Areas of endemism of mexican terrestrial mammals: a case study using species’ ecological niche modeling, parsimony analysis of endemicity and goloboff fit

Tania Escalante, Víctor Sánchez-Cordero, Juan J. Morrone and Miguel Linaje

Tania Escalante. Ph.D. in Biological Sciences, Facultad de Ciencias, Universidad Nacional Autónoma de México, (UNAM), Mexico. Postdoctoral Student, Instituto de Biología, UNAM, Mexico. Address: Circuito Exterior s/n, Ciudad Universitaria, Apdo. Postal 70-153, Coyoacán, 04510 Mexico, DF, Mexico. e-mail: tania_escalante@correo.unam.mx

Víctor Sánchez-Cordero. Ph.D. in Natural Resources, The University of Michigan, Ann Arbor MI, USA. Professor, Instituto de Biología, UNAM, Mexico. e-mail: victor@ibiologia.unam.mx

Juan J. Morrone. Ph.D. in Natural Sciences, Universidad Nacional de La Plata, Argentina. Professor, Facultad de Ciencias, UNAM, Mexico. e-mail: jjm@hp.fciencias.unam.mx

Miguel Linaje. Biologist, UNAM, Mexico. Research Assistant, Instituto de Biología, UNAM. e-mail: mlinaje@yahoo.com.mx

SUMMARY

Areas of endemism of Mexican terrestrial mammals using ecological niche modeling projected as species’ potential distributions were identified to compare its performance with a previous analysis that used point occurrence data, and to incorporate Goloboff fit to Parsimony Analysis of Endemicity (PAE) for improving identification of areas of endemism. Six PAE were performed, combined or not, with Goloboff fit (k=0 and 2) using species’ potential distributions of 429 terrestrial mammals overlaid on 248 by 232 quadrats of 1° latitude-longitude countrywide. Consistency (CI) and retention (RI) indices were used for identifying endemic, characteristic, and possibly endemic species. Based on the strict consensus cladogram with k=0, seven areas of endemism defined by two or more species were identified: the Mexican Plateau, the Baja California Peninsula (with a nested pattern of endemism in the south and north), Chiapas (with a nested pattern of endemism in the south and north), the Mexican Pacific Coast, the Isthmus of Tehuantepec, the Sierra Madre Occidental, and the Yucatan Peninsula. PAE cladograms using species’ potential distributions showed a better resolution than those produced using point occurrence data, showing consensus with fewer number of steps and higher number of synapomorphies. Goloboff fit (k) allowed individual down-weighting of "noisy" species, thus increasing the number of synapomorphies in the cladograms, and even identifying more areas of endemism. Cladograms with k=0 had the largest number of synapomorphies, whereas k=2 allowed to obtain a smaller number of cladograms.

Se identificaron áreas de endemismo de mamíferos terrestres de México, usando modelos de nicho ecológico proyectados como distribuciones potenciales de especies, con el fin de comparar su desempeño respecto a análisis previos que usan datos puntuales de ocurrencia, incorporando el ajuste de Goloboff al Análisis de Parsimonia de Endemismos (PAE) para mejorar la identificación de áreas de endemismo. Se desarrollaron seis PAE, combinados o no con el ajuste de Goloboff (k=0 y 2) usando distribuciones potenciales de especies de 429 mamíferos terrestres sobrepuestas a 248 y 232 cuadros de 1º de latitud-longitud a lo largo del país. Se utilizaron los índices de consistencia (CI) y retención (RI) para identificar especies endémicas, posiblemente endémicas y características. Con base en el cladograma de consenso estricto con k=0, se identificaron siete áreas de endemismo definidas por dos o más especies: el Altiplano Mexicano, la Península de Baja California (con un patrón de endemismo anidado en el sur y el norte), Chiapas (con un patrón de endemismo anidado en el sur y el norte), la Costa Pacífica Mexicana, el Istmo de Tehuantepec, la Sierra Madre Occidental y la Península de Yucatán. Los cladogramas de PAE usando distribuciones potenciales de las especies mostraron una mejor resolución que aquellos producidos con datos de ocurrencia puntual, mostrando consensos con menor número de pasos y alto número de sinapomorfías. El ajuste de Goloboff (k) permitió menor pesado individual de especies "con ruido", incrementando el número de sinapomorfías en los cladogramas, e identificando más áreas de endemismo. Los cladogramas con k=0 tuvieron el mayor número de sinapomorfías, mientras que k=2 permitió obtener un menor número de cladogramas.

RESUMO

Identificaram-se áreas de endemismo de mamíferos terrestres do México, usando modelos de nicho ecológico projetados como distribuições potenciais de espécies, com o fim de comparar seu desempenho em relação a análises prévias que usam dados pontuais de ocorrência, incorporando o ajuste de Goloboff a Análise de Parcimônia de Endemismos (PAE) para melhorar a identificação de áreas de endemismo. Desenvolveram-se seis PAE, combinados ou não com o ajuste de Goloboff (k=0 e 2) usando distribuições potenciais de espécies de 429 mamíferos terrestres sobrepostas a 248 e 232 quadros de 1º de latitude-longitude ao longo do país. Utilizaram-se os índices de consistência (CI) e retenção (RI) para identificar espécies endêmicas, possivelmente endêmicas e características. Com base no cladograma de consenso estrito con k=0, se identificaram sete áreas de endemismo definidas por duas ou mais espécies: o Altiplano Mexicano, a Península de Baixa Califórnia (com um padrão de endemismo aninhado no sul e no norte), Chiapas (com um padrão de endemismo aninhado no sul e no norte), a Costa Pacífica Mexicana, o Istmo de Tehuantepec, a Serra Madre Ocidental e a Península de Yucatán. Os cladogramas de PAE usando distribuições potenciais das espécies mostraram uma melhor resolução que aqueles produzidos com dados de ocorrência pontual, mostrando consensos com menor número de passos e alto número de sinapomorfias. O ajuste de Goloboff (k) permitiu menor pesagem individual de espécies "com ruído", incrementando o número de sinapomorfias nos cladogramas, e identificando mais áreas de endemismo. Os cladogramas con k=0 tiveram o maior número de sinapomorfias, enquanto que k=2 permitiu obter um menor número de cladogramas.

