsal pelage, very similar to that of H. catopterius in general aspect and in the coloration of the patch on the forearm. Notably, however, H. anomalus has distinctively rounded ears throughout its range, in contrast to H. catopterius and all other species of the genus (which have moderately pointed ears). Additionally, H. catopterius has much longer, darker mystacial vibrissae than H. anomalus (fig. 6).

Cranial differences in size and shape also distinguish Heteromys catopterius from H. anomalus. Overall, the skull of the new species is larger than that of H. anomalus and relatively wider (less elongated). Regarding proportional differences, H. catopterius generally shows a more inflated braincase, a relatively wider interorbital constriction, and a relatively wider interparietal. In addition, it tends to possess wider (often short and anteriorly tapering) incisive foramina (the incisive foramina of H. anomalus vary geographically but tend to be longer, thinner, and more parallel sided than those of H. catopterius).

The analyses of nongeographic variation in cranial measurements at Los Canales de Naiguatá documented strong age-related differences but no sexual dimorphism. The GLMs of individuals of H. anomalus in age classes 1–4 showed highly significant age-related differences for all variables but maxillary toothrow length, interparietal width, and interparietal length (P < 0.002 for significant comparisions; P  =  0.152 for IW; P  =  0.312 for IL; P  =  0.003 for MTR, not significant after correction for multiple comparisons). Sex was not significant for any variable (P  =  0.023 for ZB; P  =  0.045 for IOC, neither significant after correction for multiple comparisons; P  =  0.080–0.779 for other nonsignificant comparisons). No interaction term was significant (P  =  0.064–0.451). Furthermore, no variable showed a departure from normaility (P > 0.058). Results for H. catopterius were similar. For this species, age class was significant for all variables but interparietal width, interparietal length, and parietal breadth (P < 0.002 for significant comparisions; P  =  0.107 for IL; P  =  0.007 for IW; P  =  0.006 for PB, neither of the latter two significant after correction for multiple comparisons). Sex was never significant (P  =  0.486–0.992), nor was any interaction term (P  =  0.042 for PB, not significant after correction for multiple comparisons; P  =  0.376–0.996 for other nonsignificant comparisons). As in the tests for H. anomalus, no variable showed a departure from normaility (P  =  0.040 for SB; P  =  0.030 for MTR, neither significant after correction for multiple comparisons; P > 0.079 for other nonsignificant comparisons).

Quantitative comparisons documented differences in size and shape between Heteromys catopterius and H. anomalus. Even though the two species overlapped for mass and all external and cranial variables (for specimens in age class 4 from the primary geographic samples), means were significantly different for some external variables, mass, and most cranial variables (but no variable showed a departure from normality; P  =  0.040 for mass; P  =  0.030 for IOC, neither significant after correction for multiple comparisons; P > 0.100 for other nonsignificant comparisons). For external variables and mass, only total length, tail length, and mass were significant (P < 0.005 for significant comparisons; P  =  0.147 for head-and-body length; P  =  0.171 for ear length; P  =  0.025 for hind foot length, not significant after correction for multiple comparisons; table 1). The corresponding test for a difference in median for the tail length/head-and-body length ratio was not significant (P  =  0.090). All cranial measurements except interparietal length showed significant differences in mean between the two species (P < 0.001 for significant comparisons; P  =  0.200 for IL; table 1). For all significant comparisons, H. catopterius averaged larger than H. anomalus (see also figs. 7, 9). Regarding differences in cranial shape, H. catopterius showed a relatively wider interorbital constriction and relatively wider interparietal (IOC/ONL ratio, P < 0.001; IW/ONL ratio, P  =  0.007).

Quantitative analyses document similar differences between Heteromys catopterius and H. anomalus at a site of sympatry. At Los Canales de Naiguatá, both species and age class were important factors (but no variable showed a departure from normality; P  =  0.041 for MTR, not significant after correction for multiple comparisons; P > 0.086 for other nonsignificant comparisons). These GLMs for age classes 1–4 documented that the species were significantly different for all cranial measurements but interparietal width and interparietal length (P < 0.001 for significant comparisions; P  =  0.210 for IL; P  =  0.035 for IW, not significant after correction for multiple comparisons). Age class was significant for all measurements but IW (P < 0.004 for significant comparisons; P  =  0.005 for IW, not significant after correction for multiple comparisons), but no interaction term was significant (P  =  0.393–0.991). Heteromys catopterius averaged larger than H. anomalus for all measurements that were significantly different.

Multivariate analyses also supported the morphological distinctiveness of Heteromys catopterius from H. anomalus. In a plot of scores on the first two principal components of the PCA undertaken on specimens in age class 4 from the primary geographic samples, substantial but incomplete overlap occurred between specimens of the two species (fig. 10). The coefficients and loadings of PC1 were all of the same sign, indicating that it represented a size axis (with larger specimens having lower scores, because the signs of the coefficients were all negative; table 2). In contrast, PC2 represented a shape axis (because most coefficients were positive, but that of interparietal length was negative; table 2).

Figure 11

Plot of specimen scores on the first two axes of a principal components analysis of nine cranial measurements of specimens of Heteromys anomalus (circles) and H. catopterius (triangles) in all age classes from Los Canales de Naiguatá (localities 120 and 160). PC1 represents a size axis (with larger specimens appearing toward the left side of the plot), and PC2 portrays a difference in shape (see text and table 3).

Table 2

Results of Principal Components Analysis of Adult Specimens in Age Class 4 of Heteromys anomalus and H. catopterius from Primary Geographic Samples across Northern Venezuela (see fig. 10)

Elements of the unit eigenvector and loadings are presented for the first three principal components. Eleven cranial measurements were transformed to their natural logarithms (ln) before computation of the covariance matrix and extraction of components. Eigenvalues (multiplied by 100) and the percentage of variation among specimens explained are given for each axis.

Multivariate differences between the two species were much clearer when considering specimens from only one site. At Los Canales de Naiguatá, only very limited overlap between Heteromys catopterius and H. anomalus existed in a plot of scores on the first two principal components of the PCA undertaken on specimens in all age classes (using all cranial measurements but interparietal width and interparietal length; fig. 11). As above, the coefficients and loadings of PC1 were all of the same sign, indicating that it represented a size axis (again with larger specimens showing lower scores, since the signs of all the coefficients were negative; table 3). Here, PC2 again constituted a shape axis (because some coefficients were positive and others negative; table 3), but the axis was very different than the one for the PCA for the primary geographic samples. At Los Canales de Naiguatá, PC2 represented a contrast between occipitonasal length, rostral length, and nasal length vs. zygomatic breadth, least interorbital constriction, squamosal breadth, maxillary toothrow length, parietal breadth, and skull depth. Specimens with larger values for the former three measurements relative to their values for the latter ones showed larger (less negative) scores on this axis. This analysis indicated strong shape differences between the two species, with H. anomalus more elongated than the relatively wider H. catopterius. Accordingly, a bivariate plot of least interorbital constriction vs. nasal length (fig. 12) showed a clear difference in shape between the two species: with very little overlap, specimens of H. anomalus exhibited a longer nasal length relative to least interorbital constriction (and individuals of H. catopterius showed a wider least interorbital constriction relative to nasal length).

Figure 12

Plot of least interorbital constriction vs. nasal length for specimens of Heteromys in all age classes from Los Canales de Naiguatá (localities 120 and 160). Circles represent individuals of H. anomalus, and triangles denote those of H. catopterius. Note that specimens of H. catopterius have a wider interorbital constriction than those of H. anomalus, relative to the length of their nasals.

