SciELO - Scientific Electronic Library Online

 
vol.24 número1Terrestrial mammals of the Jequitinhonha River basin, Brazil: a transition area between Atlantic Forest and CerradoOccurrence of fin and humpback whales in the Scotia Sea and the protected marine area of the South Orkney Islands, Antarctica índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Servicios Personalizados

Revista

Articulo

Indicadores

  • No hay articulos citadosCitado por SciELO

Links relacionados

Compartir


Mastozoología neotropical

versión impresa ISSN 0327-9383versión On-line ISSN 1666-0536

Mastozool. neotrop. vol.24 no.1 Mendoza jun. 2017

 

ARTÍCULO

Habitat complexity and small mammal diversity along an elevational gradient in southern Mexico

 

José L. Mena1 and Rodrigo A. Medellín2

1 Museo de Historia Natural Vera Alleman Haeghebaert, Universidad Ricardo Palma, Lima, Perú. [Correspondencia: José L. Mena <menaa.jl@gmail.com>]
2
Instituto de Ecología, Universidad Nacional Autónoma de México, Ciudad Universitaria, D. F., México.

Recibido 30 enero 2017.
Aceptado 27 marzo 2017.
Editor asociado: J Morrone


ABSTRACT.

We tested the hypothesis that habitat complexity explains alpha diversity of nonvolant small mammals along an elevational gradient in southern Mexico. During October-November 2003, we conducted fieldwork on the Pacific slope of El Triunfo Biosphere Reserve. Small mammal trapping was conducted using standardized techniques (trap lines and pitfalls) along an elevational gradient between 500 and 2100 m elevation. Habitat assessment as indicated by vegetation complexity and diversity was conducted at each site (N = 12). Nine species and 148 individuals were captured in 8400 trap-nights. Results indicate that non volant small mammal diversity increases with habitat complexity. In addition, our study shows that the spatial pattern of diversity cannot be attributed to spatial autocorrelation.

RESUMEN.

Complejidad de hábitat y diversidad de mamíferos pequeños en un gradiente altitudinal en el sur de México.

Probamos la hipótesis de que la complejidad del hábitat explica la diversidad (diversidad alfa) de mamíferos pequeños no voladores a lo largo de un gradiente de elevación en el sur de México. El trabajo de campo se realizó entre octubre y noviembre de 2003, en la vertiente del Pacífico de la Reserva de la Biosfera El Triunfo. La captura de mamíferos pequeños se llevó a cabo utilizando técnicas estandarizadas (líneas de trampeo y trampas de caída) a lo largo del gradiente de 500 a 2100 m de elevación. El hábitat fue evaluado con base en la complejidad y diversidad de la vegetación en cada sitio evaluado (N = 12). Nueve especies y 148 individuos fueron capturados en 8400 noches-trampa. Los resultados indican que la diversidad de mamíferos pequeños aumenta con la complejidad del hábitat. Además, nuestro estudio muestra que el patrón espacial de la diversidad encontrado no está influenciado por la autocorrelación espacial.

Key words: Alpha diversity; Elevational gradient; Habitat complexity; Mexico; Small mammals.

Palabras clave: Complejidad de hábitat; Diversidad alfa; Gradiente de altitud; Mamíferos pequeños; México.


INTRODUCTION

The study of local elevational gradients has great potential for increasing our knowledge about both regional and global scale diversity processes and the implications of climate change. Indeed, elevational gradients have been reassessed in recent years (Brown, 2001; Lomolino, 2001; Mena and Vázquez-Domínguez, 2005; McCain, 2007b; Guo et al., 2013) due to changing perspectives on their interpretation. Elevational gradients can be used as natural experiments, allowing for rigorous testing of hypotheses elicited by specific questions, such as effects of small spatial scale or elevational trends in abiotic factors (Grytnes and McCain, 2007).

There are two ways to quantify elevational patterns in species richness: alpha and gamma diversity studies (McCain, 2005). Alpha diversity studies use local field sampling of plots along a transect, usually on one mountain slope, preferably with equal sampling effort at each elevational band; and gamma diversity studies use regional data from previously collected specimens and field records from an entire mountain or mountainous region (McCain, 2005; McCain, 2007a). Clearly, the observed elevational trend in species varies among groups of organisms. The most commonly observed patterns are decreasing richness with increasing elevation (amphibians, bats and reptiles) (Sánchez-Cordero, 2001; Patterson et al., 1996; Sergio and Pedrini, 2007; Chettri et al., 2010), a low plateau (high diversity across most of the lower portion of the gradient then decrease [birds, reptiles]) (McCain, 2009; McCain, 2010), or a humped pattern with a richness peak at intermediate elevations (mainly in nonvolant small mammals and plants) (Kessler, 2001; McCain, 2005; Cardelús et al., 2006).

The explanations commonly offered for elevational patterns in species richness can be grouped into four categories: climatic hypotheses based on current abiotic conditions, spatial hypotheses of area and spatial constraint, historical hypothesis invoking processes occurring across evolutionary timescales, and biotic hypotheses such as community overlap (juxtaposition), source-sink dynamics and habitat heterogeneity (Brown, 2001; Lomolino, 2001; Grytnes and McCain, 2007). However, elevational gradients in species richness result from a combination of ecological and evolutionary processes, and thus, may not reflect one overriding force (Lomolino, 2001; Wu et al., 2013).

