Biogeographic regions of North American mammals based on endemism

bs_bs_banner Biological Journal of the Linnean Society, 2013, 110, 485–499. With 6 figures Biogeographic regions of North American mammals based on endemism TANIA ESCALANTE1*, JUAN J. MORRONE1 and GERARDO RODRÍGUEZ-TAPIA2 Museo de Zoología ‘Alfonso L. Herrera’, Departamento de Biología Evolutiva, Facultad de Ciencias, Universidad Nacional Autónoma de México, Apartado Postal 70-399, 04510 Mexico, DF, Mexico 2 Unidad de Geomática, Instituto de Ecología, Universidad Nacional Autónoma de México, Apartado Postal 70-275, 04510 Mexico, DF, Mexico Received 4 April 2013; revised 22 May 2013; accepted for publication 22 May 2013 Since the 19th Century, two regions have been recognized for North American mammals, which overlap in Mexico. The Nearctic region corresponds to the northern areas and the Neotropical region corresponds to the southern ones. There are no recent regionalizations for these regions under the criterion of endemism. In the present study, we integrate two methods to regionalize North America, using species distribution models of mammals: endemicity analysis (EA) and parsimony analysis of endemicity (PAE). EA was used to obtain areas of endemism and PAE was used to hierarchize them. We found 76 consensus areas from 329 sets classified in 146 cladograms, and the strict consensus cladogram shows a basal polytomy with 14 areas and 16 clades. The final regionalization recognizes two regions (Nearctic and Neotropical) and a transition zone (Mexican Transition Zone), six subregions (Canadian, Alleghanian, Californian-Rocky Mountain, Pacific Central America, Mexican Gulf-Central America, and Central America), two dominions (Californian and Rocky Mountain), and 23 provinces. Our analysis show that North America is probably more complex than previously assumed. © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499. ADDITIONAL KEYWORDS: endemicity analysis – Mexican Transition Zone – Nearctic – Neotropical – parsimony analysis of endemicity – provinces – species distribution models. INTRODUCTION Biogeographical regionalizations are important products of biogeographical research. They are useful in conservation practices because they allow the documentation of spatial patterns of biodiversity and the proposal of criteria for identifying protected areas. Regionalizations are hierarchical schemes, based on patterns of nested areas of endemism, with provinces being the basic units (Escalante, 2009), followed by dominions, regions, and realms or kingdoms (Cabrera & Willink, 1973; Ebach et al., 2008). Areas of endemism identified under the concept of spatial homology should be based on the distributional congruence of two or more sympatric taxa restricted to *Corresponding author. E-mail: tee@ibiologia.unam.mx them (Morrone, 2001, 2009); however, several techniques of similarity that are currently used to propose hierarchical schemes were not developed explicitly for this objective, and so they lack a theoretical basis to justify their use (Murguía & Llorente-Bousquets, 2003). For example, cluster analyses group areas based on global similarity, fail to provide the endemic taxa that diagnose an area, and do not allow direct inferences on relationships of areas and historical patterns (Kreft & Jetz, 2010). There are some methods that have been developed to identify areas of endemism, although only two of these represent the most robust approximations: parsimony analysis of endemicity (PAE; Rosen, 1988; Morrone, 1994) and endemicity analysis (EA; Szumik et al., 2002; Szumik & Goloboff, 2004). PAE has been the most used method to identify areas of endemism in the last 20 years, although it has also © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 485 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 1 486 T. ESCALANTE ET AL. aiming to offer a regionalization of North America at the subregion, dominion, and province levels. MATERIAL AND METHODS We compiled a database from different sources for terrestrial mammals from Alaska and Canada to Panama: GBIF, Global Biodiversity Information Facility (http://www.gbif.org); MaNIS, Mammal Networked Information System (http://www .manisnet.org); UNIBIO, Unidad de Informática para la Biodiversidad (http://www.unibio.unam.mx); Conabio, Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (http://www.conabio.gob.mx); and Mammex, Mamíferos de México (T. Escalante, V. Sánchez-Cordero, M. Linaje & G. Rodríguez-Tapia, unpubl. data). The integrated database MamNA (Escalante & Rodríguez-Tapia, 2011) contains 245 818 records (unique combinations of scientific name and georeference) from 710 species of 11 orders of mammals. To eliminate or correct inconsistences in the database, all records were contrasted with maps of distributional areas of species from the literature and internet portals (Hall, 1981; http://www.mnh.si .edu/mna/; http://www.natureserve.org/infonatura/; http://www.conabio.gob.mx/informacion/mamiferos/ doctos/presentacion.html). We selected 652 species having more than five records in our database (Escalante, Morrone & Rodríguez-Tapia, 2013) and modelled their distributions using MAXENT, version 3.1 (Phillips, Anderson & Schapire, 2006; Phillips & Dudík, 2008). There is some controversy about the performance of models with few records; however, MAXENT works better than other algorithms with small samples (Hernandez et al., 2006; Wisz et al., 2008; Bean, Stafford & Brashares, 2012). We used 23 environmental data: 19 bioclimatic and four topographic layers at resolution of approximately 2 km (layers obtained from https://lta.cr.usgs.gov/ HYDRO1K and http://www.worldclim.org/; Hijmans et al., 2005). For each species, 25% of the records were used to validate the model internally. Model success was judged using two criteria: area under the curve (AUC) > 0.7 and P < 0.05 for at least one binomial test, as recommended by Pawar et al. (2007), both obtained from the software. The AUC is only truly informative when there