Diversity and Biogeographical Makeup of the Dung Beetle Communities Inhabiting Two Mountains in the Mexican Transition Zone

Org Divers Evol (2014) 14:105–114 DOI 10.1007/s13127-013-0148-0

ORIGINAL ARTICLE

Diversity and biogeographical makeup of the dung beetle communities inhabiting two mountains in the Mexican Transition Zone

Fredy Alvarado & Federico Escobar & Jorge Montero-Muñoz

Received: 8 February 2013 /Accepted: 9 July 2013 /Published online: 24 July 2013 # Gesellschaft für Biologische Systematik 2013

Abstract The role of horizontal and vertical colonization on the Sierra Madre Oriental mountain range and is of ancient the diversity and integration of the dung beetle fauna of two geological origin—the processes of horizontal and vertical mountains in the Mexican Transition Zone (Los Tuxtlas and colonization have had relatively different weights in terms of La Chinantla) are analyzed and compared. On each mountain their effect on the pattern of diversity and the biogeographic standardized sampling was done using pitfall traps baited integration of the beetle community, while on Los Tuxtlas, the with dung and carrion along elevation gradients. On both limited role of horizontal colonization appears to be a conse- mountains diversity decreased linearly with increasing ele- quence of its isolation and more recent geological origin. We vation. The decrease in the number of genera and species discuss the potential use of these models for studying the was not different between mountains, but the cumulative effects of climate change on elevation gradients. total number for both taxonomic levels was significantly higher on La Chinantla. There, three well-defined groups Keywords Elevation gradient . Tropical mountains . were identified for which species turnover was mainly a Diversity . Species turnover . Scarabaeinae . Aphodiinae . result of species gain. On Los Tuxtlas there was no evident Geotrupinae . Biogeographic patterns grouping pattern, and species turnover was determined by species loss. For both mountains the dominant biogeographic distribution pattern was Neotropical; however, at the Introduction higher elevations of La Chinantla, a clear replacement by lineages of Holarctic affinity was observed. We suggest that for As centers of diversity and endemism, mountain ranges are La Chinantla—a mountain that is geographically connected to systems of great heuristic value for understanding the patterns and mechanisms of the diversification of biota on the FA and FE conceived and designed the study. FA performed the study. earth (Sanders and Rahbek 2011). In general, a decrease in FA, FE and JMM analyzed the data. FA, FE and JMM wrote the species richness and changes in species composition with manuscript. increasing elevation have been documented (Rahbek 1995). Electronic supplementary material The online version of this article These patterns have been associated with a variety of con- (doi:10.1007/s13127-013-0148-0) contains supplementary material, which is available to authorized users. temporary environmental variables (Huston 1994), but the : underlying causal mechanisms behind these correlations F. Alvarado F. Escobar have, to date, been the subject of much debate (Sanders Red de Ecoetología, Instituto de Ecología A. C, Apartado Postal 63, 91000 Xalapa, Veracruz, México and Rahbek 2011), particularly the role of the biogeographic history of the mountains and their geological and geographic J. Montero-Muñoz characteristics (Lomolino 2001). Departamento de Recursos del Mar CINVESTAV- Centro de Two different but related biological processes have been Investigación y Estudios Avanzados, Instituto Politécnico Nacional, Unidad Mérida, Apartado Postal 73, 97310 Mérida, proposed to explain the composition and origin of mountain Yucatán, México beetle fauna: horizontal colonization, with elements arriving from surrounding regions at the same elevation, and vertical * F. Escobar ( ) colonization, with elements arriving from neighboring low- Instituto de Ecología, A. C, Carretera antigua a Coatepec 351, Xalapa 91070, Veracruz, México lands (Lobo and Halffter 2000). The latter authors make e-mail: [email protected] three predictions to explain the role of these colonization 106 F. Alvarado et al. processes on the diversity and makeup of the communities differ in their geological history and degree of geographic that live in the mountains: (1) mountain fauna is a rarefied isolation, both located within the Mexican Transition Zone, equivalent of lowland fauna, (2) mountain fauna is com- in order to understand the relative effects of the processes of posed of a lower number of species that are phylogenetically horizontal and vertical colonization on diversity and the way related to those inhabiting lower elevations (vertical coloni- in which the beetle communities have formed in the moun- zation model) and (3) the mountain fauna may or may not tains. We address the following questions: (1) How does the have a low number of species, but is composed of elements diversity of dung beetles change along the elevation gradi- with different evolutionary histories and geographic origins ents? (2) How does the pattern of species turnover change compared to those found in the surrounding lowlands (hor- along each elevation gradient? (3) How do the patterns of izontal colonization model). biogeographical affinity change for the species that make up Two large biotas come into contact in the Mexican Tran- the beetle communities on both mountains? sition Zone: the Nearctic and Neotropical (for further details, see Halffter 1987). This zone represents a clear example of the consequences of the processes of horizontal and vertical Materials and methods colonization, where the communities at the higher elevations are dominated by species that belong to genera with diversi- Study sites fication centers in the Holarctic realm (e.g., species with Montane Paleoamerican and Nearctic distribution patterns, The study was done on two mountains in the Mexican sensu Halffter 1987 and Halffter et al. 1995), while the set of Transition Zone (MTZ), 130 km apart: the San Martín vol- species from the lowlands are dominated by elements with a cano in the Sierra Los Tuxtlas range, in the state of Veracruz Neotropical affinity (Halffter 1976, Halffter 1987, Lobo and (18°05’–18°45’ N, 92°35’–95°13’ W) and a mountain in the Halffter 2000). The pattern of horizontal colonization would Sierra de La Chinantla range, in the state of Oaxaca (17°22'– thus be the result of the southward displacement of northern 18°12' N, 95°43'–96°58’ W; Fig. 1). These mountains have lineages in response to the climate changes of the Plio- different degrees of isolation, ages and geological histories. Pleistocene (Halffter 1976, Elias 1994), facilitated by the Los Tuxtlas has the following types of vegetation: tall stature north–south orientation of the mountains of North America rainforest (<600 m a.s.l.; temperature: 22º–25º C; annual (Halffter 1976, Lobo and Halffter 2000, Escobar et al. 2005). precipitation: 2,000 and 2,500 mm), medium stature rain- In contrast, vertical colonization would be the result of forest (650–1,000 m a.s.l.; 21º–23º C; 2,000–3,000 mm) and adaptive changes in evolutionary time as a consequence of mountain mesophyll or cloud forest (1,000–1,650 m a.s.l.; the environmental limitations imposed by the mountains. 16º–20º C; 2,500–4,500 mm). On this mountain, up to an This process has had a greater influence in the northern elevation of 1,738 m a.s.l., there are remnants of forest from Andes (Escobar et al. 2005, 2006) and on the islands of the lowlands, which are surrounded by a mosaic of areas southeastern Asia (Hanski 1983, Hanski and Niemelä used by people, environments and land use types that include 1990). Just as Lobo and Halffter (2000) propose, the relative rainforest fragments, areas in different stages of regenera- effects of each colonization model can vary depending on the tion, crops (coffee, corn and bananas) and extensive areas geological history, location, orientation and degree of geo- with pastures (Soriano et al. 1997, Guevara et al. 2004). The graphic isolation of each mountain, although to date there is San Martín volcano (hereafter, Los Tuxtlas) is isolated from no information on the role that these factors have had on the other mountains and is located on the coastal plain of the diversity patterns and the way in which mountain communi- Gulf of Mexico. Its geological origin goes back 1.5 Ma and is ties have been formed. related to the lifting of the Trans–Mexican Volcanic Belt, Dung beetles are an important component of diversity in a which occurred during the intense period of volcanic activity large number of ecosystems (Halffter and Matthews 1966). at the end of the Miocene and the beginning of the Pleisto- They belong to three well-known subfamilies: the Scara- cene (5.3–1.5 Ma; Ferrusquía–Villafranca 1993). baeinae, adapted to warm climates, and Aphodiinae and The Sierra de La Chinantla mountains are found in a Geotrupinae, which are predominantly adapted to temperate densely forested area (>5,000 ha) that has the following latitudes, and even cold temperate regions (Hanski and vegetation types: tall stature rainforest (50–700 m a.s.l; tem- Cambefort 1991, Davis and Shaw 2001). These evolutionary perature: 22º–26º C; precipitation annual: 2,000–2,500 mm), differences have resulted in a marked latitudinal segregation medium stature rainforest (500–1,000 m a.s.l.; 18º–22º C; and an analogous separation with respect to elevation as 2,500–3,000 mm), mountain mesophyll or cloud forest well, as observed in the mountains of southern Europe (900–2,000 m a.s.l.; 12º–18º C; 3,500–4,500 mm), and pine (Martín–Piera et al. 1992, Jay–Robert et al. 1997) and in and oak forests (1,900–2,800 m a.s.l.; 5º–12º C; 1,500– the Mexican Transition Zone (Halffter et al., 1995, Escobar 2,000 mm). Its lowlands are covered by pastures as well as et al. 2007). In this study we use two mountain systems that sugar cane and rubber tree crops. At intermediate elevations Diversity and biogeographical makeup of the dung beetle 107