Certainly, habitat heterogeneity has a large effect on species richness, but the relevant type of heterogeneity will depend on the species group and the scale of study (Brown, 2001; Grytnes and McCain, 2007; Rowe et al., 2015). Specifically, habitat structure appears to be the most important factor describing diversity of terrestrial small mammals (August, 1983; Medellín and Equihua, 1998; Lambert et al., 2006; Mena and Medellín, 2010); however, this hypothesis has not been tested rigorously along elevational gradients. Specifically, the habitat complexity or habitat structure hypothesis predicts that alpha diversity should vary with local habitat complexity, and peak at elevations characterized by higher habitat complexity (MacArthur, 1964; MacArthur, 1972; Rosenzweig, 1992). Here, we examine this hypothesis in order to understand if this explains species richness along an elevational gradient in southern Mexico. In addition, we explore whether the spatial autocorrelation (Koenig and Knops, 1998; Overmars et al., 2003) can influence the resulting elevational gradient in species richness. Testing spatial autocorrelation is helpful to address an essential question for this type of studies: methodologically, what elevational interval is useful to assess elevational patterns in species richness (e.g., 250 or 500 m intervals). This approach can be very helpful for selection of sites along elevational gradients where researchers will conduct small mammal inventories. Usually, sampling sites along elevational gradients are selected as vegetation changes; however, there are other ways to determine their location. In this context, spatial correlation analysis is a useful tool to investigate mechanisms operating on species richness at different spatial scales (Diniz-Filho et al., 2003). Indeed, this method is often used to assess the relationship between variables along gradients.

MATERIAL AND METHODS

Study area

The study was carried out in the El Triunfo Biosphere Reserve and its buffer zone (Fig. 1), on the Pacific slope of the Sierra Madre de Chiapas, southeastern Mexico (15° 09’10’’-15° 57’02’’N, 92° 34’04’’-93° 12’42’’W). This reserve is one of the few relatively undisturbed Mexican cloud forests. The Pacific slope descends from the highest peak, Cerro El Triunfo (2450 m a.s.l.), with steep slopes influenced by erosion and landslides. Slopes level off somewhat at mid-elevations; below 800 m, land is more influenced by agriculture (mainly coffee plantations), ranching, and human settlement. The wet season extends from May to October and the dry season from November to April. Annual mean temperature above 2000 m elevation is between 16- 20 °C and annual precipitation between 2500-4500 mm (Williams-Linera, 1991; INE, 1998; Morón-Ríos and Morón, 2001). Annual mean temperature at elevations between 1000 and 2000 m is 18-22 °C, with annual precipitation between 2000-3000 mm; below 1000 m, annual mean temperature is 26 °C, and annual precipitation between 2500-4000 mm (INE, 1998).


Fig. 1. The study area and sampling sites on the Pacific slope of the El Triunfo Biosphere Reserve. Trap lines (solid circles) and complementary trapping (white circles) are shown on the map.

Long and Heath (1991) and Williams-Linera (1991) provide detailed information on vegetation and climate at El Triunfo. This reserve protects 10 of the 19 vegetation types found in Chiapas, including large areas of the remaining stands of Central American cloud forest (Breedlove, 1981). The vegetation types that cover the Pacific slope of El Triunfo are tropical evergreen forest (< 800 m), oak forest (800-1200 m), montane rainforest (1200- 1600 m), pine forest (1600-1800 m), and upper cloud forest (> 1800 m). Tropical evergreen forest (TEF), 527-632 m elev. This forest is disturbed and fragmented (buffer zone of the reserve) and larger fragments of undisturbed forest are absent below 500 m a.s.l. In general, lower areas of the gradient are mainly cultivated with coffee plantations. Important canopy tree species in this forest are Ceiba sp. (Bombacaceae), Platymiscium sp. and Calliandra sp. (Fabaceae), Trophis sp. and Pseudolmedia sp. (Moraceae), and Eugenia sp. (Myrtaceae); understory species are Piper sp. (Piperaceae) and species of Fabaceae, Lythraceae and Urticaceae. Oak forest (OF), 945 m elev. This forest has a canopy dominated by Quercus salicifolia (Fagaceae), and an understory represented by Ternstroemia sp. (Ericaceae). Montane rainforest (MR), 1256-1318 m elev. This forest has a canopy dominated by Amphitecna sp. (Bignoniaceae), Quercus spp. (Fagaceae), Ocotea sp., Nectandra sp. Persea sp. (Lauraceae), and Eugenia sp. (Myrtaceae); and understory species of Araliaceae, Fagaceae, Melastomataceae and Piperaceae. Pine forest (PF), 1794 m elev. This forest is dominated by Cupressus sp. and Pinus sp., with understory species of Ericaceae and Rubiaceae. Upper cloud forest (UCF), 1988-2020 m elev. (mainly primary forest) is dominated by Quercus oocarpa (Fagaceae), Matudaea trinervia (Hamamelidaceae), Hedyosmum mexicanum (Chloranthaceae) and Dendropanax populifolius (Araliaceae), with understory species of Araliaceae, Melastomataceae, Piperaceae and Saurauriaceae (Long and Heath, 1991).

Small mammal trapping

During October-November 2003, we used removal trapping of nonvolant small mammals (Soricidae, Didelphidae, Heteromyidae and Cricetidae) in 12 sampling sites using trap lines for 8-9 nights (Table 1). Each trap line had 20 stations with 5-m spacing. At each station, two traps were set on the ground (1 Sherman 8 x 9 x 23 cm and 1 Victor mouse trap 16 x 9 cm) and two on fallen logs, trees, vines or lianas (ca. 2 m height). Traps were baited with a mixture of peanut butter, rolled oats and vanilla extract, a standard procedure for small mammals (Voss and Emmons, 1996).

Table 1
Non-volant small mammals along elevational sampling sites at El Triunfo Biosphere Reserve. Shaded areas indicate groups included in the generalized linear mixed model (GLMM).

