Introduction

The Sierra Madre Occidental (SMO) is the largest mountain range in Mexico, one of the largest silicic igneous provinces in the world, and the largest of the Cenozoic (^{Ferrari et al., 2007}). It constitutes the largest portion of the “Mexican mountain component” or Mexican Transition Zone, defined by ^{Morrone and Márquez (2003)} as a natural biogeographic region of Mexico. With nearly 1,400 km in length and 240 in width, it covers about 12.5% of the Mexican territory, and part of the Mexican states of Sinaloa, Sonora, Chihuahua, Durango, Nayarit, Zacatecas, Jalisco, and Aguascalientes. The SMO is a source of hydric environmental services for large areas of northwestern Mexico (Fig. 1) and produces about 60% of the standing timber in the country (^{González-Elizondo et al., 2013}). The SMO has been actively mined for iron, gold, silver, and other metals at least since the arrival of the Europeans to Mexico (^{Coll-Hurtado et al., 1990}).

Figure 1 A. Elevational gradient and main rivers in the Sierra Madre Occidental, Mexico. B. Vegetation of the Sierra Madre Occidental after ^{González-Elizondo et al. (2012)}. 

The SMO extends from the northern Sonora state up to the Mexico-U.S. border, to the Trans-Mexican Volcanic Belt and Sierra Madre del Sur mountain ranges. It follows a northwest-southeast path parallel to the Pacific coast with elevations ranging from 300 to 3,340 m above sea level, with an average height of 2,500 m (^{González-Elizondo et al., 2013}). Because of its position, it has a continental and an oceanic slope, and because of its elevation the eastern slope creates an orographic shade on the continental side, which is arid or semi-arid, whereas the western slope and highlands are more mesic. The western side is a series of canyons and ravines that can reach a depth of over 1,500 m. These characteristics result in a high topographic, climatic, and by extension, biological diversity which is seldom evident in biogeographic work (e. g. ^{Morrone, 2020}).

The relevance of the SMO as a biographic entity and as a center of biodiversity has been widely recognized by students of flora and vegetation (^{Bye, 1995}; ^{Felger & Wilson, 1995}; ^{Spellenberg et al., 1996}; ^{Felger et al., 1997}; ^{González-Elizondo, 1997}; ^{Luquín et al., 2004}; ^{Van Devender et al., 2003}; ^{Van Devender & Reina, 2005}; ^{González-Elizondo et al., 2007}; ²⁰¹³, among others). This body of work documented that the structure and diversity of the plant associations is complex and far from understood, and that there is a knowledge gap on the ecological and historical underpinnings that result on the current vegetative cover. The SMO is also a center of diversification and speciation for many taxa, a good proportion of which remain unknown to science. Understanding the diversity of the SMO becomes urgent given the ongoing rate of anthropogenic modification and climate change.

The analyses on which the planning of protected areas of Mexico is based (^{Koleff & Urquiza-Haas, 2011}) considered only five sites of extreme importance for conservation in the SMO, or approximately 1,046 km² of a total area of 251,546 km, which is little over 0.41% of its total area. This analysis was based on information on vertebrates as well as several ecological and goal-oriented criteria. The results suggest a paucity on the information available for the vertebrate biodiversity of the SMO rather than a real assessment of the biological importance of the area (^{Warshall, 1995}; ^{Flesch et al., 2016}; ^{González-Elizondo et al., 2013}).

For mammals, no comprehensive study on the SMO as a biogeographic entity exists. The available works include state-wide taxonomic treatises (^{Anderson, 1972} for Chihuahua; ^{Baker & Greer, 1962} for Durango; ^{Matson & Baker, 1986} for Zacatecas), and partial studies for selected groups (e. g. ^{Arroyo-Cabrales et al., 2008} for bats of Nayarit and ^{Carleton et al., 1982} for some rodents, also of Nayarit). There are bibliographic compilations for some areas: Sonora (^{Álvarez-Castañeda & Patton, 1999}, ²⁰⁰⁰, Aguascalientes (^{Álvarez-Castañeda et al., 2008}); or state-wide checklists (Aguascalientes, ^{de la Riva 2008}; Sonora, ^{Castillo-Gámez et al., 2010}; Nayarit, ^{Arroyo-Cabrales et al., 2015}; ^{Ramírez Silva et al., 2015}, ²⁰²¹; Chihuahua, ^{López-González & García Mendoza, 2012}; Durango, ^{García-Mendoza & López-González, 2013}; Jalisco, ^{Guerrero-Vázquez et al., 2017}; Zacatecas, ^{López-Ortega et al., 2020}). More detailed studies address mammal diversity at the regional level (e. g. ^{List & McDonald, 1998}; ^{Pacheco et al., 2000}; ^{López-González & García-Mendoza, 2006}; ^{Torres-Morales et al., 2010}; ^{Espinosa-Flores et al., 2012}; ^{López-González et al., 2014a}; ^{Coronel-Arellano et al., 2016}; ^{López-González et al., 2022}).

The main objective of this paper is to present an updated account of the mammals that occur in the SMO, based on verified museum information, our own collections, direct observations, photographic records, and specimen- or photograph-based literature data. Secondly, we described and compared the distribution of species by vegetation type and ecoregion (after ^{González-Elizondo et al., 2013}) and examined species turnover across vegetations and ecoregions describe the variation of species composition across biomes. Finally, we contrasted the species richness of the SMO with other major Mexican mountain ranges.