KEY WORDS / Ecological Niche Modeling / Endemicity / GARP / Mexico / PAE / Species’ Potential Distributions / Terrestrial Mammals /

Identification of areas of endemism constitutes an important challenge in biogeography (Henderson, 1991; Harold and Mooi, 1994; Morrone, 1994; Andersson, 1996) and is essential for subsequent cladistic biogeographic analyses. Several approaches for defining areas of endemism based on the geographical distribution of taxa have been proposed, such as clustering methods (similarity indices combined with UPGMA; Murguía and Villaseñor, 2003), parsimony methods (Rosen, 1988; Rosen and Smith, 1988; Morrone, 1994), null models (Mast and Nyffeler, 2003) and optimality methods (Szumik et al., 2002; Szumik and Goloboff, 2004). One of such methods, Parsimony Analysis of Endemicity (PAE; Rosen, 1988; Rosen and Smith, 1988; Morrone, 1994) assumes the existence of a common historical explanation for the assemblages of areas based on shared species and has been widely used to identify areas of endemism for different taxa and regions worldwide. PAE analyses have traditionally used either point occurrence data based on museum specimens or coarse distribution maps to build presence-absence species’ matrices; however, point occurrence data may be taxonomically and geographically strongly biased (Sánchez-Cordero et al., 2001; Escalante, 2005), and coarse maps built with polygons of marginal locality records usually result in significant distributional over-predictions (Hall, 1981; Illoldi-Rangel et al., 2004).

Recent approaches for modeling species’ ecological niche projected as potential distributions may provide a more robust approach to overcome the mentioned shortcomings (Escalante et al., 2005). Of the several existing methods, Genetic Algorithm for Rule-set Prediction (GARP) is an expert-system computer genetic algorithm that incorporates point occurrence data, environmental variables and GIS to model a species’ geographic ecological niche (Grinnell, 1917; MacArthur, 1972), projected onto geographical data to produce potential distributional maps (Stockwell and Peters, 1999). GARP appears robust for significantly predicting species distributions and has been widely used for mammals (Peterson et al., 1999, 2000, 2002; Illoldi-Rangel et al., 2004; Sánchez-Cordero et al., 2004, 2005a, b). Previous work incorporated species’ potential distributions into PAE. For example, Rojas-Soto et al. (2003) demonstrated that the use of point occurrence data produced limited results, leading to unsolved cladograms, whereas ecological niche modeling of birds projected as potential distributions result in better cladogram resolutions. Espadas-Manrique et al. (2003) used point occurrence data as well as maps of potential distributions produced with DOMAIN (Carpenter et al., 1993) for 162 plants, finding that the resulting cladograms had better resolution than point occurrence data. Morrone and Escalante (2002) performed a PAE with point occurrence data for terrestrial mammals in Mexico, using 230 quadrats of 1×1° (latitude-longitude) and their consensus cladogram showed a basal polytomy with about half of the quadrats and only nine synapomorphies justifying the groups. A preliminary analysis had identified eight possible areas of endemism (Espinosa et al., 2001), but given the poor resolution of the cladogram, no further analyses were performed.

Several modifications have been proposed to improve PAE. Geraads (1998) noticed that convergence is highly unlikely to be achieved in a PAE cladogram, thus recommending ACCTRAN optimization favoring reversals rather than convergences. In PAE, as usually performed, all species are assumed to have the same weight. Alternatively, for taxonomic phylogenies, Goloboff (1993) proposed a method for weighting characters known as Goloboff fit, based on a homoplasy concave function (k), which is applied to each character when the length (weight) of a tree is calculated; by increasing Goloboff fit, it results in a less severe weighting scheme (Goloboff, 1993; Thollesson, 2000). Luna-Vega et al. (2000) undertook PAE setting Goloboff fit criterion k=0, defining clades forbidding homoplasy and based generalized tracks on them. If areas of endemism are considered to be equivalent to generalized tracks (Harold and Mooi, 1994; Morrone, 2001), it is possible to use Goloboff fit to explore the sensibility of PAE cladograms for area identification.

A database of 19058 museum specimen records of 429 terrestrial mammals occurring in Mexico was used, following Villa and Cervantes’s (2003) nomenclature. Point occurrence data for each species were compiled from scientific collections (see Acknowledgments) and from CONABIO (Comisión Nacional para el Conocimiento y Uso de la Biodiversidad; www.conabio.gob.mx), and were georeferenced to the nearest 0.01° of longitude and latitude for each locality using 1:250000 topographic maps (CONABIO, 1998). To characterize species’ ecological niches, 10 environmental data layers (0.04´0.04 pixel resolution) were used, including potential vegetation types (Rzedowski, 1986); elevation, slope and aspect (US Geological Survey’s Hydro-1K data set; www.usgs.gov); and climatic parameters, including mean annual, daily and maximum daily precipitation, and mean annual, daily minimum and maximum temperature (Conabio; www.conabio.gob.mx).

The Genetic Algorithm for Rule-set Prediction (GARP; Stockwell and Peters, 1999) was employed for modeling each species’ ecological niches projected as a potential distributional map (available at www.lifemapper.org/desktopgarp). More detailed descriptions of the method are provided by Stockwell and Peters (1999). In GARP, each model is unique and slightly different (Stockwell and Peters, 1999), so an optimal subset of replicate models was produced (Anderson et al., 2003). For each species, 100 replicate models were run retaining the 20 models with lowest omission error. The 10 models with moderate commission error were then retained and the 10 models with areas predicted as "present", showing greatest deviations from the overall median areas predicted as present across all models, were discarded. This ‘best subset’ of models was summed obtaining ecological niche models for each species. These models were further refined by delimiting species’ ecological niche to physiographic provinces maps (CONABIO; see Cervantes-Zamora et al., 1990) where species had been collected to produce the final distributional models. This allowed to avoid obvious overpredictions, particularly in species with restricted distributions (Illoldi-Rangel et al., 2004). Physiographic provinces were chosen for refinement of models given their historical and geological framework for species’ distributions. Moreover, the physiographic map layer was not included in the set of layers used for GARP modeling. All final species’ distributional models were compared to the typical coarse-grained distributional maps of Hall (1981), Álvarez-Castañeda and Cortés-Calva (1999), Bogan (1999), Cervantes et al. (1999), Maldonado (1999a, b), Patton (1999), Patton and Álvarez-Castañeda (1999), Yensen and Valdés-Alarcón (1999), Arita and Rodríguez (2004), and InfoNatura (2004).

Since the aim of this work is the identification of biotic natural areas, the use of a grid was preferred instead of other natural areas (e.g. eco-regions, physiographic features), to prevent circularity in two senses: natural areas to find natural areas (they involve greater speculation), and many of these features were used in ecological niche modeling and potential distribution projection. The final distributional models for each species was generalized to a grid of 1×1° nationwide, since Morrone and Escalante (2002) previously found that PAE cladograms for Mexican mammals at this scale resulted in a better resolution than a grid of 0.5×0.5°. Furthermore, it is recommended that the number of columns (taxa) be greater than that of rows (areas or quadrats) in a PAE matrix, so it was not possible to use a smaller size of quadrat or pixels (approximately 852 quadrats of 0.5° and thousands of pixels).