As complementary methods for the small mammal’s inventory, we conducted two trappings at 2000, 1300 and 550 m. A protocol for Ichthyomyinae in small streams was conducted with 6 Victor traps for 8 days (and baited with crab meat). In addition, we installed 100-m long lines of pitfall traps to record small insectivores. Each pitfall line consisted of 10 buckets spaced 10 m apart, with one bucket at either end, for a total length of 100 m. Drift fences, consisting of a continuous barrier running the total length of each line, were made of 40 cm wide strips of hardware clear polyethylene clipped to vertical stakes hammered into the ground. Pitfall traps were operated for 8 days. However, the individuals captured with these complementary methods were not included in the analysis of our hypothesis.

Specific identification of the captured small mammals was aided by the use of a field guide (Reid, 1997). A limited number of voucher specimens of representative mammals were collected. Specimens are housed at the Colección de referencia del Laboratorio de Ecología y Conservación de Vertebrados (Instituto de Ecología, Universidad Nacional Autónoma de México).

Habitat structure assessment

We established ten circular habitat stations (4 m radius) in each trap line. Habitat stations were placed at odd trapping stations. All individual trees with > 5 cm diameter at breast height (d.b.h.) in each habitat station were identified and counted, and were assigned into four categories: 5-10 d.b.h. (T0510), 10-30 d.b.h. (T1030), 30-50 d.b.h. (T3050), > 50 d.b.h. (T>50). Other variables measured were canopy cover (PCC) with a spherical densiometer, taken five times (one in each direction of the four cardinal points and at the center of the circular station), and percent of herbaceous vegetation cover (HC) within 1-m2 quadrant (100, 50, and 0 categories).

Species richness analyses

We define alpha diversity as the number of species detected per spatially standardized survey effort (Lomolino, 2001) in one season. An alpha diversity data set included species recorded for each sampling site (N = 12 trap lines, see Table 1). In order to assess the completeness of our trapping, we used a rarefaction-extrapolation approach to extrapolate the observed accumulation curve (Chao et al., 2014; Colwell et al., 2012; Chao and Jost, 2012). In this way, we used bootstrap methods to construct confidence intervals for alpha diversity of any rarefied or extrapolated sample (trap line) and developed an individual-based (abundance) model. For each trapline, we estimated sample completeness (%) which is measured by sample coverage: the proportion of the total number of individuals that belong to the species detected in the sample (Chao and Jost, 2012). All estimates were obtained by the software iNEXT (Hsieh et al., 2015).

Autocorrelation assessment

In order to assess the spatial autocorrelation in the elevational gradient of the El Triunfo, we generated spatial correlograms for alpha diversity and elevation using Moran’s I coefficients at 8 elevational classes (SAAP 4.3; Wartenberg, 1989). Upper limits for these elevation classes were 56, 311, 476, 686, 758, 849, 1356, and 1490 m. Moran’s I usually varies between -1.0 and 1.0 for maximum negative and positive autocorrelation, respectively (Diniz-Filho et al., 2003). Non-zero values of Moran’s I indicate that richness (alpha diversity) values in sites connected at a given elevation are more similar (positive autocorrelation) or less similar (negative autocorrelation) than expected for randomly associated pairs of traplines. The spatial autocorrelogram for small mammal diversity indicated that it was positively autocorrelated up to c. 60 m (with statistical significance), followed by a continuous decrease in Moran’s I coefficients (non-significant) up to c. 1600 m, at which point there is a highly significant negative autocorrelation coefficient. After including elevation variables successively in the model (linear and quadratic function), spatial autocorrelation in the residuals disappeared. Results were not affected by number and definition of distance classes in the correlogram, and Moran’s coefficient was not significant (P > 0.90).

Hypothesis assessment

Habitat complexity describes the development of vertical strata within an habitat (August, 1983). Thus, complex habitats would have dense and tall ground cover, many large trees and shrubs, and a dense canopy and understory. In this way, we performed a principal components analysis (PCA) to identify vegetation variables that helped distinguish the trap lines in terms of their habitat complexity along the elevational gradient. The idea of PCA is to find a small number of linear combinations of the variables so as to capture most of the variation in the data as a whole, and only the first two or three components are generally used as new variables, since they often explain most of the total sample variance (Crawley, 2013). Since the variables are expressed in different measurement scales, we computed a PCA on the correlation matrix, including a standardization of the variables (Borcard et al., 2011). As index of habitat complexity by each trap line we used the scores resulting from the first and second principal components (PC1 y PC2), because these condense all of the original variables into two measures of overall size. We selected the number of axes representing the major features of the data according to the Kaiser-Guttman criterion, which consists in computing the mean of all eigenvalues and interpreting only the axes whose eigenvalues are larger than that mean (Borcard et al., 2011).

The species richness data for alpha diversity along the elevational gradient were compared with null model predictions using the Monte Carlo simulation program “Mid-Domain Null” (McCain, 2004). This program simulates species richness curves based on empirical range sizes or range midpoints within a bounded domain. The hard boundaries for the model were defined by the lowlands and highlands of the study area. The elevation midpoint for each species was drawn at random (50 000 iterations, without replacement). Regression of the empirical data on the predicted values based on the average of 50 000 iterations, provided r2 estimates to the fit of the null model (McCain, 2004).

We performed a linear regression to assess the relationship between habitat complexity and elevation. In addition, we conducted generalized linear models (GLM) with assumption of Poisson errors (Zuur et al., 2009; Crawley, 2013), to quantify the effects of habitat complexity on alpha diversity, overall abundance, and abundance of the most common species along the elevational gradient. For model selection we used the Akaike Information Criterion (AIC), and all models with AIC differences of less than 2 were retained because it is suggested that these have a substantial level of empirical support (Burnham and Anderson, 2002). For small sample sizes (n), where n/k < 40 (where k is the number of parameters) a “corrected AIC” (AICc) (Bolker, 2008) is recommended, so that, selection model was based on AICc. All analyses were executed using the R programming environment (R Core Team, 2016), with ggord (Beck, 2016), lme4 (Bates et al., 2015) and AICcmodavg (Mazerolle, 2016) libraries. We estimated the explained deviance (because we do not have an R2 in GLM models), which was calculated in terms of null deviance and residual deviance (Zuur et al., 2009).