Materials and methods

Study area The Sierra Madre Occidental is an igneous mountain range that occupies an area of 251,546 km² (Fig. 1) according to the polygon provided by S. González-Elizondo of CIIDIR Durango. Because of its position, it has a high climatic and vegetational diversity. Elevation ranges from 300 to 3,340 m, with an average elevation of 2,500 m (^{González-Elizondo et al., 2012}; ²⁰¹³; Fig. 1A). It has a rugged physiography of highland plateaus and deeply cut canyons that spans nearly 1,400 km along the Pacific coast of Mexico, from near the US border (30° 35’ N) to northern Jalisco and western Aguascalientes (21° 00’ N) (^{González-Elizondo et al., 2013}). It is a continuous range from Jalisco to Sonora, where it breaks into a set of small, isolated mountain ranges (Sky Islands). Climate is tropical dry on most of the western slope (BS climates), with winter precipitation and a marked dry season the rest of the year, but tropical mesic (Aw) climates occur in the southern portion of the western slope. Temperate climates with summer rains dominate in the highlands (Cw climates), and semiarid climates dominate on the eastern slope (BS climates, ^{González-Elizondo et al., 2012}; ^{García, 1998}). Native vegetation includes pine forest, pine-oak forest, oak woodland, grassland, tropical deciduous forest, tropical semideciduous forest, subtropical scrub, thorn forest, chaparral, evergreen scrub, and small areas of xerophytic scrub, temperate forest gaps, mesophytic forest, gallery forest (Fig. 1B). A considerable area on the eastern slope and highlands has been modified for agricultural and livestock practices (^{González-Elizondo et al., 2012}; Fig. 1B). Our analysis follows the delimitation of the SMO by ^{González-Elizondo et al. (2013)}. The limits of the SMO polygon were set by these authors at 1800 m of elevation on the eastern slope, at the transition with the Mexican Plateau; at 300 m on the Pacific slope, and at an average of 1400 m to the north. This delimitation excludes the intermontane valleys of the Sky islands.

Data collection. Data were gathered from four different sources: 1) specimens of mammals collected by us and deposited at Colección Regional Durango Mexico (CRD), CIIDIR Durango, Instituto Politécnico Nacional, and specimens also examined by us, but deposited at other collections: Natural History Museum, The University of Kansas, Lawrence, USA (KU); Museum of Michigan State University, Lansing, Michigan, USA (MSU); Museum of Texas Tech University, Lubbock, Texas USA (TTU); Biodiversity Research and Teaching Collections, Texas A&M USA (TCWC); Mammal Collection, Universidad Autónoma Metropolitana Unidad Iztapalapa (UAMI); Mammal Collection, Instituto de Biología, UNAM, Mexico (CNMA), and Mammal Collection, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, Mexico (ENCB), for a total of 2,757 records. 2) Literature records from peer-reviewed publications in which at least a point locality and species can be unequivocally identified. We only included records for which a museum voucher or photograph exists, for a total of 1,529 records. Bibliographic data and collection records of Peromyscus of the boylii species group not examined by us were omitted because the taxonomy and nomenclature has been changing in the past 50 years, and because without a specimen or tissue sample it is not possible to ascertain the identity some species. These species were included only from recent species descriptions or taxonomic revisions (e. g. ^{Bradley et al., 2004}; ^{Mathis et al., 2013a}; ^b). 3) observations and photographs of medium-sized and large mammals made by the authors by direct observation or using camera-traps, or photographs of sign and track (199 records). 4) database of mammals from Durango and Chihuahua deposited at the mammal collection, Brigham Young University, Utah, and records from the databases SNIB (^{CONABIO, 2022}) and ^{Vertnet (2022)}, containing information about specimens deposited in North American Collections. From these bases we included only records of medium-sized and large mammals (Orders Artiodactyla, Cingulata, Didelphimorphia, Carnivora, Family Sciuridae), for which most species are easily identifiable (176 records). For Geomyidae we followed the distribution of species proposed by ^{Mathis et al. (2013a}; ^b). Because the data from 4) were not verified by us, if a specimen examined by us for the same species and locality existed, we referred the data to the voucher. In all cases records with uncertain or unlikely localities were eliminated. The final base consisted of 4,659 records of 1,392 unique localities (Fig. 2). Classification of orders, families, and genera follow ^{Ramírez-Pulido et al. (2014)}, species nomenclature follows the Mammal Diversity Database of the American Society of Mammalogists (^{Mammal Diversity Database, 2022}).

Figure 2 Ecoregions of the Sierra Madre Occidental, after ^{González-Elizondo et al. (2013)}, and localities of occurrence of mammals. 

Georeferencing. Most records from CRD, field observations and camera trap records were georeferenced in the field using GPS devices with a 4-m resolution. All other records were georeferenced using Google Earth 7.3.6.9285 (^{Google LLC, 2022}). Data from CNMA and 4 had associated coordinates, however, errors were detected in some collections, so we decided to revise all localities and georeferenced the erroneous ones again. Latitude and longitude were converted to, or estimated as, decimal degrees; elevation was calculated as, or converted to, meters above sea level.

Data analysis. Records were mapped using ArcMap V.10.1 (^{ESRI Inc., 2012}) over digital layers of political division (^{CONABIO, 2003}), vegetation (^{González-Elizondo et al., 2012}), and ecoregions (^{González-Elizondo et al., 2013}) to establish the state, vegetation types, and ecoregions where each species occurs, and to visualize information gaps. One record of one species in a state, ecoregion, or vegetation type sufficed to consider that the species occurs in that state, vegetation type, or ecoregion. To facilitate the understanding of the geographic components of the Madrean Region, the subregions were labeled as Madrean North (MN), Madrean Central (MC), and Madrean South (MS) as in ^{González-Elizondo et al. (2012)} (Fig. 2). We list species richness by vegetation type and state in Supplementary Material S1a, b.