Figure 13

Plot of specimen scores on the first two axes of a principal components analysis of nine cranial measurements of adult specimens of H. catopterius in age class 4 from throughout the range of the species. Solid triangles denote specimens from Estado Carabobo and western Estado Aragua; open triangles indicate specimens from eastern Estado Aragua as well as those from Estado Miranda, Estado Vargas, and Distrito Capital; and the asterisk represents the lone specimen from Estado Monagas. PC1 represents a size axis (with larger specimens appearing toward the left side of the plot), and PC2 constitutes a shape difference (see text and table 4).

Table 3

Results of Principal Components Analysis of Interspecific Variation Between Heteromys anomalus and H. catopterius Based on Specimens in All Age Classes from Los Canales de Naiguatá (see fig. 11)

Elements of the unit eigenvector and loadings are presented for the first three principal components. Nine cranial measurements were transformed to their natural logarithms (ln) before computation of the covariance matrix and extraction of components. Eigenvalues (multiplied by 100) and the percentage of variation among specimens explained are given for each axis.

In sum, the quantitative analyses indicate that adult Heteromys catopterius were larger in overall body size than those of H. anomalus and differed in cranial proportions (figs. 7, 9, 10, 11, 12). Almost all measurements averaged larger for H. catopterius, table 1, appendices 3 and 4). Furthermore, H. catopterius showed a relatively wider, less-elongated skull (with relatively wider interorbital constriction and interparietal). These differences in shape were also evident in direct visual examination of skulls.

Geographic Variation

Quantitative analyses indicated notable geographic variation between samples of Heteromys catopterius from western vs. central portions of its range (appendix 3). Means for most variables were significantly different in univariate comparisons of specimens from Rancho Grande (locality 137) with those from Humboldt (locality 156). Externally, all variables but tail length (P  =  0.057) were significant (P < 0.003 for significant comparisons; test for mass not possible). The populations did not differ in median for the tail length/head-and-body length ratio (P  =  0.414). Comparisons of cranial measurements indicated that averages were significantly different for all measurements but rostral length and maxillary toothrow length (P  =  0.051 for MTR; P  =  0.025 for RL, not significant after correction for multiple comparisons; P < 0.006 for significant comparisons). For all variables with a significant difference, the sample from Rancho Grande averaged larger than the one from Humboldt.

Multivariate morphometric analysis also showed substantial differences between westernmost populations vs. central and eastern ones. A plot of scores on the first two principal components of the PCA undertaken on all available specimens of Heteromys catopterius in age class 4 (using all cranial measurements but interparietal width and interparietal length) summarized morphometric variation within the new species (fig. 13). Only moderate overlap existed between specimens from westernmost populations (from Estado Carabobo and western portions of Estado Aragua, in the vicinity of Rancho Grande) vs. those from the El Ávila massif (the central portion of the species' distribution; figs. 1, 4Figure 4 (Continued)). The lone available specimen from Estado Monagas in northeastern Venezuela (from the Macizo Oriental) lay within the cloud of specimens from El Ávila. Specimens from geographically intermediate populations in eastern Estado Aragua, western Estado Miranda, western Estado Vargas, and western Distrito Capital (e.g., Alto Ñeo León, Pico Codazzi, and El Laurel; appendix 1) tended to group with the central populations (from the El Ávila massif). These geographically intermediate populations lie to the west of the depression separating the El Ávila massif from the main Aragua-Carabobo chain of the Serranía del Litoral; this depression is marked by the courses of the Quebrada Tacagua and the Río Guaire. Individuals from the Serranía del Interior also fell with this group (Campamento Rafael Rangel; east-central Estado Aragua). The coefficients and loadings of PC1 all possessed the same sign, signifying that it represented a size axis (with larger specimens showing lower scores, because the signs of all the coefficients were negative; table 4). In contrast, PC2 constituted a shape axis (because some coefficients were positive and others negative; table 4). Here (as in the PCA for Los Canales de Naiguatá), PC2 represented a contrast between occipitonasal length, rostral length, and nasal length vs. zygomatic breadth, least interorbital constriction, squamosal breadth, maxillary toothrow length, parietal breadth, and skull depth. Specimens with smaller values for the former three measurements relative to their values for the latter ones showed larger (more positive) scores on this axis.

Figure 14

Habitat of Heteromys catopterius at Rancho Grande (locality 137), Estado Aragua, Venezuela, taken just above the Estación Biológica. Photographed by Eliécer E. Gutiérrez in January 2008.

Table 4

Results of Principal Components Analysis of Geographic Variation in Heteromys catopterius Based on Adult Specimens in Age Class 4 from throughout the Range of the Species (see fig. 13)

Elements of the unit eigenvector and loadings are presented for the first three principal components. Nine cranial measurements were transformed to their natural logarithms (ln) before computation of the covariance matrix and extraction of components. Eigenvalues (multiplied by 100) and the percentage of variation among specimens explained are given for each axis.

Visual morphological observations indicate that subtle differences exist between the westernmost populations of Heteromys catopterius and the central and eastern populations of the species, but geographically intermediate localities appear to show transitional morphologies. Specimens from localities in Estado Carabobo and from the Rancho Grande region of Estado Aragua are generally larger and relatively wider than those from the El Ávila massif and from the Macizo Oriental, which are more elongated in shape (but not as much as H. anomalus). In addition, the rostrum is more developed and parallel sided (not anteriorly tapering) in western populations, in contrast to the somewhat less massive and slightly tapering rostrum typical of central and eastern populations. Two further differences distinguish the Rancho Grande population from that of Humboldt: (1) the molars appear larger and more robust in Rancho Grande than in Humboldt and (2) at Rancho Grande, the anterodorsal lobe of the periotic capsule of the mastoid is large, causing a distinct undulation in the posterior border of the squamosal, whereas this lobe is small at Humboldt, not causing a notable undulation in the posterior border of the squamosal. Despite these noteworthy differences between specimens from the two regions, individuals from sites geographically between Rancho Grande and Humboldt, in eastern Estado Aragua, western Estado Miranda, western Estado Vargas, and western Distrito Capital (e.g., Pico Codazzi, Alto Ñeo León; appendix 1), exhibit intermediate morphologies but generally match the central and eastern populations more closely than they do the more-western populations. Finally, one character distinguishes the populations in the Macizo Oriental from all others: individuals there possess very soft dorsal pelage, in contrast to the spiny pelage of western and central populations.

Natural History

Published studies and field notes provide rich and varied ecological information on Heteromys catopterius. Most of the available information corresponds to two sites where it has been collected in large numbers in the Serranía del Litoral (fig. 1): the vicinity of Rancho Grande in Estado Aragua and the area surrounding Hotel Humboldt on the El Ávila massif along the border between Estado Miranda and Estado Vargas. Further information derives from sites in the Río Morón drainage on the northern slopes of the Cordillera de la Costa in Estado Carabobo (Bisbal-E., 1993). Unfortunately, the extensive morphological, reproductive, developmental, and behavioral information for Heteromys at Los Canales de Naiguatá in Estado Vargas in the Serranía del Litoral presented by Valdez et al. (1985) derives from a pooling of approximately equal numbers of H. catopterius and H. anomalus (appendix 1); future reanalysis of those data could provide fascinating insights into the ecology of the two species at this site of sympatry, but natural history information from that site is not included here.

