The study was conducted in the Sierra de San Carlos (Figure 1), which includes the municipalities of Burgos, Cruillas, Jiménez, San Carlos, San Nicolás, and Villagran located in the central-west portion of the state of Tamaulipas, and also the municipality of Linares in the extreme eastern part of the state of Nuevo León, Mexico (Arriaga et al. 2000). The Sierra de San Carlos comprises an area of 2320 km², being a polygon with the northwestern limit at 24°52.000'N, 99°12.067'W, and the southeastern limit at 24°23.050'N, 98°32.667'W. It constitutes an isolated mountain range within the Tamaulipas biogeographic province, bounded at the south by the Mexican gulf province (Morrone et al. 2002). Almost all vegetation types within the Sierra have a high conservation status, and they occur in principally temperate ecosystems (oak and pine forests) in the mountain areas, but there are also various types of tropical scrub vegetation in the lower areas, as well as other vegetational communities.

Figure 1. 

Study area. A Location of Tamaulipas in Mexico B Location of Sierra de San Carlos within Tamaulipas C Study area (red square) within Sierra de San Carlos D Details of study area: 1 = Cerro El Diente, 2 = Ejido Carricitos y Tinajas, 3 = San Nicolás.

One of the most important characteristics of Sierra de San Carlos is its designation as a Priority Conservation Area in Mexico due to its biological, ecological and physiogeographical features: it is the northern limit of the Cloud Forest vegetation in Mexico, and it has some endemic plant species; also, it is considered as a biogeographical island (“sky island”) due to its isolation from other nearby mountain ranges, such as the Sierra Madre Oriental and Sierra de Tamaulipas (Arriaga et al. 2000). Accordingly, some areas with median, high and extreme conservation priority are located within Sierra de San Carlos (CONABIO 2007).

We selected seven sampling sites distributed in three localities within Sierra de San Carlos (Figure 1D), including various elevations and vegetation types, as well as various conservation priorities. The first locality was Cerro El Diente which included four sampling sites: Site 1) containing submountain scrub vegetation, at a mean elevation of 550 masl; Site 2) consisting of Tamaulipan thorny scrub vegetation, at a mean elevation of 760 masl; Site 3) had oak forest vegetation, at a mean elevation of 960 masl; and Site 4) with Cloud Forest vegetation, at a mean elevation of 1080 masl. The second locality was Ejido Carricitos y Tinajas which had two sampling sites. These were: Site 5) with secondary elements of riparian vegetation, at a mean elevation of 730 masl; and Site 6) containing oak and pine forest vegetation, at a mean elevation of 820 masl. The third locality was San Nicolás with a single sampling site: Site 7) with submountain and Tamaulipan thorny scrub vegetation, at a mean elevation of 500 masl. Both the Cerro El Diente and Ejido Carricitos y Tinajas are located within areas with extreme conservation priority; the locality of San Nicolás belongs to a median conservation category (CONABIO 2007).

Twelve sampling plots of 400 m² each (20 × 20 m) were established within each of the seven sampling sites. Plot dimensions were determined with the species-area curve method, using the nested quadrat type (Scheiner 2003). The number of plots sampled was established by Clench analysis, with 70% as a minimum limit of completeness (for a detailed analysis, see Jiménez-Valverde and Hortal 2003). The plots were previously and randomly located within each of these sites, using GIS software. This was accomplished by: 1) creating a polygon shape for the selected area of the sampling site, and 2) delimiting a square graticule inside this polygon using the Repeating Shapes tool, with the sides of the squares 20 meters in length; both procedures were made using ArcView GIS (1992–1999). 3) The graticule was exported to IDRISI Selva (1987-2012) and converted to a .kml format archive; 4) the new archive was displayed in Google Earth Pro software, permitting the visualization of the graticule of real dimension plots displayed over actual satellite imagery; and finally, 5) this graticule was treated as a coordinate system (rows and columns), and the location of each one of the 12 plots was selected by the random numbers tool in Microsoft Office Excel. These methods were employed for the seven sites, and the only difference was the dimensions of the polygon shape of the sample area which were adjusted to the vegetational communities. Plots were georeferenced and later located in the field. When a plot was located in areas not accessible or impossible to sample (i.e. steep hills, areas of bare soil), it was moved to the closest location where vegetational cover was available. Detailed geographical data (latitude and longitude coordinates, elevation) for each plot within each site are presented in Tables 1–3; detailed spatial arrangement of sampling plots within each site is presented in Figures 2–5.

Figure 2. 