To answer the question about how strongly alpha diversity is related to habitat complexity, or to predict alpha diversity from habitat complexity, we drawn inferences from a Bayesian framework, where is key the joint posterior distribution of β (vector that contains β0 and β1) and σ2, the residual variance. The posterior distributions describe the range of plausible parameter values given the data and the model, an estimate of our uncertainty about the model parameters. In this way, the 2.5% and 97.5% quantiles of the marginal posterior distributions can be used as 95% credible intervals (CrI) of the model parameters, thus, the interpretation of the 95% credible interval is straightforward, in this way, we are 95% sure that the true regression line is within the credible interval (Korner-Nievergelt et al., 2015). First, we obtained parameter estimates and then, we used the sim function in the package arm (Gelman and Su, 2015), which uses the results from the model fit to calculate the posterior distribution assuming flat prior distributions (Gelman and Hill, 2007). We use the function sim to draw 5000 random values from the joint posterior distribution of the model parameters; that is, we draw 5000 values for each parameter while taking the correlation between the parameters into account. We obtained a graphical output using a frequentist method with the predict function (Crawley, 2013; Korner-Nievergelt et al., 2015), which simulated data to estimate confidence intervals around the predicted line (Zuur et al., 2009). This provides a similar result with Bayesian methods, but it is simpler in R (Korner-Nievergelt et al., 2015). Finally, we conducted a generalized linear mixed model (GLMM) grouping nearby sampling sites (< 60 m apart), in three groups (low, middle and high elevation), following recommendations provided by the autocorrelation analysis. GLMM are useful when we model spatial correlation, where it can be accommodated by adding correlated spatial random effects, with the correlation being a function of their distance in space (Kéry and Royle, 2016). This model was performed with assumption of Poisson errors (Zuur et al., 2009; Crawley, 2013), to quantify the effects of habitat complexity on alpha diversity. Habitat complexity was treated as a fixed factor and elevational replicates as a random one. We performed a dispersion test of the mixed model recommend in Zuur et al. (2013).

RESULTS

One hundred forty-eight individuals representing 9 species were captured in 8400 trap-nights (Table 1). The estimated completeness of the small mammal fauna for each trap line was similar along the elevational gradient (> 76%). Nearly all species were distributed across all elevations, with the exception of those restricted to higher elevations (i.e. Reithrodontomys mexicanus and Peromyscus guatemalensis). Most species occurring in the lowlands had elevational ranges of between 700 and 1500 m. Sorex veraepacis (at 2000 m) and Rheomys thomasi (at 1300 m) were recorded with pitfall and Ichthyomyinae protocol respectively. Alpha diversity did not fit the predictions of a null model (see McCain, 2004).

Our PCA model was performed with 8 habitat variables (see Table 2), and the first two principal components explaining > 65% of the variance of the data (Fig. 2). Trap lines at high elevations displayed the largest value of number of trees and basal area (Table 3). Trap lines at low elevations (and middle elevations) showed the largest values in herbaceous cover and the lowest values in canopy cover and basal area.

Table 2
Average of vegetation variables for each trapping site along the Pacific slope of the El Triunfo, Chiapas, México.


Fig. 2. PCA biplot of the habitat variables (BA, CC, HC and trees) along the elevational gradient at the El Triunfo. The plot shows sampling sites.

Table 3
Summary of the first two principal components

Habitat complexity showed a significant, positive modest association with elevation (y = -2.203 + 0.002x, R2 = 0.37, F1, 10 = 5.92, P = 0.035). Alpha diversity was poorly associated with elevation; however, it showed a positive modest association with habitat complexity (see Table 4 and Fig. 3). The uncertainty measurements for the parameter estimates (alpha diversity vs. habitat complexity) were obtained from the posterior distribution simulated by sim. The 95% credible interval of (Alpha diversity ~ β0 + β1*habitat complexity) was -0.089 – 0.253 (β1= 0.084). Fig. 3 shows alpha diversity with a fitted Poisson GLM curve with 95% confidence bands. Moreover, we found a distinctive positive relationship between habitat complexity and abundances of some common species of small mammals along the elevational gradient (see Table 4). GLMM provides similar results than GLM, a positive relationship between habitat complexity with alpha diversity, and we did not find problems of overdispersion in each case.

Table 4
Summary of model (GLM) selection for data on small mammal communities along the elevational gradient at El Triunfo. Only models with a model weight (w) > 0.1 are shown.


Fig. 3. Observed species richness with a fitted Poisson GLM (solid line) and 95% confidence bands (dotted lines) based on estimates of model parameters ( = -5.265, β0 ^ = 0.084)

DISCUSSION

Our study provides new data on inventories of nonvolant small mammals on the Pacific slope of the El Triunfo Biosphere Reserve, one of the very few remaining complete elevational gradients in Mesoamerica. Previous inventories in this reserve have been conducted mainly in the upper cloud forest and along the eastern slope of the Sierra Madre de Chiapas. Species accumulation curves showed that our sampling protocol was efficient and our inventory nearly completes (see Table 1). Complementary methods provide records of species at middle and high elevations (S. veraepacis and R. thomasi), and two additional records (Mena, in litt.), Neotoma mexicana at 1611 m in a previous pilot assessment and an anecdotic record of Oligoryzomys fulvescens at 500 m provide some evidence of an increase in species at middle and high elevations.

