2 July 2002 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON 115(2):249-272. 2002. A new species of small-eared shrew from Colombia and Venezuela (Mammalia: Soricomorpha: Soricidae: Genus Cryptotis) Neal Woodman USGS Patuxent Wildlife Research Center, Smithsonian Institution, P.O. Box 37012, National Museum of Natural History, NHB 378, MRC-111, Washington, DC 20013-7012, U.S.A. Abstract.?Populations of small-eared shrews inhabiting the northern Cor- dillera Oriental of Colombia and adjoining Venezuelan highlands in the vicinity of P?ramo de Tama have been referred alternatively to Cryptotis thomasi or Cryptotis meridensis. Morphological and morphometiical study of this popu- lation indicates that it belongs to neither taxon, but represents a distinct, pre- viously unrecognized species. I describe this new species as Cryptotis tamensis and redescribe C meridensis. Recognition of the population at P?ramo de Tama as a separate taxon calls into question the identities of populations of shrews currently represented only by single specimens from Cerro Pintado in the Sierra de Perij?, Colombia, and near El Junquito in the coastal highlands of Vene- zuela. All South American shrews (Family Sor- icidae) are members of the genus Cryptotis, a taxon that is widespread in eastern and central North America, Central America, and northwestern South America. Soricids inhabit exclusively montane regions in South America, being restricted to high- lands above at least 1200 m in the Andes north of the Huancabamba Depression, Peru. Shrews were first reported from South America near the end of the 19th century, when Cryptotis thomasi (Merriam, 1897) was described from near Bogot? in the Cor- dillera Oriental of Colombia. Since then, several additional taxa of shrews have been named from the continent (e.g., Thomas 1898, 1912, 1921; Allen 1912; An?iony 1921; Woodman & Timm 1993; Vivar et al. 1997), although the number of recognized species has varied from one (Hershkovitz 1969, 1972) to as many as ten (T?te 1932). The taxonomy of North American and Cen- tral American species of Cryptotis has been studied in some detail, but a comprehensive revision of South American taxa has not been attempted since T?te (1932) summa- rized their distributions. The taxonomy, systematics, and geographic distributions of South American soricids remain poorly un- derstood. The first species of shrew recorded from Venezuela was Cryptotis meridensis (Thomas, 1898), which was described based on specimens from Venezuela's Cor- dillera de los Andes, near M?rida. Although the taxonomic status of C meridensis has fluctuated between that of a subspecies of a widespread, poorly-defined C. thomasi (Ca- brera 1958; Handley 1976; Corbet & Hill 1980, 1986, 1991; Honacki et al. 1982; Du- rant & Pefaur 1984; Eisenberg 1989) and that of a distinct species (Cabrera 1925; T?te 1932; Cabrera & Yepes 1940, 1960; Hutterer 1986, 1993; Linares 1998), its po- sition as a definable taxon has appeared to remain relatively stable. In reality, its per- ceived membership has varied consider- ably. The geographic range of C meridensis tiaditionally has included both the Cordil- lera de los Andes (including Sierra de M?- rida and Sierra del Norte [= Sierra de La Culata]) in Venezuela and the P?ramo de Tama (= Macizo de El Tama) on the Co- 250 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON lombia-Venezuela border along Colombia's Cordillera Oriental (Osgood 1912, T?te 1932, Cabrera 1958, Durant & Pefaur 1984, Linares 1998). The first individuals from the latter locality were collected for the Field Museum by W. H. Osgood and S. G. Jewett in 1911. These specimens were iden- tified as Blarina meridensis with assistance from Oldfield Thomas (Osgood 1912), who first described the species. P?ramo de Tama is physiographically separated from the Cordillera de los Andes by the T?chira De- pression, an arid valley that, according to Durant & Pefaur (1984), drops to ca. 800 m elevation. (Most large-scale maps of the region show this feature as above 1000 m.) Possibly owing to this disjunct distribution and the location of P?ramo de Tama in the Cordillera Oriental, some authors more re- cently have referred shrews from Paramo de Tama to C. thomasi, restricting C. mer- idensis mainly to populations in the Cor- dillera de los Andes and the coastal high- lands near Caracas (Hutterer 1986, 1993; Soriano et al. 1999). Recent redefinition of C. thomasi (Wood- man 1996, Vivar et al. 1997) has made it much easier to compare this species with other South American shrews. My mor- phometrical and character analyses of Cryp- totis from P?ramo de Tama and the Cordil- lera de los Andes indicate that neither pop- ulation is referable to C. thomasi. These studies also indicate, however, that C. mer- idensis s. 1. is polytypic. The purpose of this paper is to describe a previously unrecog- nized species of small-eared shrew from P?ramo de Tama and to redescribe Cryp- totis meridensis (Thomas, 1898) in the con- text of our greater understanding of varia- tion within and among species of Cryptotis. Materials and Methods The current work relied on the tech- niques and methods of previous studies of Cryptotis (Woodman & Timm 1992, 1993, 1999). Additional information regarding methodology is presented there. Species synonomies list only first use of published names. Regional names, place names, and map coordinates derive from original col- lector tags, field notes, and field catalogs supplemented with additions, corrections, and alternative readings based on review of published localities, maps, and gazetteers of the region (Paynter & Tray lor 1981; Payn- ter 1982; USBGN 1988, 1993). Localities for specimens from the Smithsonian Vene- zuelan Project (SVP) are based on original field tags supported by information from Norman E. Peterson's unpublished field cat- alog and journal for 1968, reposited at the National Museum of Natural History. SVP localities had been recalculated prior to be- ing published by Handley (1976), and I note his equivalent localities in brackets. Capitalized lifezone names follow Holdrid- ge (1947), Ewel et al. (1965), and Ewel & Madriz (1968). In the descriptions of the new species of Cryptotis and C meridensis, I compare characteristics of the species primarily with other members of the genus Cryptotis, un- less stated otherwise. This context is re- flected in the comparative terms used for measurements and ratios. A feature referred to as of "moderate" length, for example, is one whose mean value for the species falls within the range of plus or minus one stan- dard deviation of the mean for the genus. The mean for Cryptotis was established by averaging the means of all known species. A "long" or "wide" feature is one greater than the mean plus one standard deviation; a "short" or "narrow" feature is less than the mean minus one standard deviation. Terminology of dentition and dental characteristics follows Choate (1970). An- atomical terminology of the hum?rus and other aspects of the postcranial skeleton fol- lows Reed (1951). Formal color names for pelage hues are from Ridgway (1912). Pel- age coloration was determined only from museum specimens. Because such speci- mens may exhibit varying degrees of fox- ing, colors may not accurately represent those of live or newly-captured specimens. VOLUME 115, NUMBER 2 251 Measurements used in my analyses fol- low Woodman & Timm (1993, 1999); ab- breviations are explained in Table 1. In ad- dition, crown-to-rump length (CR) is re- ported for embryos. External measurements are those taken by the original collector, ex- cept for head-and-body length, which was calculated by subtracting length of tail from total length. Measurements of the skull were taken to the nearest 0.1 mm using ei- ther an ocular micrometer in a binocular microscope or a hand-held dial caliper (for CBL and CB). Univariate statistics include mean ? SD and total range. Ratios of mea- surements were multiplied by 100 to cal- culate percentages. Orthogonal principal components were calculated from matrices of 18 untransformed craniomandibular var- iables (CBL, ZP, lO, UIB, M2B, PL, TR, UTR, MTR, LMl, ML HCP HCV, HAC, BAC, AC3, TRM, Lml) to understand the relationships in overall shape of the skull among 36 Cryptotis meridensis, 16 C. tho- masi, and 13 individuals of the new species with complete crania. Male and