At Rancho Grande near the crest of the main range of the Serranía del Litoral in Estado Aragua (fig. 14), field notes and specimen tags indicate that the species was captured commonly in various forested habitats but not in grassy and brushy areas. In the region surrounding the Estación Biológica de Rancho Grande (localities 131, 132, 136, and 137, Estado Aragua), Charles O. Handley, Jr. and Daniel I. Rhymer (assisted occasionally by John P. Rood) caught H. catopterius frequently in several habitats in March and April of 1960 (C.O. Handley, Jr. field notes, 1960; specimen tags). These habitats include: (1) an area of secondary vegetation below the Estación Biológica in an abandoned coffee and banana farm (although traps were placed there in areas grown up with weeds, grass, and brush, as well as in “dry” forest, evidently Heteromys were captured only in the forested areas, termed “deciduous forest” on specimen tags); (2) a forested (presumably evergreen) area on the slope behind the Estación Biológica along the “water trail” (see map in Beebe and Crane, 1947: 48; (3) forested areas (including cloud forest) at El Portachuelo, slightly below and west of the Estación Biológica; and (4) cloud forest on Pico Limón. Although traps were placed in grassy areas near the Estación Biológica, Heteromys was not captured there.

The month following completion of work at Rancho Grande by Handley and Rhymer, John P. Rood conducted capture-recapture sampling and laboratory experiments at Rancho Grande that provide extensive information regarding the ecology, behavior, and development of Heteromys catopterius (Rood, 1963; Rood and Test, 1968). Rood collected a few voucher specimens from these trapping grids and adjacent areas (see Rood and Test, 1968); all that we have examined represent H. catopterius (appendix 1). Approximately half derive from cloud forest, but others were captured in a forest near a small stream, a second-growth forest, second growth at the edge of tall grass, and a clump of tall grass near second-growth forest (J.P. Rood field notes, 1960). Although some of the individuals in Rood's studies may have been H. anomalus (which has been captured occasionally at Rancho Grande; see Sympatry and Zones of Contact with Heteromys anomalus), at least the vast majority probably were the species now known as H. catopterius.

Rood sampled two areas using traps arranged in grids (Rood and Test, 1968: 90–92). Area 1 comprised a mosaic of secondary habitats in a former coffee plantation, including grass, Heliconia, bamboo, short trees, and Cecropia, as well as primary cloud forest. In contrast, area 2 was mature cloud forest. There,

The two areas showed differing patterns of capture of Heteromys. In area 1, trap success was highest in cloud forest and Heliconia, although moderate numbers of captures were also recorded in Cecropia and grass and a few in bamboo and secondary forest. In area 2, captures were more evenly distributed across the grid (which was entirely within cloud forest), and the trapping there yielded a higher number of unique individuals (Rood and Test 1968: 93–94).

Rood's fieldwork and observation of individuals in captivity provide some information regarding reproduction and development of Heteromys catopterius. Although trapping occurred only over a few weeks, the observed patterns indicated a possible peak of reproductive activity in April and May (corresponding to the onset of the rainy season; Rood, 1963: 189; see also Rood and Test, 1968: 94–96). Copulation was observed once in captivity (Rood, 1963: 189). Based on a small sample size, litter size varied from 1 to 3, and mass at birth averaged 3.3 g. Major events in the development of H. catopterius (measured in captivity) include: ears opened on day 5, biting (when handled) commenced on day 10, dorsal hair became apparent on day 11 (when the individuals weighed 13.7 g), walking and eating solid food began at 20 days (when individuals had full juvenal pelage, but their eyes were still closed; average weight at the time was 21.4 g), but nursing continued to day 48 (when observation ceased; Rood, 1963: 190).

Heteromys catopterius consumed a wide variety of native and exotic foods. In captivity, it ate corn especially readily but also consumed avocados, tomatoes, bread, dried cereals, rolled oats, cabbage, and cheese. The nut of the stilt palm (Iriartea) was found in live-traps along with Heteromys, presumably brought there in the animal's cheek pouches. This nut was eaten readily by captive animals, as well as several other foods found naturally in the surrounding environment: “spiny palm roots (Bactris), seeds and blades of an unidentified sedge, roots and blades of grasses, and fruits of Passiflora, the candela tree (Gyranthera caribensis Pittier), and several others which could not be identified” (Rood, 1963: 188). In captivity, the mice used their cheek pouches to bring corn to nest (Rood, 1963: 188–189), and while eating, “the rats did not hold the food in their front paws, as do many other rodents, but rested it on the ground supported by the forepaws” (Rood, 1963: 189).

A few individuals were brought into captivity in the Estación Biológica and housed in elaborate experimental conditions, where nesting and digging comprised noteworthy activities. After housing individuals in cages, Rood moved some to a study room measuring 17 × 18 feet,

In that environment, Heteromys catopterius was observed making nests and digging. Individuals

Observations by Rood also offer information regarding behavior and locomotion of the species. Adult males showed aggression toward humans, but neither adult females nor juveniles of either sex did. Adult males exhibited aggression toward females unless the latter were in estrus (Rood, 1963: 187, 191). A clear dominance structure existed, especially in the context of intraspecific competition for nests in captivity: adult males were dominant over adult females and juvenal males; and adult females over juvenal males. Dominance among adult males was determined via brief fights (Rood, 1963: 191). The mice were typically nocturnal (Rood, 1963: 188; Rood and Test, 1968: 96). Information regarding locomotion is contradictory, including the observation that individuals “run with rapid bounds and the tail does not drag” (Rood, 1963: 188) but that locomotion was “remarkably inept or phlegmatic. Their movements, mostly hops, were usually slow.” (Rood and Test, 1968: 98).

For areas near Hotel Humboldt in Estado Miranda and Estado Vargas (locality 156) near the crest of the El Ávila massif of the Serranía del Litoral, field notes provide detailed ecological information for Heteromys catopterius. There, the species inhabits primary cloud forests, especially in association with palms. Merlin D. Tuttle and several other workers sampled this region in August and September of 1965 (see also Handley, 1976). The macrohabitat of the species at that site can be characterized as follows (M.D. Tuttle field notes, 1965):

Furthermore, Tuttle noted that: “Heteromys [were] especially associated with areas of palms … some were in rock cover, and about the same number were in bamboo cover below the palm trees … [many were found] in the forest where palms were dense to at least scattered … (in contrast to O. albigularis [ =  Nephelomys caracolus], which seem to occur wherever there is forest).” Indeed, several species of small nonvolant mammals (but no Heteromys) were captured in trap lines at similar elevations in (1) forest with bamboo understory (but few or no palms); (2) forest with disturbed understory (ferns, grass, moss, etc., but few or no palms); and (3) grassy areas.

Detailed microhabitat information for Heteromys catopterius at Hotel Humboldt is available as well (M.D. Tuttle field notes, 1965). The species was most commonly captured in association with palms, especially at the bases of palm trees with stiltlike root clumps: in a dense growth of palms (2 individuals); in a dense palm forest 15 ft [5 m] tall; on the ground around the palms (in a palm forest about 30 ft [9 m] tall, where there were also other species of trees that reached at least 60 ft [18 m]; (3 individuals); at the base of a palm tree (2 individuals); at the bases of palm trees in the forest; at the base of a palm tree in forest 30 ft [9 m] tall; at the base of a palm where there was a dense growth of young palms and saplings 15 ft [9 m] tall; at the bases of palm trees in the forest (there are large root masses above the ground at the bases of each palm providing good cover; 3 individuals); in the root clump of a palm tree; in the roots of a palm in dense palm forest; at the root clump base of a palm tree in the forest; at the root clump bases of palms (2 individuals); in the root clumps of palms in the forest (2 individuals).

The species was also trapped in areas with bamboo, and these areas often also contained palms (M.D. Tuttle field notes, 1965): in dense bamboo; in dense bamboo in forest; in dead leaves in bamboo about 3 ft [0.9 m] tall in forest; under rocks in grass and bamboo in forest; in a dense mixture of bamboo, vines, and ferns; in a dense mixture of bamboo, vines, and ferns in the forest; in dense bamboo beneath palms; in dead leaves and low bamboo in scattered palms; in a root clump of a palm in a dense mixture of grass and bamboo; in a dense mixture of bamboo, vines, ferns, and other plants about 4 ft [1.2 m] tall; in a dense mixture of bamboo, vines, ferns, etc., about 5 ft [1.5 m] tall in the forest; in a mixture of bamboo, vines, ferns, and young palms in the dense forest; in mixture of bamboo, vines, ferns, and young palms about 6 ft [1.8 m] tall in forest of scattered palms mixed with other trees.