Two matrices for PAE were built, one with 248 quadrats and another with 232 quadrats, the latter lacking 16 problematic quadrats (those in which terrestrial surface was less than 20% and, thus, with very few species). All PAE were performed in PAUP 4.0b10 (Swofford, 2002), using ACCTRAN optimization (Geraads, 1998), without and with Goloboff fit (k=0 and k=2). Following Escalante et al. (2003), consistency (CI) and retention (RI) indices were used to choose important species. CI=1.0 denotes synapomorphies (endemic taxa: present in a monophyletic group of areas) or autapomorphies (characteristic taxa: present in a single area). CI=0.50 may indicate possible endemic taxa (only when they are synapomorphies with a posterior reversal, indicating a probable extinction). In some cases, CI=0.33 was used to identify possible synapomorphies. The resulting monophyletic groups of quadrats were analyzed, and important species were mapped in Arc View 3.2 (ESRI, 1999) for identifying areas of endemism. Maps are presented in Lambert conic conformal projection.

The matrix with 248 quadrats and 429 species was run applying the heuristic search, resulting in 3100 cladograms with 3773 steps, CI=0.11 and RI=0.83 (Table I). The strict consensus cladogram had 4021 steps, CI=0.11 and RI=0.81, and included 39 characteristic species (Ammospermophilus insularis, A. interpres, Centronycteris maximiliani, Cratogeomys neglectus, Cryptotis goodwini, Dipodomys compactus, Euderma maculatum, Geomys personatus, Habromys lophurus, Lepus insularis, Lichonycteris obscura, Liomys salvini, Megadontomys nelsoni, Molossops greenhalli, Myotis ciliolabrum, Nelsonia goldmani, Neotoma anthonyi, N. bryanti, N. bunkeri, N. martinensis, N. nelsoni, N. varia, Orthogeomys lanius, Pappogeomys alcorni, Peromyscus caniceps, P. crinitus, P. hooperi, P. interparietalis, P. pembertoni, P. pseudocrinitus, Rheomys thomasi, Rhogeessa genowaysi, R. mira, Scotinomys teguina, Sorex arizonae, S. emarginatus, S. stizodon, Tamiasciurus mearnsi, and Tylomys tumbalensis), 7 geographic synapomorphies, and 19 possible synapomorphies. The matrix with 232 and 429 species resulted in 3100 cladograms (which corresponded to the top number of cladograms able to be contained by our computer) with 3548 steps, CI=0.11 and RI=0.83. The strict consensus cladogram had 3798 steps, CI=0.11 and RI=0.82, and included 36 characteristic species (all of 248 quadrats except D. compactus, N. bryanti, and P. pembertoni), 7 geographic synapomorphies, and 18 possible synapomorphies.

The matrix with 248 quadrats run using Goloboff fit (k=2) resulted in 36 cladograms with 4069 steps, CI=0.10 and RI=0.81. The strict consensus cladogram had 4078 steps, CI=0.11 and RI=0.82, and had 24 geographic synapomorphies and 29 possible synapomorphies. The matrix with 232 quadrats using Goboloff fit (k=2) resulted in 53 cladograms with 3818 steps, CI=0.11 and RI=0.82. The strict consensus cladogram had 3826 steps, CI=0.11 and RI=0.82, and included 22 geographic synapomorphies and 23 possible synapomorphies (Table I).

By eliminating 16 problematic quadrats the number of cladograms, synapomorphies, CI or RI did not improve. Therefore, their deletion was irrelevant, despite that their terrestrial area was incomplete and they included few species. In addition, three characteristic species: D. compactus, N. bryani, and P. pembertoni were lost.

Geographic synapomorphies were essentially the same for the strict consensus cladograms of k=0 and k=2, even though some changed their status, from synapomorphies to possible geographic synapomorphies and vice versa (22 were synapomorphies consistently, 20 were possible synapomorphies consistently, and 24 changed from synapomorphies to possible ones and vice versa). Cladograms showed three main clades: 1) quadrats including Baja California Peninsula + Sonora + Sierra Madre Occidental, 2) quadrats including Chiapas + Yucatan Peninsula + Isthmus of Tehuantepec, and 3) quadrats from the rest of country (Figure 1b). Within the latter, no smaller clades were particularly supported, probably due to the presence of few endemic species. The 248 quadrats cladogram with k=0 was chosen for identifying areas of endemism, because it had the largest number of geographic synapomorphies. Seven areas of endemism were identified, defined by two or more geographic synapomorphies and apomorphies: Mexican Plateau, Baja California Peninsula (with a nested pattern of endemism in the south and north), Chiapas (with a nested pattern of endemism in the south and north), Mexican Pacific Coast, Isthmus of Tehuantepec, Sierra Madre Occidental and Yucatán Peninsula.

The Mexican Plateau is located north, in the states of Sonora (northern part), Chihuahua, Coahuila, Durango, Nuevo León, Tamaulipas, Zacatecas, San Luis Potosí, and parts of Aguascalientes, Jalisco, Guanajuato, Querétaro, and Hidalgo (Figure 2a). Its species were A. interpres, Bison bison, Cratogeomys castanops, D. compactus, Microtus pennsylvanicus, Myotis planiceps, Onychomys arenicola, P. hooperi, and P. nasutus. Also, Neotoma micropus and Onychomys leucogaster, with CI=0.33.

Figura 2. a: Mexicab Plateau area of endemism, supported by 11 species; b: Pacific Coast area of endemism, supported by 2 especies.

Figura 3. Nested pattern of endemism of Baja California Peninsula. a: All Peninsula, b: North, c: South

Chiapas (Figure 4) showed a nested pattern of endemism: a) All Chiapas: the state of Chiapas, eastern Oaxaca, and southern Veracruz and Tabasco, based on Cabassous centralis, Saccopteryx leptura and Micronycteris brachyotis; b) Northern Chiapas: only the northern part of Chiapas and Oaxaca, and the states of Veracruz and Tabasco, based on C. goodwini, H. lophurus, L. obscura, Macrophyllum macrophyllum, Microtus guatemalensis, Peromyscus zarhynchus, Sorex sclateri, S. stizodon, Scotinomys teguina, Tylomys bullaris and T. tumbalensis, and c) Southern Chiapas: only the southern part of Chiapas, based on Molossus coibensis, R. thomasi, R. genowaysi, Sciurus variegatoides and Thryroptera tricolor.