female Cryptotis are difficult to distinguish externally, and many museum specimens either lack sexual identification or may have been sexed incorrectly in the field. Previous tests for sexual dimorphism in Cryptotis, including C meridensis (Du- rant & Pefaur 1984), indicate that males and females generally are not significantly different in size or shape (Choate 1970, Woodman 1992, Woodman & Timm 1993). For these reasons, both sexes were grouped together for all statistical analyses. In my investigations of Cryptotis, I op- erate under the philosophical influence of the evolutionary species concept as rede- fined by Wiley (1978, 1981), which re- quires that the species category (and tax- onomy in general) be logically consistent with reconstructed phylogenetic history. In application, I consider a species to be the largest monophyletic entity whose constit- uent parts interact and are not on different evolutionary trajectories (Alternative 2 of Frost & Hillis 1990). In practice, I use unique distributional patterns of morpho- logical characters among populations to dis- tinguish individual, presumably genetically- cohesive groups from other genetically-co- hesive groups. This results in a clear, sup- portable, and testable framework for additional studies. Specimens from the following institu- tions were used in this study: American Museum of Natural History, New York (AMNH); Natural History Museum, Lon- don (BM); Field Museum, Chicago (FMNH); Instituto de Ciencias Naturales, Universidad Nacional de Colombia, Bogot? (ICN); University of Kansas Natural His- tory Museum, Lawrence (KU); Museum of Comparative Zoology, Cambridge (MCZ); Museo de la Universidad de Antioquia, Me- dell?n (MUA); Royal Ontario Museum, To- ronto (ROM); National Museum of Natural History, Washington (USNM). Systematic Framework Ongoing studies of morphological vari- ation, species diversity, evolutionary rela- tionships, and biogeography of Cryptotis (Choate 1970; Hutterer 1986; Vivar et al. 1997; Woodman 1992, 1996; Woodman & Timm 1992, 1993, 1999, 2000) have added to our knowledge of Central American and South American members of the genus. Within Cryptotis, I recognize four infor- mal, but definable, groups of species that correspond closely to groupings originally delineated by Choate (1970): Cryptotis mexicana-gioup, containing nine species occurring from northeastern Mexico to central Honduras (Woodman & Timm 1999, 2000); C. nigrescens-group, consist- ing of six species distributed from southern Mexico to Colombia (Woodman & Timm 1993; 1999); C parva-gioup, comprising at least two diagnosable species (C. oro- phila and C. parva) and having a discon- tinuous distribution from southeasternmost Canada and the eastern United States to the Central Valley of Costa Rica; C thomasi- group, comprised entirely of Andean spe- 252 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON Table 1.?Measurements of selected Cryptotis from Colombia and Venezuela. Statistics are mean range. Measurements follow Woodman & Tinmi (1993). SD, and C. meridensis C. tamensis C. medeltinia External measurements: n = 65 n = 21 head-and-body length (HB) 89 ? 5 86 ? 4 76-102 80-91 tail length (TL) 34 ? 3 36 ? 2 25^1 32-39 Skull measurements: n = 35 n = 13 condylobasal length (CBL) 21.5 ? 0.7 21.5 ? 0.5 20.5-23.5 20.8-22.9 cranial breadth (CB) 10.4 ? 0.3 10.5 ? 0.2 10.0-11.0 10.1-10.8 (n = 11) breadth of zygomatic plate (ZP) 2.0 ? 0.2 2.1 ? 0.1 1.6-2.4 1.9-2.3 interorbital breadth (lO) 5.0 ? 0.2 5.0 ? 0.2 4.7-5.4 4.8-5.3 breadth across first unicuspids (UIB) 2.9 ? 0.1 3.0 ? 0.1 2.6-3.1 2.7-3.1 breadth across second molars (M2B) 6.4 ? 0.2 6.4 ? 0.1 6.0-7.0 6.1 -6.6 palatal length (PL) 9.5 ? 0.4 9.5 ? 0.2 8.8-10.2 9.1-9.8 length of upper toothrow (TR) 1 8.2 ? 0.3 8,3 ? 0.3 7.3-8.8 7.9-8.7 length of unicuspid toothrow (UTR) 2.8 ? 0.1 2.8 ? 0.1 2.6-3.0 2.6-3.0 length of molariform toothrow (MTR) 5.8 ? 0.2 5.9 ? 0.2 5.5-6.3 5.6-6.1 posterior width of first upper molar (MIW) 2.0 ? 0.1 2.0 ? 0.1 2.0-2.2 1.9-2.1 n = 22 86 ? 6 69-96 24 ? 2 20-27 n = 28 21.7 ? 0.5 20.4-22.6 (n = 23) 10.6 ? 0.2 10.2-11.0 in = 18) 2.0 ? 0.2 1.7-2.4 5.0 ? 0.2 4.6-5.4 2.7 ? 0.1 2.5-2.9 6.2 ? 0.2 5.9-6.5 9.3 ? 0.3 8.6-9.7 8.2 ? 0.2 7.7-8.5 2.8 ? 0.1 2.4-3.0, 5.8 ? 0.1 5.5-6.0 1.9 ? 0.1 1.8-2.1 n = 13 85 ? 7 77-98 37 + 6 29-46 n = 12 22.0 ? 1.0 20.0-23.2 in = 11) 10.9 ? 0.4 10.3-11.4 (n = 10) 2.1 ? 0.2 1.7-2.4 5.4 ? 0.2 5.0-5.8 3.0 ? 0.2 2.7-3.2 6.7 ? 0.4 6.0-7.2 9.8 ? 0.4 9.2-10.4 8.7 ? 0.4 8.1-9.2 3.0 ? 0.1 2.8-3.2 6.2 ? 0.3 5.7-6.5 2.1 ? 0.1 2.0-2.3 VOLUME 115, NUMBER 2 253 Table 1.?Continued. C. meridensis C. tamensis C. thomasi C. medellinia length of mandible (ML) 6.9 ? 0.3 7.0 ? 0.2 7.0 ? 0.2 7.5 ? 0.3 6.3-7.6 6.9-7.4 6.6-7.4 7.0-8.0 height of coronoid process (HCP) 4.8 ? 0.2 4.6 ? 0.2 4.6 ? 0.1 5.1 ? 0.3 4.5-5.2 4.3^.8 4.3^.9 4.5-5.5 height of coronoid valley (HCV) 3.3 ? 0.1 3.1 ? 0.1 3.1 ? 0.1 3.3 ? 0.2 3.1-3.5 3.0-3.2 2.8-34 2.9-3.6 height of articular condyle (HAC) 4.4 ? 0.2 4.4 ? 0.1 4.3 ? 0.1 4.5 ? 0.3 4.0-4.8 4.2^.6 4.1^.6 4.0-4.9 breadth of articular condyle (BAC) 3.6 ? 0.2 3.5 ? 0.1 3.6 ? 0.1 3.7 ? 0.1 3.2-4.0 3.3-3.6 3.4-3.8 3.4-3.8 articular condyle to m3 (AC3) 5.6 ? 0.3 54 ? 0.1 5.7 ? 0.2 5.5 ? 0.2 5.1-6.3 5.2-5.7 5.0-6.0 5.1-5.8 length of lower toothrow (TRM) 6.5 ? 0.2 6.6 ? 0.2 64 ? 0.2 6.8 ? 0.3 6.1-6.9 6.3-6.9 6.1-6.7 6.4-7.2 length of first lower molar (Lml) 2.0 ? 0.1 1.9 ? 0.1 1.9 ? 0.1 2.0 ? 0.1 1.8-2.1 1.8-2.1 1.7-2.0 1.9-2.2 Weight (g): 12.3 ? 2.0 13.9 ? 2.2 10, 10 15.5, 16, 17 8.5-18.0 10.8-16.1 (n = 48) (n = 6) (n = 2) (n = 3) cies (Hutterer 1993, Woodman 1996, Vivar et al. 1997). AU South American shrews are members of either the C. thomasi- group or the C. nigrescens-gronp. Eisen- berg (1989) incorrectly suggested that C. parva might occur in northern Colombia. In fact, the southernmost taxon of the C. parva-group, C. orophila, is not known from south of central Costa Rica, and there is no evidence that it ever extended into South America. Specimens of C. parva previously reported from Panama were shown to be misidentified C. nigrescens (Woodman & Timm 1993). With the exceptions of C. mera and C. colombiana, which are members of the mostly Central American C. nigrescens- group (Woodman & Timm 1993, Wood- man 1996), South American taxa of shrews are members of the C. thomasi-group. The C. thomasi-group currently includes only South American taxa, but the membership of the group may change as additional South American and southern Central American taxa are analyzed and phyloge- netic relationships are understood more completely. This group appears to share a number of apomorphies with the C. mexi- cana-gxonp (Woodman & Timm 1999, 2000) and may prove to be a subset of that 254 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON group. Within this framework, all of the species considered in this study (C. med- ellinia, C. meridensis, C. thomasi, and the new species from P?ramo de Tama) are members of the Cryptotis thomasi-gTOup of shrews. Some portions of this view of relation- ships of South American shrews were chal- lenged by Vivar et al. (1997), who restrict- ed membership of the C. thomasi-group to C. thomasi and possibly C. medellinia. They suggested that C. colombiana be re- moved from the C. nigrescens-group, but provided no guidance as to its possible al- ternative relationships. The C. thomasi-gToup and the C. ni- grescens-gToup can be distinguished using a combination of characters. Members of the C. thomasi-gToup all have longer, more luxuriant fur than members of the C. ni- grescens-gToup. Unlike other New World soricids, males of the C. thomasi-gioup lack obvious bare patches marking the lo- cation of the lateral glands. The forefeet of these shrews are somewhat enlarged, and the foreclaws are elongate (but not broad- ened, as in the C. mexicana-gioup), unlike the C. nigrescens-gToup in which neither forefeet nor foreclaws are enlarged. Ul-3 typically are relatively narrow and concave to very concave on the posteroventral mar- gin, rather than cone-shaped with a straight or convex posteroventral margin. The an- terior element of ectoloph of Ml is re- duced relative to the posterior element, rather than being of approximately equal size. The anterior border of the coronoid process of the mandible joins the horizon- tal ramus at a relatively low angle, rather than at a high angle. The articular process