To examine species turnover, using PAST 3.20 (^{Hammer et al., 2001}) we conducted a Principal Coordinates Analysis (^{Manly, 2005}) on a matrix of ecoregions-by species using the Whittaker turnover index (^{Tuomisto, 2010}) as a measure of distance. An analogous analysis was performed in a vegetation-by-species matrix using the categories of ^{González-Elizondo et al. (2012)}. We excluded agricultural areas, areas without vegetation cover, secondary chaparral, urban areas, and human-induced grasslands from the analysis because they occur throughout the SMO and their inclusion may conceal the patterns of species richness in native biomes. ^{Gonzalez-Elizondo et al. (2012)} recognized three kinds of oak forests: “bosque bajo abierto” (of temperate affinities), an open canopy oak or oak-pine woodland at mid elevations on the eastern slope; a temperate oak forest on the highlands of the eastern slope; and a low, open forest of tropical affinities on the western slope. Because of their ecological importance as transitional biomes, we followed this distinction. Finally, we contrasted the species richness at the SMO with other major mountain ranges in Mexico for which general accounts exist, namely the Sierra Madre Oriental (SMOr, ^{León-Paniagua et al., 2004}), and the Trans-Mexican Volcanic Belt (TMVB, ^{Escalante et al., 2007a}).

Results

Species richness and conservation status. Eight orders or 72.7% of the 11 orders of terrestrial mammals occurring in Mexico, considering Erinaceomorpha as part of Eulipotyphla (^{Mammal Diversity Database, 2022}), 21 families (58.3%), 82 genera (48.5%), and 171 species (34.3%, ^{Ramírez-Pulido et al., 2014}; ^{Sánchez-Cordero et al., 2014}) occur in the SMO (Appendix). Only the orders Sirenia, Primates and Perissodactyla are not represented. The SMO encompasses part of eight Mexican states. Durango had the highest richness (133), followed by Chihuahua (106), Zacatecas (101) and Nayarit (84). Aguascalientes, the smallest state had the lowest richness (37, Supplementary material S1a). From the 171 species recorded, 47 are endemic of Mexico (27.5%; Table 1), and ten are endemic of the SMO: the ground squirrels Callospermophilus madrensis Merriam, 1901, Neotamias bulleri (J.A. Allen, 1889), the woodrats Nelsonia neotomodon (Merriam, 1897) and Neotoma palatina Goldman, 1905, the deermice Peromyscus polius Osgood, 1904, P. schmidlyi, and P. carletoni; the harvest mouse Reithrodontomys zacatecae Merriam, 1901, and the pocket gophers Thomomys sheldoni and T. nayarensis. All but N. palatina are distributed in temperate, forested areas.

The NOM-ECOL-059 (^{SEMARNAT, 2019}) includes 25 species with some status of concern that occur on the SMO: seven under special protection (Leptonycteris yerbabuenae Martínez & Villa-R., 1940, Euderma maculatum [J. A. Allen, 1891], Myotis carteri, Cynomops mexicanus, Callospermophyllus madrensis, Dipodomys phillipsii J. E. Gray, 1841, and N. neotomodon), and the populations of Sciurus aberti Woodhouse, 1853 from the SMO (considered endemic to Mexico by ^{SEMARNAT, 2019}). Ten are considered as threatened: the shrews Megasorex gigas (Merriam, 1897), Notiosorex crawfordi (Coues, 1877), Notiosorex evotis (Coues, 1877), Sorex monticola Merriam, 1890, the nectar-feeding bats Choeronycteris mexicana Tschudi, 1844, and Leptonycteris nivalis (Saussure, 1860), the squirrel Sciurus arizonensis Coues, 1867, the yaguaroundi Herpailurus yagouaroundi (É. Geoffroy Saint-Hilaire, 1803), the river otter Lontra longicaudis (Olfers, 1818), and the American badger Taxidea taxus. Six species are considered as endangered: the shrew Sorex arizonae Diersing & Hoffmeister, 1977, the margay Leopardus wiedii (Schinz, 1821), the ocelot Leopardus pardalis (Linnaeus, 1758), the jaguar Panthera onca (Linnaeus, 1758), the wolf Canis lupus Linnaeus, 1758, and the black bear Ursus americanus Pallas, 1780. The brown bear Ursus arctos Linnaeus, 1758 is listed as extinct in the wild, but because there are extant populations in northern North America (^{Wilson & Ruff, 1999}), it should be considered as extirpated from Mexico (^{Ceballos & Navarro, 1991}). There were recent attempts of reintroduction of the Mexican wolf to the SMO, with mixed results (^{Cruz-Romo et al., 2013}). More recent, unpublished reports estimate the number of Mexican wolves in the wild in Mexico from at least 44 (^{SEMARNAT, 2023}), to 33 (^{USFWS, 2017}), to 4 (^{Western Watersheds Project, 2023}). It remains unclear whether an established population exists in the SMO. The IUCN classification lists as endangered only one species (L. nivalis), and 14 as near threatened or vulnerable (^{IUCN, 2022}, Appendix). The remaining 156 are considered as not evaluated (22) or of least concern (134). The number of Mexican endemics by region was highest in the Tropical region (T, 37), followed by Madrean South (MS, 24), Madrean Tropical (MT, 22), Madrean Xerophylous (MX, 16), Madrean Central (MC, 15), Madrean North (MN, 11) and Sky Islands (SI, 1).