Figura 4. Nested pattern of endemism of Chiapas. a: All Chiapas, b: North, c: South

The Mexican Pacific Coast (Figure 2b), in the western coast of Nayarit, Jalisco and Colima, is based on Cratogeomys fumosus and Xenomys nelsoni. The Isthmus of Tehuantepec, in eastern Oaxaca and a small part of Puebla, Veracruz and Chiapas (Figure 5a), is based on Centronycteris maximiliani, Enchistenes hartii –although Arita and Rodríguez (2004) and Hall (1981) considered the latter more widespread compared with our model–, Habromys lepturus, Megadontomys cryophilus, Microtus oaxacensis, M. umbrosus, Molossops greenhalli (similar to Enchistenes hartii), Orthogeomys cuniculus and Rheomys mexicanus.

Figura 5. a: Isthmus of Tehuantepec area of endemism, identified by 9 species; b: Sierra Madre Occidental area of endmism, defined by 3 species; c:Yucatan Peninsula area of endemism, supported by 8 species.

he Sierra Madre Occidental, including parts of Sonora, Chihuahua, Sinaloa, Durango, Nayarit, Zacatecas, and Jalisco (Figure 5b), is based on Neotoma palatina, Sciurus aberti and Tamias dorsalis, the latter two species with CI=0.33. The Yucatan Peninsula covered Campeche, Yucatan and Quintana Roo (Figure 5c), based on Heteromys gaumeri, Mimon crenolatum (in the southern part only), Micronycteris schmidtorum, Molossus bondae, Otonyctomys hatti, Peromyscus yucatanicus, Procyon pygmaeus (insular) and Rhogeessa aeneus.

Some synapomorphic or possible synapomorphic species did not allow proper identification of areas of endemism: Ammospermophilus harrisii, Corynorhinus mexicanus, Dermanura watsoni, Dasyprocta mexicana, Geomys tropicalis, Megadontomys thomasi, Phyllostomus stenops, Sorex macrodon, Tapirus bairdii, and Zygogeomys trichopus. None of the following autapomorphies shaped any pattern: C. neglectus, G. personatus, L. salvini, M. nelsoni, M. ciliolabrum, N. goldmani, N. nelsoni, N. varia, O. lanius, P. alcorni, P. crinitus, P. pembertoni, R. mira and S. emarginatus.

Cladograms using species’ distribution models of terrestrial mammals showed better resolution than cladograms obtained with point occurrence data. The latter had the largest number of steps, the fewest number of synapomorphies, and a basal polytomy with 56 quadrats (see Espinosa et al., 2001; Morrone and Escalante, 2002). Conversely, the cladograms of this study had no polytomies despite rather low CI and RI. The results caution against using only raw point occurrence data for extrapolating species distributions (Sánchez-Cordero et al., 2001, 2004; Escalante, 2005), and support modeling approaches of species’ distributions for biogeographic inferences (Manrique et al., 2003; Rojas-Soto et al., 2003; Sánchez-Cordero et al., 2004; Escalante, 2005). Species’ ecological niche modeling projected onto geographic distributions may help understand the geographic history and speciation events (Peterson et al., 1999) that led to areas of endemism, as it has been suggested that species tend to conserve their ecological niches over evolutionary times (Peterson et al., 1999; Wiens and Graham 2005). For example, Peterson et al. (1999) showed that vicariance speciation occurred in the same ecological niches for sister species of birds, mammals and butterflies in Oaxaca and Chiapas, two regions with high endemism in the present study. Conversely, the modeling approach used herein does not include biotic interactions or ecological and geographical barriers known to be causal factors for delimiting species’ distributions under certain circumstances (Soberón and Peterson, 2005). To overcome such shortcomings for future biogeographic studies is a challenge (Sánchez-Cordero et al., 2004; Soberón and Peterson, 2005).

PAE groups areas (analogous to taxa) based on their shared taxa (analogous to characters), assigning all species equal weights. All species of Mexican terrestrial mammals were used in this study, including different orders, dispersal abilities and sizes of distributional areas; and some of these characteristics are due to homoplasy ("noise", for example widespread taxa, taxa that have dispersed, etc.). In a PAE cladogram, Goloboff fit (k) allows individual down-weighting of "noisy" species, thus increasing the number of synapomorphies in the cladograms, and even identifying more areas of endemism. Cladograms with k=0 had the largest number of synapomorphies, whereas k=2 resulted in fewer cladograms. Goloboff fit has been used to obtain taxonomic phylogenies (Anderson, 2000; Thollesson, 2000; Hunt et al., 2001; Blum et al., 2003; Montresor et al., 2003) and test the stability of taxonomic cladograms, using different k values (Bossuyt and Milinkovitch, 2000; Cassens et al., 2000; Takami, 2000), which may show different species arrangements than analyses performed without k. Some differences between area relationships in cladograms without Goloboff fit, and with k=0 and 2 were observed, particularly in the Mexican Plateau and the Isthmus of Tehuantepec areas of endemism. Although some authors have proposed criteria to choose the function and value of the concavity (such as stability and sensitivity analysis and jackknife, among others), there are no conclusions about their use (e.g. Turner and Zandee, 1995; Goloboff, 1997; Arias and Miranda-Esquivel, 2004). For the areas of endemicity, the present authors prefer to choose the concavity which offers more synapomorphies (k=0).

Morrone and Escalante (2002) performed a PAE using point occurrence data for Mexican terrestrial mammals resulting in several unresolved cladograms. Cladograms of this study were more informative for identifying areas of endemism at the same scale of 1×1° grid, and it is likely that a smaller grid may have revealed less resolved cladograms than those obtained herein. A matrix using the thousands of pixels of species’ distributional models will be almost impossible to run in a parsimony program. Escalante et al. (2003) using point occurrence data found five areas of endemism for terrestrial mammals in Mexico, depicted in ecoregions: the Northern High Plateau, Baja California, Chiapas, Isthmus, and the Yucatan Peninsula. The present results were similar, but added the Mexican Pacific Coast and the Sierra Madre Occidental areas of endemism (Table II).

The identification of patterns of endemism is a first step for understanding biogeographic historical processes as vicariance, isolation, speciation, dispersal, and extinction (Morrone and Escalante, 2002; Escalante et al., 2003).

The Mexican Plateau has no morphotectonic continuity (Ferrusquía-Villafranca, 1998), although the endemic species identified here are distributed over the entire province, and there is controversy over whether it represents either one or two provinces (Ramírez-Pulido and Castro-Campillo, 1990; Arriaga et al., 1997; Morrone et al., 2002).