are true instances of absence available and the objective is to estimate the realized distribution and this is not the case. Therefore, the values of AUC reported in the present study do not necessarily reflect the accuracy of the models because the AUC is highly prudent for evaluation of model accuracy, especially when true absences are unavailable for testing the models (Jiménez-Valverde, 2012). Although controversy © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 been somewhat controversial (Escalante, 2011; Donato & Miranda-Esquivel, 2012). PAE is used to identify areas of endemism based on the clades supported by two or more species or supraspecific taxa in a cladogram obtained by a parsimony analysis of a presence-absence data matrix. It has been compared to cluster analysis, having more advantages over the unweighted pair group method with arithmetic mean (Trejo-Torres & Ackerman, 2002; Carine et al., 2009), even when PAE can mistakenly be applied basing areas of endemism on paraphyletic groups (Moreno Saiz et al., 2013). EA uses an optimality criterion to assign a score of endemicity to a given area, based on the adjustment of species to that area, so that areas with a score above 2.0 are candidates to be recognized as areas of endemism. EA only uses grids as unit of analyses and is sensible to scale (Casagranda, Roig-Juñent & Szumik, 2009). Some studies have suggested that EA performs better than PAE (Carine et al., 2009; Escalante, Szumik & Morrone, 2009; Casagranda, Taher & Szumik, 2012), although we consider that both offer different advantages that can be complementary: EA is able to recognize best supported areas of endemism and PAE is able to hierarchize them. The Holarctic realm comprises basically the areas in the Northern Hemisphere, corresponding to the palaeocontinent of Laurasia. It has been divided in two regions: the Nearctic in the New World and the Palearctic region in the Old World (Morrone, 2002). The Nearctic region occupies almost all North America, and has been recognized since Sclater and Wallace in the 19th Century. It has been recently redefined using the concept of endemism and mammal distribution maps generalized to 4° latitude-longitude grid (Escalante et al., 2010). Some recent regionalizations of the world at regional levels (Procheş, 2005; Kreft & Jetz, 2010; Procheş & Ramdhani, 2012; Holt et al., 2013) used methods based on global similarity, and so the areas recognized cannot be interpreted in terms of their endemic taxa. The Nearctic region (Escalante et al., 2010) reaches 26° latitude, where it borders with the Neotropical region of the Holotropical realm (Morrone, 2002), and both regions overlap in a wide area named the Mexican Transition Zone (Halfter, 1962, 1964a, b; Morrone, 2005). Endemism patterns within the Nearctic region are rather unexplored. Moreover, which provinces are included in the transition zone is a matter of dispute (Escalante, Rodríguez & Morrone, 2004; Morrone, 2005; Espinosa et al., 2008). Our aims are twofold: (1) to identify areas of endemism of North America, in particular in the Nearctic region, using species distribution models of mammal species at a grid of 2° latitude-longitude with EA, and (2) to arrange them hierarchically using PAE, BIOGEOGRAPHICAL REGIONALIZATION OF NORTH AMERICA sus at 40% of similarity of species with any area. All consensus areas were analyzed based on their endemic species and sets included. To classify areas of endemism in subregions and dominions, the consensus areas were grouped undertaking a PAE. We built a new matrix of consensus areas (rows) and their genera (columns) because supraspecific taxa may be informative and allow to incorporate phylogenetic information to the analysis (Cracraft, 1991; Porzecanski & Cracraft, 2005; Morrone, 2009). The matrix was run in TNT, version 1.1 (Goloboff, Farris & Nixon, 2011), using the New Technology algorithm, finding the minimum length 100 times and 1000 trees, and using sectorial search, ratchet, and drift with default options. To penalize homoplasy, we performed implied weighting with k = 3.0 (Goloboff, 1993; Luna-Vega et al., 2000; Escalante et al., 2007b). A strict consensus cladogram was obtained and analyzed regarding its clades. The consensus areas and the models of the species diagnosing particular clades were mapped with ARCVIEW, version 3.1 (ESRI, 1998). Because one or more consensus areas may represent the same province (i.e. they have the same geographical location), some are not shown on maps. RESULTS The analysis with NDM led to 329 sets (candidate areas of endemism), with scores recalculated (substracting species with score less than 0.4) from 1.57 to 17.49. Based on these, 76 consensus areas were calculated, with maximum scores between 3.25 and 17.49 (Table 1; see also Supporting information, Table A1). There are 420 species that give score to some area, where Bauerus dubiaquercus and Eptesicus brasilensis were the species endemic to more sets (N = 21), followed by Trachops cirrhosus (N = 20). The PAE of the consensus areas and their genera resulted in 146 cladograms with a length of 396, a consistency index of 0.38, and a retention index of 0.38. The strict consensus cladogram had a length of 417, a consistency index of 0.36, and a retention index of 0.32. It shows a basal polytomy with 14 areas and 16 clades (Fig. 1). We consider that each consensus area is a province, and that several consensus areas of a clade are a subregion or dominion. Several clades occupying a similar geographical area were considered to belong to the same subregion. For the 16 clades, we found four larger patterns (Table 2), which correspond to the Eastern and Western subregions of the Nearctic region, the Neotropical region, and the Mexican Transition Zone. Also, there are three subregions within the Nearctic region and three subregions within the Neotropical region (Figs 2, 3, 4, 5, 6). © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 exists, the AUC values that we obtained may be