Fig. 1 Location of the different elevations studied in the Mexican Transition Zone. The black dots indicate the sampling stations

there are corn crops, coffee plantations and fruit orchards, and 50 m (see Larsen and Forsyth 2005). In Los Tuxtlas 128 higher up (>1800 m a.s.l.) there are dairy farming, cereal and traps were set and in La Chinantla, 208. Each trap consisted legume crops (soya, peas and beans; García–Mendoza et al. of a 1-L container (11.5 cm in diameter, 13.5 cm deep) buried 2004). This mountain is part of the Sierra Madre Oriental flush with the soil and baited with 25 g of fresh human range, which is considered one of the oldest mountain systems excrement or rotting fish in a 25-ml cup suspended by a wire in the country, having formed during the geological events of over a larger container that was one-quarter full of a solution the Miocene (20 Ma; Ferrusquía–Villafranca 1993). of water, salt and soap to prevent the beetles from escaping. All of the traps were left in place for 48 h, after which the Sampling beetles that had been captured were placed in a 70 % alcohol solution for later identification. The study was done between June and September 2010 during the rainy season when the dung beetles are most active. An elevation transect was set up on each mountain Data analysis across areas of continuous forest and passing through the different natural vegetation formations present along each The representativeness of the beetle inventory for each gradient. For this reason, on both mountains the sampling mountain and sampling station was evaluated using the stations were set up at elevation intervals of 200 m. In Los proportion of species observed relative to the values predict-