The Baja California Peninsula has been recognized as an important area of endemism for several taxa (Morrone et al., 2002; Rojas-Soto et al., 2003; Álvarez-Mondragón and Morrone, 2004). Escalante et al. (2004) described a generalized track in the northern Baja California Peninsula, with Microtus californicus, Sorex ornatus, and Sylvilagus bachmani supporting it. The high number of characteristic species appears as a result of the evolution of new taxa on islands in the vicinity of Baja California, and vicariance processes have been proposed based on phylogeographic studies (Riddle et al., 2000).

Chiapas harbors some relictual Nearctic taxa, despite being classified in the Neotropical region, which may be due to events of expansion and isolation during the Pleistocene (Fa and Morales, 1998; Conroy et al., 2001; Escalante et al., 2004). Generalized tracks of Heterolinus and Homalinus (Coleoptera: Staphylinidae; Márquez and Morrone, 2003), birds (Álvarez-Mondragón and Morrone, 2004), and Lampetis (Coleoptera: Buprestidae; Corona and Morrone, 2005) were found from central Chiapas to Central America. Moreover, the Altos de Chiapas has been recognized as an important biogeographic and conservation region, harboring high species richness, endemicity, and endangered species; it is thus a biogeographic node (Escalante, 2003; Álvarez-Mondragón and Morrone, 2004; Escalante et al., 2004). The Isthmus of Tehuantepec has been considered as a bridge linking the biotas of the eastern and western coastal areas (Leopold, 1983), but more studies are needed to interpret its history (Carleton et al., 2002). Corona and Morrone (2005) found a generalized track for Lampetis (Coleoptera: Buprestidae), and Márquez and Morrone (2003) identified a biogeographic node in northeastern Oaxaca.

The Yucatan Peninsula had a different geological history from the rest of the country, reflected in a morphotectonic unit (Ferrusquía-Villafranca, 1998), which is congruent with the presence of endemic taxa. The Mexican Pacific Coast is a small area, likely nested in the larger Mexican Pacific Coast province sensu Morrone et al. (2002); other taxa as the insect Lampetis showed a similar pattern (Corona and Morrone, 2005). The Sierra Madre Occidental is the province of the Mexican Transition Zone more closely related to the Nearctic region (Escalante et al., 2005). It was identified for other taxa as well (Heterolinus and Homalinus by Márquez and Morrone, 2003; Lampetis by Corona and Morrone, 2005) and its origin is related to the Farallon Plate subduction (Ferrusquía-Villafranca, 1998).

According to Rojas-Soto et al. (2003), the absence of biogeographical definition of areas of endemism may result from either low endemicity and/or the presence of geographical homoplasies caused by local population extinction or dispersal. Interestingly, the Transmexican Volcanic Belt and the Sierra Madre del Sur have been recognized as important areas of endemism for terrestrial mammals in Mexico (Fa and Morales, 1998), but they were not identified by the present analyses; both provinces could be biogeographic nodes shaping the boundaries between regions, being part of the Mexican Transition Zone (Escalante et al., 2004). The Transmexican Volcanic Belt includes many characteristic species only occurring therein (Fa and Morales, 1998; Sánchez-Cordero et al., 2004), but their distributions do not cover the entire province (each has a highly restricted distributional area) and it may not be a homogeneous unit (Escalante et al., in press). An alternative is that PAE failed to identify the Transmexican Volcanic Belt and Sierra Madre del Sur as important areas of endemism because they were present in cladograms as paraphyletic or polyphyletic groups of quadrats. Further analyses with different methods (Szumik and Goloboff, 2004) could test whether both provinces are areas of endemism.

By incorporating ecological niche modeling projected as species’ potential distributions into PAE, a robust approach for identifying areas of endemism is provided, based on the analysis of geographic synapomorphies of cladograms under weighting schemes. Although PAE and panbiogeography have been criticized (Humphries and Parenti, 1999; Brooks and Van Veller, 2003; Santos, 2005), this approach is still useful for inferring biogeographic patterns, particularly in the absence of robust phylogenies for many taxa (Morrone and Escalante, 2002).

Point occurrence data from museum specimen localities were obtained from the Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (Conabio) and the following museum collections were consulted: Colección Nacional de Mamíferos, Universidad Nacional Autónoma de México; Colección de Mamíferos, Universidad Autónoma Metropolitana-Iztapalapa; Centro Interdisciplinario de Investigación y Desarrollo Regional de Oaxaca; University of Kansas Natural History Museum; American Museum of Natural History, New York; National Museum of Natural History, Washington, DC; Field Museum of Natural History, Chicago, IL; Museum of Zoology, University of Michigan, Ann Arbor; Michigan State University Museum, East Lansing; Museum of Vertebrate Zoology, University of California, Berkeley; Texas Tech University Museum, Lubbock; and Texas Cooperative Wildlife Collections, Texas A and M University, College Station. The authors thank P. Illoldi for comments and G. Rodríguez for verification of the potential distributional maps. T. Escalante acknowledges the postdoctoral scholarship granted by DGAPA-Universidad Nacional Autónoma de México, 2004-2005. This work was partially supported by the Consejo Nacional de Ciencias y Tecnología and the Secretaría del Medio Ambiente y Recursos Naturales (SEMARNAT-CONACyT project 2002-C01-314-A1) and PAPIIT-UNAM IN218706 project, both to VS-C).

1. Álvarez-Castañeda ST, Cortés-Calva P (1999) Familia Muridae. In Álvarez-Castañeda ST, Patton JL (Eds.) Mamíferos del Noroeste de México. CIBNOR. La Paz, Mexico. pp. 446-568.        [ Links ]

2. Álvarez-Mondragón E, Morrone JJ (2004) Propuesta de áreas para la conservación de aves de México, empleando herramientas panbiogeográficas e índices de complementariedad. Interciencia 29: 112-120.        [ Links ]

3. Anderson FE (2000) Phylogenetic relationships among loliginid squids (Cephalopoda: Myopsida) based on analyses of multiple data sets. Zool. J. Linn. Soc. 130: 603–633.        [ Links ]

4. Anderson RP, Lew D, Peterson AT (2003) Evaluating predictive models of species’ distributions: Criteria for selecting optimal models. Ecol. Model. 162: 211-232.        [ Links ]

5. Andersson L (1996) An ontological dilema: Epistemology and methodology of historial biogeography. J. Biogeogr. 23: 269-277.        [ Links ]

6. Arias JS, Miranda-Esquivel DR (2004) Profile Parsimony (PP): An analysis under implied weights (IW). Cladistics 20: 56-63.        [ Links ]