is relatively high and narrow, rather than low and broad. The hum?rus in the C. tho- masi-gToup is moderately long, quite ro- bust, and slightly curved with a prominent medial epicondyle and a well-defined teres tubercle that is more centrally located. The hum?rus in the C. nigrescens-gioup is sim- ilar to that of the C parva-gioup: it is long, straight, and narrow, with the teres tubercle more proximally located. I consider Cryptotis thomasi s. s. to be restricted entirely to the central portion of the Eastern Cordillera of Colombia around Bogot?, where it is known only from ele- vations above 2700 m (Woodman 1996). This region has a long history of scientific collection, and, for shrews, it is one of the better represented parts of the Andes. The southern limit of C. thomasi's distribution probably is north of the Paso de las Cruces (3?17'N, 74?37'W), which drops below 1900 m. There are no specimens known from between the northern border of Cun- dinamarca Department and northern San- tander Department, and the northern limit of the species may be associated with head- water valleys and tributaries of the Rio Chi- camocha and Rio Upia. Cryptotis medellinia is distinct from C. thomasi and is known mainly from the northern portion of the Central Cordillera of Colombia. More recently, I identified specimens of this species from the north- ern part of the Western Cordillera of Co- lombia as well. Its elevational range is from about 2000 to 3800 m. The southern limits of its distribution in each of the cor- dilleras is uncertain, but no specimens are known from south of Paso de Quind?o (= Boquer?n de Quind?o, 4?38'N, 75?32'W) in the Central Cordillera, nor south of P?r- amo Frontino (6?28'N, 76?04'W) in the Western Cordillera. Cryptotis tamensis, new species Figs. 1, 3-7 Blarina meridensis: Osgood, 1912:62. C[ryptotis] meridensis: Thomas, 1921:354 (in part). Cryptotis- thomasi meridensis: Cabrera, 1958:48 (in part). Cryptotis meridensis meridensis: Linares, 1998:106 (in part). Holotype.?Skin, skull (Fig. 1), and flu- id-preserved body of adult female (lactat- ing, gravid: 2 fetuses, CR = 19), USNM VOLUME 115, NUMBER 2 A 255 Fig. 1. Holotype of Cryptotis tamensis (USNM 418567): (A) palatal view of cranium; (B) lateral view of the cranium and mandible; and (C) posterior view of articular process of the mandible. 418567, taken in a snap trap baited with sardines, rolled oats, and raisins on 8 March 1968 by a Smithsonian Venezuelan Project (SVP) team led by Norman E. Peterson, field number SVP 21863. Skin in good con- dition, some matting of fur; skull has well worn dentition, lacking left tympanic, some damage to right pterygoid; mandible lacks both angular processes. Standard field mea- surements: total length =118; length of tail = 35; length of hind foot = 17; length of ear slit = 8; weight = 15.9 g. Type /oca/??y.?VENEZUELA: State of T?chira: Buena Vista, 7?27'N, 72?26'W, 2415 m; near P?ramo de Tama; 35 km S, 22 km W of San Crist?bal [reported as "41 km SW San Crist?bal and 12 km SSE Las Delicias" by Handley 1976:64]. Referred specimens (21).?COLOMBIA: Norte de Santander: P?ramo de Tama, head of R?o T?chira [7?25'N, 72?26'W, 2450 m?Paynter & Traylor 1981] (FMNH 18571, 18614, 18615); P?ramo de Tama [7?25'N, 72?26'W, 3329 m?Paynter & Traylor 1981] (FMNH 18572, 18608- 18611, 18613, 18621; MCZ 21004 [= 256 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON Fig. 2. Map of the northern Andes showing the distributions of four species of Cryptotis in Colombia and Venezuela: C. medeUinia (closed squares); C. meridensis (dotted open circles); C. tamensis (closed circles); C. thomasi (closed triangles). Question marks indicate the locations of two reported collections of Cryptotis whose specific identities are unknown (see text). The 1000 m contour is shown. FMNH 18616]; USNM 260747 [= FMNH 18612]). Santander, above Surat? [7?22'N, 73?00'W?USBGN 1988], 2750 m (LACM 56101). VENEZUELA: T?chira: Buena Vista, 7?27'N, 72?26'W, 2380-2415 m [type locality] (USNM 418566, 418569, 418570, 490534); P?ramo de Tama [7?25'N, 72?26'W?USBGN 1988] (FMNH 18617-18620). Additional record.?COLOMBIA: San- tander: Finca El Rasg?n, 6?55'30"N, 73?02'40"W, 2350 m. Vereda La Cristalina, Municipio de Piedecuesta (MUA no num- ber). Etymology.?Tam(?) + ensis [Latin end- ing denoting place], "from Tama." Distribution.?Montane Rain Forest, Montane Wet Forest, Lower Montane Wet Forest, disturbed "cloud forest," and pas- ture margins, from the Tama highlands of western T?chira, Venezuela, and southeast- em Norte de Santander Dept., Colombia, to northeastern Santander Dept., Colombia (Fig. 2); known elevational distribution, 2385-3329 m. Also may occur in Lower Montane Rain Forest and Lower Montane Moist Forest life zones. Diagnosis.?A large Cryptotis with long. VOLUME 115, NUMBER 2 257 chocolate-brown dorsal pelage that con- trasts subtly with paler, olive-brown venter; moderately long tail, averaging 42% of head-and-body length; enlarged forepaws and elongate foreclaws; indistinct lateral glands. Distinguished within C. thomasi- group by possessing bulbous dentition; re- duced U4 (averaging 29% of U3), but hav- ing four unicuspids visible in lateral view of cranium; slender Ul-3 with concave postero ventral margins; low coronoid pro- cess; short posterior mandible behind m3; simple M3; often an obvious foramen on posterior edge of tympanic process of one or both petromastoids; distinct pigmenta- tion in protoconal basin (occasionally pale pigmentation in hypoconal basin) of P3, M1-2; shallow inferior sigmoid notch; no obvious entoconid on m3. Description.?A large Cryptotis, head- and-body length averaging 86 mm (Table 1). Tail moderately long, averaging 42% of head-and-body length (Table 2). Dorsal hairs about 6-7 mm long. Dorsal pelage Mummy Brown to Clove Brown; venter Ol- ive Brown to Buffy Brown; dorsal hairs 2- banded, with grey base and mid-section, and 1-1.5 mm brown distal tip. Forepaws somewhat enlarged (Fig. 3). Foreclaws sim- ilar to those of C thomasi: elongate, but not broadened; similar in length to those of C phillipsii and C obscura, but relatively nar- rower. Lateral glands of males indistinct. Usually two obvious dorsal foramina along the suture between the fron tais (95%, n = 19); generally unequal in size (67%). Typically lacking lateral branch of sinus ca- nal and associated foramen (see Woodman & Timm 1999) posterior to the dorsal artic- ular facet (94%, n = 18); this foramen was found on only one specimen (USNM 418566), in which it is well developed, but present only on one side of the cranium. A foramen often present dorsal to one or both dorsal articular facets (44%, n = 18); some- times only on one side (18%, n = 18) of the cranium; typically minute, but can be well developed. Typically a foramen on the posterior edge of the tympanic process of one or both petromastoids (88%, n = 16 individuals; Fig. 4); typically on both tym- panic processes (73%, n = 15 individuals); foramen usually small (71%, n = 24 foram- ina), but can be medium-sized and obvious (29%); often varies in size between left and right sides of a single individual (46%, n = 13; 38% of all individuals have at least one foramen that is medium-sized or larger, n = 16; Fig. 4); foramen never as large as in C thomasi or C medellinia. Rostrum of moderate length (PL/CBL = 44.0%). Inter- orbital area moderately broad (lO/CBL = 23.2%). Zygomatic plate of moderate breadth in proportion to CBL (9.6%) and PL (21.9%); anterior border of zygomatic plate at mesostyle-metastyle valley to me- tastyle of Ml; posterior border from metas- tyle of M2 to parastyle-mesostyle valley of M3, and from about middle to posterior edge of maxillary process. Palate of mod- erate breadth for the genus (M2B/PL = 68.0%). Anterior process of petromastoid low and narrow. Dentition bulbous. Teeth moderately pig- mented: dark red to nearly black on tips of cones, styles, and cristae; pale to medium pigment typically extends into protoconal basin (and occasionally into hypoconal ba- sin) of Ml and M2. Unicuspid toothrow of moderate length (UTR/CBL = 13.2%); un- crowded, U4 generally visible in lateral view of skull (74%, n = 19; Fig. 5). Ul-3 relatively slender and concave on the pos- teroventral margin. Posterolingual cuspules typically minute (61%, n = 18) or obvious (39%) on cingulum of Ul-3. U4 reduced, averaging 29% the surface area of U3 (Fig. 6). Posterior borders of P4, Ml, M2 only slightly recessed. Anterior element of ec- toloph of Ml reduced relative to posterior element; protoconal basin of Ml about equal in size to hypoconal basin. M3 gen- erally appears simple: hypocone usually ab- sent; postcentrocrista shortened; metacone minute (50%, n = 16) and often lightly pig- mented, or absent (50%); reduced hypocone occasionally present, resulting in a larger, more squared M3 that appears complex. 