used in this way according to Smith (2013), only assuming an apparent accuracy positively correlated with the current accuracy because generally the test sites are a sample of all possible sites. In the same way, the strategies used to evaluate the models need to be different because the weight of commission errors is definitively lower in the case of the potential distribution than in the case of the realized distribution (Lobo, Jiménez-Valverde & Real, 2008; Peterson, Papes & Soberón, 2008). Another important consideration for maintaining the spatial independence of separated localities is to consider a buffer exhibiting sufficient potential variation or difference between localities (Pearson et al., 2007). Models were generated in ascii format, and exported directly to the ARCVIEW GIS, version 3.1 (ESRI, 1998). Although there are many methods to select thresholds to transform outputs of models to a binary map, there are few comparisons and evaluation of them for MAXENT (Pearson et al., 2007; Aranda & Lobo, 2011; Bean et al., 2012), regarding the comparisson of thresholds that exist for other algorithms (Manel, Williams & Omerod, 2001; Liu et al., 2005; Jiménez-Valverde & Lobo, 2007; Freeman & Moisen, 2008; Nenzén & Araújo, 2011; Liu, White & Newell, 2013). Choice of threshold is mostly determined by the proposed application of the model and depends on the type of data that are available (Peterson, 2006; Peterson et al., 2011). Moreover, because the selection of a threshold should be attentive to the relative importance of the two primary forms of omission and commission errors (Liu et al., 2013), and data of several sources tend to have more errors of georeference, omission of 10% of data training may be considered. We refined each model with the the tenth percentile training presence threshold because this threshold also recovered better than others the patterns of endemism (Escalante et al., 2013). Additionally, we used maps and the literature to eliminate overpredictions (Hall, 1981; http://www.mnh.si.edu/ mna/; http://www.natureserve.org/infonatura/; http:// www.conabio.gob.mx/informacion/mamiferos/doctos/ presentacion.html). Each model was overlain with a grid of 2° latitude– longitude grid-cells and we assigned ‘1’ to the presence and ‘0’ to the absence of a species in a cell, aiming to build a binary matrix. The final matrix contains 652 species in 2280 cells. An EA was performed in NDM/VNDM, version 3.0 (Goloboff, 2011), saving all sets with a score > 2.0 and two or more species. We ran the search 100 times with 50% of unique species each time using edge proportions. From the sets obtained, we chose those species with a minimum score of 0.4 and we calculated the consen- 487 488 T. ESCALANTE ET AL. Table 1. Consensus areas of endemism in North America Consensus area Number of endemic species Geographical location 0, 78, 130, 132, 142, 189, 225 1, 88, 133 2 3 4, 12, 13, 20, 26, 71, 82, 87, 93, 102, 112, 116, 117, 119, 124, 128, 146, 147, 151, 168, 169, 181, 187, 191, 195, 203, 206, 213, 218, 223, 233, 234, 235, 245, 252, 254, 256, 259, 261, 262, 266, 267, 269, 270, 279, 280, 289, 294, 297, 306, 307, 312, 314, 318, 324, 326 5 6, 98, 222, 224, 227, 305 7, 121, 205, 228, 238, 325 8, 120, 143 9, 48, 53 3.2 4.94 14.91 14.50 8.14 Baja California peninsula Chiapas/Tehuantepec Isthmus Southern Central America Southern Central America Mexico 10, 24, 79, 106, 161 11, 67 14, 49, 72, 95, 96, 125, 144, 157, 159, 175, 177, 188, 215, 236, 241, 244, 249, 273, 274, 284, 300, 301, 311, 323, 15, 37, 141, 185 16, 58, 73, 85 17, 43, 81, 91, 149, 152, 220 18, 126 2.57 3.57 6.67 0 1 2 3 4 8 7 19 18 45 5 6 7 8 9 5 7 7 6 18 10 11 12 6 5 26 13 14 15 16 8 20 9 18 17 18 20 20 19 20 5 10 21 28 22 23 24 5 3 32 25 26 27 28 29 30 31 6 5 5 5 4 7 27 32 33 34 3 4 6 25, 40, 60, 68, 97, 100, 129, 136, 153, 160, 162, 173, 194, 202, 214, 229, 240, 257, 260, 264, 286, 287, 296, 298, 319, 321 27, 154 28 29, 69, 70, 114, 118, 134, 140, 178, 190, 242, 268, 278, 282 30 31, 108, 258 32 33 34, 59 35 36, 63, 74, 86, 94, 109, 113, 145, 156, 174, 193, 196, 201, 251, 292, 315, 38 39 41, 308 35 7 42, 166 5.39 36 37 38 5 7 3 44, 77, 165, 232, 290 45 47, 92 2.75 6.35 2.76 19 21, 46, 103, 123, 192, 208, 221, 281, 295, 304, 317 22, 99, 209 23, 105, 4.05 4.60 3.38 3.25 9.07 3.14 10.79 3.90 9.86 15.68 8.55 3.24 5.24 9.99 2.98 2.66 17.49 Northern Mexican Plateau Eastern USA Mexican Southeastern Western USA Mountains of Central and Southern Mexico Central USA Western USA Mexican Gulf/Southern Mexico/Central America Chihuahua/Sonora California Western Coast of USA Lowlands of Mexico and Central America Southern Central America Lowlands of Mexico and Central America Sonora Lowlands of southern Mexico and Pacific of Central America Western Coast of USA Sonora Mexico Chiapas and Central America 4.54 3.02 4.01 2.85 2.65 5.41 9.94 Mexican Gulf and Yucatan peninsula South Pacific/Central America Florida Nevada/California Baja California peninsula Northern Pacific Coast of Mexico Pacific and Central of Mexico 2.83 3.18 3.60 Western USA Deserts of North America Lowlands of Mexico and Central America Lowlands of Mexico and Central America Mexico and Western USA Yucatan peninsula Tamaulipas © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 Score (maximum) Areas included BIOGEOGRAPHICAL REGIONALIZATION OF NORTH AMERICA 489 Table 1. Continued Number of endemic species Areas included Score (maximum) 39 40 3 13 50 51, 135, 155, 172, 255, 322 2.84 7.14 41 42 43 9 4 11 3.97 2.38 3.83 44 45 46 47 48 49 50 5 3 5 3 8 3 11 52, 76, 104, 182, 210, 212, 226, 247 54 55, 83, 89, 150, 170, 186, 198, 253, 285, 288, 299, 310, 316, 327 56 57 61 62 64, 65, 75, 271 66, 171 80, 272, 283, 303 51 10 84, 101, 139, 163 3.98 52 4 90, 313 2.78 53 54 55 56 57 3 3 5 6 7 107 110 111, 204, 265 115, 179 122, 137, 217, 230, 246, 291 2.35 2.40 3.72 3.75 3.22 58 59 60 61 62 63 64 65 66 67 4 4 6 4 4 7 5 3 4 2 127 131, 138, 148, 158, 164 167, 176, 180, 184, 3.12 2.37 2.85 3.52 2.40 4.89 3.28 2.63 2.93 1.85 68 69 70 71 