Only three relatively rare species known from the area such as Cryptotis goodwini, Habromys lophurus and Reithrodontomys megalotis (Medellín, 1988; Espinoza-Medinilla et al., 1998) were not represented in our samples. McCain (2004) emphasized the importance of replication in examining spatial diversity but noted that if only single surveys are feasible, sampling small mammals in Central American forests should be during the wet season (but see Wen et al., 2014). Although, expansion of species elevational ranges and seasonal differences in species trappability due to season remain to be assessed, we do not think that our results were strongly affected. Overall, it is becoming increasingly rare to find complete and intact elevational gradients ranging from sea level to high-elevation mountaintops due to habitat disturbance (Nogues-Bravo et al., 2008).

A quantitative analysis of species richness patterns (plants, invertebrates and vertebrates) along elevational gradients world-wide (includ-ing 204 data sets) showed that about 50% were hump-shaped, 25% were monotonically-decreasing, and 25% had other distributions (Rahbek, 2005). A mid-elevational peak in species richness has been suggested to be the rule for terrestrial small mammals (McCain, 2005). However, some studies have shown an increase of species richness with elevation. On the western slope of the Andes, diversity increases with elevation, probably as a result of increased rainfall (and vegetation), and more speciation events there (Pearson and Ralph, 1978; Marquet, 1994); a similar increase has been reported in the Philippine Islands (Rickart et al., 1991; Balete et al., 2009; Rickart et al., 2011) and, associated with highlands being centers of mammalian diversity. Our results suggest that alpha diversity is associated with habitat complexity along the Pacific slope of El Triunfo, but a clear relationship with elevation was not evident. Furthermore, it is important to acknowledge that Highlands of Chiapas have been recognized as an important biogeographic region, harboring high levels of species richness and endemicity (Escalante et al., 2007). Endemic species restricted to upper cloud forest include Cryptotis goodwini, Habromys lophurus, Peromyscus aztecus, P. guatemalensis and Sorex veraepacis which occur there or in high elevation sites in adjacent regions of Chiapas (Espinoza-Medinilla et al., 1998; Vázquez et al., 2001; Carraway, 2007). Elsewhere, the Chiapas highlands and north-western Middle America are implicated as an important area for diversification of small mammals (Woodman, 2005; León-Paniagua et al., 2007; Rogers et al., 2007), and provide support for a hypothesis that species richness (alpha diversity) increases in areas with high rates of speciation (see Sánchez-Cordero, 2001).

Few studies have integrated habitat structure in the assessment of elevational patterns in alpha diversity, and none in Mesoamerica (Patterson et al., 1990; Kok et al., 2012). Indeed, this paper suggests that habitat complexity is positively related to alpha diversity. In temperate Andean rainforest, Patterson et al. (1990) found that elevation variation in habitat and specific habitat associations provided a plausible explanation for correlations of mammalian distribution and abundance with elevation. Similarly, Batin et al. (2002) found that elevation significantly influenced habitat variables along an elevational gradient in Mount Nuang (Malaysia), suggesting that declining habitat structure may reduce resource availability with increasing elevation, and thus explain declining small mammal diversity. Similar results were described for Mount Kilimanjaro, Tanzania (Mulungu et al., 2008) and Oaxaca, México (Sánchez- Cordero, 2001) where small mammal diversity and distribution patterns were influenced by habitat complexity at different elevations. In our study site, traplines at the upper cloud forest (> 1800 m) was characterized by higher tree basal areas, and abundance and species richness of trees (Table 2; Williams-Linera, 1991; Martínez-Melendez et al., 2008), lending support to the habitat complexity hypothesis. In general, this forest, the wettest habitat on the Pacific versant of Chiapas, and indeed, of Mexico, had greater species richness compared with other habitats along the gradient. This is consistent with the hypothesis that species richness increases with increasing rainfall and primary productivity (Rahbek, 1997; Lomolino, 2001; Sánchez-Cordero, 2001). Moreover, we found some relationships between habitat and species abundance. For example, Heteromys desmarestianus is associated with several vegetation types including tropical rainforest, coffee plantations and agricultural lands from sea level up to ~ 1860 m a.s.l. (Ceballos, 2014), but despite their apparent generalist habitat requirements, we found a relationship with habitat complexity. Similarly, Marmosa mexicana appears to be associated to forest habitats and disturbed areas (Alonso-Mejía and Medellín, 1992; Ceballos, 2014), and we found an association with habitat complexity.

Our analyses suggest that spatial patterns of species richness cannot be attributed primarily to spatial autocorrelations in elevation. Most positive autocorrelations were for sites with close elevational proximity and are likely a result of sampling within areas with similar habitat characteristics; this is mainly a problem when explicit causal effects are being tested (Legendre et al., 2002). The main positive, small-scale spatial autocorrelations in our data suggest that some replicates from nearby trap lines were not truly independent (< 60 m between elevations). Thus, studies analyzing species richness should not ignore spatial effects, and sampling sites along elevational gradients should be distant enough from each other to ensure that spatial autocorrelation is minimal.

We did not address how the anthropic disturbance of the lowlands may have influenced small mammal distributions and elevational patterns. In particular, lowland forests ( < 800 m a.s.l.) have been largely fragmented and deforested on much of the Pacific versant of Chiapas; cleared areas are now coffee agroecosystems, cattle ranches and farmlands (Macip-Ríos and Muñoz-Alonso, 2008). We recognized this effect (i.e., that of disturbed lowlands) by initiating our elevational transect at 527 m instead of 0 m; similarly, the maximum elevation sampled was 2020 instead of 2450 m. In the latter case, we expect higher species richness above 2000 m since the upper cloud forest extends to 2300 m; however, the summit of Cerro el Triunfo (2300 -2450 m) is covered by evergreen shrub ca. 2 m in height and with mosses and bracken ferns (Long and Heath, 1991), and we would expect a precipitously decreased species richness there.