The species was occasionally found associated with rocks or logs, but only rarely in grass or other microhabitats (M.D. Tuttle field notes, 1965): in rocks; in rocks in the forest; under a rock in dense forest; on mossy rocks in the forest (3 individuals); under large mossy rocks in the forest near the stream; under fallen trees (2 individuals); in a hollow tree in dense forest; on the trunk of a leaning tree in dense forest 45 ft [14 m] tall; on a dead log 8 ft [2.4 m] above the stream and spanning a distance of 25 ft [8 m] from one side to the other, on one side there was dense bamboo, and on the other a mixture of dense ferns and young bamboo (the forest above was up to 60 ft [18 m] tall); in dense grass a short distance from the palm forest; in dense grass about 1.5 ft (0.5 m) tall near the stream in a slightly open part of the forest; in dense vines 6 ft (1.8 m) from the stream; at the entrance to a hole under a root ledge.

At Hotel Humboldt, the species only occasionally was recorded to have held items in its cheek pouches (M.D. Tuttle field notes, 1965). Two individuals were taken carrying palm nuts 26 × 22 mm in their pouches. Each had one nut, and one of them also had half a millipede in its pouch with the nut. A third individual had a palm nut 27 × 20 mm in its pouch.

Further information regarding the natural history of Heteromys catopterius derives from the results of surveys conducted in the Río Morón drainage on the northern slopes of the Cordillera de la Costa in Estado Carabobo. Francisco J. Bisbal-E. and colleagues (Sergio Bermúdez, José Daniel P. Bora, Arnaldo Ferrer, Alfredo Lander, Alejo Ospino, and Ramón Rivero) undertook a general inventory of vertebrates in the Río Morón drainage in 1990–1991, including trapping for small nonvolant mammals (Bisbal-E., 1993). Much of the region had been cultivated for many years, but in 1971 an effort began to return the area to a more natural state for the purpose of restoring water supply to small towns and a large industrial complex on the coast. Although not all captured individuals were collected, all the voucher specimens of Heteromys that we have examined from this work represent H. catopterius (from La Justa and Palmichal; appendix 1). Heteromys catopterius was one of the most commonly captured species in both cloud forest and semideciduous forest in this area, and it inhabited some secondary areas in addition to primary forest (Bisbal-E., 1993).

At La Justa (300–400 m; locality 141), Heteromys catopterius and Proechimys guairae were the most commonly captured small, nonvolant mammals. The macrohabitat at La Justa was recovering semideciduous forest. The area had been cultivated for coffee, cocoa, and other crops for many years, but the croplands had been abandoned, leading to the formation of secondary forest before the inventory work commenced. The forest had a “fairly dense but irregularly distributed shrub layer, due to differing light conditions within the forest, and [overall, the forest had] two or three dense strata with an average height of 15–20 m” (Bisbal-E., 1993: 366; translation ours).

At Palmichal (700–1000 m; localities 143 and 144), Heteromys catopterius and Oryzomys albigularis ( =  Nephelomys caracolus) constituted the most commonly captured small, nonvolant mammals. The macrohabitats at Palmichal included both primary and secondary cloud forests. The latter were abandoned coffee plantations and small plots of subsistence agriculture that had regrown by the time of the inventories. The cloud forest “has an average height of 20–25 m with emergent trees reaching 30–40 m; it has 2–3 strata and a well-developed understory. Epiphytes are abundant on higher branches and treetops, whereas lianas are rare” (Bisbal-E., 1993: 366; translation ours). At both La Justa and Palmichal, H. catopterius was captured only on the ground, despite the placement of traps on tree branches as well.

Sympatry and Zones of Contact with Heteromys anomalus

Presently available specimens indicate that Heteromys catopterius and H. anomalus show largely complementary distributions in north-central and northeastern Venezuela. Apparently, the distributions of the two species are generally parapatric, with only very narrow zones of sympatry. In general, H. catopterius inhabits higher, wetter areas than H. anomalus. Known sites of contact between the species occur in the Rancho Grande region in the western portion of the Serranía del Litoral, the El Ávila massif of the Serranía del Litoral, the Tiara region of the Serranía del Interior, and the Caripe region of the Macizo Oriental (figs. 1, 4Figure 4 (Continued)). Similar parapatric distributions have been observed for several other pairs of species in the subfamily (Genoways, 1973; Rogers and Engstrom, 1992; Anderson, 1999; Anderson and Jarrín-V., 2002; Anderson and Timm, 2006).

The most detailed distributional information for an area of geographic contact between Heteromys catopterius and H. anomalus derives from the Rancho Grande region (in the western portion of the Serranía del Litoral). There, the distributions of the two species meet on both the southern and northern slopes. On the southern slope, the species' distributions come into contact at ca. 1050 m. Many specimens of H. catopterius from the Rancho Grande region correspond to areas near the Estación Biológica de Rancho Grande and El Portachuelo. El Portachuelo (locality 136) lies at the crest of the range and represents a pass between Pico Periquito and Pico Guacamaya. The Estación Biológica (localities 47 and 137) lies on the southern slope, slightly higher than and just to the east of El Portachuelo. Near the Estación Biológica and El Portachelo, H. catopterius was captured in large numbers in cloud forests and secondary forests (see Natural History; C.O. Handley, Jr. field notes, 1960). In contrast, only a few specimens of H. anomalus exist from areas near the Estación Biológica. One was collected by William Beebe; it was taken in an unspecified habitat within 1 km of the Estación Biológica (Tate, 1947). The Smithsonian Venezuelan Project collected three H. anomalus from areas slightly below the Estación Biológica on the southern slope; most of the sampling was in secondary habitats, especially in and near a “banana grove with thick underbrush and tall tress” (N.E. Peterson field notes, 1965). The tags for those specimens indicate that one was captured “in dense herbs—coffee and banana growth” and the other two in “dense herbs and banana plants.” In addition, the Smithsonian Venezuelan Project also collected an individual (USNM 371016) that shows a mixture of morphological characters of the two species and may represent a hybrid between them. We provisionally identify it as H. catopterius, but genetic studies are necessary to examine the possibility of limited hybridization between the two species in this region intimated by the morphological data. This specimen derives from the same macrohabitat as the three individuals of H. anomalus, and the tag indicates the microhabitat as at the “base of sapling, low herbs.” Interestingly, H. catopterius was captured in an area below the Estación Biológica on the southern slope along a stream in rainforest habitat (habitat described and pictured in Voss, 1988: 417–420; R.S. Voss field notes, 1979). Taken together, the available natural history information suggests that the two species likely segregate by habitat type (H. catopterius in wetter and/or more mature habitats and H. anomalus in drier and/or more disturbed areas) at a local scale in this area.

Records also document the presence of both species from the northern slope of the Serranía del Litoral just north of El Portachuelo and Rancho Grande. Researchers from the Museo del Instituto de Zoología Agrícola (Universidad Central de Venezuela) in Maracay surveyed small mammals at many elevations along the highway that passes by Rancho Grande and leads to Ocumare de la Costa. Both Heteromys catopterius and H. anomalus have been documented from 900–920 m on the northern slope (localities 45 and 130); in addition, H. anomalus is known from 630 m (locality 44), and H. catopterius from 680 m (locality 129). No habitat information accompanies these records.