258 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON Table 2.?Characters of selected Colombian and Venezuelan Cryptotis. C. meridensis C. medellinia Petromastoid foramen present one or both petromastoids 87% 88% (n = 55) (n = 16) 72% both 73% both 28% one 27% one (n = 46) (n = 15) Postero-lingual cuspules on cingulae of Ul-3 100% (n = 27) 100% both 100% (n = 14) 100% both 78% absent 61% minute 30% absent 15% absent 22% minute 39% obvious 30% minute 70% minute (n = 40) (n = 18) 40% obvious 15% obvious (n = 10) (n = 13) U4 lacking 25% 0% 0% 0% (n = 51) (n = 19) (n = 13) (n = 12) Size of U4 (% of U3) 19 ? 7% 29 ? 8% 50 ? 9% 40 ? 9% (2-37) (12^2) (36-64) (27-53) (n = 48) (n = 19) (n = 13) (n = 12) Emargination of posterior borders of P4, Ml V. s?ght V. slight slight to V. slight to slight to slight moderate to slight Complexity of M3 98% simple 50% simple 8% simple 73% simple 2% minute 50% complex 92% complex 27% complex (n = 45) (n = 16) (n = 13) (n= 11) Lower sigmoid notch shallow shallow shallow to V. shallow mod. deep Entoconid of m3 present (absent, minute, obvious) 15% minute 32% minute 30% obvious 53% minute 85% absent 68% absent 70 absent 47% absent (n = 41) (n = 19) (n = 23) (n = 15) Relative tail length (TL/HB X 100) 38 ? 4 42 ? 3 29 ? 4 43 ? 6 28^9 35^8 21-36 34-56 (n = 65) (n = 21) (n = 22) (n = 13) Relative length of rostrum (PL/CBL X 100) 44.2 ? 0.9 44.0 ? 1.4 43.2 ? 0.6 44.3 ? 0.6 40.9^5.3 40.2^5.6 42.2^14.3 43.4-^5.3 (n = 35) (n = 13) (n = 23) (n = 10) Relative breadth of zygomatic plate (ZP/PL X 100) 21.2 ? 2.0 21.9 ? 1.4 21.3 ? 1.7' 21.2 ? 2.1 15.8-25.0 19.6-25.0 18.5-25.3 17.7-24.0 (n = 50) (n = 17) (n = 27) (n = 10) Relative breadth of zygomatic plate (ZP/CBL X 100) 9.5 ? 0.9 9.6 ? 0.6 9.3 ? 0.6 9.4 ? 1.0 7.1-10.9 8.7-10.8 7.8-10.6 7.7-10.6 (n = 35) (n = 13) (n = 23) (n = 10) VOLUME 115, NUMBER 2 259 Table 2.?Continued. C. meridensis C. tamensis C. medellinia Breadth of interorbital area (lO/CBL) 23.5 ? 0.8 23.2 ? 0.7 22.0-24.9 22.0-24.4 (n = 35) (n = 13) Relative length of unicuspid toothrow (UTR/CBL X 100) 13.1 ? 0.5 13.2 ? 0.5 12.2-14.3 12.3-13.9 (n = 35) (n = 13) Relative palatal breadth (M2B/PL X 100) 67.2 ? 2.1 68.0 ? 2.0 62.5-72.7 64.3-71.0 (n = 50) (n = 17) Relative height of coronoid process (HCP/ML X 100) 69.8 ? 2.5 65.0 ? 2.0 63.4-76.2 62.3-68.1 (n = 50) (n = 17) Relative length of posterior portion of the mandible (AC3/ML 81.0 ? 2.6 77.2 ? 1.8 73.7-86.6 75.0-80.9 (n = 50) (n = 17) 23.0 ? 1.0 24.3 ? 0.6 20.6-24.6 23.3-25.4 (n = 23) (n = 10) 13.1 ? 0.5 13.5 ? 0.4 11.8-14.0 12.9-14.2 (n = 23) (n = 10) 66.6 ? 2.5 68.3 ? 2.4 60.8-70.5 64.5-71.7 (n = 27) (n = 10) 66.0 ? 2.2 68.5 ? 2.5 60.6-70.6 64.3-72.6 (n = 27) (n = 17) X 100) 80.9 ?3.0 75.3 ? 3.0 71.4-86.8 71.3-78.1 (n = 27) (n = 17) Coronoid process low (HCP/LM = 65.0%); anterior border joins the horizontal ramus at a relatively low angle. Inferior sig- moid notch variable, but typically shallow; not as shallow as in C nigrescens, in which the notch typically is absent. Posterior man- dible behind m3 moderately long (AC3/ML = 77.2%). Articular process relatively high and narrow. Third lower premolar long and low. Entoconid minute (32%, n = 19) or absent (68%) from talonid of m3. Hum?rus similar to that of C thomasi: moderately long, robust, slightly curved; the narrowest portion of the shaft is broader in anterior aspect than in lateral aspect; head more or less rounded; proximal and distal ends broad; medial epicondyle and teres tubercle prominent (Fig. 7). Compared to members of the C. nigrescens-gioup and C. parva-gToup, the hum?rus is shorter rel- ative to overall body size, more robust, and more curved, and teres tubercle more cen- trally located. Posterior edge of the falci- form process of the tibia deeply pocketed. Comparisons.?Cryptotis meridensis: Cryptotis tamensis is most readily distin- guished by having U4 present, much less reduced, and typically visible in lateral view of cranium; lower coronoid process relative to length of mandible (Table 2); and shorter posterior mandible (behind m3). In addition, C. tamensis averages smaller in head-and-body length (Table 1) and has a relatively longer tail; more commonly has an obvious foramen on the posterior edge of the tympanic process of the petromastoid (Fig. 4); dental pigmentation tends to be more extensive (pigment typically darker in hypoconal basin and more commonly found in protoconal basin); small posterolingual cuspule more commonly present (and more distinct) on cingulum of Ul, U2, or U3; M3 occasionally complex. Cryptotis thomasi: Cryptotis tamensis is most immediately and easily distinguished by its longer tail (Table 2); smaller, less evenly rounded foramen (when present) on the tympanic process of the petromastoid (Fig. 4); and more reduced U4. The new species also has a relatively longer rostrum 260 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON 1 mm Fig. 3. Ventral views of forepaws of C. tamensis: (A) right forepaw of a fluid-preserved specimen (USNM 490534) showing plantar pads, but with atypically short claws; (B) left forepaw from dried skin (USNM 418570) showing typically long claws. and broader palate (Table 2); less emargin- ate upper dentition; slightly deeper inferior sigmoid notch; shorter posterior mandible; less obviously complex M3. Cryptotis medellinia: Cryptotis tamensis is most easily distinguished by its shorter tail (Table 2); smaller, less evenly rounded foramen (when present) on the tympanic process of the petromastoid; and more re- duced U4. It also has a relatively narrower interorbital area (Table 2); relatively shorter unicuspid toothrow; relatively lower coro- noid process; shallower inferior sigmoid notch. Results of multivariate analysis.?Princi- pal components analysis (PCA) of craniom- andibular variables from C meridensis, C. tamensis, and C. thomasi yielded greatest separation on factors axes 2 and 3 (Fig. 8; Table 3). The three species overlap each oth- er to a considerable and similar extent on both axes, despite the ease with which C. thomasi can be distinguished from the other two species using both external and cran- iomandibular characters (see comparisons). The broad intersections among the three groups reflect the generally conservative shape of the skull within the genus Cryptotis VOLUME 115, NUMBER 2 261 Fig. 4. Foramen on the posterior edge of the tympanic process of the petromastoid: (A) posterior view of cranium, illustrating the location and size of the foramen; (B) a minute foramen in C. meridensis; (C) a small foramen in C. meridensis; medium foramina in (D) C. meridensis and (E) C. tamensis; (F) a huge foramen typical of C. thomasi. in general and among members of the C thomasi-gioup in particular (note results of similar analyses for other species of Cryp- totis in Woodman and Timm 1993, 1999, 2000). A plot of scores on factor axes 1 and 2 from PCA of C. meridensis and C. tamensis supports the separation of the two species (Fig. 9). This plot shows the lack of differ- ence in overall craniomandibular size (fac- tor axis 1?Table 4) despite the larger head- and-body-length of C. meridensis (Table 1). Instead, the two species are separated by the combined effects of several, mostly man- dibular, variables that include HCV, HCP, and AC3. Remarks.?Based on the distribution of unpolarized, shared characteristics (e.g., re- duction of U4) and on overall similarity, C. tamensis and C. meridensis appear to be more closely related to each other than ei- ther is to any other species. This opinion contrasts with that of Hutterer (1986, 1993) and Soriano et al. (1999), who inferred that populations in the region of P?ramo de Tama were conspecific with (hence, more closely related to) C. thomasi. Cryptotis thomasi and C medellinia nearly always have a very large, rounded foramen on the tympanic process of both petromastoids (an apomorphic character within the C. tho- wasi-group), and most likely, the