7. Arita HT, Rodríguez G (2004) Patrones geográficos de diversidad de los mamíferos terrestres de América del Norte. Instituto de Ecología, UNAM. SNIB-Conabio Q068 database project. Mexico.        [ Links ]

8. Arriaga L, Aguilar C, Espinosa D, Jiménez R (1997) Regionalización ecológica y biogeográfica de México. CONABIO. Mexico. 99 pp.        [ Links ]

9. Blum MJ, Bermingham E, Dasmahapatra K (2003) A molecular phylogeny of the Neotropical butterfly genus Anartia (Lepidoptera: Nymphalidae). Mol. Phylogenet. Evol. 26: 46–55.        [ Links ]

10. Bogan MA (1999) Family Vespertilionidae. In Álvarez-Castañeda ST, Patton JL (Eds.) Mamíferos del Noroeste de México. CIBNOR. La Paz, Mexico. pp. 141-181.        [ Links ]

11. Bossuyt F, Milinkovitch MC (2000) Convergent adaptive radiations in Madagascan and Asian ranid frogs reveal covariation between larval and adult traits. Proc. Natl. Acad. Sci. 97: 6585-6590.        [ Links ]

12. Brooks DR, Van Veller MGP (2003) Critique of parsimony analysis of endemicity as a method of historical biogeography. J. Biogeogr. 30: 819–825.        [ Links ]

13. Carleton MD, Sánchez O, Urbano-Vidales G (2002) A new species of Habromys (Muroidea: Neotominae) from Mexico, with generic review of species definitions and remarks on diversity patterns among Mesoamerican small mammals restricted to humid montane forests. Proc. Biol. Soc. Wash. 115: 488-533.        [ Links ]

14. Carpenter G, Gillison AN, Winter J (1993) DOMAIN: A flexible modeling procedure for mapping potential distributions of plants and animals. Biodiv. Cons. 2: 667-680.        [ Links ]

15. Cassens I, Vicario S, Waddell VG, Balchowsky H, Van Belle D, Ding W, Fan C, Mohan RSL, Simões-Lopesi PC, Bastida R, Meyer A, Stanhope MJ, Milinkovitch MC (2000) Independent adaptation to riverine habitats allowed survival of ancient cetacean lineages. Proc. Natl. Acad. Sci. 97: 11343-11347.        [ Links ]

16. Cervantes FA, Lorenzo C, Vargas J (1999) Familia Leporidae. In Álvarez-Castañeda ST, Patton JL (Eds.) Mamíferos del Noroeste de México. CIBNOR. La Paz, Mexico. pp. 199-237.        [ Links ]

17. Cervantes-Zamora Y, Cornejo-Olgín SL, Lucero-Márquez R, Espinoza-Rodríguez JM, Miranda-Viquez E, Pineda-Velázquez A (1990) Provincias Fisiográficas de México, Clasificación de Regiones Naturales de México II, IV.10.2. Atlas Nacional de México. V. II. Scale 1:4 000 000. Instituto de Geografía, UNAM. México.        [ Links ]

18. Conabio (1998) Topografía de México. Escala 1:250 000. Instituto Nacional de Estadística, Geografía e Informática (INEGI). Comisión Nacional para el Conocimiento y Uso de la Biodiversidad. México.        [ Links ]

19. Conroy VCJ, Hortelano Y, Cervantes FA, Cook JA (2001) The phylogenetic position of southern relictual species of Microtus (Muridae: Rodentia) in North America. Mamm. Biol. 66: 332-344.        [ Links ]

20. Corona AM, Morrone JJ (2005) Track analysis of the species of Lampetis (Spinthoptera) Casey, 1909 (Coleoptera: Buprestidae) in North America, Central America, and the West Indies. Caribb. J. Sci. 41: 37-41.        [ Links ]

21. Escalante T (2003) Determinación de prioridades en las áreas de conservación para los mamíferos terrestres de México, empleando criterios biogeográficos. An. Inst. Biol. UNAM, Ser. Zool. 74: 211-237.        [ Links ]

22. Escalante T (2005) Las bases de datos curatoriales y el estudio espacial de la biodiversidad: Un ejemplo con los mamíferos terrestres de México. In Llorente J, Morrone JJ (Eds.) Regionalización biogeográfica en Iberoamérica y tópicos afines: Primeras Jornadas Biogeográficas de la Red Iberoamericana de Biogeografía y Entomología Sistemática (RIBES XII.I-CYTED). CYTED-UNAM-Conabio. Mexico. pp. 339-350        [ Links ]

23. Escalante T, Espinosa D, Morrone JJ (2003) Using Parsimony Analysis of endemism to analyze the distribution of Mexican land mammals. Southwest. Nat. 48: 563-578.        [ Links ]

24. Escalante T, Rodríguez G, Morrone JJ (2004) The diversification of Nearctic mammals in the Mexican Transition Zone. Biol. J. Linn. Soc. 83: 327-339.        [ Links ]

25. Escalante T, Rodríguez G, Morrone JJ (2005) Las provincias biogeográficas del componente Mexicano de Montaña desde la perspectiva de los mamíferos continentales. Rev. Mex. Biodiv. 76: 199-205.        [ Links ]

26. Escalante T, Rodríguez G, Gámez N, Sánchez-Cordero V, Barrera O, León L. Biogeografía y conservación de los mamíferos de la Faja Volcánica Transmexicana. In Luna I, Morrone JJ, Espinosa D (Eds.) Biodiversidad de la Faja Volcánica Transmexicana. Mexico. (In press)        [ Links ]

27. Espadas-Manrique C, Durán R, Argáez J (2003) Phytogeographic analysis of taxa endemic to the Yucatan Peninsula using geographic information systems, the domain heuristic method and parsimony analysis of endemicity. Div. Distr. 9: 313-330.        [ Links ]

28. Espinosa D, Aguilar C, Escalante T (2001) Endemismo, áreas de endemismo y regionalización biogeográfica. In Llorente J, Morrone JJ (Eds.) Introducción a la biogeografía en Latinoamérica: Teorías, conceptos, métodos y aplicaciones. Las Prensas de Ciencias. Mexico. pp. 31-37.        [ Links ]

29. ESRI (1999) Arc View GIS 3.2. Environmental Systems Research Institute. California, USA.        [ Links ]

30. Fa JE, Morales LM (1998) Patrones de diversidad de mamíferos de México. In Ramamoorthy T P, Bye R, Lot A, Fa J (Eds.) Diversidad biológica de México: Orígenes y distribución. Instituto de Biología, UNAM. Mexico. pp. 315-352.        [ Links ]