72 3 3 4 11 7 197 211 239 243 248 2.01 2.34 3.03 6.46 4.13 73 6 275, 302 3.29 74 7 276 5.40 75 3 309, 328 2.45 216, 277, 320 237, 263 183, 207 293 231 200, 219 199 250 DISCUSSION NEARCTIC REGION The boundaries of the Nearctic region were not recovered in our analysis. This may be a result of the scale and the grid-cell size. Casagranda et al. (2009) suggested that the use of small cells allows identification 4.03 2.93 3.83 2.78 4.33 2.49 6.49 Geographical location Eastern USA Lowlands of Mexico and Central America Chiapas/Central America Central America/Yucatan peninsula Mexico/Central America Sonora/Baja California peninsula Mexican Plateau Southern Central America Central USA Central America Western Coast of USA Lowlands of Mexico and Central America Mountains of Central and Southern Mexico Mountains of Central and Northern Mexico Chihuahua Sierra Madre del Sur/Oaxaca Gulf of Mexico and Central America Central Mexico Lowlands of Mexico and Central America Central America Northwestern Mexico Central America Pacific of Mexico and Central America Western USA Central Mexico Western Coast of USA/Canada Mexico Western USA Lowlands of Mexico and Central America Central Mexico Western USA Central America Chiapas and Central America Lowlands of Mexico and Central America Lowlands of Mexico and Central America Lowlands of Mexico and Central America Western Coast of USA of small and disjunct areas of endemism, whereas big cells allow identification of broad areas that may appear as fragmented when smaller cells are analyzed. Escalante et al. (2010) recovered the Nearctic region at 4° latitude-longitude but, when each cell was partitioned into four cells, the pattern was not recovered. © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 Consensus area 490 T. ESCALANTE ET AL. © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 Figure 1. Strict consensus cladogram from 76 areas of endemism; numbers represent the 16 clades of Table 2. Escalante et al. (2010) identified the southern boundary of the Nearctic region between 22–30° latitude, mainly around 26° latitude, and diagnosed it by endemic genera and species. The lack of a precise boundary is because the Nearctic region overlaps with the Neotropical region in the Mexican Transition Zone. Kreft & Jetz (2010) found that the transition from both regions is relatively gradual, although there was a visible discontinuity at approximately 29° latitude in northern Mexico, coinciding with the southernmost boundary of Escalante et al. (2010). Procheş & Ramdhani (2012) defined the Nearctic region as including central Mexico but excluded the northern areas of Canada and Alaska. Holt et al. (2013) found that the Nearctic did not include the northern areas of North America, and its southern boundary lies near the Istmus of Tehuantepec (thus including the Mexican Transition Zone). Two main patterns were recovered within the Nearctic region (eastern and western), which coincide with those found by Escalante et al. (2010). Although we did not recover the northern pattern of that study, probably because of the size of the grid-cells, we considered it as the Canadian subregion sensu Wallace (1876). The clade 5 (Fig. 1) includes as monophyletic the two patterns of the Nearctic region found by us: consensus areas 6 (in the east) and 32 (in the west; Figs 2, 3). Nelson (2008) described similar disjunct patterns for stoneflies; for mussels, this region shows five patterns (Graf & Cummings, 2007), although, in general, there is a coincidence in a west– east division (see also the analysis dealing only with mammals of Holt et al., 2013). Nelson (2008) proposed that this disjunction pattern may reflect geological processes from late Cretaceous or late Miocene. Consensus area 6 (c6; Fig. 2) corresponds to the eastern pattern of Escalante et al. (2010), and we assigned to it the subregion rank (Table 3). It was named the Alleghanian subregion by Wallace (1876), who considered that this area extends eastward to the Atlantic, including the Mississippi Valley, the Alleghany Mountains, and the eastern USA. Also, our pattern is similar to the Gulf Coastal subregion (NA3) of Graf & Cummings (2007). This subregion contains two provinces (clade 7; Fig. 1): Floridian and Tamaulipas (c27 and c38, respectively; Fig. 2). The Floridian province was named by Cooper (1859) and the Tamaulipas province was named by Smith (1941). Consensus area 32 (c32; Fig. 3), which corresponds to the western pattern of Escalante et al. (2010), was given the rank of subregion. Wallace (1876) did not recognize it as a subregion, although he distinguished two subregions: Californian and Rocky Mountain, with a boundary at 100° longitude. Thus, it may be named the Western or Californian-Rocky Mountain subregion, and it is represented by clades 6, 9, 15, BIOGEOGRAPHICAL REGIONALIZATION OF NORTH AMERICA 491 Table 2. General patterns corresponding to regions, subregions, and dominions in North America Clade Consensus areas Geographical location Pattern 1 2 43, 65 40, 55 Mexican Transition Zone Mexican Gulf-Central America subregion 3 26, 50 4 6, 32 5 6 7 8 9 0, 49 56, 68 70, 48, 58 44, 52, 53 13, 27, 38 10 61, 16, 74 11 72, 73, 18, 57 12 41, 60, 24, 71 13 14 46, 3, 2, 17 37, 12, 25, 7, 20, 51 15 33, 45, 66, 62, 69, 22, 14, 19 54 11, 64 28, 15, 8, 21 1, 9, 63, 4, 31 Mexico and northern Central America Gulf of Mexico coast, Yucatan peninsula, Isthmus of Tehuantepec, south of Mexican Pacific coast and Central America Gulf of Mexico coast, Yucatan peninsula, Isthmus of Tehuantepec, south of Mexican Pacific coast and Central America Alopatric patterns of eastern and western USA Pacific coast of Mexico and USA Central Mexico Central America Northwestern Mexico Allopatric patterns of eastern and western USA Mexican Pacific coast Yucatan peninsula, Isthmus of Tehuantepec, south of Gulf of Mexico coast, and Central America Mexican Pacific coast Yucatan peninsula, Isthmus of Tehuantepec, Gulf of Mexico coast, and Central America Chiapas, southern Yucatan