two are sister species. In commenting on P?ramo de Tama, 262 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON 1 mm Fig. 5. Lateral view of upper first incisor and unicuspids of Cryptotis: (A) fourth unicuspid not readily visible, typical for C. meridensis; (B) fourth unicuspid visible (arrow), typical for C. tamensis. where he worked from 12 February to 6 March 1911, Osgood (1912) noted that the word "p?ramo" was used to designate the entire highland area (rather than the open, high-elevation vegetational formation with which it is more commonly associated to- day). In fact, there appears to be no true p?ramo vegetation present in this area, and most specimens were collected in dense cloud forest: The large forest trees, all of unfamiliar species, lace their tops together at an average height of some- thing over 100 feet. Mists and drizzling rains prevail a great part of the year and mosses, air plants, and climbing vines thrive, but the growth in general is not so profuse as that of the hot tropics. A beautiful climbing bamboo is exceedingly abundant, and graceful tree ferns raise their spreading tops at pleasing intervals (Osgood 1912:36). Seven of the eight specimens of C. ta- mensis captured by the Smithsonian Vene- VOLUME 115, NUMBER 2 263 B 1 mm Fig. 6. Occusal view of unicuspid toothrows of Cryptotis illustrating size variation of U4. Percentages represent approximate surface area of U4 relative to U3. The U4 of C. tamensis (29 ? 8%) averages larger than that of C. meridensis (19 ? 7%) and smaller than that of C. thomasi (50 ? 9%): (A) U4 ca. 10% of U3 (C. meridensis, USNM 579288); (B) 20% (C. meridensis. USNM 579296); (C) 25% (C. tamensis. USNM 418570); (D) 38% (C. tamensis, FMNH 18621); (E) 49% (C. thomasi. USNM 80906). zuelan Project at Buena Vista were taken in cloud forest and one in a cut and burned pasture bordered on all sides by cloud forest (N. E. Peterson 1968 field notes, USNM). Handley (1976:64-65) described Buena Vista as: Hilly terrain on north facing slope at head of large, wide Andean valley. Numerous small (2-6 m wide), swift, rocky streams; Rio T?chira nearby; occasion- al swampy and marshy areas. Mature cloud forest with discontinuous canopy of scattered trees 25-30 m high; subcanopy at 10-15 m, open or closed; many tree ferns, stilt palms, and thick clusters of tree bamboo; vines few and thin, or hanging from trees in abundance; shrub stratum 1.5 m high; moss and other epiphytes very abundant on trees and ground; abundant litter of logs, fallen trees, dead bamboo, and leaves. Virtually impenetrable clumps of vinelike bamboo dominant (choking out other vegetation), scattered throughout the forest, around fields, and forming thick canopies over streams. Clearings (formed by cutting and burning forest) with thick grass and clover, patches of neddielike rushes, and clumps of ferns, herbs, low shrubs, and Fig. 7. Anterior aspect of left humeri of Cryptotis: (A) C. orophila (KU 142693), (B) C. meridensis (USNM 385102), (C) C. tamensis (FMNH 18621), and (D) C. thomasi (FMNH 71027). 264 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON CO 1 H O ^ 0- O C. meridensis ? C. tamensis A C. thomasi ?r -3 ?r -2 0 factor 2 Fig. 8. Plot of scores on factor axes 2 and 3 from principal components analysis (PCA) of 18 cranioman- dibular measurements (Table 3) from C. meridensis, C. tamensis, and C. thomasi. blackberries. Collecting at upper edge of agricultur- al clearing (pasture and cropland); cloud forest con- tinuous on steep slopes from this point up to ridg- etop p?ramo, 5 hours' walk distant. Reproductive data for C. tamensis are scarce. Gravid females were taken on 2 March (2 fetuses, no measurements, USNM 418565; 1 fetus, CR = 13 mm, USNM 418566) and 8 March (2 fetuses, CR = 19, USNM 418567). A lactating female was captured on 25 March (USNM 418570). Cryptotis meridensis (Thomas, 1898) Figs. 4-7 Blarina meridensis Thomas, 1898:457. C[ryptotis^. meridensis: Thomas, 1921:354 (in part). Cryptotis thomasi meridensis: Cabrera, 1958:48 (in part). Cryptotis meridensis meridensis: Linares, 1998:106 (in part). Holotype.?Skin and skuU, British Mu- seum no. 98.5.15.5. Type locality.?"Montes del Valle Mer- ida 2165 m," M?rida State, Venezuela. This locality is from the skin tag attached to the holotype. Thomas (1898:45) reported the type locality simply as "Merida, alt. 2165 m." Paynter (1982) equated Montes del Valle with "Valle" [ca. 8?40'N, 71?06'W]. See his discussion of possible locations; also may be "Valle Grande," a valley at 8?40'N, 71?08'W (USBGN 1993). Distribution.?Cloud forest and p?ramo in the Cordillera de los Andes in Trujillo, M?rida, and eastern T?chira, Venezuela (Fig. 2); reported elevational range is 1640? 3950 m. Durant & Pefaur (1984:7) gave the VOLUME 115, NUMBER 2 265 Table 3.?Factor loadings for the first three factor axes from principal components analysis of (PCA) of 18 craniomandibular measurements from Cryptotis meridensis, C. tamensis, and C. thomasi. Variables are listed in descending order by their loadings on the second axis. Abbreviations of measurements are explained in Table 1. Correlations Variable PCI PC2 PC3 AC3 -0.213 0.500 0.028 BAC -0.253 0.330 -0.028 HCP -0.249 0.204 -0.350 ML -0.249 0.187 0.247 CBL -0.284 0.173 0.204 HAC -0.262 0.158 -0.230 HCV -0.212 0.101 -0.496 PL -0.286 0.079 0.080 lO -0.164 0.042 0.142 ZP 0.067 0.023 0.342 TR -0.253 -0.085 0.312 UTR -0.219 -0.103 0.232 MTR -0.279 -0.173 0.158 Lml -0.205 -0.234 -0.034 TRM -0.273 -0.252 0.163 M2B -0.259 -0.299 -0.129 UIB -0.217 -0.324 -0.010 MIW -0.196 -0.361 -0.333 Eigenvalue: 9.0633 1.7975 1.3586 Proportion of variation explained: 50.4% 10.0% 7.5% elevational limits as 1800-3600 m, al- though they also reported a specimen from 1640 m at Rio Santo Domingo. Otherwise, the lowest reported records are at 1890 m at Rio El Chama site E-II and Rio Motat?n site E-II (Durant & D?az 1995). No subspe- cies are recognized. Description.?Among the largest-bodied Cryptotis, head-and-body length averaging 89 mm (Table 1). Tail of moderate length for the genus, averaging 38% of head-and- body length (Table 2). Dorsal hairs about 4-6 mm long. Dorsal pelage Olive Brown, Mummy Brown to Clove Brown; venter Ol- ive Brown to Buffy Brown; dorsal hairs 2- banded, with grey base and mid-section, and 1?1.5 mm brown distal tip. Forepaws somewhat enlarged, similar to those of C. tamensis (Fig. 3). Foreclaws elongate, but not broadened; similar overall to C. thomasi and C. tamensis. Lateral glands of males indistinct. Usually two well-developed dorsal fo- ramina (72%, n = 39), typically of similar size (64%). Usually lacking lateral branch of sinus canal and associated foramen pos- terior to the dorsal articular facet (94%, n = 53); this foramen minute when present, and may be on only one side of cranium. Usually no foramen dorsal to dorsal artic- ular facet (81%, n = 53); when present this foramina can be well-developed, but is more typically minute and only on one side of the skull (70%, n = 10). Typically a fo- ramen on the posterior edge of the tympanic process of one or both petromastoids (87%, n = 55 individuals; Fig. 4); typically on both petromastoids (72%, n = 46 individ- uals); foramina usually quite small (80%, n = 80 foramina), but can be of medium size and obvious (20%); often varies in size be- tween left and right sides of a single indi- vidual (30%, n = 33; 20% of all individuals have at least one obvious foramen that is medium-sized or larger, n = 55; Fig. 4); foramen never as large as in Cryptotis tho- masi or C. medellinia. Rostrum of moderate length (PL/CBL = 44.2%). Interorbital area 266 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON CM 0- O 45-1 . , 1 1 < i , o 0 0 0 0 0 1 1 1 1 1 1 1 . 1 i 1 1 1 , o o o cp 0 0 0 ? 0 0 0 0 ? 0 ? ? > ? 0 0 ? 0 0 ? ? 0 ? ? ? ? ? ? ? 