Species turnover. The Tropical region was the richest in species with 125, followed by MS (100), MX (95), MT (89), MC and MN with 75 each, and SI (43). Turnover (as distance in ordination plot, Fig. 3A) was the highest between the Sky Islands and the rest of the ecoregions, scoring on the extreme positive side of the ordination plot. The tropical regions (MT and T) clustered on the negative side of PCo1. A “temperate” cluster including all the Madrean regions (MX, MN, MC, and MS), scored in the central portion of PCo1. Along PCo2, the tropical regions score on the positive side, whereas the temperate regions score on the negative side. The first axis shows the turnover of species of tropical affinities, MS, which shares many tropical species with the temperate areas, scores on the negative side, whereas MC, MX, and MN have progressively less tropical influence. The SI have an extreme score because data available include species that can occur in the SI and tropical areas, such as mid-sized mammals and some bats, but there are few rodent records. Adding the rodent fauna to the species list presented here likely will change the position for this ecoregion to a more realistic one. Along the second axis of variation (PCo2) the trend is similar but inverse.

Figure 3 Principal Coordinates Analysis of the distribution of 171 species of mammals that occur in the SMO. A) by ecoregion: Tropical (T), Madrean Tropical (MT), Madrean North (MN), Madrean Central (MC) Madrean South (MS), Madrean Xerophylous (MX), Sky Islands (SI); B) by vegetation type: Oak or pine-oak woodlands (OPW), Pine forest (PF), Pine-oak forest (POF), Western oak forest (WOF), Estern oak forest (EOF), Mesophytic forest (MF), Gallery forest (GF), Chaparral (CH), Evergreen scrub (ES), Temperate forest gaps (TFG,) Tropical deciduous forest (TDF), Tropical semi-deciduous forest (TSF), Subtropical scrub (SS), Xerophytic scrub (XS), Thorn forest (TF), Grassland (GRS), (after ^{González-Elizondo et al., 2012}, ²⁰¹³). 

The PCoA analysis by vegetation (Fig. 3B) suggests two different sets of species: the first occurs in arid lands (xerophytic scrub, thorn forest, and chaparral) on the positive side of PCo1 and negative of PCo2. The second group scores on the negative side of PCo1 and is distributed along PCo2. This second group illustrates the turnover of species from west to east and from tropical to temperate across the SMO: on the positive side of PCo2 the species that occur in vegetations of the Pacific slope and western piedmont, and the transitional western oak forest and subtropical scrub. On the negate.ve side score the species that occur in temperate vegetations, from pine and pine-oak to grasslands and oak forests of the eastern slope. Thus, PCo1 is related with aridity, whereas PCo2 is related with elevation and slope (Fig. 3B). The position of mesophytic forest and gallery forests on the positive side of PCo1 likely reflects the paucity of data for these areas, which are very small and not well sampled. The small areas of mesophytic forest occur on the W slope of the southern SMO, the gallery forests throughout the SMO.

Comparison with other mountain ranges. Species numbers by order were similar between the three mountain ranges for most orders, except for Chiroptera (Fig. 4A); the TVMB has only 47 species of bats compared with the 59 of the SMO and 69 of the SMOr. The SMOr is the only range that harbors primates (Ateles geoffroyi Kuhl). The TVMB had the highest number of endemic species (50), followed by the SMO (47) and SMOr (39; Fig. 4B). The number of local endemic species differs for each mountain range, with SMO having 10, TVMB 15, and SMOr six.

Figure 4 A) Comparison of the species richness by order between the Sierra Madre Occidental (SMO, yellow) and the two next largest Mexican mountain ranges, Sierra Madre Oriental (SMOr, blue) and Trans Volcanic Mexican Belt (TVMB, red). Numbers on top of columns are species richness. B) Species richness by orders, families, species, Mexican endemics, and local endemics by mountain range. Data for TVMB from ^{Escalante et al. (2007a)}, data for SMOr from ^{León-Paniagua et al. (2004)}. 

Discussion

Sampling biases and limitations. We documented 171 species of mammals for the SMO, or 34.3% of the terrestrial mammals known for Mexico. ^{Ramírez-Pulido and Castro-Campillo (1990)} reported 169 species in a map of Mexican mammals. Our analysis of species distribution is based only on records of mammals that fall within the limits of the SMO as defined by ^{González-Elizondo et al. (2013)}. This means that records of species occurring on the valleys between the Sky Islands were not included. Therefore, the species count can be larger under a different definition of the SMO. The valleys that surround this system of biogeographic isolates warrants examination on their own right, given that the arid lands and the northern edge of the subtropical vegetations mingle in this area, and important conservation areas, such as the Janos Biosphere reserve, and the Campo Verde area of protection of flora and fauna occur there. No other range-wide work is available.

Some comprehensive inventories have been conducted in some CONABIO terrestrial priority areas (^{Arriaga et al., 2000}), Alta Tarahumara-Barrancas, Chihuahua (RTP 30; ^{López-González & García-Mendoza, 2006}), San Juan de Camarones, Durango (RTP-23; ^{Torres-Morales et al., 2010}), San Pedro-Mezquital river basin, Durango and Nayarit (RTP 57; ^{López-González et al., 2014a}), and Jesús María River basin, Nayarit (RTP 59; ^{García-Mendoza et al., 2021}; ^{López-González et al., 2022}). Other inventories have focused on protected areas like la Michilía, Durango (^{Álvarez & Polaco, 1984}); Área de Protección de Flora y Fauna Bavispe, Sonora (^{Coronel-Arellano et al., 2016}; ²⁰¹⁸); and Cerro Mohinora, Chihuahua (^{Webb & Baker, 1984}).