Tuxtlas collecting was done at 8 stations between 200 and ed by two estimators (Chao1 and Chao2), recommended for 1600 m a.s.l. and for La Chinantla at 13 stations between 200 relatively small sample sizes (Colwell and Coddington and 2600 m a.s.l. At each station, 16 pitfall traps were set up, 1994). Additionally, we calculated MaoTau+CI 95 % as a alternating along two lines with eight traps each (four baited measure of the variation in the observed richness. Both the with excrement and four with carrion) and separated by richness estimators and the confidence interval (CI 95 %) 100 m. The distance between the traps on each line was were calculated using EstimateS v.8.0 (Colwell 2008). 108 F. Alvarado et al.

With a simple linear regression analysis, we determined speciation in the zone (Halffter, 1987). We did a goodness- whether the slopes of the decrease and increase in the num- of-fit test to evaluate whether the proportion of individuals in ber of genera and species along each elevation gradient were each BDP was the same between elevations. different from zero. After this, the slopes for each mountain and taxonomic category (genus and species) were compared using a t test for independent samples (Zar 1996). Results A multivariate analysis of variance was used, based on permutations (PERMANOVA, Anderson 2001,Anderson Diversity patterns et al. 2006) to evaluate the percentage of variance in species composition explained by elevation. This method analyzes the In Los Tuxtlas, 2,296 specimens belonging to 31 species (14 dispersal of species composition within and among elevations genera, all in the subfamily Scarabaeinae, Table S1) were on the basis of any measure of distance or dissimilarity collected, and in La Chinantla, 3,077 specimens belonging to (Anderson 2001). Variation in the rate of species turnover 40 species (16 genera in three subfamilies: Scarabaeinae, was determined using the BETADISPER function of the VEGAN Aphodiinae and Geotrupinae, Table S2). Of the 47 species statistical program available in R (Oksanen 2010). This func- found on both mountains, 24 (51 %) are shared, 7 (15 %) are tion examines the differences in the homogeneity between exclusive to Los Tuxtlas and 16 (34 %) exclusive to La groups and is analogous to Levene’s equality of variances test. Chinantla. The evaluation of inventory completeness indi-

We used Bray-Curtis as a measure of dissimilarity with log10 cates that close to 90 % of the species present on each transformed data to reduce the effect of abundant species. mountain was captured, while at the scale of the sampling Principal coordinate analysis (PCoA) was used to visualize stations it was 88 %. Only at 1 of the 8 Los Tuxtlas stations the results indicating differences in composition and its vari- and 2 of the 13 La Chinantla stations were the estimated ation within and between elevations, assuming that the mea- species richness values <70 % (Table 1). sure of dissimilarity is monotonically related to ecological On both mountains the richness of genera and species distance, thus providing a robust method for ranking that is decreased linearly with increasing elevation (Fig. 2). The simple to interpret (Quinn and Keough 2002). slope of the loss of genera was greater for Los Tuxtlas We evaluated the rate of genus and species replacement (−6.7 genera/1,000 m) than for La Chinantla (−4.5 between contiguous pairs of elevations along the gradient as genera/1,000 m) but was not different between mountains (qDβ–1)/(N–1), where N is the number of samples and qDβ (t=−1.016, df=19, p=0.32). On La Chinantla there was a is the q order beta diversity, which is obtained from the gain of 1.6 genera/1,000 m, while on Los Tuxtlas the rate of quotient qDγ/qDα (with qDγ gamma diversity and qDα genus accumulation was zero, indicating that the species at alpha diversity, Jost 2007). A value of q=0 was used to higher elevations belong to the same genera as those found at evaluate the rate of replacement based exclusively on neighboring lower elevations (Fig. 2). Although the decrease presence-absence data. The rate of replacement ranges from in the number of species with increasing elevation was 0 to 1 and can be expressed as a percentage, with zero greater for Los Tuxtlas (−14.5 species/1,000 m) than for La indicating there is no species replacement between samples Chinantla (−8.9 species/1,000 m), the slopes were not statis- and 1 or 100 % indicating that samples are completely tically different (t=−0.80, df=19, p=0.43; Fig. 2). The rate different in their species composition (Jost 2007). Given that of species accumulation was significantly greater on La the rate of replacement does not reveal whether the turnover Chinantla (6.0 species/1,000 m) than on Los Tuxtlas (4.4 rates result from the loss or gain of species, and in order to species/1,000 m; t=3.39, df=19, p=0.003; Fig. 2). understand the relative influence of each process, we calculated the number of genera and species gained and lost for Variations in composition, turnover in genera and species each comparison. We grouped beetle species according to their biogeo- The degree of species turnover with increasing elevation was graphical distribution pattern (BDP) using Halffter’s(1987) different on each mountain. On Los Tuxtlas elevation proposed classification, which is based on the origin, distri- explained 35 % of the variation (Fperm=6.4, df =7, p<0.01) bution and penetration of species in the MTZ: Nearctic and on La Chinantla, 57 % (Fperm=16.0, df=12, p<0.001). (NEA), Tropical Paleoamerican (PAT), Mountain Paleo- However, on Los Tuxtlas variation in species composition american (PAM), Mountain Mesoamerican (MAM) and within each elevation was high and did not differ between