0 C. mer?densis '- ? C. tamensis ' 1 1" -^ ?T--^--r--'? ' 1 1 1 1 1 1 1 -9-8-7-6-5-4-3-2-1 O 1 2 3 4 5 factor 1 Fig. 9. Plot of scores on factor axes 1 and 2 from PCA of 18 craniomandibular measurements (Table 4) from C. meridensis and C. tamensis. moderately broad (lO/CBL = 23.2). Zy- gomatic plate of moderate breadth in pro- portion to CBL (9.5%) and PL (21.2%); an- terior border of zygomatic plate at mesos- tyle-metastyle valley to metastyle of Ml; posterior border from mesostyle-metastyle valley of M2 to parastyle of M3, and from about middle to posterior one-half of max- illary process. Palate of moderate breadth (M2B/PL = 67.2%). Anterior process of petromastoid typically low and narrow. Dentition bulbous. Teeth moderately pigmented: dark red to nearly black on tips of cones, styles, and cristae; pale to me- dium color often extending into protoconal basin (but not hypoconal basin) of Ml and M2. Unicuspid toothrow of moderate length (UTR/CBL =13.1%) and uncrowd- ed; however, U4 generally not visible in lateral view of skull (73%, n = 44; Fig. 5), obscured from view by U3 and P4. U4 of- ten missing on one (19%; n = 51) or both (6%) sides of cranium; reduced when pre- sent, averaging 19% the surface area of U3 (Fig. 6); U4 often a simple, slender, round- ed peg (38%, n = 48), occasionally smaller in diameter than the root of a neighboring unicuspid. The tendency to lose U4 is more prevalent than observed in Cryptotis phillipsii (12%?Woodman & Timm 2000). Ul-3 relatively slender and con- cave on the posteroventral margin. Poster- olingual cuspules typically absent (81%, n = 37) from cingulum of Ul?3, or minute (19%). Posterior borders of P4, Ml, M2 unrecessed or slightly recessed. M3 com- monly appears simple: hypocone usually absent (84%, n = 45), postcentrocrista VOLUME 115, NUMBER 2 267 Table 4.?Factor loadings for the first two factor axes from principal components analysis of (PCA) of C. mer- idensis and C. tamensis. Variables are listed in descend- ing order by their loadings on the second axis. Abbre- viations of measurements are explained in Table 1. Correlations Variable PCI PC2 HCV -0.188 0.496 HCP -0.232 0.422 AC3 -0.239 0.340 BAC -0.267 0.199 HAC -0.259 0.192 PL -0.270 0.028 ro -0.168 -0.002 Lml -0.208 -0.011 MIW -0.226 -0.034 CBL -0.288 -0.039 ZP 0.117 -0.053 ML -0.267 -0.077 M2B -0.261 -0.154 MTR -0.270 -0.198 UTR -0.211 -0.243 TR -0.237 -0.245 TRM -0.260 -0.305 UIB -0.208 -0.310 Eigenvalue: 10.021 1.744 Proportion of variation explained: 55.7% 9.7% shortened, metacone absent (76%, n = 34) or minute (24%). Coronoid process of moderate height (HCP/LM = 69.8%). Inferior sigmoid notch variable, but typically shallow; simi- lar to that of C. tamensis. Posterior portion of mandible (behind m3) quite long (AC3/ ML = 81.0%). Articular process relatively high and narrow. Third lower premolar long and low. Entoconid of m3 absent (85%, n = 41) or minute (15%). Hum?rus similar to those of C. thomasi and C. tamensis: moderately long, robust, slightly curved; the narrowest portion of the shaft is broader in anterior aspect than in lateral aspect; head more or less rounded; proximal and distal ends broad; medial epi- condyle and teres tubercle prominent (Fig. 7). Posterior edge of the falciform process of the tibia deeply pocketed. Comparisons.?Cryptotis meridensis is a large member of the genus with long, choc- olate-brown dorsal pelage contrasting sub- tly with paler, olive-brown venter; moder- ately long tail; enlarged forepaws and elon- gate foreclaws. Within the C. thomasi- group, C. meridensis is distinguished by having bulbous dentition; absent or reduced U4 (averaging 19% of U3); U4 not visible in lateral view of cranium; slender Ul-3 with concave posteroventral margins; high coronoid process; long posterior mandible behind m3; simple M3; occasionally an ob- vious foramen on posterior edge of tym- panic process of one or both petromastoids; distinct pigmentation in protoconal basin (but not hypoconal basin) of P3, Ml?2; shallow inferior sigmoid notch; no entoco- nid on m3. Cryptotis thomasi: Cryptotis meridensis is most readily distinguished by its longer tail (Table 2); and smaller foramen (when present) on the tympanic process of the pe- tromastoid; absent or more reduced U4; and relatively higher coronoid process. Crypto- tis meridensis averages larger in head-and- body length and has a relatively longer ros- trum and broader palate; less extensive dis- tribution of dental pigmentation (in C. tho- masi, the protoconal basin and hypocone typically are darkly pigmented, and the hy- poconal basin has pale- to medium-red pig- mentation); slightly deeper lower sigmoid notch; less complex M3. Cryptotis medellinia: Cryptotis meriden- sis is most easily distinguished by its short- er tail (Table 2); smaller foramen (when present) on the tympanic process of the pe- tromastoid; more reduced or absent U4; and relatively shorter posterior portion of the mandible. It averages larger in head-and- body length; has a relatively narrower in- terobital area; relatively shorter unicuspid toothrow; shallower inferior sigmoid notch. Remarks.?Durant & P?faur (1984) not- ed differences in pelage color and texture between a series of C. meridensis from high elevation p?ramo (3300-3600 m) and those from cloud forest (1640-2500 m). They de- scribed the former as having longer, softer, paler, two-banded hairs on the dorsum and 268 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON a distinctly paler venter than dorsum. Shrews from cloud forest had shorter, coarser, unicolor dorsal hairs with more uniform dorsal and ventral coloration. So- riano et al. (1990) were unable to corrob- orate these differences. Among the dried skins I inspected, specimens from p?ramo and "Coloradito" (see below) vegetational associations typically had long (6-7 mm), pale dorsal hairs and a paler venter. Indi- viduals from cloud forest and disturbed cloud forest included individuals with shorter (4-5 mm), darker dorsal hairs and less contrast between the dorsum and ven- ter, but also included individuals indistin- guishable from p?ramo and Coloradito specimens. I had insufficient specimens to determine whether the variation might be attributable to seasonal changes in the en- vironment. Overall, the differences in pel- age are subtle, and variation in coloration among museum specimens may be con- founded by foxing and molt. Cryptotis meridensis primarily occupies cloud forest and p?ramo vegetational as- semblages, in which it is relatively abun- dant. Four of the 13 specimens of C. mer- idensis captured by the Smithsonian Vene- zuelan Project were taken in cloud forest, four in disturbed (previously burned) cloud forest, one in a thicket of bamboo and low trees bordering Quebrada Mucuy, one in previously burned cloud forest-paramo transition, and three in p?ramo (N. E. Pe- terson 1966 field notes, USNM). These re- cords came from three Holdridge life zones: Lower Montane Moist Forest, Montane Rain Forest, and Subalpine P?ramo (Han- dley 1976). Similarly, Durant & D?az (1995) recorded C. meridensis at fiye sites in p?ramo, Montane Rain Forest, and Low- er Montane Rain Forest. Aagaard (1982) found C. meridensis to be most abundant in Subalpine P?ramo, Subalpine Rain P?ramo, and a vegational assemblage he identified as "Coloradito," consisting of stands of ar- borescent Polylepis sericea and Espeletia humertii in sheltered areas surrounded by p?ramo vegetation. In these three assem- blages, shrews were the most common or second most common species of small mammal captured using Museum Special mouse traps and Victor rat traps during a nine-month period. Individual shrews also were taken in Montane Wet Forest and Lower Montane Wet Forest, but not in Montane Rain Forest. Diaz de Pascual (1993, 1994) found C. meridensis to aver- age second most common of eight species of small mammals taken in pitfall traps dur- ing during the first four years of a study in cloud forest at Monte Zerpa. Durant & D?az (1995) showed two cap- ture peaks during the year, in March-April and September-October, coinciding with the time periods leading up to biannual peaks in rainfall in many parts of the Ve- nezuelan Andes (April-May and October; see Diaz de Pascual 1994, Durant & D?az 1995:89, 90, table 5, fig. 3; Aagaard 1982: 7, fig. 1). The two peaks in abundance of C. meridensis suggest two periods of repro- duction and/or two periods of increased ac- tivity during the two seasons of greatest rainfall. In contrast, Aagaard (1982) report- ed significantly more captures of C. meri- densis in dry months [January (? = 0 cap- tures), February (15), March (7), July (5), August (2)] than in wet months [April (6), May, (0), June (0), September (0)]. Little has been reported on reproduction in C. meridensis. A lactating female (USNM 385106) was captured on 15 April 1966 at Middle Refugio. Non-gravid fe- males have been taken in February (? = 5), March (? = 5), April (? = 4), July (? = 2), and August (? = 2). Aagaard (1982) re- ported one gravid female from 16 July (4 fetuses, crown-to-rump-length =11 mm). Based on testes sizes, he suggested that males were sexually active during both wet and dry months. Specimens examined (62).?VENEZUE- LA: M?rida: near Laguna Mucubaj? [ca. 8?48'N, 70?48'W?Paynter 1982], 3600 m, 3.25 km ESE of Apartaderos (USNM 579287-579298); near Laguna Negra [ca. 8?48'N, 70?46'W?Paynter 1982], 3500 m, 269 5.75 km ESE of Apartaderos (USNM 579273-579286); P?ramo de [La] Culata [ca. 8?45'N, 71?05'W?Paynter 1982], 9000 ft, R?o Mucuj?n (FMNH 21837, 21838); Montes de La Culata [La Culata, ca. 8?45'N, 71?05'W?Paynter 1982], 2000 m, M?rida (BMNH 98.7.1.13); Rio Mucu- j?n [8?45'N, 7r04'W?from map], 9000- 12,500 ft [Osgood & Conover 1922:23] (FMNH 21839, 21840, 21842, 21844; BMNH 29.11.7.16 [= FMNH 21841]; USNM 260756 [= FMNH 21843]); Montes del Valle [ca. 8?40'N, 71?06'W?Paynter 1982], 2125-2165 m, M?rida (BMNH 98.7.1.11, 98.7.1.12, 98.5.15.5?holotype); P?ramo Tambor [at Hacienda La Carbonera (8?38'N, 7r32'W?Paynter 1982), near head of a western branch of R?o Guachi (Osgood & Conover 1922:23)] (FMNH 21845); near Santa Rosa, 8?38'N, 7r09'30"W, 1980-1990 m, 1 km N, 2 km W of M?rida [reported as "1-2 km N M?