Many collections have been conducted along the Durango-Mazatlán, or the Chihuahua-Hermosillo roads (Fig. 2), the only paved roads that until early 2000 crossed the SMO from east to west. The construction of two more roads across the SMO has opened the opportunity for sampling in the SMO, unfortunately, at the same time the insecurity of these areas has increased, precluding the sampling of mammals in the area. The mid elevations (about 500-2000 m) of the western slope are the least sampled areas, particularly from central Chihuahua to central southern Durango. These areas are of difficult access, pronounced slopes and deep canyons, and therefore sampling is difficult. Large portions of the highlands (2000-3000 m) of northwestern Chihuahua and central north Durango are underexplored, even though most of the timber activity occurs there. In general, these results suggest that half of the SMO is under-sampled or has not been sampled at all (Fig. 2). Yet, between 2000 and 2020, three new species have been discovered using morphological or molecular means (Peromyscus schmidlyi^{Bradley, Carroll, Muñiz-Martínez, Hamilton & Kilpatrick, 2004}; Peromyscus carletoni^{Bradley, Ordóñez-Garza, Sotero-Caio, Huynh, Kilpatrick, Iñiguez-Dávalos & Schmidly, 2014}; and Thomomys nayarensis^{Mathis, M. S. Hafner, D. J. Hafner, & Demastes, 2013b}), and some populations formerly named Thomomys umbrinus (Richardson, 1829), are now recognized as a different taxon (T. sheldoni Bailey, 1915; ^{Mathis et al., 2013a}), which suggests that as species groups are studied using integrative taxonomy approaches, more species that so far remain cryptic will be discovered, either in the field or in collections. In addition, further research is being conducted using camera-traps in the northwestern edge of the SMO that is documenting medium-sized and large mammals in selected protected areas (^{Espinosa-Flores et al., 2012}; ^{Coronel-Arellano et al., 2016}; ^{Gallo-Reynoso et al., 2021}).

Endemic species. The Madrean ecoregions (MS, MC, MT) harbor the highest numbers of endemic species. Some are widely distributed along the highlands from Sonora to Nayarit (e. g. P. schmidlyi, R. zacatecae), but others are microendemic. For instance, the ground squirrel C. madrensisis is distributed only on the highland, mesic pine forests of northern Durango and southern Chihuahua above 2,400 m. Its distribution appears patchy, but the populations seem to be large in the patches it occupies (CLG, pers. obs.). However, no estimation of population size is available, and therefore it is not possible to ascertain its conservation risk. MS contained four of the 10 endemics of the SMO, which is explained by the presence of the Mezquital-San Pedro and Santiago River basins. Resulting from their environmental complexity, numerous micro-environments are created, which promote the diversification of populations, as documented by the pocket gophers (Thomomys) and the P. boylii species group (^{Mathis et al., 2013}a; b; ^{Bradley et al., 2014}). Other vertebrates for which data are available, also have high numbers of endemic species in the SMO. For herps, ^{Rodríguez-Gutiérrez (2023)} reported 116 species endemic of Mexico, and ^{Johnson et al. (2017)}, considered the SMO as one of the most important areas of endemism for Mexico. For birds, ^{Kobelkowsky-Vidrio et al. (2014)} reported 67 endemic or quasi-endemic of Mexico. Thus, it is likely that the number of endemic species, particularly of organisms like salamanders, shrews, and aquatic or semi-aquatic organisms is under-sampled, and that new endemic species are yet to be discovered.

Distribution and turnover. The general latitudinal gradient in species richness expected from south to north is not evident for the SMO. The southern states (Jalisco and Aguascalientes) do not have the highest species richness, nor the northernmost ones have the lowest richness. The states that encompass the largest proportions of the SMO are the northern ones (Chihuahua, Durango), whereas the ones with the smallest areas are located in the South (Jalisco, Aguascalientes). It could be possible that the effects of area and latitude could cancel each other out, and that the species composition were similar along the 10 degrees of latitude encompassed within the SMO. This is not the case, Durango, on a central-south position, has the highest species richness (Supplementary material S1a), whereas Zacatecas, to the south, and Chihuahua, to the north, have the second highest richness. Factors additional to area or latitude explain the species distributions in the SMO.

Variation in species richness throughout the SMO can be explained by the heterogeneity of its biomes, resulting from a combination of latitudinal, topographic, elevational, and longitudinal (east-to-west) variation. In terms of ecoregions, the area with the highest species richness is the western tropical slope (Fig. 3A). Although considered as a single ecoregion (Tropical, Fig. 2) by ^{González-Elizondo et al. (2013)}, it responds to latitudinal and humidity gradients. The forests on the western slope, approximately from Ruiz, Nayarit to Escuinapa, Sinaloa, are the most mesic and harbor the tallest tropical forests along the western slope. These forests occur only at the lowest elevations and not farther than 60 km E from the edge of the Marismas Nacionales (Fig. 2). The highlands of these areas contain small areas or mesophytic forests. To the northwest, there occur other tropical forests, drier and lower. From Escuinapa northwestward, the western slope of the SMO is dominated by tropical deciduous forest that gradually become subtropical scrub, to finally turn into desert scrub at the northern end of the SMO in western Sonora (^{González-Elizondo et al., 2012}). As a result, the mammalian fauna in the southern extreme of the SMO has more in common with the faunas of the coastal plain of Jalisco to the south, than with the coastal plain of northern Sinaloa (^{López-González et al., 2014a}; ^{López-González et al., 2022}).

Another dimension of complexity in the SMO is related to elevation and slope (roughly east-to-west variation). The extreme examples are the Mezquital and Santiago River basins in MS, which traverse the SMO from east to west, through a system of canyons that can reach over 1,500 m in depth (^{González-Elizondo et al., 2007}). The result is that the tropical forests of the Pacific slope can penetrate the Sierra, and as elevation rises and humidity diminishes, become subtropical scrub, and finally xerophylous scrub and chaparral on the eastern slope. The change is gradual, and vegetation responds to the local conditions at each point. Because the canyons have a convoluted course throughout, the result is a mosaic of vegetation and climatic conditions that change in short distances, both horizontally and vertically (Fig. 1A, B). A similar effect is observed in all the western slope, where canyons are also deep, creating corridors of tropical vegetation that intermingle with the highland conifer forests. Along the eastern, continental slope, a similar phenomenon occurs, albeit more gradually: the semiarid vegetations of the Mexican Plateau on the south and central SMO, and the Chihuahuan desert and northern prairies on the north, are replaced by temperate forests as elevation increases.