Typical Neotropical (NTT). These distribution patterns have elevations (Fperm=0.72, df=7, p=0.61). This result was cor- been used to understand the origin and makeup of insects in roborated by the ordination plot, which suggests there is no the MTZ and assume that each group has been subjected to clustering pattern of sites with respect to species composi- similar macroecological forces over long periods of time tion (Fig. 3). For La Chinantla the analysis indicated that and, as such, have a common history of colonization or there are differences in species composition between elevations Diversity and biogeographical makeup of the dung beetle 109

Table 1 Observed and estimated species abundance and rich- Mountain elevation Individuals No. species MauTao+95 % CI Chao1 Chao2 Estimate (%) ness based on two non-paramet- (m a.s.l.) ric estimators (Chao1 and Chao2) along each elevation gradient. La Chinantla 3,077 40 44 43 42 93 MauTao+95 % CI values are in- 200 451 24 28 27 26 89 cluded as a measure of the vari- 400 591 24 31 27 34 80 ation in the observed richness. 600 240 14 20 24 19 67* Percent completeness for each sampling station was obtained 800 452 15 18 15 16 92 from the mean of the two esti- 1,000 665 14 17 14 16 89 mators and the MauTao+95 % 1,200 180 7 7 7 7 100 CI value 1,400 155 4 4 4 4 100 1,600 125 3 3 3 3 100 1,800 95 3 3 3 3 100 2,000 10 2 2 2 2 100 2,200 14 4 7 4 7 67* 2,400 24 4 5 4 4 92 2,600 75 4 4 4 4 100 Los Tuxtlas 2,296 31 36 35 35 89 200 1,382 24 26 25 25 95 400 116 16 19 21 18 83 600 182 19 25 28 33 66* 800 236 13 17 18 16 76 1,000 158 10 12 10 10 94 1,200 190 8 10 8 8 92 1,400 30 5 7 6 5 83 1,600 2 2 4 3 3 86 *Sites with estimates <75 %

(Fperm=2.274, df=12, p<0.01; Fig. 3). For this mountain the ordination plot reveals the formation of three well-differentiated sets of species: one at low elevations (200 to 1,000 m a.s.l.), another at intermediate elevations (1,200 to 1,800 m a.s.l.) and a third at elevations above 2,000 m a.s.l. The turnover of genera and species was different between mountains (Fig. 4). The rate of replacement at Los Tuxtlas for both taxonomic levels was similar, with maximum replacement values at intermediate elevations (800–1,000 m a.s.l.) and at the top of the gradient (1,400–1,600 m a.s.l.). Along the slope of this mountain the dominant process was species loss, particularly of species belonging to the same genera, with some gains at lower elevations (Fig. 4, Table S1). On La Chinantla, in contrast, there was a decrease in the rate of species replacement from the lowlands to the mid-elevations of the mountain (1,400–1,600 m a.s.l.), and above this there was an increase in species replacement with two well- defined peaks (400–600 and 2,200–2,400 m a.s.l.). For genera, replacement was higher around the middle of the gradient. This mountain is characterized by the gain of species belonging to different genera and subfamilies along Fig. 2 Relationship between the decrease (circles and black squares) the gradient and particularly above 1,800 m a.s.l. (Fig. 4, and the accumulation (circles and open squares) of genera and species Table S2). with increasing elevation in Los Tuxtlas and La Chinantla 110 F. Alvarado et al.

Fig. 3 Principal coordinates graph from the permutational multi- the differences in the distribution of means (distance to the variate analysis of variance (PERMANOVA) for Los Tuxtlas and centroid) and quartiles for each elevation La Chinantla. For both mountains the box and whisker plot shows

Biogeographic distribution patterns Discussion

On both mountains the proportion of individuals exhibiting As a consequence of the complex geological and evolution- a given biogeographical distribution pattern depended on ary history of mountains (van der Hammen 1995), it is elevation (Los Tuxtlas: chi2=516.70, df=14, p<0.0001; La possible to say that each represents a unique unrepeatable Chinantla: chi2=5,764.28, df=48; p<0.0001). On Los scenario in which ecological and historical factors have had a Tuxtlas the pattern that dominated along the entire gradient similar impact on the biota (Halffter et al. 1995, Körner was NTT (82 % of all individuals, Fig. 5), while PAM and 2000). Because of this and the fact that the evolutionary PAT were more prevalent at intermediate elevations (Fig. 5). history of each taxon prevents the occurrence of a general Although on La Chinantla 72 % of all the individuals pattern of variation (Rahbek 1995, Lomolino 2001), the had the NTT pattern, this group dominated the lowlands search for an explanation of the diversity patterns on moun- below 1,000 m a.s.l.; they were replaced by MAM species at tains might not make sense in the absence of knowledge mid-elevations and by NEA species at higher elevations about the biogeographic history of the region in which the (Fig. 5). mountains occur. This emphasizes the importance of Diversity and biogeographical makeup of the dung beetle 111