31. Ferrusquía-Villafranca I (1998) Geología de México: Una sinopsis. In Ramamoorthy TP, Bye R, Lot A, Fa J (Eds.) Diversidad biológica de México: Orígenes y distribución. Instituto de Biología, UNAM. Mexico. pp. 3-108.        [ Links ]

32. Geraads D (1998) Biogeography of circum-Mediterranean Miocene-Pliocene rodents; a revision using factor analysis and parsimony analysis of endemicity. Palaeogeogr. Palaeoclim. Palaeoecol. 137: 273-288.        [ Links ]

33. Goloboff P (1993) Estimating character weights during tree search. Cladistics 9: 83–91.        [ Links ]

34. Goloboff PA (1997) Self-weighted optimization: tree searches and character state reconstructions under implied transformation costs. Cladistics 13: 225-245.        [ Links ]

35. Grinnell J (1917) Field tests of theories concerning distributional control. Amer. Nat. 51: 115-128.        [ Links ]

36. Hall ER (1981) The mammals of North America. Vols. I and II. Wiley. New York, USA. 1181 pp.        [ Links ]

37. Harold AS, Mooi RD (1994) Areas of endemism: Definition and recognition criteria. Syst. Biol. 43: 261-266.        [ Links ]

38. Henderson IM (1991) Biogeography without area? Austr. Syst. Bot. 4: 59-71.        [ Links ]

39. Humphries CJ, Parenti LR (1999) Cladistic biogeography: Interpreting patterns of plant and animal distributions, 2^nd ed. Oxford University Press. London, UK. 187 pp.        [ Links ]

40. Hunt JS, Bermingham E, Ricklefs RE (2001) Molecular systematics and biogeography of Antillean thrashers, tremblers, and mockingbirds (Aves: Mimidae). Auk 118: 35–55.        [ Links ]

41. Illoldi-Rangel P, Sánchez-Cordero V, Peterson AT (2004) Predicting distributions of Mexican mammals using ecological niche modeling. J. Mammal. 85: 658–662.        [ Links ]

42. InfoNatura (2004) InfoNatura: Birds, mammals, and amphibians of Latin America. Version 3.1. Arlington, VI, USA. NatureServe. www.natureserve.org/infonatura.        [ Links ]

43. Leopold AS (1983) Fauna silvestre de México. Prax. Mexico. 568 pp.        [ Links ]

44. Luna-Vega I, Alcántara Ayala O, Morrone JJ, Espinosa Organista D (2000) Track analysis and conservation priorities in the cloud forests of Hidalgo, Mexico. Divers. Distrib. 6: 137-143.        [ Links ]

45. MacArthur RH (1972) Geographical Ecology. Harper/Row. New York, USA. 269 pp.        [ Links ]

46. Maldonado JE (1999a) Family Soricidae. In Álvarez-Castañeda ST, Patton JL (Eds.) Mamíferos del Noroeste de México. CIBNOR. La Paz, Mexico. pp. 39-52.        [ Links ]

47. Maldonado JE (1999b) Family Talpidae. In Álvarez-Castañeda ST, Patton JL (Eds.) Mamíferos del Noroeste de México. CIBNOR. La Paz, Mexico. pp. 53-55.        [ Links ]

48. Márquez J, Morrone JJ (2003) Análisis panbiogeográfico de las especies de Heterolinus y Homalolinus (Coleoptera: Staphylinidae: Xantholinini). Acta Zool. Mex. (n. s.) 90: 15-25.        [ Links ]

49. Mast AR, Nyffeler R (2003) Using a null model to recognize significant co-occurrence prior to identifying candidate areas of endemism. Syst. Biol. 52: 271-280.        [ Links ]

50. Montresor M, Lovejoy C, Orsini L, Procaccini G, Roy S (2003) Bipolar distribution of the cyst-forming dinoflagellate Polarella glacialis. Polar Biol. 26: 186–194.        [ Links ]

51. Morrone JJ (1994) On the identification of areas of endemism. Syst. Biol. 43: 438-441.        [ Links ]

52. Morrone JJ (2001) Homology, biogeography and areas of endemism. Div. Distr. 7: 297-300.        [ Links ]

53. Morrone JJ, Escalante T (2002) Parsimony Analysis of Endemicity (PAE) of Mexican terrestrial mammals at different area units: When size matters. J. Biogeogr. 29: 1095-1104.        [ Links ]

54. Morrone JJ, Espinosa-Organista D, Llorente-Bousquets J (2002) Mexican biogeographic provinces: Preliminary scheme, general characterizations, and synonymies. Acta Zool. Mex. (n. s.) 85: 83-108.        [ Links ]

55. Murguía M, Villaseñor JL (2003) Estimating the effect of the similarity coefficient and the cluster algorithm on biogeographic classifications. Ann. Bot. Fenn. 40: 415-421.        [ Links ]

56. Patton JL (1999) Family Geomyidae. In Álvarez-Castañeda ST, Patton JL (Eds.) Mamíferos del Noroeste de México. CIBNOR. La Paz, Mexico. pp. 321-350.        [ Links ]

57. Patton JL, Álvarez-Castañeda ST (1999) Family Heteromyidae. In Álvarez-Castañeda ST, Patton JL (Eds.) Mamíferos del Noroeste de México. CIBNOR. La Paz, Mexico. pp. 351-443.        [ Links ]

58. Peterson AT, Ball LG, Cohoon KP (2002) Predicting distributions of Mexican birds using ecological niche modelling methods. Ibis 144 (on-line): E27–E32.        [ Links ]

59. Peterson AT, Egbert SL, Sánchez-Cordero V, Price KP (2000) Geographic analysis of conservation priority: Endemic birds and mammals in Veracruz, Mexico. Biol. Cons. 93: 85–94.        [ Links ]

60. Peterson AT, Soberón J, Sánchez-Cordero V (1999) Conservatism of ecological niches in evolutionary time. Science 285: 1265-1267.        [ Links ]

61. Ramírez-Pulido J, Castro-Campillo A (1990) Provincias Mastofaunísticas 1:4000000 Map IV.8.8A. Atlas Nacional de México. Vol. 2. UNAM. Mexico.        [ Links ]

62. Riddle BR, Hafner DJ, Alexander LF, Jaeger JR (2000) Cryptic vicariance in the historical assembly of a Baja California Peninsular Desert biota. Proc. Natl. Acad. Sci. 97: 14438-14443.        [ Links ]