Middle American montane forests are severely threatened by deforestation and other anthropic impacts at lower elevations, and this may have already affected the ability to detect processes which explain diversity there (see Nogues-Bravo et al., 2008). In fact, complete elevational gradients in reasonably intact state are virtually absent in the region. Thus, we need to preserve as many such tropical gradients as well as cloud and montane forests as possible. Understanding the factors that explain diversity along elevational transects remains a key question in ecology and conservation, and El Triunfo is a primary example that provides important insights on this.

ACKNOWLEDGEMENTS

This work was supported by the SRE scholarship from the Mexican government to José Luis Mena, by a grant from Rufford Small Grant and by the Instituto de Ecología, Universidad Nacional Autónoma de México. We thank Lisette Adrianzén, Marco Hernández, Susana Maza, and Alejandro Gómez Nisino for their support during the field work. We also thank Osiris Gaona for technical assistance. Thanks are extended to the El Triunfo Biosphere Reserve for permission to conduct this study. This is a contribution of the Wildlife Trust Alliance and BIOCONCIENCIA, A.C.

LITERATURE CITED

1. ALONSO-MEJÍA A and RA MEDELLÍN. 1992. Marmosa mexicana. Mammalian Species 421:1-4.         [ Links ]

2. AUGUST PV. 1983. The role of habitat complexity and heterogeneity in structuring tropical mammal’s communities. Ecology 64:1495-1507.

3. BALETE DS, LR HEANEY, MJ VELUZ and EA RICKART. 2009. Diversity patterns of small mammals in the Zambales Mts., Luzon, Philippines. Mammalian Biology 74:456-466.         [ Links ]

4. BATES D, M MAECHLER, B BOLKER and S WALKER. 2015. Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software 67:1-48.         [ Links ]

5. BATIN B, SM NOR and MY AHMAD. 2002. Influence of elevational habitat changes on non-volant small mammal species distribution and diversity on Mount Nuang, Hulu Langat, Selangor, Malaysia. Journal of Biological Sciences 5:819-824.         [ Links ]

6. BECK MW. 2016. ggord: Ordination Plots with ggplot2. R package version 0.11.9000.         [ Links ]

7. BOLKER BM. 2008. Ecological Models and Data in R. Princeton University Press.         [ Links ]

8. BORCARD D, F GILLET and L LEGENDRE. 2011. Numerical ecology with R. Springer, New York.         [ Links ]

9. BREEDLOVE DE. 1981. Flora of Chiapas. Part I. Introduction to the flora of Chiapas. California Academy of Sciences, San Francisco, California.         [ Links ]

10. BROWN JH. 2001. Mammals on mountainsides: elevational patterns of diversity. Global Ecology & Biogeography 10:101-109.         [ Links ]

11. BURNHAM KP and DR ANDERSON. 2002. Model Selection and Multimodal Inference: a practical information-theoretic approach. Springer.         [ Links ]

12. CARDELÚS CL, RK COLWELL and JEJ WATKINS. 2006. Vascular epiphyte distribution patterns: explaining the mid-elevation richness park. Journal of Ecology 94:144-156.         [ Links ]

13. CARRAWAY LN. 2007. Shrews (Eulypotyphla: Soricidae) of Mexico. Monographs of the Western North American Naturalist 3:1-91.         [ Links ]

14. CEBALLOS G. 2014. Mammals of Mexico. JHU Press.         [ Links ]

15. COLWELL RK, A CHAO, NJ GOTELLI, S-Y LIN, CX MAO, RL CHAZDON and JT LONGINO. 2012. Models and estimators linking individual-based and sample-based rarefaction, extrapolation and comparison of assemblages. Journal of Plant Ecology 5:3-21.         [ Links ]

16. CRAWLEY MJ. 2013. The R book. Wiley & Sons, Ltd., Chichester, West Sussex, UK.         [ Links ]

17. CHAO A, NJ GOTELLI, TC HSIEH, EL SANDER, KH MA, RK COLWELL and AM ELLISON. 2014. Rarefaction and extrapolation with Hill numbers: a framework for sampling and estimation in species diversity studies. Ecological Monographs 84:45-67.         [ Links ]

18. CHAO A and L JOST. 2012. Coverage-based rarefaction and extrapolation: standardizing samples by completeness rather than size. Ecology 93:2533-2547.         [ Links ]

19. CHETTRI B, S BHUPATHY and BK ACHARYA. 2010. Distribution pattern of reptiles along an eastern Himalayan elevation gradient, India. Acta Oecologica 36:16-22.         [ Links ]

20. DINIZ-FILHO JAF, LM BINI and BA HAWKINS. 2003. Spatial autocorrelation and red herrings in geographical ecology. Global Ecology & Biogeography 12: 53-64.         [ Links ]

21. ESCALANTE T, V SÁNCHEZ-CORDERO and JJ MORRONE. 2007. 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.

22. ESPINOZA-MEDINILLA E, A ANZURES and E CRUZ. 1998. Mamíferos de la Reserva de la Biósfera El Triunfo, Chiapas. Revista Mexicana de Mastozoología 3: 79-94.         [ Links ]

23. GELMAN A and J HILL. 2007. Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge University Press.         [ Links ]

24. GELMAN A and Y-S SU. 2015. arm: Data Analysis Using Regression and Multilevel/Hierarchical Models. R package version 1.8-6. https://CRAN.R-project.org/package=arm.         [ Links ]

25. GRYTNES JA and CM MCCAIN 2007. Elevational trends in biodiversity. 1-8, in: Encyclopedia of Biodiversity (LEVIN, S (ed.         [ Links ])

26. GUO QF, DA KELT, ZY SUN, HX LIU, LJ HU, H REN and J WEN .2013. Global variation in elevational diversity patterns. Scientific Reports 3: 1-7.         [ Links ]

27. HSIEH TC, KH MA and A CHAO. 2015. iNEXT: An R package for interpolation and extrapolation of species diversity (Hill number). http://chao.stat.nthu.edu.tw/blog/software-download/.         [ Links ]

28. INE. 1998. Programa de Manejo de la Reserva de la Biosfera “El Triunfo”. Instituto Nacional de Ecología, México D. F.