Fig. 4 Turnover (open circles) of genera and species between pairs of contiguous elevations along each elevation gradient. The number of genera and species lost (black bars) and gained (gray bars) for each pair was compared comparative studies between and within the taxa of different intermediate elevations (Escobar et al. 2005, García–López mountains for understanding the underlying causes of diver- et al. 2011). The decrease in diversity along elevation gradi- sity gradients (Sanders and Rahbek 2011). ents has been associated with a progressive reduction in area The results of this study suggest that the processes of (Lomolino 2001) and ecological filters such as climate (i.e., horizontal and vertical colonization have a different effect low temperatures) and energetic variables (i.e., limited food on diversity patterns and the way in which the beetle com- availability) imposed by the mountain on the establishment munities came into being on the two mountains studied in the of populations (Janzen 1967). Such conditions require phys- MTZ. Similarly, the results corroborate the predictions of iological adjustments by species if the higher elevations of Lobo and Halffter (2000), and provide more information about the mountains are to be colonized (Navas 2003). This is the impact that the characteristics particular to each mountain particularly true of the subfamily Scarabaeinae, a monophy- can have on the communities that inhabit them. As noted by letic group that depends on dung, mainly that of mammals, Rahbek (1995) and Lomolino (2001), our results point to the for food and reproduction. This family is comprised of general conclusion that the degree of geographic isolation and species that are adapted to warm environments in tropical differences in the geological ages of the mountains, and par- and subtropical regions (Scholtz 1990), for which the verti- ticularly the evolutionary history of the taxon, are all determi- cal colonization of mountains from the lowlands could thus nant factors in the patterns of richness and changes in species be limited. composition within and between mountains. The rate of decrease in the number of species with in- Decreasing species richness with elevation is common, creasing elevation was 38 % higher in Los Tuxtlas (32 % for but the patterns vary among taxonomic groups (Rahbek genera) than in La Chinantla, while species accumulation 1995). The general pattern observed in insects is that of a was 26 % higher in La Chinantla than in Los Tuxtlas. On Los linear decrease in species richness (McCoy 1990), although Tuxtlas no exclusive genera were observed at higher eleva- for dung beetles peaks in richness have also been reported at tions, and this results in a species-poor beetle community that 112 F. Alvarado et al.