Additionally, two sites of sympatry and one site of near parapatry exist in the central portion of the range of Heteromys catopterius. The two species come into geographic contact on the northern slopes of the El Ávila massif (in the central portion of the Serranía del Litoral). There, both species have been collected in very large numbers at Los Canales de Naiguatá, at 720–800 m (localities 120 and 160; see Valdez et al., 1985). Because these water-gathering canals span several kilometers, it is not clear whether or not the species coexist in the same macrohabitat or, alternatively, are separated by habitat and/or elevation. Specimens also document an area of geographic contact between the two species in the Tiara region of the Serranía del Interior. The species are sympatric at Campamento Rafael Rangel ( =  Loma de Hierro/Minas de Niquel; localities 53 and 138) at 1260 m, but no habitat information exists for any of the specimens. Heteromys catopterius exists from 1200 m at the same site, and H. anomalus occurs at nearby localities of La Horqueta (locality 51) and Agua Amarilla (locality 52), at slightly lower elevations (ca. 1080–1100 m). At the eastern end of the main Aragua-Carabobo chain of the Serranía del Litoral, the two species come into close geographic contact, although no known site of sympatry exists. Here, H. catopterius inhabits El Laurel (locality 155) at 1355 m, and H. anomalus occurs at the nearby sites Turgua (1144 m; locality 76) and Bosquecillo, on the campus of the Universidad Simón Bolívar (ca. 1100 m; locality 75).

Finally, the two species come into geographic contact in the Caripe region of the Macizo Oriental. There, two very nearby sites of sympatry exist: La Laguna (1330 m for Heteromys anomalus and 1170–1335 m for H. catopterius; localities 84 and 159 for H. catopterius) and an area sampled near San Agustín (1180–1260 m for H. anomalus and 1170–1180 m for H. catopterius; localities 84 and 159). In addition, H. catopterius occurs at two nearby sites (San Agustín, 1270 m and Hacienda San Fernardo, 1320 m; localities 157 and 158), and closeby, H. anomalus is known from the Cueva del Guácharo (1060–1065 m; locality 86). Finally, two specimens (USNM 409820 and 409821) from Cueva del Guácharo (1010 m) display a mixture of characters of the two species; although provisionally we ascribe them to H. anomalus, they may constitute rare hybrids between the two species. Future genetic studies in the region are necessary to address this possibility. In contrast to the well-documented area of contact between the two species in the region of Rancho Grande (see above), the few available records from near Caripe do not allow for any clear characterization of habitat-related, successional, and/or elevational separation between the two species.

Conservation Status

Heteromys catopterius probably faces minimal risk of extinction or even population-level decline in most of its range. Like most species of the genus, data available from field collectors indicate that the species inhabits only closed-canopy evergreen forests (i.e., not deciduous forests, xerophytic vegetation, natural grasslands, or open agricultural areas). Among the evergreen montane forests it inhabits, the species appears especially associated with cloud forests and with areas holding an abundance of palms. Fortunately, it shows substantial tolerance for disturbance within forested ecosystems. In addition to particularly frequent collection in mature cloud forests, the species also has been trapped in large numbers in secondary forests (e.g., at Rancho Grande; see Natural History). In addition, H. catopterius exhibits a moderately widespread distribution (throughout most of the Cordillera de la Costa). Finally, many national parks and other governmentally protected reserves harbor known localities of the species: Parques Nacionales El Ávila, El Guácharo, Henri Pittier, Macarao, Miguel José en Sanz San Esteban; and Monumentos Naturales Cueva del Guácharo and Pico Codazzi). Heteromys catopterius also has been documented from private lands conserved as a natural watershed in the Río Morón drainage (Bisbal-E., 1993). The protection afforded these reserves varies greatly, but large areas of intact forest remain in several of them (perhaps most notably the Parque Nacional Henri Pittier and Parque Nacional El Ávila).

Heteromys anomalus (Thompson, 1815)

Caribbean Spiny Pocket Mouse

Distribution

Distributed across much of northern Colombia and Venezuela and in the upper Río Magdalena Valley of central Colombia, from near sea level to middle elevations; also inhabits the nearby Caribbean islands of Margarita, Trinidad, and Tobago (figs. 1, 4Figure 4 (Continued); Anderson, 2003b; appendices 1 and 2). In addition to regions dominated by forest, H. anomalus also inhabits restricted forested areas (riparian gallery forests) in the llanos (savanna) of Venezuela (August, 1984; Soriano and Clulow, 1988). The species typically ranges from sea level to ca. 1600 m in elevation, although its upper distributional limit is lower in some mountain ranges (e.g., the Cordillera de la Costa in Venezuela; see Biogeography of Heteromys catopterius and H. anomalus); a population in the semi-isolated Macizo de Guaramacal in western Venezuela inhabits an anomalously high elevation (2430 m; Anderson, 2003b). Note also seven localities of H. anomalus in the central and western llanos of Venezuela reported by Utrera et al. (2000); we have examined vouchers only from one of those sites (Caño Hondo, locality 66, appendix 1); the remainder probably represent new records of H. anomalus.

Diagnosis

A species of spiny pocket mouse with adults showing the following combination of characters (figs. 6, 7, 9): p4 (lower permanent premolar) with three lophids; P4 (permanent upper premolar) with straight, moderately long fold in anterior margin of posterior loph; tubercle or swelling at posteroventral border of infraorbital foramen absent; mesopterygoid fossa formed by long, thin hamular processes of pterygoids; optic foramen especially small, with posterior margin formed by strong bar of bone; parietomastoid suture dipping well ventral to parietal crest posterior to its widest point, then ascending dorsally to rejoin crest near its posterior termination; rostrum moderately long and generally moderately tapered anteriorly, without anterodorsal flare; anterior portion of premaxillary convex (inflated), forming a smooth (not stepped) lateral border of rostrum; interorbital constriction narrow; braincase narrow and not inflated; interparietal variable in size and shape; skull average for genus (ONL 31.06–39.02 mm in adult specimens of age class 4; table 1, appendix 3); dorsal pelage typically pale brown strongly grizzled with thin ochraceous hairs intermixed among spines, but occasionally almost uniformly dark drab gray (in wet lowlands of Estado Zulia, in some parts of the Cordillera de Mérida, in mesic areas of Estado Monagas and Estado Sucre, and on the island of Tobago), usually harsh and spiny, but often softer in montane regions; weak patch of dark coloration present on dorsal and external surfaces of forearms, continuous with dark coloration of flanks; ears distinctively rounded, brown to gray and large relative to body size; orange band on flanks absent; tail typically strongly bicolored and slightly to moderately longer than head-and-body length; plantar surface of hind feet naked.

Karyology

Engstrom et al. (1987) reported the standard karyotype of 2n  =  60, FN  =  68 for 20 individuals of H. anomalus from Estado Miranda (24 km N Altagracia de Orituco and 25 km N Altagracia de Orituco, localities 79 and 80), Estado Monagas (Caripito, locality 88) and Estado Sucre (40 km SW Caripito, locality 104), with voucher specimens deposited at the TCWC and CM. Although those authors did not provide museum catalog numbers for the specimens that were karyotyped, we have examined specimens from three of these localities in those museum collections (appendix 1); all represent H. anomalus.

Comparisons

See above for comparisons with Heteromys catopterius, and Anderson (2003b) and Anderson and Timm (2006) for comparisons with other species of the genus.

Geographic Variation

Although a comprehensive review of geographic variation in the widespread Heteromys anomalus lies beyond the scope of the current work, some noteworthy patterns exist. Externally, a few populations in mesic areas (e.g., wet regions of Estados Lara, Monagas, Sucre, and Zulia in Venezuela, and on the island of Tobago) show darker, less grizzled dorsal pelage than populations in most parts of the species' range (see also Anderson, 2003b). One doubly insular population (Bush Bush Forest; locality 22) exhibits markedly smaller size than other examined populations (Anderson, 2003b). It corresponds to the tiny island of the same name, which lies within the Nariva Swamp in central-eastern Trinidad (Downs et al., 1968).