The mammalian fauna responds to this complexity as a function of the breadth of habitat, mobility capabilities, and ecological niche of each species (^{López-González & Lozano 2015}; ^{López-González et al., 2014a}, ²⁰²²). The species-by vegetation ordination (Fig. 3B) illustrates the species turnover across vegetation types, arranging mammalian species composition in each vegetation type according to slope, elevation and aridity gradients. For bats, species of tropical distribution are a subset of the Neotropical species pool plus a few endemics from Mexico. This species set can be further subdivided based on their ability for dispersing throughout the mountain range. One group includes species that occur only in the lowland (<1,000 m), more mesic tropical forests of the south (e. g. Rhogeessa parvula H. Allen, 1866, Dermanura phaeotis Miller, 1902, Glyphonycteris sylvestris Thomas, 1896). Another group is widely distributed in the Neotropics, occurs both in mesic and dry tropical areas, and also is able to penetrate the SMO through the canyons of the western slope, reaching elevations of over 2,000 m, e. g. Dermanura azteca and Anoura geoffroyi J. E. Gray, 1838 (^{Sánchez-Hernández et al., 2002}; ^{Ortega & Alarcón-D., 2007}). The distribution of this subgroup roughly coincides with the T and MT regions. It is possible that these species occur in the highlands only part of the year, and that they migrate locally along the elevational gradient following the canyons and avoiding the low temperatures and lack of resources in the highland winters. ^{López-Segoviano et al. (2019)}, proposed that something similar might be occurring for birds at El Palmito, Sinaloa, on the western slope of the Sierra.

The mammal data partially support the subdivision of the highlands in SI, MN, MC, and MS. Three areas are distinguishable in terms of species richness and distribution: The SI, MS, and MC + MN. Considering the latitudinal gradient only, a nested pattern of species turnover could be expected from south to north. However, the contrastingly lower richness of the Sky Islands (43 species) may be the result of collection deficit rather than latitudinal or environmental gradients. This area is one of the least sampled regions of the SMO (^{Van Devender et al., 2013}), whereas for Chihuahua, Durango, and Zacatecas there are state-wide, relatively comprehensive inventories. Moreover, because of their very quality of islands, a higher degree of diversification and turnover could be expected for non-flying, small mammals (^{Warshall, 1995}), compared with the continuous range to the south. Comprehensive inventories of the SI are needed to document the yet unknown diversity of the region, particularly of rodents, using integrative taxonomic approaches.

The highlands of the SMO (SI, MN, MC, and MN) include species of Nearctic distribution that reach the southern edge of their distribution in north-central Mexico. This fauna is a subset of those species widely distributed in the temperate areas of North America, except for Corynorhinus mexicanus, an endemic of the Mexican highlands. Many of these species occur both in MX and in the more mesic highlands (MN, MC, MS, Fig. 2). A few have been collected only on the arid lands, for instance Antrozous pallidus (LeConte, 1856), or in the eastern piedmont and northern arid lands on both slopes (e. g. Corynorhinus townsendii Cooper, 1837). A few species are widely distributed on the highlands of both slopes, e. g. Aeorestes cinereus (Palisot de Beauvois, 1796). Some of these species are probably able to effect seasonal elevational movements as well. It is even possible that, at least part of the year, in some areas with a marked elevational gradient, bats ascend to the forested areas during the night, and roost in lower, warmer places during the day. These hypotheses are yet to be formally tested. Other bats of temperate origin are able to migrate (like A. cinereus or T. brasiliensis I. Geoffroy Saint-Hilaire, 1824) or hibernate (Corynorhinus spp.; ^{López-González & Torres-Morales, 2004}) to escape the cold winters of the SMO highlands. Species like the nectar-feeding bats Leptonycteris nivalis and Choeronycteris mexicana are widely distributed in tropical and arid regions of México, and in the SMO occur on both slopes, but are limited in elevation by the distribution of the plants on which they feed (Bombacaceae, some legumes, Agave spp., Lemaireocereus spp., Ceiba spp., Ipomoea spp., Myrtillocactus spp.; ^{Arita & Humphrey, 1988}; ^{Álvarez & González Quintero, 1969}; ^{Stoner et al., 2003}).

For non-flying mammals, the patterns are more complex and differ between animals that move considerable distances during their lifetime (in the order of kilometers), and species that have small home ranges (like small rodents). For instance, the jaguar (Panthera onca) occurs throughout the western slope SMO, and has been recorded as far as northern Chihuahua, northwestern Sonora and Southern Arizona (^{Girmendonk, 1994}; ^{Brown & López-González, 2000}; ^{López-González & Brown, 2002}; ^{Navarro-Serment et al., 2005}). In contrast, some species of rodents inhabit only the highland conifer and mixed forests of the SMO, generally distributed above 2,300 m (e. g. the squirrel S. aberti, the deermouse P. schmidlyi, and the vole Microtus mexicanus (Saussure, 1861)) or small forest patches in the southern edge (e. g., P. carletoni). Other species have more complex distribution patterns: for instance, the cactus mice of the P. eremicus species group effectively surround the SMO from Nayarit to Sonora to Zacatecas (^{Riddle et al., 2000}), with tropical populations entering from the western slope as far as Mesa del Nayar in Nayarit (1,500 m of elevation) and entering from the eastern Mexican Plateau to the Mezquital River Basin at least as far as Agua Zarca, Durango (1,500 m). It is possible that populations of this species enter in contact in the southern portion of the SMO, in the Mezquital or Jesús María basin, but molecular data are not available for this part of the Sierra (^{Riddle et al., 2000}). This distributional pattern, and the presence of P. merriami Mearns, 1896, a closely related species, in the lowlands of eastern Nayarit (^{García-Mendoza et al., 2021}) suggests a more complex, yet to be investigated phylogeographic story for the Haplomylomys species group in western Mexico.