- rida" by Handley 1976:83; Handley (1976: 10) incorrectly reported USNM 385110 as from "La Carbonera, 12 km SE La Azulita, 1990 m"] (USNM 385110-385112); near La Mucuy [8?38'N, 71?02'30"W?from map], 2450 m, 2.9 km E of Tabay (USNM 579305); area 37, 8?37'00"N, 71?0r30"W, 2630 m, near Middle Refugio, 2 km S, 5.5 km E of Tabay [reported as "5 to 9 km SE and ESE Tabay," and "6 km ESE Tabay" by Handley 1976:10, 77] (USNM 385106); M?rida [8?36'N, 71?08W?Paynter 1982], 2165 m (USNM 94165?"paratype 0[ldfield] T[homas]" written on skin tag); La Monta?a, 2250 m, 3.1 km SE of M?rida (USNM 579307); La Aguada, near Laguna La Fr?a [P?ramo La Frfa, ca. 8?36'N, 71?02W?Paynter 1982], 3600 m, 7 km SE of M?rida (USNM 579299-579304); near Loma Redonda cable car station, 4100 m, 8.8 km SE of M?rida (USNM 579306); El Tambor [P?ramo Tambor, 8?36'N, 71?24'W?Paynter 1982], 8800 ft (AMNH 96156-96158); P?ramo de los Conejos [ca. 8?35'N, 7ri4'W?from map], 9600 ft (AMNH 96159); La Coromoto, 8?35'45"N, 71?01'W, 3160-3175 m, 4 km S, 6.5 km E of Tabay [reported as "5 to 9 km SE and ESE Tabay," and "7 km SE Tabay" by Handley 1976:10, 75] (USNM 385101, 385104); near Laguna Verde, 8?34'30"N, 7rOO'30"W, 3533-3545 m, 7.5 km E, 6 km S of Tabay [reported as "5 to 9 km SE and ESE Tabay," and "9 km SE Tabay" by Handley 1976:10, 76] (USNM 385102, 385103, 385105). Additional records.?VENEZUELA: Merida: Rio Motat?n site E-I, 8?49'N, 70?59'W, 3890 m (Durant & D?az 1995); P?ramo de Mucubaj?, 8?48'06"N, 70?48'57"W, 3420-3800 m. Distrito Rangel (Durant & Pefaur 1984); R?o Chama site E- I, 8?48'N, 70?48'W, 3950 m (Durant & D?az 1995); Monte Zerpa cloud forest, 8?37'N, 71?10'W, 2000-2800 m, Sierra del Norte de La Culata (D?az de Pascual & de Ascen?ao 2000); area 51, 8?37'45"N, 71?01'30"W, 2640 m, near Middle Refugio, 1 km S, 5 km E of Tabay [reported as "5 to 9 km SE and ESE Tabay" and "6 km ESE Tabay" by Handley (1976:10, 77)]; R?o Chama site E-II, 8?36'N, 71?10'W, 1890 m (Durant & D?az 1995); Right bank of the R?o Santo Domingo, 1640 m, near Quebrada de la Virgen, km 36 of the Apartadero-Barinas Highway, Distrito Rangel (Durant & Pefaur 1984). M?rida-Tachira border: Rio Esca- lante site E-I [ca. 8?18'N, 71?55'W], 2600 m. P?ramo de Marino (Durant & D?az 1995). Trujillo: R?o Motat?n site E-II, 9?05'N, 70?39'W, 1890 m (Durant & D?az 1995); Parque Nacional General Cruz Car- rillo (Guaramacal), ca. 2200-3100 m (So- riano et al. 1990). Relevant Additional Records of Shrews from Venezuela and Colombia Two records of Cryptotis from Venezuela and Colombia are noteworthy here because of their vicinity to the known distributions of C. meridensis and C. tamensis. Both re- cords are based on individuals preserved as a skull without skin or skeleton. I have not had the opportunity to inspect either spec- imen. The two collection localities are rep- 270 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON resented on the map in Figure 2 by question marks. It is possible that these two records may provide geographic range extensions of known species as defined herein, or one or both may be undescribed taxa. Each pop- ulation represented by these records will re- quire study of a large series of complete specimens to adequately document local variation in characters and to correctly es- tablish its relationship with other South American species. Ojasti & Mondolfi (1968) reported a sin- gle specimen identified as "Cryptotis tho- masi meridensis" from a forested area near El Junquito (10?28'N, 67?05') in the coastal highlands west of Caracas. This locality is separated from the main distribution of C. meridensis in the Cordillera de los Andes by an expanse of more than 300 km. In be- tween are extensive lowlands below 500 m occupied by Tropical Dry Forest (Ewel et al. 1965, Ewel & Madriz 1968) that cur- rently present a barrier to dispersal for An- dean shrews. No description, photograph, or illustration of the specimen was provid- ed, making it difficult to conjecture as to the identity of the population it represents. The specimen is reposited at the Museo de Biolog?a, Universidad Central de Venezue- la. Duarte & Viloria (1992) reported a spec- imen of ''Cryptotis thomasV from the vi- cinity of Cerro Pintado (10?24'N, 72?53'W) in the Sierra de Perij?, northeastern Cordil- lera Oriental, Colombia. Cryptotis tamensis, C. thomasi, and C. colombiana are the only species known to occur in the Cordillera Oriental. The closest record of a shrew to this locality is that of C. tamensis from P?r- amo de Tama, which is more than 300 km south of Duarte & Viloria's (1992) collec- tion site. Photographs of the skull accom- panying the report are unclear, and most characters are indistinct. However, the shape of the mandible, particularly at the junction of the coronoid process with the horizontal ramus, strongly suggests that it is a member of the C. thomasi-group rather than the C. nigrescens-group. This charac- ter, together with the large size of the spec- imen, indicate that it is unlikely to be C colombiana. The specimen appears to lack the very large foramen on the tympanic process of the petromastoid, an absence that would rule out both C. colombiana and C. thomasi. The low elevation (2100 m) at which the specimen was discovered, the depth of lower sigmoid notch, and the short posterior portion of the mandible also sug- gest that it is not C thomasi. The specimen from Sierra de Perij? thus may represent C tamensis. However, if measurements I ob- tained from the photographs are correct, the specimen has a relatively shorter zygomatic plate and a relatively broader palate and in- terorbital area than any measured specimen of either C. tamensis or C. thomasi. The specimen is in the collection of the Museo de Biolog?a, Universidad del Zulia (MBLUZ M-105). The two records from Sierra de Perij? and El Junquito exemplify the continued poor state of knowledge of South American shrews, and they communicate the need for more complete mammalian surveys even near well-populated areas. Regardless of whether the occurrence of shrews in the coastal highlands of Venezuela represents dispersal or vicariance, it testifies to their ability at some time in the past to negotiate, or even inhabit, lowlands that are currently inhospitable. It also hints at the probability that shrews may be found in suitable habi- tats elsewhere in the region, where they should be sought using appropriate meth- ods. For example, the presence of Cryptotis in northernmost portion of the Cordillera Oriental, less than 100 km from the isolated Sierra Nevada de Santa Marta, suggests an- other place where shrews are likely to be discovered. Acknowledgments 1 thank the following curators and col- lections managers for loans or for permis- sion to examine specimens under their care: Guy G. Musser and Robert Voss (AMNH); VOLUME 115, NUMBER 2 271 Paula Jenkins (BM); Lawrence R. Heaney, Bruce D. Patterson, John Phelps, and Wil- liam Stanley (FMNH); Alberto Cadena (ICN); Robert M. Timm and Thor Holmes (KU); Maria E. Rutzmoser (MCZ); Carlos A. Cuartos and Carlos A. Delgado (MUA); Mark D. Engstrom (ROM). Earl M. J. Aa- gaard provided access to unpublished field notes on C. meridensis and William Stanley to notes on Osgood and Jewett's collections in Colombia and Venezuela. Jerry R. Choate, Julian C. Kerbis, and Robert M. Timm provided much appreciated com- ments on earlier versions of this paper. Literature Cited Aagaard, E. M. J. 1982. Ecological distribution of mammals in the cloud forests and p?ramos of the Andes, M?rida, Venezuela. Ph.D. disserta- tion, Colorado State University, 277 pp. Allen, J. A. 1912. Mammals from western Colom- bia.?Bulletin of the American Museum of Nat- ural History 31:71-95. Anthony, H. E. 1921. Preliminary report on Ecuado- rean mammals. No. 1.?American Museum Novitates 20:1-6. Cabrera, A. 1925. Generum Mammalium. Insect?vora. Galeopithecia. Madrid: Museo Nacional de Ciencias Naturales, 232 pp. . 