Natural History

Field notes and published studies provide substantial ecological information on Heteromys anomalus from a few sites where it was collected in large numbers in eastern Venezuela and Trinidad and Tobago. However, as mentioned above (see account of H. catopterius), the extensive information provided by Valdez et al. (1985) for the genus at Los Canales de Naiguatá in Estado Vargas in the Serranía del Litoral corresponds to pooled data for H. anomalus and H. catopterius (localities 120 and 160; appendix 1); we do not include natural history information from that site here. See Anderson (2003b) for summaries of natural history information for sites in Colombia and western Venezuela.

On the slopes of the Macizo Oriental (fig. 1) in Estado Monagas and Estado Sucre, field notes indicate that Heteromys anomalus was captured commonly in intact tropical rainforest adjacent to cropland but only occasionally in drier areas more dominated by agriculture. G.H.H. Tate and H.J. Clement sampled several sites in this region in March–May 1925 (see also Tate, 1931). The area at Neverí (locality 111) where H. anomalus was common, was:

Latal (locality 113), another site where H. anomalus was common, “is surrounded by high hills all partly cleared in their lower slopes and planted to coffee. However there is much fine forest.” (G.H.H. Tate field notes, 1925).

In contrast, Heteromys anomalus was captured only occasionally by Tate and Clement at San Antonio de Maturín (locality 89), a drier region with denser human settlement. The region as a whole holds “xerophytic vegetation, but far up the valley on either hand one can descry good looking forest.” (G.H.H. Tate field notes, 1925). The valley floor was densely populated, and many pastures were present. The valley was drained by “a stony, broad, shallow stream meandering between alternate small bluffs and playas. Along this stream are also large stands of caña lata, the tall reed-like cane often used for walling in native houses where no strength is needed” (G.H.H. Tate field notes, 1925).

At Finca Vuelta Larga (locality 101) in the lowlands of Estado Sucre, field notes indicate that Heteromys anomalus was captured in areas of secondary evergreen forest as well as adjacent small agricultural plots, but not in inundated savannas of grasses and sparse shrubs. Robert S. Voss, Hernán C. Castellanos, and Carlos Gauveca-C. sampled these managed and seminatural habitats in August 1987 (R.S. Voss and H.C. Castellanos field notes, 1987; see also Voss, 1991). The area comprised a mosaic of inundated savannas (estero) and secondary forests (mato) at slightly higher elevations. Similar numbers of traps were set in inundated savanna and in forest, as well as a lesser number in small recently established agricultural plots (conuco). The savanna was flooded by use of a series of dikes and fenced for heavy grazing; shrubs and tall herbs grew along the fences, where most of the savanna traps were set. The forest was secondary growth, cleared ca. 14 years prior for small agricultural plots and allowed to regrow (abandoned conuco). The vegetation there was “a dense tangle of second-growth trees (mostly unbuttressed) and shrubs with many viny tangles” (R.S. Voss field notes, 1987). This secondary forest was evergreen and dominated by palms, with a variable and discontinuous canopy ca. 30 m high (H.C. Castellanos field notes, 1987). Heteromys anomalus was captured in secondary forest and recently established agricultural plots (conuco) but not in traps set in the wet, marshy savanna margins under fences, tall grass, and occasional shrubs.

Some information regarding the microhabitats of Heteromys anomalus at Finca Vuelta Larga is also available (R.S. Voss, H.C. Castellanos, and C. Gauveca-C. field notes, 1987; translations ours, where relevant): on the ground in leaf litter under dense undergrowth in forest; on bare ground under tangled lianas with leaf litter on top, and under stilt roots of a Cecropia; on ground in wet leaf litter under tall grass in old [abandoned] conuco (grass, bushes, plantains, and small trees); on ground in leaf litter and below a group of lianas that surround a palm (Corozo) 665 mm in dbh; on ground in tall grasses and shrubs (2 individuals).

The species may be an agricultural pest in at least some regions. In the original description, Thompson (1815) noted that “The habits of this tribe of rats are singular and curious: where numerous, they do incalculable mischief in barns and granaries; for, not satisfied with what they can eat on the spot, they stow away and carry off in their cheek pouches no inconsiderable quantity [of grain], to be deposited in their retreats for times when food is not to be procured from without” (Thompson, 1815: 162). We know of no studies quantifying the extraction of agricultural grain by Heteromys anomalus. Whereas the species does not inhabit large fields of mechanized agriculture (see Utrera et al., 2000; Anderson, 2003b), it does exist in many areas containing smaller fields in subsistence agriculture settings (see above), where it may be a pest to humans.

Substantial natural history information exists for Heteromys anomalus at a site in Trinidad where large numbers of the species were captured during surveys that spanned several years. Teams from the Trinidad Regional Virus Laboratory surveyed Bush Bush Forest (locality 22) for small, nonvolant mammals from 1960–1963 (Worth et al., 1968; see also Aitken et al., 1968). Downs et al. (1968) provide detailed information on the region, which we summarize here. Bush Bush Forest is the name given to the areas sampled on Bush Bush Island, a low, forested island present in the Nariva Swamp of central-eastern Trinidad. The island has sandy, well-drained soils and is approximately 0.25 mi (0.40 km) wide and 2–3 mi (3–5 km) long. Although most of the island itself is never flooded, it merges into areas that are seasonally inundated that support swamp forest comprised of palms and a few hardwoods. Most of the surrounding open swamp, however, is almost continuously inundated, covered by grass, cane, or aeroids. Before 1960, the island received only periodic visits by humans for fishing during the dry season, as well as sporadic visitation associated with hunting or selective extraction of timber. Both removal trapping and mark-recapture studies were undertaken; all voucher specimens that we examined from this work represent H. anomalus (appendix 1).

Heteromys anomalus was consistently one of the most commonly collected small, nonvolant mammals in nonflooded forested areas on Bush Bush Forest, and a variety of information regarding its natural history is available (Worth et al., 1968, from which the following information derives). Removal trapping spread over several parts of the island over four years yielded a total of 423 individuals of H. anomalus, with an overall trap success that varied from 0.3 % to 2.5 % in individual years. A general decline in numbers was observed during the study. Minimal trapping in swampy habitats failed to capture any Heteromys. Mark–recapture studies in the Restan area of Bush Bush Island captured 183 individuals (including recaptures) over three years, with much higher trap success (8.2%). The species showed a small home range, “rarely entering traps in more than two adjacent stations” that were set an average of ca. 100 ft (30 m) apart (Worth et al., 1968: 271). No individual of the species was found to have moved even 300 ft (91 m). Immature individuals were present in every month of the year but January, but were most frequent from May through October (the wet season). Some individuals were followed for more than six months. The species “tended to begin foraging earlier in the evening than other species” (of rodents and marsupials; Worth et al., 1968: 271).