peninsula, Central America Southern Central America Mexican Pacific coast Yucatan peninsula, Isthmus of Tehuantepec, Gulf of Mexico coast, and Central America Western USA, California, Nothern Mexico Oaxaca Western USA Western USA Mexico (a) (b) (c) (d) Alleghanian and Californian-Rocky Mountain subregions Californian dominion Mexican Transition Zone Central America subregion Rocky Mountain dominion Alleghanian and Californian-Rocky Mountain subregions Pacific Central America subregion Neotropical region Central America subregion Central America dominion Mexican Gulf-Central America subregion Rocky Mountain dominion Mexican Transition Zone Californian dominion Californian dominion Mexican Transition Zone Figure 2. Alleghanian subregion (c6) and its provinces: Floridian (c27) and Tamaulipas (c38). © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 16 16 16 16 Pacific Central America subregion 492 T. ESCALANTE ET AL. Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 Figure 3. Californian-Rocky Mountain subregion (c32) and its provinces: Californian (c0), Oregonian (c21), and Nevadian (c49), belonging to the Californian dominion; and Artemisian (c8), Sierra Nevada (c28), Mohavian (c33), Gulf of California (c44), Mexican Plateau (c45), and Sierra Madre Occidental-Central Plateau (c53), belonging to the Rocky Mountain dominion. © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 BIOGEOGRAPHICAL REGIONALIZATION OF NORTH AMERICA 493 and 16c (Fig. 1), with two dominions: Californian (Western Coastal) and Rocky Mountain (Western Central). The Californian dominion is situated in the western coast of Canada and USA from 50° latitude North to Baja California Peninsula, and it is constituted by three patterns (c0, c21, and c49; Fig. 3, Table 3). The Rocky Mountain (Western Central) dominion corresponds to the subregion of Wallace (1876), located at that mountain chain with its plateaus, and central plains and prairies to 100° longitude, including the Mexican Plateau. We found seven endemism patterns, which were assigned province rank (c8, c28, c33, c44, c45, and c53; Fig. 3, Table 3). The two provinces of the Californian dominion are Californian (c0) and Oregonian (c21), and there is a third pattern that includes both partially (c49). Californian and Oregonian were named by Cooper (1859) and Dice (1943), respectively, and we conserve these names. However, although the third pattern was not recovered in those studies, Cooper (1859) recognized another hierarchical level that groups Californian and Oregonian with two other provinces, which was named Nevadian. Thus, the third pattern may be named the Nevadian province as a particular case of partially overlapping areas. The Rocky Mountain dominion has six provinces: Artemisian (c8), sensu Dice (1943); Sierra Nevada (c28); Mohavian (c33), although it is bigger than the province of Dice (1943), we maintained this name; Gulf of California (c44), which is partially overlapped with Californian; Mexican Plateau (c45), sensu Morrone (2005); and Sierra Madre OccidentalCentral Plateau (c53), defined by Marshall & Liebherr (2000). MEXICAN TRANSITION ZONE Areas of endemism do not necessarily always represent allopatric patterns, as a result of different processes (vicariance and posterior dispersal, areas of congruence as a result of geological events, etc). They may overlap partially (Szumik et al., 2002), representing transition zones. This overlap can occur at any level of the biogeographical regionalization (Escalante, 2009), although it is more evident at a regional level. When two or more regions overlap, it may be the case that: (1) each region maintains its endemic taxa but shares some taxa with the other region; (2) each region has its endemic taxa and shares taxa with the other region but also new taxa © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 Figure 4. Mexican Transition Zone (c4, c65) and its provinces: Isthmus of Tehuantepec/Chiapas (c1), Mexican Pacific/ Central Mexico (c31), Sierra Madre del Sur (c54), Transmexican Volcanic Belt (c56), and Mesoamerican (c63). 494 T. ESCALANTE ET AL. have speciated in the area of overlap, generating a new area of endemism; or (3) the area of overlap has a significantly lower diversity compared to both regions. Wallace (1876) recognized a Mexican subregion, which he considered to be a transition to the Nearctic region (‘Table of Regions and Sub-regions’, p. 82). Later, Darlington (1957) and Halffter (1962, 1964a, b, 1965, 1972, 1974, 1976, 1978, 1987) developed the concept of the Mexican Transition Zone. In our case, clades 1, 10, 16a, and 16d (Fig. 1) are located on Mexico, where we found consensus areas occupying almost all of Mexico, except the Baja California peninsula (c4, c65; Fig. 4). Moreover, there are some areas nested in them (c1, c31, c54, c56, c63; Fig. 4), showing five patterns corresponding to provinces (Table 3). Then, we can define the Mexican Transition Zone as a wide area in Mexico, where the Nearctic and Neotropical regions overlap, even including part of the Mexican Plateau province belonging to the Nearctic region and Chiapas province corresponding to Neotropical region. However, this transition zone may be also an area of endemism per se. There are many explanations for this ‘splendid mixture’ (Morrone & Márquez, 2008), including the Great American Biotic Interchange, Pleistocenic climatic changes, and other older historical processes (see Escalante et al., 2007a), and those processes also have generated speciation in situ. The provinces belonging to the Mexican Transition Zone as identified in the present study are (Fig. 4, Table 3): Isthmus of Tehuantepec/Chiapas (c1); Mexican Pacific/Central Mexico (c31); Sierra Madre del Sur (c54), sensu Morrone (2005); Transmexican Volcanic Belt (c56) by Morrone (2005); and Mesoamerican (c63) sensu Morrone & Márquez (2003). © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 Figure 5. Neotropical region (c20, c72) with two subregions: Pacific-Central America (c26, c74) and