63. Rojas-Soto OR, Alcántara-Ayala O, Navarro AG (2003) Regionalization of the avifauna of the Baja California Peninsula, México: A parsimony analysis of endemicity and distributional modeling approach. J. Biogeogr. 30: 449-461.        [ Links ]

64. Rosen BR (1988) From fossils to earth history: Applied historical biogeography. In Myers AA, Giller P (Eds.) Analytical biogeography: An integrated approach to the study of animal and plant distributions. Chapman/Hall. London, UK. pp. 437-481.        [ Links ]

65. Rosen BR, Smith AB (1988) Tectonics from fossils? Analysis of reef-coral and sea-urchin distributions from late Cretaceous to Recent, using a new method. In Audley-Charles MG, Hallam A (Eds.) Gondwana and Tethys. Special Publication of the Geological Society of London 37. London, UK. pp. 275-306.        [ Links ]

66. Rzedowski J (1986) Vegetación de México. Limusa. Mexico. 432 pp.        [ Links ]

67. Sánchez-Cordero V, Peterson AT, Escalante-Pliego P (2001) El modelado de la distribución de especies y la conservación de la diversidad biológica. In Hernández HMA, García-Aldrete N, Álvarez F, Ulloa M (Eds.) Enfoques contemporáneos para el estudio de la biodiversidad. Ediciones Científicas Universitarias / Fondo de Cultura Económica / Academia Mexicana de Ciencias / UNAM. Mexico. pp. 359-379.        [ Links ]

68. Sánchez-Cordero V, Munguía M, Peterson AT (2004) GIS-based predictive biogeography in the context of conservation. In Lomolino MV, Heaney LR (Eds.) Frontiers of biogeography: New directions in the geography of nature. Sinauer. Sunderland, MA, USA. pp. 311-323.        [ Links ]

69. Sánchez-Cordero V, Cirelli V, Munguía M, Sarkar S (2005a) Place prioritization for biodiversity representation using species’ ecological niche modeling. Biodiv. Inform. (on line) 2: 11-23.        [ Links ]

70. Sánchez-Cordero V, Illoldi-Rangel P, Linaje M, Sarkar S, Peterson AT (2005b) Deforestation and extant distributions of Mexican endemic mammals. Biol. Cons. 126: 465-473.        [ Links ]

71. Santos CMD (2005) Parsimony analysis of endemicity: time for an epitaph? J. Biogeogr. 32: 1284-1286.        [ Links ]

72. Soberón J, Peterson T (2005) Interpretation of Models of Fundamental Ecological Niches and Species’ Distributional Areas. Biodiv. Inform. (on line) 2: 1-10.        [ Links ]

73. Stockwell DBR, Peters D (1999) The GARP modeling system: Problems and solutions to automated spatial prediction. Int. J. Geogr. Inf. Sci. 13: 143-158.        [ Links ]

74. Swofford DL (2002) PAUP* v. 4.0b10. Phylogenetic analysis using parsimony (*and other methods), ver. 4. Sinauer. Sunderland, MA, USA.        [ Links ]

75. Szumik AC, Goloboff P (2004) Areas of endemism: an improved optimality criterion. Syst. Biol. 53: 968-977.        [ Links ]

76. Szumik CA, Cuezzo F, Goloboff PA, Chalup AE (2002) An optimality criterion to determine areas of endemism. Syst. Biol. 51: 806-816.        [ Links ]

77. Takami Y (2000) Phylogeny of the subgenus Ohomopterus (Coleoptera, Carabidae, genus Carabus): A morphological aspect. Tokyo Metrop. Univ. Bull. Nat. Hist. 4: 1-32.        [ Links ]

78. Thollesson M (2000) Increasing fidelity in parsimony analysis of Dorid nudibranchs by differential weighting, or a tale of two genes. Mol. Phylogenet. Evol. 16: 161–172.        [ Links ]

79. Turner H, Zandee M (1995) The behaviour of Goloboff’s tree fitness measure F. Cladistics 11: 57-72.        [ Links ]

80. Villa B, Cervantes FA (2003) Los mamíferos de México. UNAM / Iberoamérica. Mexico. 140 pp.        [ Links ]

81. Wiens JJ, Graham CH (2005) Niche conservatism: Integrating evolution, ecology, and conservation biology. Annu. Rev. Ecol. Syst. 36: 519-539.        [ Links ]

82. Yensen E, Valdés-Alarcón M (1999) Family Sciuridae. In Álvarez-Castañeda ST, Patton JL (Eds) Mamíferos del Noroeste de México. CIBNOR. La Paz, Mexico. pp. 239-320.        [ Links ] ]]> 1 1999 446-568 2 2004 29 112-120 3 2000 130 603-633 4 2003 162 211-232 5 1996 23 269-277 6 2004 20 56-63 7 2004 8 1997 99 9 2003 26 46-55 10 1999 141-181 11 2000 97 6585-6590 12 2003 30 819-825 13 2002 115 488-533 14 1993 2 667-680 15 2000 97 11343-11347 16 1999 199-237 17 1990 18 Conabio 1998 19 2001 66 332-344 20 2005 41 37-41 21 2003 74 211-237 22 2005 339-350 23 2003 48 563-578 24 2004 83 327-339 25 2005 76 199-205 26 27 2003 9 313-330 28 2001 31-37 29 ESRI 1999 30 1998 Ed s. 315-352 31 1998 3-108 32 1998 137 273-288 33 1993 9 83-91 34 1997 13 225-245 35 1917 51 115-128 36 1981 I and II 1181  37 1994 43 261-266 38 1991 4 59-71 39 1999 2nd 187 40 2001 118 35-55 41 2004 85 658-662 42 InfoNatura 2004 43 1983 568 44 2000 6 137-143 45 1972 269 46 1999 39-52 47 1999 53-55 48 2003 90 15-25 49 2003 52 271-280 50 2003 26 186-194 51 1994 43 438-441 52 2001 7 297-300 53 2002 29 1095-1104 54 2002 85 83-108 55 2003 40 415-421 56 1999 321-350 57 1999 351-443 58 2002 59 2000 93 85-94 60 1999 285 1265-1267 61 1990 2 62 2000 97 14438-14443 63 2003 30 449-461 64 1988 437-481 65 1988 275-306 66 1986 432 67 2001 359-379 68 2004 311-323 69 2005 2 11-23 70 2005 126 465-473 71 2005 32 1284-1286 72 2005 2 1-10 73 1999 13 143-158 74 2002 75 2004 53 968-977 76 2002 51 806-816 77 2000 4 1-32 78 2000 16 161-172 79 1995 11 57-72 80 2003 140 81 2005 36 519-539 82 1999 239-320