the relative influence of historical processes on the regional scale (Martín–Piera et al. 1992,Jay–Robert et al. 1997, Lobo and Halffter 2000). Specifically, the horizontal colonization (or the latitudinal turnover of the biota) by the Holarctic taxa southward along the upper parts of the mountains (>2,000 m a.s.l.) during the cold periods of the Plio- Pleistocene appears to have been strongly limited by geographic isolation and recent geological activity, as occurred in the northern Andes (Escobar et al. 2005, 2006). Despite the differences in elevation between the mountains under study, this same phenomenon—the absence of elements from out- side the tropics at higher elevations in Los Tuxtlas—has been observed in high elevation mountain systems (>3,000 m a.s.l.) that are geographically isolated from the mountains located to the north and those that are recent in geological terms, such as the Talamanca Mountain Range in Costa Rica (Halffter et al. 1995) and the Sierra Nevada de Santa Marta Mountain Range in Colombia (Noriega et al. 2007). Although we do not have a definitive picture of horizontal colonization in tropical mountains, studies of plants and birds (Vuilleumier 1986, Gentry 2001) suggest that the importance of this process depends on the taxonomic group (i.e., dispersal type and capacity). As a tool for analyzing the complex biogeographic scenario of the MTZ, Halffter (1964, 1976, 1987) proposed five distribution patterns that have been widely adopted (Thomas 1993, Marshall and Liebherr 2000, Morrone 2005, Morrone and Márquez 2008, Escobar et al. 2007, Morrone 2010, Fig. 5 Changes in the proportion of individuals belonging to species with different biogeographical distribution patterns (sensu Halffter Jones et al. 2012). These distribution patterns are based on 1987). NEA Nearctic; PMM Montane Paleoamerican; PAT Tropical the origin and phylogeny of the biota as well as the climate Paleoamerican; MAM Montane Mesoamerican; NTT Typical Neotropical and geological history of the region. In the MTZ not only do lineages of different origins (north and south) overlap, they also entered the zone during different geological periods and is made up of species that are phylogenetically related under different paleogeographic scenarios. In general terms, to those in the adjacent lowlands, all of which belong to we can identify genera or lines of species for which penetra- the subfamily Scarabaeinae (i.e., the genera Canthidium, tion into the MTZ is ancient (Miocene) and more recently Canthon, Eurysternus, Dichotomius, Ontherus, Phanaeus arrived genera (Pliocene), and this is reflected in the species and Uroxys, all abundant below 200 m a.s.l., where they composition observed in the elevation gradients studied. represent 60 % of all beetles captured; see Table S1). Al- These distribution patterns are useful for understanding ver- though, relatively speaking, the beetle community of the upper tical and horizontal colonization by the beetle communities elevations of La Chinantla was also poor in species, it is of Los Tuxtlas and La Chinantla. composed of extratropical lineages, different from those found At La Chinantla there are three well-differentiated sets of in the lowlands and belonging to the genera of two different beetle species with two transition elevations for the fauna: subfamilies: Geotrupes (Geotrupinae), Gonaphodiellus and the first in the area where the forests of the lowlands are in Agrilinellus (Aphodiinae). Most of the lineages of the Scara- contact with cloud forest (1,000–1,200 m a.s.l.) and the baeinae would have difficulty colonizing and inhabiting cold, second between the cloud forest and pine-oak forest temperate climates while some Aphodiinae and Geotrupinae (1,800–2,000 m a.s.l.). These transition zones clearly corre- would have difficulty in warm or tropical climates. This ex- spond to changes in the dominance of species with different plains the marked substitution at the subfamily level above biogeographic distribution patterns. In the first transition 2,500 m a.s.l., as observed for other mountains in the MTZ zone, the change is from the NTT pattern to MAM, the andinsouthernEurope(Martín–Piera and Lobo 1993,Lobo former composed of species typical of the lowland forests, and Halffter 2000). of South American origin and of recent penetration into the The differences in the rates of gain and loss with elevation MTZ, while the latter is made up of species of both northern for both genera and species can be interpreted as indicative of affinity or originating in Mexico and ancient species of Diversity and biogeographical makeup of the dung beetle 113 southern affinity (Halffter 1987, Moron 1983). In the second change in the mountains and its relationship to human activ- transition zone, there is a decrease in the dominance of ities. In this context, the hypothesis of Lobo and Halffter MAM species, which are replaced by species with Holarctic (2000) regarding how tropical mountains continue to be or Nearctic distributions, restricted to the mountains of Mex- colonized by lineages typical of lower elevations is very ico and northern Central America, many of them associated interesting. This phenomenon can be reinforced by changes with coniferous forests or the arid and semiarid regions of the in land use along elevation gradients. Under this hypothesis Mexican High Plain (Halffter 1987). In spite of the relevance we would expect not only changes in the elevational distri- of horizontal colonization in the formation of the beetle bution