Mexican Gulf-Central America (c25, c37, c40). BIOGEOGRAPHICAL REGIONALIZATION OF NORTH AMERICA 495 Table 3. Biogeographical regionalization of North America Regions Subregions Nearctic Canadian (Northern) Alleghanian (Eastern) Californian-Rocky Mountain (Western) Dominions Californian (Western Coastal) Rocky Mountain (Western Central) Mexican Transition Zone Neotropical Pacific Central America Mexican Gulf-Central America Central America Provinces Floridian Tamaulipas Californian Nevadian Oregonian Artemisian Gulf of California Mexican Plateau Mohavian Sierra Madre Occidental-Central Plateau Sierra Nevada Isthmus of Tehuantepec/Chiapas Mesoamerican Mexican Pacific/Central Mexico Sierra Madre del Sur Transmexican Volcanic Belt Mexican Pacific/Central America Yucatan peninsula Yucatan peninsula/Mexican Gulf Southestern Mexico Chiapas/Central America Central America Talamancan Cordillera © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 Figure 6. Central American subregion (c2, c17, c24, c48, c60, and c70), belonging to the Neotropical region. 496 T. ESCALANTE ET AL. NEOTROPICAL REGION CONCLUSIONS The biogeography of North America is highly complex. Several processes have shaped the distributional patterns of its biota, mainly the mixture in central Mexico. Two regions are represented in North America: the Nearctic and the Neotropical, as well as a transition zone. The Nearctic region, although not a rich region, is probably more complex than previously assumed. Moreover, the Mexican Transition Zone acts as an endemism area, and not simply as an overlap area. Future studies with other plant and animal taxa will allow a better understanding of the natural regionalization of North America. ACKNOWLEDGEMENTS Claudia Szumik and Pablo Goloboff provided support for analysis using NDM. This work was supported by Conacyt project 80370. We thank two anonymous reviewers for helpful comments. REFERENCES Aranda SD, Lobo JM. 2011. How well does presence-onlybased species distribution modelling predict assemblage diversity? A case study of the Tenerife flora. Ecography 34: 31–38. Bean WT, Stafford R, Brashares JS. 2012. The effects of small sample size and sample bias on threshold selection and accuracy assessment of species distribution models. Ecography 35: 250–258. Cabrera AL, Willink A. 1973. Biogeografía de América Latina. Washington, DC: Secretaría General de la Organización de Estados Americanos, Serie de Biología, Monografía 13. Carine MA, Humphries CJ, Guma R, Reyes-Betancort JA, Santos Guerra A. 2009. Areas and algorithms: evaluating numerical approaches for the delimitation of areas of endemism in the Canary Islands archipelago. Journal of Biogeography 36: 593–611. Casagranda MD, Roig-Juñent S, Szumik C. 2009. Endemismo a diferentes escalas espaciales: Un ejemplo con Carabidae (Coleoptera: Insecta) de América del Sur austral. Revista Chilena de Historia Natural 82: 17–42. Casagranda MD, Taher L, Szumik C. 2012. Endemicity analysis, parsimony and biotic elements: a formal comparison using hypothetical distributions. Cladistics 28: 645– 654. Cooper JG. 1859. On the distribution of the forests and trees of North America, with notes on its physical geography. Annual Report of the Broad of Regents of the Smithsonian Institution, Showing the operations, expenditures, and condition of the Institution for the year 1858. Cracraft J. 1991. Patterns of diversification within continental biotas: hierarchical congruence among the areas of endemism of Australian vertebrates. Australian Systematic Botany 4: 211–227. Darlington PJ. 1957. Zoogeography: the geographical distribution of animals. New York, NY: John Wiley and Sons. Dice LR. 1943. The biotic provinces of North America. Ann Arbor, MI: Univ. Michigan Press. Donato M, Miranda-Esquivel DR. 2012. Respuesta a Escalante (2011) ‘De cómo el análisis de parsimonia de endemismos (PAE) tampoco explica la selección natural. Revista Mexicana de Biodiversidad 83: 892–896. Ebach MC, Morrone JJ, Parenti LR, Viloria ÁL. 2008. International code of area nomenclature. Journal of Biogeography 35: 1153–1157. Escalante T. 2009. Un ensayo sobre regionalización biogeográfica. Revista Mexicana de Biodiversidad 80: 551– 560. Escalante T. 2011. De cómo el Análisis de Parsimonia de Endemismos tampoco explica la selección natural. Revista Mexicana de Biodiversidad 82: 1057–1059. Escalante T, Morrone JJ, Rodríguez-Tapia G. 2013. Data from: biogeographic regions of North American mammals © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 The Neotropical region is only partially represented by clades 2, 3, 4, 8, 11, 12, 13, and 14 (Fig. 1). Wallace (1876) maintained this name given by Sclater (1858), and he divided it into four subregions: Brazilian, Chilean, Mexican, and Antillean. The part of the Neotropical region considered here corresponds to the Mexican subregion, located sensu Wallace (1876) in Central America and southern Mexico (c20, c72; Fig. 5). Although there is overlap between the patterns of this region, we can distinguish three possible subregions, which should be analyzed using a more wide area of study, also including South America. The subregions proposed by us are: Pacific-Central America, Mexican Gulf-Central America, and Central America (Table 3). The first includes consensus areas in lowlands of the Pacific coast in Mexico and Central America (c26, c74; Fig. 5). The second includes provinces mainly in the lowlands of the Yucatan peninsula, Mexican Gulf-coast, and Central America (c25, c37, c40; Fig. 5). Finally, the third subregion is located in Chiapas and Central America, and probably extends to South America (c2, c17, c24, c48, c60, and c70; Fig. 6). Although, in North America, there is only part of the Neotropical region, our subregions and provinces coincide with those of other studies. Marshall & Liebherr (2000) considered a province named Talamancan Cordillera, which corresponds to our consensus areas c2 and c17 in the Central America subregion, and we maintain that name. For other