Detailed position of sampling plots in Sierra de San Carlos. Cerro El Diente locality: A Site 1 B Site 2. Dotted lines shows elevation curves.

Figure 3. 

Detailed position of sampling plots in Sierra de San Carlos. Cerro El Diente locality: A Site 3 B Site 4. Dotted lines shows elevation curves.

Figure 4. 

Detailed position of sampling plots in Sierra de San Carlos. Ejido Carricitos y Tinajas locality: A Site 5 B Site 6. Dotted lines shows elevation curves.

Figure 5. 

Detailed position of sampling plots in Sierra de San Carlos. San Nicolás locality, Site 7. Dotted lines shows elevation curves.

Systematic sampling was conducted between 10:00 and 17:00 h, using a standard entomological sweep net of 40 cm diameter. Individual samples consisted of 120-200 sweeps of the shrub and herbaceous vegetation in each plot. Contents of the net were emptied into a plastic bag, adding 60% ethanol and an indelible label with corresponding data. Each plot (12) within the seven sites was sampled monthly, from February 2013 to January 2014, comprising 1,008 total samples at the end of the study; sweeping was conducted by the same person during the whole study to reduce sampling error. Each sample and the specimens obtained were processed according to the method described by Sánchez-Reyes et al. (2014). Also, leaf beetles encountered independent of the standardized sweeps were added to the species checklist. Specimens are deposited in the collection of the Facultad de Ingeniería y Ciencias at the Universidad Autónoma de Tamaulipas, Ciudad Victoria, Tamaulipas, Mexico.

Identification of specimens was made using available literature on Chrysomelidae (Wilcox 1965, White 1968, Wilcox 1972, Scherer 1983, White 1993, Flowers 1996, Riley et al. 2002, Staines 2002). Additionally, material was compared with identified specimens deposited in the collection of Chrysomelidae of the Facultad de Ingeniería y Ciencias, Universidad Autónoma de Tamaulipas. However, those specimens that could not be identified to the species level were compared with other unidentified specimens and carefully grouped into morphospecies, and so the designation of “species” in this study includes both morphospecies and determined species. Taxonomical arrangement follows the categories employed by Riley et al. (2003), except that the subfamily Bruchinae is not included in this study.

We obtained environmental data from two meteorological stations located in the municipalities of San Carlos and San Nicolás in the study area. Historical data of total monthly rainfall and monthly average temperature (only the average from 1951-2010 was available) were plotted to visually analyze the fluctuation of these parameters. On this basis, four seasons were defined: Early dry season (EDS: November, December, January), Late dry season (LDS: February, March, April), Early rainy season (ERS: May, June, July), and Late rainy season (LRS: August, September, October). Data of precipitation and temperature were correlated with species richness and abundance, using a Spearman correlation analysis in STATISTICA 8.0 (StatSoft Inc. 2007).

All the following analyses were made only with the data obtained through systematic sampling (i.e., by sweeping in 12 plots at each site). Species collected otherwise were excluded from the analysis, but are included in the checklist of species.

As a measure of species richness, we used the total number of species present throughout the Sierra de San Carlos, and at each site and season. Significant differences in the number of species were assessed through permutation tests in PAST 3.07 (Hammer et al. 2001). Estimated species richness at each level was measured by means of the nonparametric estimators ACE, Chao 1, and Jackknife 1 (Magurran 2004, Hortal et al. 2006, Gotelli and Colwell 2011), and was calculated with the software EstimateS 8.2 (Colwell 2013), using 100 randomizations without replacement and the abundance data of each species found per sampling unit (plot). Also, we used the Clench model to determine the sampling efficiency through the estimated species richness and the slope of the species accumulation curve, which measures the inventory quality; calculations were made according to the method described by Jiménez-Valverde and Hortal (2003), and were performed in STATISTICA 8.0 (StatSoft Inc. 2007).

Species from Sierra de San Carlos were divided into five categories, according to their total abundance: 1) very common (more than 70 individuals); 2) common (11 to 70 specimens); 3) rare (3 to 10 specimens); 4) doubletons (two specimens); and 5) singletons (one specimen only). These categories were used because they have been implemented in similar studies with Chrysomelidae, also in the state of Tamaulipas (Sánchez-Reyes et al. 2014). Also, abundance was measured at each site and season, and differences between these values were analyzed with Kruskal-Wallis and Mann-Whitney nonparametric tests, as data did not meet the assumptions of normality. We used the Simpson dominance index (D) and the Shannon entropy index (H’) to measure both principal components of alpha diversity (Magurran 2004). Values obtained were transformed to a true diversity value according to Jost (2006, 2007), and were calculated for the whole study area and for each site and season. Differences in species composition between each pairwise comparison of sites and months were assessed through PERMANOVA analysis, using the Bray-Curtis index as distance measure, with 9999 random permutations. Beta diversity was measured as the faunistic similarity between sites and months, using the Bray-Curtis index of similarity; also, a Cluster analysis was performed to define groups of sites and months according to species composition, using the Bray-Curtis index as a distance measure and the Ward method as an amalgamation algorithm. All calculations were made in PAST 3.07 (Hammer et al. 2001) and STATISTICA 8.0 (StatSoft Inc. 2007).