of species, but also an increase in heliophile species fauna of La Chinantla, there are several phyletic lines whose that reach the upper elevations of mountains by crossing presence at the mid- and higher elevations of this mountain pastures, as reported for the mountains of Mexico (Halffter are the result of vertical colonization, such as those classified et al. 1995), Colombia (Escobar et al. 2007) and Peru (Larsen as having the MAM (Deltochilum mexicanum, Onthophagus 2012). We would also expect a change – upward displace- subcancer) and PAM (Copris sallei, Onthophagus ment – of the transition zone between lineages of Holarctic chevrolati) patterns. and Neotropical affinity in the mountains of the MTZ and At Los Tuxtlas, in contrast, it was not possible to distin- other regions where these two lineages come into contact. As guish a fauna specific to the higher elevations owing to the a consequence, this could result in the loss of phylogenetic high variation and high degree of overlap among the species diversity and the homogenization of mountain fauna. assemblages at each elevation range. However, as revealed by the ordination analysis, the stations above 1,200 m a.s.l. Acknowledgments The authors thank Mario Zunino (Onthophagus), tended to be different from those located below 1,000 m a.s.l. Francisco José Cabrero (Aphodiinae) and Nuria Trotta (Geotrupinae) (see Fig. 3). On this mountain the high degree of overlap in for their valuable help identifying the specimens collected, and are grateful to Fernando Escobar for assistance with the taxonomy of the species composition is related to the presence of lineages Scarabaeinae and for his collaboration during the field work of this exhibiting the NTT pattern with broad elevation distribution study. We are grateful to Gonzalo Halffter for his contribution to the ranges, such as Canthidium ardens, Canthon vasquezae, discussion about the biogeographic patterns in the Mexican Transition Dichotomius satanas, Onthophagus batesi and Uroxys Zone. Fredy Alvarado gratefully acknowledges the award of a graduate studies scholarship from CONACyT (no. 40081). The authors thank platypyga, all of which are common in the tropical lowland Bianca Delfosse for translating and editing the English version of this forests. The preponderance of vertical colonization at Los manuscript. Tuxtlas is confirmed by the dominance of the PAT and PAM distribution patterns, both northern affinity and ancient penetration into the MTZ, that reach intermediate elevations on References the mountain (800 to 1200 m a.s.l.; see Fig. 5) from the lower elevations. According to Lobo and Halffter (2000), the more Anderson, M. J. (2001). A new method for non–parametric multivariate isolated a mountain is, the less suitable it is as a refuge and analysis of variance. Austral Ecology, 26,32–46. the lower its capacity to maintain allochthonous elements Anderson, M. J., Ellingsen, K. E., & McArdle, B. H. (2006). Multivariate dispersion as a measure of beta diversity. Ecology Letters, 9,683–693. during the climate fluctuations that occurred during periods Colwell, R. K., & Coddington, J. A. (1994). Estimating the extent of of glaciation. Consequently, under these conditions, horizon- terrestrial biodiversity through extrapolation. Philosophical Trans- tal colonization would have less of an effect and species actions of the Royal Society, 345, 101–118. assemblages would be the result of vertical colonization. Colwell, R. K. (2008). EstimateS: Statistical estimation of species richness and shared species from samples. Version 8.0. User's According to Hewitt (2001), the species recorded at higher Guide and application published at: http://purl.oclc.org/estimates. elevations in the American tropics may be recent and geo- Davis, M. B., & Shaw, R. G. (2001). Range shifts and adaptive graphically limited, likely indicating the existence of an responses to Quaternary climate change. Science, 292, 673–678. association between modern diversification processes and Escobar, F., Lobo, J. M., & Halffter, G. (2005). Altitudinal variation of recent geological activity. This could explain the absence dung beetle (Scarabaeidae: Scarabaeinae) assemblages in the Co- lombian Andes. Global Ecology and Biogeography, 14, 327–337. of extratropical lineages of ancient origin from Los Tuxtlas. Escobar, F., Lobo, J. M., & Halffter, G. (2006). Assessing the origin of Neotropical mountain dung beetle assemblages (Scarabaeidae: Conservation perspective Scarabaeinae): the comparative influence of vertical and horizontal colonization. Journal of Biogeography, 33, 1793–1803. Escobar, F., Halffter, G., & Arellano, L. (2007). From forest to pasture: an The results of this study support the idea that along elevation evaluation of the influence of environment and biogeography on the gradients community diversity results from a combination, to structure of dung beetle (Scarabaeinae) assemblages along three alti- varying degrees, of vertical colonization (regulated mainly tudinal gradients in the Neotropical region. Ecography, 30, 193–208. by ecological processes) and horizontal colonization (histor- Elias, S. A. (1994). Quaternary insects and their environments. Wash- ington, DC: Smithsonian Institution Press. ical processes). Keeping this conceptual framework in mind, Ferrusquía–Villafranca, I. (1993). Geology of México: a synopsis. In T. P. we believe it is possible to explore the impact of climate Ramamoorthy, R. Bye, A. Lot, & J. Fa (Eds.), Biological diversity of 114 F. Alvarado et al.