provinces, we follow the nomenclature of Morrone (2005) (Yucatan peninsula; c37), although we considered that it should be necessary to enlarge the study area to South America to obtain complete patterns. BIOGEOGRAPHICAL REGIONALIZATION OF NORTH AMERICA Halffter G. 1962. Explicación preliminar de la distribución geográfica de los Scarabaeidae mexicanos. Acta Zoológica Mexicana (nueva serie) 5: 1–17. Halffter G. 1964a. La entomofauna americana, ideas acerca de su origen y distribución. Folia Entomológica Mexicana 6: 1–108. Halffter G. 1964b. Las regiones Neártica y Neotropical desde el punto de vista de su entomofauna. Anais do II Congresso Latinoamericano de Zoologia, São Paulo 1: 51–61. Halffter G. 1965. Algunas ideas acerca de la zoogeografía de América. Revista de la Sociedad Mexicana de Historia Natural 26: 1–16. Halffter G. 1972. Eléments anciens de l’entomofaune neotropicale: Ses implications biogéographiques. In: Biogeographie et liasons intercontinentales au cours du Mésozoique, 17me Congres International de Zoologie, Monte Carlo, 1–40. Halffter G. 1974. Eléments anciens de l’entomofaune neotropicale: Ses implications biogéographiques. Quaestiones Entomologicae 10: 223–262. Halffter G. 1976. Distribución de los insectos en la Zona de Transición Mexicana: Relaciones con la entomofauna de Norteamérica. Folia Entomológica Mexicana 35: 1–64. Halffter G. 1978. Un nuevo patrón de dispersión en la Zona de Transición Mexicana: El mesoamericano de montaña. Folia Entomológica Mexicana 39–40: 219–222. Halffter G. 1987. Biogeography of the montane entomofauna of Mexico and Central America. Annual Review of Entomology 32: 95–114. Hall ER. 1981. The mammals of North America, Vols I and II. New York, NY: John Wiley and Sons. Hernandez PA, Graham CH, Master LL, Albert DL. 2006. The effect of sample size and species characteristics on performance of different species distribution modeling methods. Ecography 29: 773–785. Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. 2005. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25: 1965–1978. Holt BG, Lessard JP, Borregaard MK, Fritz SA, Araújo MB, Dimitrov D, Fabre PH, Graham CH, Graves GR, Jonsson KA, Nogués-Bravo D, Wang Z, Whittaker RJ, Fjeldsa J, Rahbek C. 2013. An update of Wallace’s zoogeographic regions of the world. Science 339: 74–78. Jiménez-Valverde A. 2012. Insights into the area under the receiver operating characteristics curve (AUC) as a discrimination measure in species distribution modeling. Global Ecology and Biogeography 21: 498–507. Jiménez-Valverde A, Lobo JM. 2007. Threshold criteria for conversion of probability of species presence to either-or presence absence. Acta Oecologica 31: 361–369. Kreft H, Jetz W. 2010. A framework for delineating biogeographical regions based on species distributions. Journal of Biogeography 37: 2029–2053. Liu C, Berry M, Dawson TP, Pearson RG. 2005. Selecting thresholds of occurrence in the prediction of species distributions. Ecography 28: 385–393. © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 based on endemism. Dryad Digital Repository. doi: 10.5061/ dryad.18j3d. Escalante T, Rodríguez G, Cao N, Ebach MC, Morrone JJ. 2007a. Cladistic biogeographic analysis suggests an early Caribbean diversification in Mexico. Die Naturwissenschaften 94: 561–565. Escalante T, Rodríguez G, Morrone JJ. 2004. The diversification of Nearctic mammals in the Mexican Transition Zone. Biological Journal of the Linnean Society 83: 327–339. Escalante T, Rodríguez-Tapia G. 2011. Base de datos geoespacial de mamíferos terrestres de América del Norte: Una aproximación a sus patrones biogeográficos y conservación. In: Mas JF, Cuevas G, González R, eds. Memorias de la XIX Reunión Nacional SELPER-Mexico. Mexico, DF: Centro de Investigaciones en Geografía Ambiental, UNAM, 110–113. Escalante T, Rodríguez-Tapia G, Linaje M, Illoldi-Rangel P, González-López R. 2013. Identification of areas of endemism from species distribution models: threshold selection and Nearctic mammals. TIP Revista Especializada en Ciencias Químico-Biológicas 16: 15– 17. Escalante T, Rodríguez-Tapia G, Szumik C, Morrone JJ, Rivas M. 2010. Delimitation of the Nearctic region according to mammalian distributional patterns. Journal of Mammalogy 91: 1381–1388. Escalante T, Sánchez-Cordero V, Morrone JJ, Linaje M. 2007b. Areas of endemism of Mexican terrestrial mammals: a case study using species’ ecological niche modeling, parsimony analysis of endemicity and Goloboff fit. Interciencia 32: 151–159. Escalante T, Szumik C, Morrone JJ. 2009. Areas of endemism of Mexican mammals: re-analysis applying the optimality criterion. Biological Journal of the Linnean Society 98: 468–478. Espinosa D, Ocegueda S, Aguilar C, Flores VO, Llorente-Bousquets J. 2008. El conocimiento biogeográfico de las especies y su regionalización natural. In: Sarukhán J., ed. Capital natural de México, Vol. I/Mexico, DF: Comisión Nacional para el Conocimiento y Uso de la Biodiversidad, 33–65. ESRI. 1998. ArcView 3.1 GIS. Redlands, CA: Environmental Systems Research Institute, Inc. Freeman EA, Moisen GG. 2008. A comparison of the performance of threshold criteria for binary classification in terms of predicted prevalence and kappa. Ecological Modelling 217: 48–58. Goloboff P. 1993. Estimating character weights during tree search. Cladistics 9: 83–91. Goloboff P. 2011. NDM/VNDM v. 3.0 Programs for identification of areas of endemism. Available at: http://www .zmuc.dk/public/phylogeny/endemism Goloboff P, Farris J, Nixon K. 2011. T.N.T.: tree analysis using new technology. Available at: http://www.cladistics.org Graf DL, Cummings KS. 2007. Review of the systematics and global diversity of freshwater mussel species (Bivalvia: Unionoida). Journal of Molluscan Studies 73: 291– 314. 