29. KÉRY M and JA ROYLE. 2016. Applied Hierarchical Modeling in Ecology. Academic Press, Boston.         [ Links ]

30. KESSLER M. 2001. Patterns of diversity and range size of selected plant groups along an elevational transect in the Bolivian Andes. Biodiversity and Conservation 10:1897-1921.         [ Links ]

31. KOENIG WD and JMH KNOPS. 1998. Testing for spatial autocorrelation in ecological studies. Ecography 21:423-429.         [ Links ]

32. KOK AD, DM PARKER and NP BARKER. 2012. Life on high: the diversity of small mammals at high altitude in South Africa. Biodiversity and Conservation 21:2823-2843.         [ Links ]

33. KORNER-NIEVERGELT F, T ROTH, S VON FELTEN, J GUÉLAT, B ALMASI and P KORNER-NIEVERGELT. 2015. Bayesian Data Analysis in Ecology Using Linear Models with R, BUGS, and Stan. Elsevier.         [ Links ]

34. LAMBERT TD, JR MALCOLM and BL ZIMMERMAN. 2006. Amazonian small mammal abundances in relation to habitat structure and resource abundance. Journal of Mammalogy 87:766-776.         [ Links ]

35. LEGENDRE P, MRT DALE, M-J FORTIN, J GUREVITCH, M HOHN and D MYERS. 2002. The consequences of spatial structure for the design and analysis of ecological field surveys. Ecography 25:601-615.         [ Links ]

36. LEÓN-PANIAGUA L, AG NAVARRO-SIGÜENZA, BE HERNÁNDEZ-BAÑOS and JC MORALES. 2007. Diversification of the arboreal mice of the genus Habromys (Rodentia: Cricetidae: Neotominae) in the Mesoamerican highlands. Molecular Phylogenetics and Evolution 42: 653-664.         [ Links ]

37. LOMOLINO MV. 2001. Elevational gradients of species-density: historical and prospective views. Global Ecology & Biogeography 10:3-13.         [ Links ]

38. LONG A and M HEATH. 1991. Flora of the El Triunfo Biosphere Reserve, Chiapas, Mexico: a preliminary floristic inventory and the plant communities of poylgon I. Anales del Instituto de Biología, Universidad Nacional Autónoma de México, Serie Botánica 62:133-172.         [ Links ]

39. MACARTHUR RH. 1964. Environmental factors affecting bird species diversity. The American Naturalist 98:387-397.         [ Links ]

40. MACARTHUR RH. 1972. Geographical ecology: patterns in the distribution of species. Harper & Row, New York.         [ Links ]

41. MACIP-RÍOS R and A MUÑOZ-ALONSO. 2008. Diversidad de largartijas en cafetales y bosque primario en el Soconusco chiapaneco. Revista Mexicana de Biodiversidad 79:185-189.         [ Links ]

42. MARQUET PA. 1994. Diversity of small mammals in the Pacific Coastal desert of Peru and Chile and in the adjacent area: Biogeography and community structure. Australian Journal of Zoology 42:527-542.         [ Links ]

43. MARTÍNEZ-MELENDEZ J, MÁ PÉREZ FARRERA and O FARRERA. 2008. Inventario florístico del cerro el Cebú y zonas adyacentes en la Reserva de la Biósfera el Triunfo (Polígono V), Chiapas, México. Boletín de la Sociedad Botánica de México 82:21-40.         [ Links ]

44. MAZEROLLE MJ. 2016. AICcmodavg: Model selection and multimodel inference based on (Q)AIC(c). R package version 2.0-4. http://CRAN.R-project.org/package=AICcmodavg.         [ Links ]

45. MCCAIN CM. 2004. The mid-domain effect applied to elevational gradients: species richness of small mammals in Costa Rica. Journal of Biogeography 31:19-31.         [ Links ]

46. MCCAIN CM. 2005. Elevational gradients in diversity of small mammals. Ecology 86:366-372.         [ Links ]

47. MCCAIN CM. 2007a. Area and mammalian elevational diversity. Ecology 88:76-86.         [ Links ]

48. MCCAIN CM. 2007b. Could temperature and water availability drive elevational species richness patterns? A global case study for bats. Global Ecology & Biogeography 16:1-13.         [ Links ]

49. MCCAIN CM. 2009. Global analysis of bird elevational diversity. Global Ecology & Biogeography 18:346-360.         [ Links ]

50. MCCAIN CM. 2010. Global analysis of reptile elevational diversity. Global Ecology & Biogeography 19:541-553.         [ Links ]

51. MEDELLÍN RA. 1988. Prey of Chrotopterus auritus, with notes on feeding behaviour. Journal of Mammalogy 69:841-844.         [ Links ]

52. MEDELLÍN RA and M EQUIHUA. 1998. Mammal species richness and habitat use in rainforest and abandoned agricultural fields in Chiapas, Mexico. Journal of Applied Ecology 35:13-23.         [ Links ]

53. MENA JL and RA MEDELLÍN. 2010. Small mammal assemblages in a disturbed tropical landscape at Pozuzo, Peru. Mammalian Biology 75:83-91.         [ Links ]

54. MENA JL and E VÁZQUEZ-DOMÍNGUEZ. 2005. Species turnover on elevational gradients in small rodents. Global Ecology & Biogeography 14:539-547.         [ Links ]

55. MORÓN-RÍOS A and MA MORÓN. 2001. La fauna de coleoptora Melolonthidae de la Reserva de la Biósfera “El Triunfo”, Chiapas, México. Acta Zoologica Mexicana (n.s.) 84:1-25.