Some behavioral information in captivity is available for individuals of Heteromys anomalus from the island of Trinidad. John F. Eisenberg conducted a landmark comparative study of behavior across the family Heteromyidae (Eisenberg, 1963). Most of his investigation of the genus Heteromys focused on H. lepturus (from Finca La Selva, Veracruz, Mexico;  =  H. desmarestianus; see Anderson et al., 2006, for geographic variation within the H. desmarestianus complex). In addition, however, four individuals of H. anomalus were surveyed for individual behavior and some aspects of intraspecific social behavior. These individuals were provided by the Trinidad Regional Virus Laboratories from an unspecified location, presumably in Trinidad. Individuals were held in terraria at high humidity with soil covering the floor. Observations were performed at night using a red light for illumination. In male-female encounters, H. anomalus was less aggressive than species of other heteromyine genera, with “chasing” constituting the primary aggressive behavior of H. anomalus (Eisenberg, 1963: 50, see also 68, 94). Both naso-nasal and naso-anal behaviors were observed between individuals of the opposite sex (Eisenberg, 1963: 51–52). Subsequently, preliminary sexual behaviors were observed, but copulation was not: “Driving [chasing and attempting to mount the female], perineal investigation, grooming, patting, and mounting [gripping of flanks of female with forelimbs, pressing ventrum against back of female and attempting intromission] were displayed by the males. A typical lordosis [freezing movement, allowing the male to mount as the female raises her hindquarters] was shown by the females … . The male Heteromys [anomalus] uses a neck grip [grasping with incisors the fur over the dorsal gland of the female during a mount]” (Eisenberg, 1963: 51, see also 47, 58).

On the island of Tobago, field notes document the capture of Heteromys anomalus near Charlottesville on the slopes of Pigeon Peak in several forested habitats (especially montane forest) but not in areas under active cultivation (locality 1; R.S. Voss and M.E. Holden field notes, 1989). Robert S. Voss and Mary Ellen Holden sampled a variety of habitats in this region in January 1989 (see also habitat descriptions and photographs in Voss, 1991: 73–76). Heteromys anomalus was the most commonly captured small nonvolant mammal in montane forests, whereas Zygodontomys brevicauda was the most common in most other sampled habitats. Heteromys anomalus was also captured occasionally in lowland primary rainforest (Voss, 1991) and abandoned cocoa plantations with an intact original canopy. The species was not caught in agricultural clearings (cocoa, plantains, pigeon peas, papaya, and cassava [manioc]) or in active cocoa plantations (R.S. Voss field notes, 1989). The montane forest where H. anomalus was especially common can be characterized as follows:

Information regarding the microhabitats of Heteromys anomalus near Charlottesville is also available (R.S. Voss and M.E. Holden field notes, 1989). All were captured on the ground in forest (all but four in montane forest). Individual microhabitats follow: under fallen vegetative litter and leaves; under wet vegetative litter and leaves beside tree; under fallen palm fronds; under rotting palm fronds beside large tree; beside rotting log under dead palm frond, fallen branches and other vegetative litter; under log; under inclined tree trunk (2 individuals); under roots of fallen tree; under stilt roots of Cecropia; under stilt roots of small tree; under tree ferns and aroids; beside trees under tangled vines; beside buttress of large tree with vine coverage.

Conservation Status

Heteromys anomalus faces no substantial risk of extinction, but local populations are vulnerable to declines or extirpation in areas undergoing deforestation. Available data indicate that H. anomalus inhabits principally closed-canopy forests, but not xerophytic vegetation, natural or anthropogenic grasslands, or large, open agricultural areas (e.g., mechanized agriculture; Utrera et al., 2000). However, in contrast to most species of the genus, it is also known from small plots of subsistence agriculture; see Natural History; see also Anderson, 2003b: 24–25). Furthermore, in addition to evergreen forests typical of the genus, it also inhabits areas where deciduous forests constitute the predominant vegetational type. In such regions, however, it remains unclear whether the species occupies deciduous forests per se, or only smaller areas of evergreen vegetation in more mesic situations, such as along watercourses. The species exhibits a very widespread distribution throughout most of northern Colombia and northern Venezuela, as well as much of Trinidad and Tobago. In central and eastern Venezuela, the species has been collected in many national parks (Parques Nacionales Cerro El Copey-Jóvito Villalba, El Ávila, El Guácharo, Guatopo, Henri Pittier, Miguel José en Sanz San Esteban, and Península de Paria) and other governmentally protected areas (Monumentos Naturales Cerros Matasiete-Guayamurú and Cueva del Guácharo). These reserves receive varying levels of protection; whereas some face substantial human pressure, others (e.g., Parque Nacional Henri Pittier) contain large extents of intact forest.

Endemism in the Cordillera de la Costa

Current information suggests that Heteromys catopterius is endemic to the Cordillera de la Costa (figs. 1, 4Figure 4 (Continued)). Records derive from throughout the Serranía del Litoral, and the species is also known from one area of the Serranía del Interior (near Tiara, in Estado Aragua). Because this latter range is lower and drier, it probably offers much less extensive areas of suitable habitat for the species. Heteromys catopterius is not known from mountain ranges to the west of the Cordillera de la Costa. No records of the species exist in the Sierra de Aroa to the west of the Cordillera de la Costa and separated from it by the Depresión de Yaracuy (SAGCN, 1995). The floral composition of the Sierra de Aroa shows strong similarities to the Cordillera de la Costa (Huber, 1997), but the upper portions of that range have received very little sampling by mammalogists, and conclusions regarding its mammalian fauna must await further fieldwork (see Anderson, 2003a, for methods of testing hypotheses regarding a species' absence). The presence of H. catopterius seems much less likely in the Cordillera de Mérida or Serranía de San Luis even farther to the west. Extensive sampling for small nonvolant mammals has occurred in many portions of the Cordillera de Mérida, yielding many records of H. anomalus (Anderson, 2003b). The Serranía de San Luis occupies a very isolated position far from other mesic montane regions in northern Venezuela, and available records document the presence of H. anomalus there as well (Anderson, 2003b).

Similarly, several mountain ranges present near the eastern end of the known distribution of Heteromys catopterius lack records of the species. Two subranges comprise the Macizo Oriental (the easternmost portion of the Cordillera de la Costa). All documented populations of H. catopterius in the Macizo Oriental correspond to the Caripe region in the eastern subrange, which is lower and wetter (Huber, 1997). None derive from the higher and drier Cerro Turimiquire subrange just to the west, an area that has received little sampling by mammalogists (Tate, 1931; G.H.H. Tate field notes, 1925). The only specimen of Heteromys known from the higher areas on Cerro Turimiquire represents H. anomalus (locality 114). Similarly, no records of H. catopterius exist from moderately high montane regions on the Península de Paria in extreme northeastern Venezuela, where H. anomalus is known, or on Trinidad or Tobago, two continental islands with plentiful records of H. anomalus. Interestingly, the habitat of H. anomalus at Pigeon Peak in Tobago (see Natural History of H. anomalus) holds many similarities with that of sites on the mainland where H. catopterius has been collected in large numbers (e.g., Rancho Grande and Hotel Humboldt; see Natural History of H. catopterius).

Following current taxonomy, the Cordillera de la Costa harbors only three endemic species of mammals, out of the 327 (Linares, 1998) to 351 (Ochoa-G. and Aguilera-M., 2003) mammals estimated to inhabit Venezuela. An oryzomyine rodent, Nephelomys caracolus (formerly included in the Oryzomys albigularis species complex) shows a distribution somewhat similar to that of Heteromys catopterius, inhabiting the central portion of the Cordillera de la Costa (both the main Aragua-Carabobo chain and the El Ávila massif) at elevations between 1050 and 2300 m; however, in contrast to H. catopterius, no records of N. caracolus exist from the Serranía del Interior or the Macizo Oriental (Percequillo, 2003; see also Aguilera et al., 1995a; Márquez et al. 2000; Anderson, 2003a). Even more restricted, an ichthyomyine rodent, Ichthyomys pittieri, remains known from only a few specimens from the Serranía del Litoral (Handley and Mondolfi, 1963; Voss, 1988). Whereas several species of nonmammalian tetrapods show distributions restricted to montane areas on the Península de Paria in northeastern Venezuela (Estado Sucre) as well as in the Cordillera de la Costa (see below), currently available information indicates that no mammal shows that pattern. Not surprisingly, the mammals endemic to the Cordillera de la Costa are rodents, which typically show small home ranges and low dispersal abilities. Future taxonomic revisions may alter this preliminary information by bringing to light other endemic species currently considered to belong to widespread species (as has happened with revisions of the common rodents Heteromys and Nephelomys). On the contrary, however, new fieldwork may extend the known ranges of these three species, altering the present conclusions regarding their endemism.