As previously documented for some areas of the SMO (^{López-González et al., 2014a}; ^{López-González et al., 2022}), its biological and evolutionary importance resides not only in its absolute richness, but on its high species turnover rate both at large and small scales (^{Rodríguez et al., 2003}). At the range scale, the east-west and north-south, effects of orographic shade and latitude, respectively, produce the Nearctic-Neotropical divide described by numerous biogeographers before (e. g. ^{Morrone & Márquez, 2003}; ^{Morrone, 2020}). At the local scale, environmental complexity, and physiographic variation due to latitude, longitude, canyon depth, elevation, aspect, and slope, create distinct local environments (Fig. 1B) in which habitat-restricted populations may diversify, regardless of their biogeographic affinities or origins. Turnover is only one of the components of beta diversity (^{Soininen et al., 2018}); disentangling the effects of turnover and nestedness in the SMO is beyond the scope of this paper, though some nestedness should be expected along elevational gradients on both the E and W slopes. The available phylogeographic work hints at the effects of scale and beta diversity, in addition to historic processes, to explain the diversity and distribution of the SMO biota (e. g. ^{Bryson et al., 2011}; ^{Gugger et al., 2011}; ^{Sánchez-Sánchez et al., 2012}; ^{López-González et al., 2014b}).

Comparisons with other mountain ranges. The other two major mountain ranges in Mexico, considered as part of the “Componente Mexicano de Montaña” or Mexican Transition Zone (^{Morrone & Márquez, 2003}; ^{Morrone, 2020}), and for which comparable data exist, are the Sierra Madre Oriental (SMOr) and Transvolcanic Mexican Belt (TVMB). In the SMOr, 201 species have been recorded (^{León-Paniagua et al., 2004}), at the TVMB, 152 (^{Escalante et al., 2007a}). Nonetheless, a relatively recent book on the biodiversity of northern Mexico, devotes very little attention to the SMO (^{Cartron et al., 2005}).

The diversity of tropical orders (Didelphimorphia, Cingulata) and of deer (Artiodactyla) is lower in the SMO than in the other mountain ranges because only the white-tailed deer (Odocoileus virginianus Zimmermann, 1780) and the peccari Dycotiles tajacu (Linnaeus, 1758) are able to reach the highlands, whereas other artiodactyls from northern Mexico like the pronghorn Antilocapra americana (Ord, 1815), the mule deer Odocoileus hemionus (Rafinesque, 1817), American bison (Bos bison Linnaeus, 1758) and bighorn sheep (Ovis canadensis Shaw, 1804), seldom have been recorded in the SMO, and all records are from the intermontane valleys of the Sky Islands (^{Pacheco et al., 2000}; ^{Medellín et al., 2005}; ^{Pelz-Serrano et al., 2006}; ^{Gallina-Tessaro et al., 2019}). Chiroptera and Rodentia have similar or higher species richness in the SMO because many species of temperate distribution that can occur in the highlands and lowlands, and because tropical species reach their distributional limits on the western slope. In contrast, in the TMVB and SMOr the species composition has more tropical and endemic components (^{López-Wilchis et al., 1990}; ^{León-Paniagua et al., 2004}; ^{Escalante et al., 2007a}).

The TVMB, one of the best studied areas of Mexico, harbors 50 mammals endemic of Mexico, whereas the SMO, a mountain range about three times larger (SMO 251,648 km² not including the Sky Islands, vs. 175,700 km²; ^{González-Elizondo et al., 2013}; ^{Valencia, 2007}) and much less explored, has 47. The proportions of local endemics (15 vs 10, respectively) suggest that some of the endemic fauna of the SMO is yet to be discovered. ^{Escalante et al. (2007b)}, using Parsimony Analysis of Endemicity, concluded that the SMO was an important area of endemism. However, in a second analysis (^{Escalante et al., 2009}), using a different methodology, the SMO was not considered as an area of endemism, except as part of the Mexican Transition Zone. These results probably reflect lack of information about SMO. These authors report three endemic species for the SMO (^{Escalante et al., 2007a}) and later they reported nine (^{Escalante et al., 2009}). Since these papers were published, at least four more endemic species have been described (P. schmidlyi, P. carletoni, T. nayarensis and T. sheldoni). This evinces the need for continuous sampling and inventory, as well as analysis of materials already in collections, to improve our understanding of this mountain range. We expect that in the next years, and if there are consistent efforts to sample the Sky Islands and the western slope of the SMO, particularly on the north, the number of species will become similar to that of the SMOr.

Conservation. The Sierra Madre Occidental is also complex in its human composition. In addition to the mestizo population, in the SMO inhabit six indigenous groups (^{Serrano Carreto, 2006}), some of which occur on the north (Tarahumara o Rarámuri, northern Tepehuano), and others on the south (Cora, Huichol, southern Tepehuano or Odam, and Mexicanero). In some areas one or more of these groups constitute most of the human population. Also, a Mormon (^{Waterbury, 2013}) and a Menonite population (^{le Polain de Waroux et al., 2021}), inhabit northern and central Chihuahua. The main occupations timber logging, agriculture, mining, livestock, and more recently hunting, environmental services, tourism, production of illicit drugs. Most of the population participates in one or more of these activities. Although the population density is relatively low in most of the SMO (about 27.5 inhabitants per squared km, estimated from ^{INEGI, 2020}), the SMO has been intensively modified in the last 100 years, so much so that old growth forest, for instance, was practically gone (99.4%) by the end of the 20^th century (^{Lammertink et al., 1996}).