56. MULUNGU LS, RH MAKUNDI, AW MASSAWE, RS MACHANG’U and NE MBIJE. 2008. Diversity and distribution of rodent and shrew species associated with variations in altitude on Mount Kilimanjaro, Tanzania. Mammalia 72:178-185.

57. NOGUES-BRAVO D, MB ARAUJO, T ROMDAL and C RAHBEK. 2008. Scale effects and human impact on the elevational species richness gradients. Nature 453:216-219.         [ Links ]

58. OVERMARS KP, GHJ DE KONING and A VELDKAMP. 2003. Spatial autocorrelation in multi-scale land use models. Ecological Modelling 164:257-270.         [ Links ]

59. PATTERSON BD, PL MESERVE and BK LANG. 1990. Quantitative habitat associations of small mammals along an elevational transect in temperate rainforests of Chile. Journal of Mammalogy 71:620-633.         [ Links ]

60. PATTERSON BD, V PACHECO and S SOLARI. 1996. Distributions of bats along an elevational gradient in the Andes of south-eastern Peru. Journal of Zoology 240:637-658.         [ Links ]

61. PEARSON OP and CP RALPH. 1978. The diversity and abundance of vertebrates along an altitudinal gradient in Peru. Memorias del Museo de Historia Natural “Javier Prado” 18:1-97.

62. R CORE TEAM. 2016. R: A language and environment for statistical computing. R Foundation for Statistical Computing URL. URL https://www.R-project.org/. Vienna, Austria.         [ Links ]

63. RAHBEK C. 1997. The relationship among area, elevation, and regional species richness in Neotropical birds. The American Naturalist 149:875-902.         [ Links ]

64. RAHBEK C. 2005. The role of spatial scale and the perception of large-scale species-richness patterns. Ecology Letters 8:224-239.         [ Links ]

65. REID FA. 1997. A field guide to the Mammals of Central America and Southeast Mexico. Oxford University Press, New York.         [ Links ]

66. RICKART EA, LR HEANEY, DS BALETE and BR TABARANZA. 2011. Small mammal diversity along an elevational gradient in northern Luzon, Philippines. Mammalian Biology 76:12-21.         [ Links ]

67. RICKART EA, LR HEANEY and RCB UTZURRUM. 1991. Distribution and ecology of small mammals along an elevational transect in Southeastern Luzon, Philippines. Journal of Mammalogy 72:458-469.         [ Links ]

68. ROGERS D, C FUNK, J MILLER and M ENGSTROM. 2007. Molecular Phylogenetic Relationships among crested-tailed mice (Genus Habromys). Journal of Mammalian Evolution 14:37-55.         [ Links ]

69. ROSENZWEIG ML. 1992. Species diversity gradients: we know more and less than we thought. Journal of Mammalogy 73:715-730.         [ Links ]

70. ROWE RJ, LR HEANEY and EA RICKART. 2015. Scale effects on the pattern and predictors of small mammal diversity along a local elevational gradient in the Great Basin. Journal of Biogeography 42:1964-1974.         [ Links ]

71. SÁNCHEZ-CORDERO V. 2001. Elevational gradients of diversity for rodents and bats in Oaxaca, Mexico. Global Ecology & Biogeography 10:63-76.         [ Links ]

72. SERGIO F and P PEDRINI. 2007. Biodiversity gradients in the Alps: the overriding importance of elevation. Biodiversity and Conservation 16:3243-3254.         [ Links ]

73. VÁZQUEZ LB, GN CAMERON and RA MEDELLÍN. 2001. Peromyscus aztecus Mammalian Species 649:1-4.         [ Links ]

74. VOSS RS and LH EMMONS. 1996. Mammalian diversity in Neotropical lowland rainforests: A preliminary assessment. Bulletin of the American Museum of Natural History 230:1-115.         [ Links ]

75. WARTENBERG D. 1989. SAAP 4.3: spatial autocorrelation analysis program. Exeter software Setauket, New York.         [ Links ]

76. WEN Z, Y WU, Y DU, L XIA, D GE, Q YANG and L CHEN. 2014. Seasonal change of species diversity patterns of non-volant small mammals along three subtropical elevational gradients. Biotropica 46:479-488.         [ Links ]

77. WILLIAMS-LINERA G. 1991. Nota sobre la estructura del estrato arbóreo del bosque mesófilo de montaña en los alrededores del campamento “El Triunfo”, Chiapas. Acta Botánica Mexicana 13:1-7.

78. WOODMAN N 2005. Evolution and biogeography of Mexican small-eared shrews of the Cryptotis mexicana-group (Insectivora: Soricidae). Cap. 42:523-534, in: Contribuciones mastozoológicas en homenaje a Bernardo Villa (V Sánchez-Cordero y RA Medellin, eds.         [ Links ]).

79. WU YJ, QS YANG, ZX WEN, L XIA, Q ZHANG and HM ZHOU. 2013. What drives the species richness patterns of non-volant small mammals along a subtropical elevational gradient? Ecography 36:185-196.         [ Links ]

80. ZUUR AF, JM HILBE and EN IENO. 2013. A Beginner’s Guide to GLM and GLMM with R: A Frequentist and Bayesian Perspective for Ecologists. Highland Statistics Newburgh.

81. ZUUR AF, EN IENO, NJ WALKER, AA SAVELIEV and GM SMITH. 2009. Mixed effects models and extensions in ecology with R. Springer.         [ Links ]

Creative Commons License Todo el contenido de esta revista, excepto dónde está identificado, está bajo una Licencia Creative Commons