Mexico. Origins and Distribution (pp. 3–107). New York: Oxford comparative study. Annals of the Entomological Society of Amer- University Press. ica, 93,115–126. García–Mendoza, A., Ordóñez–Díaz, M. J., & Briones-Salas, M. Lomolino, M. V. (2001). Elevation gradients of species–density: his- (2004). Biodiversidad de Oaxaca. México: Instituto de Biología, torical and prospective views. Global Ecology and Biogeography, Universidad Nacional Autónoma de México. 10,3–13. García–López, A., E. Micó., & Galante, E. (2011). From lowlands to Martín–Piera, F., Veiga, C. M., & Lobo, J. M. (1992). Ecology and biogeo- highlands: searching for elevational patterns of species richness graphy of dung-beetle communities (Coleoptera, Scarabaeoidea) and distribution of scarab beetles in Costa Rica. Diversity and in an Iberian mountain range. Journal of Biogeography, 19, Distibutions,1–11. 677–691. Gentry, A. H. (2001). Patrones de diversidad y composición florística Martín–Piera, F., & Lobo, J. M. (1993). Altitudinal distribution patterns en los bosques de las montañas tropicales. In K. Kappelle & A. D. of copro-necrophage Scarabaoidea (Coleoptera) in Veracruz, Mex- Brown (Eds.), Bosques Nublados del Neotropico (pp. 63–84). ico. The Coleopterists Bulletin, 47, 321–394. Costa Rica: Instituto Nacional de Biodiversidad. Marshall, C. J., & Liebherr, J. K. (2000). Cladistic biogeography of the Guevara, S., Laborde, J., & Sánchez–Ríos, G. (2004). Los Tuxtlas. El Mexican Transition Zone. Journal of Biogeography, 27, 203–216. paisaje de la Sierra. México. Xalapa Ver: Instituto de Ecología A.C. McCoy, E. D. (1990). The distribution of insects along elevational Halffter, G. (1964). La entomofauna Americana. Ideas acerca de su gradients. Oikos, 58, 313–322. origen y distribución. Folia Entomologica Mexicana, 6,1–108. Moron, M. A. (1983). A revision of the subtribe Heterosternina (Cole- Halffter, G., & Matthews, E. G. (1966). Natural history of dung beetles optera: Melolonthidae: Rutelinae). Folia Entomologica Mexicana, of the subfamily Scarabaeinae (Coleoptera: Scarabaeidae). Folia 55,31–101. Entomológica Mexicana, 12–14,1–132. Morrone, J. J. (2005). Hacia una síntesis biogeográfica de México. Halffter, G. (1976). Distribución de los insectos en la Zona de Transición Revista Mexicana de Biodiversidad, 76, 207–252. Mexicana. Relaciones con la Entomofauna de Norteamérica. Folia Morrone, J., & Márquez, J. (2008). Biodiversity of Mexican terrestrial Entomológica Mexicana, 35,1–64. arthropods (Arachnida and Hexapoda): a biogeographical puzzle. Halffter, G. (1987). Biogeography of the montane entomofauna of Mexico Acta Zoológica Mexicana (n.s), 24,15–41. and Central America. Annual Review of Entomology, 32,95–114. Morrone, J. J. (2010). Fundamental biogeographic patterns across the Halffter, G., Favila, M. E., & Arellano, L. (1995). Spatial distribution of Mexican Transition Zone: an evolutionary approach. Ecography, three groups of Coleoptera along and altitudinal transect in the 33, 355–361. Mexican Transition Zone and its biogeographical implications. Navas, C. (2003). Herpetological diversity along Andean elevational Elytron, 9, 151–185. gradients: links with physiological ecology and evolutionary phys- Hanski, I. (1983). Distributional ecology and abundance of dung and iology. Comparative Biochemistry and Physiology, 133, 469–485. carrion-feeding beetles (Scarabaeidae) in tropical rain forest in Noriega, J. A., Solis, C., Escobar, F., & Realpe, E. (2007). Escarabajos Sarawak, Borneo. Acta Zoologica Fennica, 167,1–45. coprófagos (Coleoptera: Scarabaeidae) de la provincia de la Sierra Hanski I., & Niemelä J. (1990). Elevational distributions of dung and carrion Nevada de Santa Marta. Biota Colombiana, 8,77–86. beetles in northern Sulawesi. In: Knight WJ, Holloway JD, editors. Oksanen, J. (2010). Multivariate analyses of ecological communities In Insects and the Rain Forests of South East Asia (Wallacea). p 145–152. R: Vegan tutorial. (pp. 35–40). http://cc.oulu.fi/_jarioksa/opetus/ Hanski, I., & Cambefort, Y. (1991). Dung Beetle Ecology. Princeton: metodi/vegantutor.pdf. Princeton University Press. Quinn, G. P., & Keough, M. J. (2002). Experimental design and data Hewitt, G. M. (2001). Speciation, hybrid zones and phylogeography – or analysis for biologists. Cambridge University Press. (pp. 401–471). seeing genes in space and time. Molecular Ecology, 10,537–549. Rahbek, C. (1995). The elevational gradient of species richness: a Huston, M. A. (1994). Biological Diversity: The Coexistence of Species uniform pattern? Ecography, 18, 200–205. on Changing Landscape. Cambridge: Cambridge University Press. Sanders, N. J., & Rahbek, C. (2011). The patterns and causes of eleva- Janzen, D. (1967). Why mountain passes are higher in the tropics. tional diversity gradients Ecography, Virtual Special Issue: 000:001– American Naturalist, 112, 225–229. 003. http://issuu.com/wblifesci/docs/ecography-elevational_diversity_ Jay–Robert, P., Lobo, J. M., & Lumaret, J. P. (1997). Altitudinal turnover gradients?mode=window&viewMode=doublePage and species richness variation in European montane dung beetle Scholtz, C. H. (1990). Phylogenetic trends in the Scarabaeoidea assemblages. Arctic, Antarctic, and Alpine Research, 29,196–205. (Coleoptera). Journal of Natural History, 24, 1027–1066. Jones, R. W., Niño–Maldonado, S., & Luna–Cozar, J. (2012). Diversity and Soriano, E., Dirzo, R., & Vogt, R. C. (1997). Historia Natural de Los biogeographic affinities of Apionidae (Coleoptera: Curculionoidea) Tuxtlas. México: Instituto de Biología, Universidad Nacional along an altitudinal gradient in El Cielo Biosphere Reserve of north- Autónoma de México. eastern Mexico. Revista Mexicana de Biodiversidad, 83, 100–109. Thomas, D. B. (1993). Scarabaeidae (Coleoptera) of the Chiapanecan Jost, L. (2007). Partitioning diversity into independent alpha and beta forests: a faunal survey and chorographic analysis. The Coleoperists components. Ecology, 88, 2427–2439. Bulletin, 47,363–408. Körner, C. H. (2000). Why are there global gradients in species rich- van der Hammen, T. (1995). Global change, biodiversity, and conserva- ness? Mountains may hold the answer. Trends in Ecology & tion of Neotropical montane forest. In S. P. Churchill, H. Balslev, E. Evolution, 15,5–13. Forero, & J. L. Luteyn (Eds.), Biodiversity and conservation of Larsen, T., & Forsyth, A. (2005). Trap spacing and transect design for Neotropical montane forests. Proceedings of the Neotropical mon- dung beetle biodiversity studies. Biotropica, 37, 322–325. tane forest biodiversity and conservation symposium (pp. 603–607). Larsen, T. H. (2012). Upslope range shifts of Andean dung beetles in New York: New York Botanical Garden. response to deforestation: compounding and confounding effects Vuilleumier, F. (1986). Origins of the tropical avifaunas in the high Andes. of microclimatic change. Biotropica, 44,82–89. In F. Vuilleumier & M. Monasterio (Eds.), High altitude tropical Lobo, J. M., & Halffter, G. (2000). Biogeographical and ecological biogeography (pp. 586–622). Oxford: Oxford University Press. factors affecting the altitudinal variation of mountainous commu- Zar, J. (1996). Biostatistical Analysis. Englewood Cliffs, New York: nities of coprophagous beetles (Coleoptera: Scarabaeoidea): a Prentice Hall.