Despite the magnitude in area and biological diversity of the SMO, the environmental services it provides (^{Monárrez-González et al., 2018}) and its high rate of modification, it contains very few protected areas. Some areas of the SMO have been designated as “Regiones Terrestres Prioritarias para la Conservación” (RTPs) or prioritary terrestrial regions for conservation by CONABIO, the Mexican agency in charge of biodiversity knowledge and management (^{Arriaga et al., 2000}). There are 35 RTPs (of 152 for Mexico) that are included, all or in part, in the SMO, covering approximately 71,776 km² or 28.52% of the SMO polygon as defined here. These areas, however, do not have any legal standing as protected areas, but they are considered important because of their biological characteristics. Only eight RTPs are in a protected area.

There are ten federal protected areas in the SMO: the Michilía Biosphere reserve; the highlands of the Janos Biosphere Reserve; three national parks, the Cascada de Basaseachic, Sierra de Órganos, and Cumbres de Majalca; six areas of flora and fauna protection: Cerro Mohinora, Campo Verde, Papigochic, Bavispe, Tutuaca and the highlands of Sierra de Álamos-Río Cuchujaqui. There are three state reserves, Sierra Fría in Aguascalientes, and el Tecuan state park, and the Area de Protección de Recursos Naturales Quebrada de Santa Bárbara in Durango, covering approximately 12,215 km² or 4.8% of the SMO polygon as to 2022 (^{Lozano-Román & Estrada-Aguilera, 2008}; ^{Rentería-Arrieta & Montiel-Antuna, 2017}; ^{CONANP, 2022a}). According to ^{CONANP (2022b)}, the Cuenca Alimentadora del Distrito Nacional de Riego 043 is a protected area that includes several municipalities and parts of five states (Aguascalientes, Jalisco, Durango, Nayarit, and Zacatecas), covering an area of 2,329,026.75 ha. However, this area includes several towns, agricultural areas, mountains, as well as part of the Rio Grande de Santiago Basin. Although some CONANP personnel is assigned in some regions of this vast area, it does not have a management plan, and because of its size and complexity, it is unlikely that the decree under which it was created can ever be enforced beyond some local, manageable areas. The currently protected areas do not include mesophytic forests, and only small areas of Pseudotsuga or Picea forests are included. Of the four priority areas for the conservation of bats (AICOMS, ^{Barquez et al., 2022}) that exist in the SMO, only a small portion of the Mezquital River Basin occurs in a protected area. It is remarkable the absence of RAMSAR sites in most of the SMO (^{RAMSAR, 2022}), probably explained by the lack of large water bodies (^{Sandoval et al., 2020}). Nonetheless, the SMO is winter refuge of numerous terrestrial North American bird species (^{Wehenkel et al., 2017}).

There SMO harbors 800 UMAs (units for management and exploitation of resources, ^{SEMARNAT, 2020}), a model of conservation that allows the use of wildlife through the sustainable management of their populations (^{SEMARNAT, 2018}). There were 1,722 UMAS in Mexico in 2018 (^{SEMARNAT, 2018}) of which almost half (46.5%) fell in the polygon of the SMO, covering 25,829 km² (10.3%, ^{SEMARNAT, 2020}). Most UMAs in the SMO are hunting ranches devoted to the harvesting and management of game species, either native (most) or introduced (a few). A recent assessment of the UMA system nationwide (^{CONABIO, 2012}) was conducted by regions, two of which include part of the SMO. Durango (with 279 UMAs in the SMO), Chihuahua (63), Sinaloa (14), and Sonora (146) were included in the Northwestern-North-Central region (NW-NC), whereas Nayarit (19), Jalisco (32), Zacatecas (196) and Aguascalientes (51) were included in the West-Central Region (^{CONABIO 2012}). According to this report, for the NW-NC zone, the one with the most UMAs, only 1.7% of them had objectives and management plans, and a system of indicators of success. Even these few had no clear objectives, or quantifiable measures of success of the management or monitoring goals. Furthermore, only 30.9% of applications for species harvesting included an estimation of population density (^{CONABIO, 2012}). Yet, 27.2% of the UMAs in the NW-NC zone harbored CITES species, and 18.3% harbored species protected by Mexican law (^{SEMARNAT, 2019}). In this information void, it is difficult to assess the impact of these entities as conservation units.

The clearing of the forests by the timber industry has resulted in the gradual loss of litter and therefore humidity of the naturally thin soils that constitute a large proportion of the SMO (^{Novo-Fernández et al., 2018}). The loss of organic matter together with the higher incidence of sunlight, is resulting in a gradually drier SMO. In this scenario, all species that require high levels of soil humidity and a thick litter cover to survive, such as voles and shrews, are probably threatened. Because of the pronounced slopes and shallow soils that characterize most of the SMO, it is likely that these species were never widely distributed or abundant, except on the relatively flat mesas of the central portion of the highlands. Yet, the already small areas that meet the requirements for their existence are bound to continue shrinking, and probably disappear as logging and climate change continue. ^{Vázquez et al. (2009)} concluded that the northern portion of the SMO was one of the three most important areas for conservation of Mexican mammals in Mexico. Based on our results, we propose that the southern portion is also a very important area for protection because of its biological complexity (^{González-Elizondo et al., 2013}; ^{López-González et al., 2014a}; ^{Kobelkowsky-Vidrio et al., 2014}; ^{López-González et al., 2022}). An informed reassessment of the conservation priorities that include the participation of all stakeholders and the new data available, is urgently needed for the SMO.