497 498 T. ESCALANTE ET AL. prioritization of areas in northeast India: priorities for amphibians and reptiles. Biological Conservation 136: 346– 361. Pearson RG, Raxworthy CJ, Nakamura M, Peterson T. 2007. Predicting species distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar. Journal of Biogeography 34: 102–117. Peterson AT. 2006. Uses and requirements of ecological niche models and related distributional models. Biodiversity Informatics 3: 59–72. Peterson AT, Papes M, Soberón J. 2008. Rethinking receiver operating characteristic analysis applications in ecological niche modeling. Ecological Modelling 213: 63– 72. Peterson AT, Soberón J, Pearson RG, Anderson RP, Martínez-Meyer E, Nakamura M, Bastos Araújo M. 2011. Ecological niches and geographic distributions. Princeton, NJ: Princeton University Press. Phillips SJ, Anderson RP, Schapire RE. 2006. A maximum entropy modelling of species geographic distributions. Ecological Modelling 190: 231–259. Phillips SJ, Dudík M. 2008. Modelling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecogeography 31: 161–175. Porzecanski AL, Cracraft J. 2005. Cladistic analysis of distributions and endemism (CADE): using raw distributions of birds to unravel the biogeography of the South American aridlands. Journal of Biogeography 32: 261–275. Procheş Ş. 2005. The world’s biogeographical regions: cluster analyses based on bat distributions. Journal of Biogeography 32: 607–614. Procheş Ş, Ramdhani S. 2012. The world’s zoogeographical regions confirmed by cross-taxon analyses. BioScience 62: 260–270. 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. London: Chapman and Hall, 437– 481. Sclater PL. 1858. On the general geographical distribution of the members of the class Aves. Proceedings of the Linnean Society (Zoology) 2: 130–145. Smith AB. 2013. On evaluating species distribution models with rndom background sites in place of absences when test presences disproportionately sample suitable habitat. Diversity and Distributions 19: 867–872. Smith HM. 1941. Las provincias bióticas de México, según la distribución geográfica de las lagartijas del género Sceloporus. Anales de la Escuela Nacional de Ciencias Biológicas 2: 103–110. Szumik C, Cuezzo F, Goloboff PA, Chalup AE. 2002. An optimality criterion to determine areas of endemism. Systematic Biology 51: 806–816. Szumik C, Goloboff PA. 2004. Areas of endemism: an improved optimality criterion. Systematic Biology 53: 968– 977. Trejo-Torres JC, Ackerman JD. 2002. Composition patterns of Caribbean limestone forests: are parsimony, © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 Liu C, White M, Newell G. 2013. Selecting thresholds for the prediction of species occurrence with presence-only data. Journal of Biogeography 40: 778–789. Lobo JM, Jiménez-Valverde A, Real R. 2008. AUC: a misleading measure of the performance of predictive distribution models. Global Ecology and Biogeography. 17: 145– 151. Luna-Vega I, Alcántara Ayala O, Morrone JJ, Espinosa D. 2000. Track analysis and conservation priorities in the cloud forests of Hidalgo, Mexico. Diversity and Distributions 6: 137–143. Manel S, Williams HC, Omerod DJ. 2001. Evaluating presence-absence models in ecology: the need to account for prevalence. Journal of Applied Ecology 38: 921–931. Marshall CJ, Liebherr JK. 2000. Cladistic biogeography of the Mexican transition zone. Journal of Biogeography 27: 203–216. Moreno Saiz JC, Donato M, Katinas L, Crisci JV, Posadas P. 2013. New insights into the biogeography of south-western Europe: spatial patterns from vascular plants using cluster analysis and parsimony. Journal of Biogeography 40: 90–104. Morrone JJ. 1994. On the identification of areas of endemism. Systematic Biology 43: 438–441. Morrone JJ. 2001. Homology, biogeography and areas of endemism. Diversity and Distributions 7: 297–300. Morrone JJ. 2002. Biogeographical regions under track and cladistic scrutiny. Journal of Biogeography 29: 149– 152. Morrone JJ. 2005. Hacia una síntesis biogeográfica de México. Revista Mexicana de Biodiversidad 76: 207–252. Morrone JJ. 2009. Evolutionary biogeography: an integrative approach with case studies. New York, NY: Columbia University Press. Morrone JJ, Márquez J. 2003. Aproximación a un Atlas Biogeográfico Mexicano: Componentes bióticos principales y provincias biogeográficas. In: Morrone JJ, Llorente Bousquets J, eds. Una perspectiva latinoamericana de la biogeografía. Mexico, DF: Las Prensas de Ciencias, UNAM, 217–220. Morrone JJ, Márquez J. 2008. Biodiversity of Mexican terrestrial arthropods (Arachnida and Hexapoda): a biogeographical puzzle. Acta Zoológica Mexicana (nueva serie) 24: 15–41. Murguía M, Llorente-Bousquets J. 2003. Reflexiones conceptuales en biogeografía cuantitativa. In: Llorente J, Morrone JJ, eds. Una perspectiva latinoamericana de la biogeografía. Mexico, DF: Las Prensas de Ciencias, UNAM, 133–140. Nelson CH. 2008. Hierarchical relationships of North American states and provinces: an area cladistic analysis based on the distribution of stoneflies (Insecta: Plecoptera). Illiesia 4: 176–204. Nenzén HK, Araújo MB. 2011. Choice of threshold alters projections of species range shifts under climate change. Ecological Modelling 222: 3346–3354. Pawar S, Koo MS, Kelley C, Ahmed FM, Choudhury S, Sarkar S. 2007. Conservation assessment and BIOGEOGRAPHICAL REGIONALIZATION OF NORTH AMERICA classification, and ordination analyses congruent? Biotropica 34: 502–515. Wallace AR. 1876. Geographical distribution of animals: with a study of the relations of living and extinct faunas as elucidating the past changes of the earth’s surface. London: Macmillan. 499 Wisz MS, Hijmans RJ, Peterson AT, Graham CH, Guisan A, NCEAS Predicting Species Distributions Working Group. 2008. Effects of sample size on the performance of species distribution models. Diversity and Distributions 14: 763–773. ARCHIVED DATA SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Table A1. Information on consensus areas and keys of species © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 110, 485–499 Downloaded from https://academic.oup.com/biolinnean/article-abstract/110/3/485/2415643 by guest on 16 June 2020 Data and shapefiles of areas of endemism deposited at Dryad: Escalante et al. (2013).