Rates of avian endemism in the Cordillera de la Costa appear grossly similar to the situation in mammals, but estimates range much higher for amphibians and reptiles. Although taxonomic studies and inventory work remain incomplete (especially for amphibians and reptiles), we provide information from available biogeographic summaries of nonmammalian tetrapods (note that regions of endemism in Cracraft [1985] and Péfaur and Rivero [2000] include the surrounding lowlands; therefore, the information provided in those sources cannot be used in the present comparisons of montane endemism). Among a total of 1360 species of birds considered to occur in Venezuela, seven or eight are endemic to the Cordillera de la Costa (Lentino-R., 2003); broadening the definition of the region of endemism to include montane areas on the Península de Paria would add five other species. Note, however, that Hilty (2003) indicated more extensive distributions for five of these 13 species, including other montane systems in northern Venezuela, most commonly the Sierra de Aroa and Serranía de San Luis. Of the 284 species of amphibians documented for the country, 27 are endemic to a region including the Cordillera de la Costa and the Península de Paria; information for amphibian endemism in the Cordillera de la Costa itself is not available (La Marca, 2003a). In total, 341 species of reptiles are estimated to inhabit Venezuela, but published information regarding Venezuelan reptiles in the Cordillera de la Costa remains lacking (La Marca, 2003b). Ongoing research indicates that at least 27 species of reptiles are endemic to the Cordillera de la Costa and montane areas on the Península de Paria; at least 20 of those are endemic to the Cordillera de la Costa itself (G. Rivas, in litt.). As is the case for mammals, future fieldwork and taxonomic research may alter these conclusions.

Ecogeography

Although most localities of Heteromys catopterius correspond to wet highland sites (including many cloud forests), the species also has been documented at intermediate elevations in sufficiently wet areas. The lowest occurrence record of the species on the southern versant of the main Aragua-Carabobo chain lies at 1050 m (Rancho Grande; locality 137), but the species inhabits wet forests at 680 m elevation on the moister northern slope in Estado Aragua (Carretera Estación Biológica Rancho Grande–Cata, km 29; locality 129) and even as low as 350 m in the Río Morón drainage in Estado Carabobo (La Justa; locality 141; figs. 1, 4Figure 4 (Continued)). On the El Ávila massif, the species descends to 1050 m on the southern slope along the border between Distrito Captial and Estado Miranda (Quebrada Chacaito; locality 147), but records of the species are known from as low as 720 m on the wetter northern slope (Los Canales de Naiguatá; locality 160; Estado Vargas). These altitudinal patterns coincide with known vegetational differences. Although much local variation exists (related to factors such as exposure to wind), mesic vegetational zones generally descend farther on the northern (Caribbean-facing) slope of the Serranía del Litoral than on the southern, interior versant (Huber, 1997; Ataroff-S., 2003; Meier, 2004).

We conclude that Heteromys catopterius has a disjunct distribution within the Cordillera de la Costa. At the very least, the low, dry Depresión de Unare surely isolates the eastern populations in the Macizo Oriental from those in the western and central part of the range of the species. In addition, the central populations on the El Ávila massif may be disjunct from the western populations in the main Aragua-Carabobo chain, due to the drier and moderately low saddle drained by the Quebrada Tacagua on the northwest and the Río Guaire on the southeast (through the city of Caracas). It remains even less clear whether populations in the Serranía del Interior are contiguous with those of the Serranía del Litoral. In any case, the species constitutes a system propitious for fascinating phylogeographic work, especially in a comparative context (i.e., in comparison with other codistributed species, such as Nephelomys caracolus) and in conjunction with niche-based distributional modeling in a Geographic Information Systems (GIS) framework (Graham et al., 2004; Phillips et al., 2006).

The ecogeographic patterns of records of Heteromys anomalus also merit comment and future research. Heteromys anomalus only rarely inhabits localities over 1400 m in regions of the Cordillera de la Costa where H. catopterius is present. Heteromys anomalus shows a distribution complementary to that of H. catopterius, inhabiting higher elevations on the southern (interior) slopes of the central portions of the Cordillera de la Costa than it does on the northern versants. Eight localities above 1000 m exist for H. anomalus in the central portion of the Cordillera de la Costa, but none reaches 1400 m (localities 47, 51–53, 62, 63, 75, and 76, Estados Aragua, Carabobo, and Miranda, 1050–1350 m). All pertain to either the southern slopes of main Aragua-Carabobo chain or the Serranía del Interior. In contrast, the highest records of the species on the northern slope of the main Aragua-Carabobo chain and the El Ávila massif are only 920 m and 800 m, respectively (locality 45, Estado Aragua; locality 120, Estado Vargas). In the eastern portion of the Cordillera de la Costa (the Macizo Oriental), H. anomalus occurs at elevations above 1000 m at five localities. Four lie in the eastern portion of the Macizo near Caripe and known records of H. catopterius, and only one of those corresponds to an elevation above 1400 m (localities 82–84, 86, Estado Monagas, 1060–1550 m). The last is from Cerro Turimiquire at a much higher elevation (locality 114, Estado Sucre, 6000 ft [1829 m]) in the western portion of the Macizo, which is drier (Huber, 1997) and where no record of H. catopterius is known.

In contrast to its distribution in the Cordillera de la Costa (where it very rarely exceeds 1400 m), Heteromys anomalus is documented from localities above 1400 m in several ranges to the west, outside the range of H. catopterius. These include the Sierra Nevada de Santa Marta (San Miguel, Departamento de La Guajira, Colombia, 1700 m), Serranía de Perijá (Sierra Negra, Departamento de la Guajira, Colombia, 1500 m), Serranía de San Luis (Cerro La Danta, Estado Falcón, 1300–1470 m), and Cordillera de Mérida (Cerro Alto, Estado Barinas, 1460–1600 m; El Blanquito and other localities near Sanare, Estado Lara, 1500–1700 m; near Estánquez, Estado Mérida, 1500 m; and near Boconó [in the semi-isolated Macizo de Guaramacal, where several species of mammals show atypically high distributions, Soriano et al., 1990], Estado Trujillo, 2430 m; Anderson, 2003b). Notably, in the isolated Sierra de Aroa between the Cordillera de Mérida and the Cordillera de la Costa, records of H. anomalus reach only 800 m (Finca El Jaguar, 680–800 m; Anderson, 2003b). Future fieldwork at higher elevations in that range is likely to discover populations of some species of Heteromys in the mesic montane forests there (IGVSB, 2004), which have received very little sampling by mammalogists.

The ecogeographic patterns of Heteromys catopterius and H. anomalus described above beg for studies assessing the relative importance of climate, biotic interactions, and history in determining their distributions. Likely, all three classes of factors will prove important. Clearly, the localities that the species inhabit differ in climate. In addition, however, occurrence records intimate the possibility of competitive exclusion of H. anomalus from some areas of suitable habitat by H. catopterius. Finally, the absence of H. catopterius from some ranges likely derives from dispersal limitations. To distinguish between these hypotheses in particular regions, niche-based distributional modeling and considerations of sampling adequacy in a GIS framework are needed (e.g., Anderson et al., 2002a, 2002b; Anderson, 2003a; Graham et al., 2004; Phillips et al., 2006).