1958. Cat?logo de los mam?feros de Am?rica del Sur. Vol. 1.?Revista del Museo Argentino de Ciencias Naturales "Bernardino Rivadavia" 4:1-307. , & J. Yepes. 1940. Mam?feros Sud Americanos (vida, costumbres y descripci?n). Compa??a Ar- gentina de Editores, Buenos Aires, 370 pp. , & . 1960. Mam?feros Sud Americanos, vol. 1, 2nd edition. Ediar S. A., Buenos Aires, 187 pp. Choate, J. R. 1970. Systematics and zoogeography of Middle American shrews of the genus Crypto- tis.?University of Kansas Publications, Muse- um of Natural History 19:195-317. Corbet, G. B., & J. E. Hill. 1980. A World List of Mammalian Species. Cornell University Press, Ithaca, 226 pp. , & . 1986. A world list of mammalian species, 2nd edition. Facts On File Publications, New York, 254 pp. , & . 1991. A world list of mammalian species, 3rd edition. Oxford University Press, Oxford, 243 pp. Diaz de Pascual, A. 1993. Caracterizaci?n del habitat de algunas especies de peque?os mam?feros de la selva nublada de Monte Zerpa, M?rida.? Ecotropicos 6:1-9. . 1994. The rodent community of the Venezue- lan Andes.?Polish Ecological Studies 20:155- 161. , & A. de Ascen?ao. 2000. Diet of the cloud forest shrew Cryptotis meridensis (Insect?vora: Soricidae) in the Venezuelan Andes.?Acta Theriologica 45:13-24. Duarte, M. A., & A. L. Viloria. 1992. Nuevo hallazgo de Cryptotis thomasi (Merriam, 1897) (Mam- malia: Insect?vora) en la Sierra de Perij?, nores- te de Colombia.?Acta Cient?fica Venezolana 43:240-242. Durant, P., & A. D?az. 1995. Aspectos de la ecolog?a de roedores y musara?as de las cuencas hidro- gr?ficas Andino-Venezolanas.?Caribbean Jour- nal of Science 31:83-94. , & J. E. Pefaur 1984. Sistem?tica y ecolog?a de la Musara?a de M?rida. Soricidae: Insect?v- ora. Cryptotis thomasi.?Revista de Ecolog?a, Conservaci?n y Ornitolog?a Latinoamericana 1(2):3-14. Eisenberg, J. F 1989. Mammals of the Neotropics. The northern Neotropics. The University of Chicago Press, Chicago, 449 pp. Ewel, J. J., & A. Madriz. 1968. Zonas de vida de Ven- ezuela: memoria explicativa sobre el mapa ecol- ?gico. Ministerio de Agricultura y Cria, Direc- ci?n de Investigaci?n, Caracas, 264 pp. , , & J. A. TOS?. 1965. Rep?blica de Venezuela. Mapa Ecol?gico. Ministerio de Agricultura y Cria, Direcci?n de Investigaci?n, Caracas. Frost, D. R., & D. M. Hillis. 1990. Species in concept and practice: herpetological applications.?Her- petologia 46:87-104. Handley, C. O., Jr 1976. Mammals of the Smithsonian Venezuelan Project.?Brigham Young Science Bulletin, 20:1-91. Hershkovitz, P. 1969. The evolution of mammals on southern continents. VI. The recent mammals of the Neotropical region: a Zoogeographie and ecological review.?Quarterly Review of Biol- ogy 44:1-70. . 1972. The Recent mammals of the Neotropi- cal region: a Zoogeographie and ecological re- view. Pp. 311-431 in A. Keast, F C. Erk, & B. Glass, eds.. Evolution, mammals, and southern continents. State University of New York Press, ?Albany, 543 pp. Holdridge, L. R. 1947. Determination of world plant formations from simple climatic data.?Science 105:367-368. Honacki, J. H., K. E. Kinman, & J. W. Koeppl. 1982. Mammal Species of the World. Lawrence; Allen Press, Inc. & the Association of Systematics Collections, 694 pp. 272 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON Hutterer, R. 1986. S?damerikanische Spitzm?use: Cryptotis meridensis und C. thomasi als ver- schiedene Arten.?Zeitschrift f?r S?ugetierkun- de 51, Sonderheft: 33-34. . 1993. Order Insect?vora. Pp. 69-130 in D. E. Wilson & D. M. Reeder, eds.. Mammal species of the world. Smithsonian Institution Press, Washington, 1206 pp. Linares, O. J. 1998. Mam?feros de Venezuela. Socie- dad Conservacionista Audubon de Venezuela, Caracas, 691 pp. Merriam, C. H. 1897. Descriptions of five new shrews from Mexico, Guatemala, and Colombia.?Pro- ceedings of the Biological Society of Washing- ton 11:227-230. Ojasti, J., & E. Mondolfi. 1968. Esbozo de la fauna de mam?feros de Caracas. Pp. 441-461 in Estudio de Caracas, vol. 1. Ecolog?a vegetal y fauna. Universidad Central de Venezuela, Caracas. Osgood, W H. 1912. Mammals from western Vene- zuela and eastern Colombia.?Field Museum of Natural History, Zoological Series 10(5):33-66. , & B. Conover. 1922. Game birds from north- western Venezuela.?Field Museum of Natural History, Zoological Series 12(3): 19-47. Paynter, R. A., Jr. 1982. Ornithological gazetteer of Venezuela. Bird Department, Museum of Com- parative Zoology, Harvard University, Cam- bridge, 245 pp. , & M. A. Traylor. 1981. Ornithological gaz- etteer of Colombia. Bird Department, Museum of Comparative Zoology, Harvard University, Cambridge, 311 pp. Reed, C. A. 1951. Locomotion and appendicular anat- omy in three soricoid insectivores.?American Midland Naturalist 45:513-671. Ridgway, R. 1912. Color standards and color nomen- clature. Washington: published privately, 43 pp., 53 pi. Soriano, P. J., A. Utrera, & M. Sosa. 1990. Inventario preliminar de los mam?feros del Parque Nacion- al General Cruz Carrillo (Guaramacal), Estado Trujillo, Venezuela.?BioUania 7:83-99. , A. D?az de Pascual, J. Ochoa G., & M. Aguil- era. 1999. Biogeographic analysis of the mam- mal communities in the Venezuelan Andes.? Interciencia 24:17-25. T?te, G. H. H. 1932. Distribution of the South Amer- ican shrews.?Journal of Mammalogy 13:223- 228. Thomas, O. 1898. On seven new small mammals from Ecuador and Venezuela.?Annals and Maga- zine of Natural History, Series 7, 1:451-457. ?. 1912. Three new small mammals from S. America.?Annals and Magazine of Natural History, Series 8, 9:408-410. -. 1921. New Cryptotis, Thomasomys, and Ory- zomys from Colombia.?Annals and Magazine of Natural History, Series 9, 8:354-357. USBGN [U.S. Board on Geographical Names]. 1988. Gazetteer of Colombia, 3rd ed. Defense Map- ping Agency, Washington, D.C., 859 pp. . 1993. Gazetteer of Venezuela, 2nd ed. De- fense Mapping Agency, Washington, D.C., 2 vols., 701 pp. Vivar, E., V. Pacheco, and M. Valqui. 1997. A new species of Cryptotis (Insect?vora: Soricidae) from northern Peru.?American Museum Nov- itates 3202:1-15. Wiley, E. O. 1978. The evolutionary species concept reconsidered.?Systematic Zoology 27: 17-26. . 1981. Phylogenetics. New York: John Wiley & Sons. Woodman, N. 1992. Biogeographical and evolutionary relationships among Central American small- eared shrews of the genus Cryptotis (Mamma- lia: Insect?vora: Soricidae). Unpublished Ph.D. dissertation. University of Kansas, Lawrence, 386 pp. . 1996. Taxonomic status of the enigmatic Cryptotis avia (Mammalia: Insect?vora: Sorici- dae), with comments on the distribution of the Colombian small-eared shrew, Cryptotis col- ombiana.?Proceedings of the Biological Soci- ety of Washington 109:409-418. , & R. M. Timm. 1992. A new species of small- eared shrew, genus Cryptotis (Insect?vora: Sor- icidae), from Honduras.?Proceedings of the Biological Society of Washington 105:1-12. , & . 1993. Intraspecific and interspe- cific variation in the Cryptotis nigrescens spe- cies complex of small-eared shrews (Insect?vo- ra: Soricidae), with the description of a new species from Colombia.?Fieldiana: Zoology, new series 1452:1-30. , & . 1999. Geographic variation and evolutionary relationships among broad-clawed shrews of the Cryptotis goldmani-gtoup (Mam- malia: Insect?vora: Soricidae).?Fieldiana: Zo- ology, new series 1497:1-35. , & . 2000. Taxonomy and evolutionary relationships of Phillips' small-eared shrew, Cryptotis phillipsii (Schaldach, 1966), fromOa- xaca, Mexico (Mammalia: Insect?vora: Sorici- dae).?Proceedings of the Biological Society of Washington 113:339-355.