Finally, the association of abundance and species richness obtained at each elevational site during each month was measured with a Correspondence analysis. This is a multivariate technique based on contingency tables and count data, where the significant statistical dependence between rows (sites) and columns (months) is tested by a chi-square test (Beh 2004). The analysis was conducted in STATISTICA 8.0.

In total, 3,081 specimens, belonging to 109 species, 63 genera and six subfamilies, were obtained. An additional four species were obtained by independent collecting in various sites from study area, resulting in a total species richness of 113 species and 65 genera within the study period (2013-2014) in the Sierra de San Carlos (Appendix 1, Figures 6–7). According to ACE, Chao 1, and Jackknife indexes, the estimated species richness was between 115.88 and 128.98 species; also, the estimation using the Clench Model was 120 species. These values indicate that the proportion of observed species richness in relation to the richness estimates was 84.5 to 94% (Table 4). Slope value of the Clench Model suggests that faunistic inventory in Sierra de San Carlos was complete and reliable (slope = 0.012). The Dominance value for the Simpson index (D) was 0.066, with a diversity value (1/D) of 15.064. For the Shannon index (H´), the value was 3.41, with a diversity value (e^H´) of 30.29.

Figure 6. 

Examples of leaf beetle biodiversity from Sierra de San Carlos, Mexico. A Lema balteata LeConte, 1884 B Lema opulenta Harold, 1874 C Helocassis clavata (Fabricius, 1798) D Plagiodera semivittata Stål, 1860 E Miraces aeneipennis Jacoby, 1888 F Malacorhinus acaciae (Schaeffer, 1906)G Cyclotrypema furcata (Olivier, 1808) H Acrocyum dorsalis Jacoby, 1885 I Colaspis melancholica Jacoby, 1881 J Griburius montezuma (Suffrian, 1852) K Cryptocephalus trizonatus Suffrian, 1858 L Coscinoptera tamaulipasi Medvedev, 2012 M Diplacaspis prosternalis (Schaeffer, 1906).

Figure 7. 

Examples of leaf beetle biodiversity from Sierra de San Carlos, Mexico. A Trichaltica scabricula (Crotch, 1873), new country record. In decreasing order from B to I, the most abundant species in the current study. B Syphrea sp. 2 C Diachus sp. 1 D Xanthonia sp. 1 E Centralaphthona diversa (Baly, 1877) F Chrysogramma sp. 1 G Pachybrachis sp. 1 H Sumitrosis inaequalis (Weber, 1801) I Margaridisa atriventris (Melsheimer, 1847).

The most abundant subfamily in the study area was Galerucinae (including Alticini), with 53.6% (1,652 specimens) of the total abundance of Chrysomelidae in the Sierra de San Carlos, followed by Cryptocephalinae with 26.9% (828 specimens). Less abundance was found in Eumolpinae (12.6%, 388 specimens), Cassidinae (4.9%, 152 specimens), Chrysomelinae (1.1%, 33 specimens) and Criocerinae (0.9%, 28 specimens). Species counts per subfamily were greatest for Galerucinae, with 54 species (49.5%) representing half of the total species richness recorded in Sierra de San Carlos. Lower values were recorded for Cryptocephalinae with 27 species (24.8%), Eumolpinae with nine species (8.3%), Cassidinae and Criocerinae both with eight species (7.3%), and Chrysomelinae with only three species (2.8%).

Eight species were categorized as “very abundant species,” each with over 70 specimens that accounted for 60.3% (1,859 total specimens) of the total abundance obtained from Sierra de San Carlos. Of these eight species, Syphrea sp. 2 (475 specimens) and Diachus sp. 1 (418) were the most abundant, followed by Xanthonia sp. 1 (276), Centralaphthona diversa (Baly, 1877) (244), Chrysogramma sp. 1 (193), Pachybrachis sp. 1 (103), Sumitrosis inaequalis (Weber, 1801) (78), and Margaridisa atriventris(Melsheimer, 1847) (72) (Figure 7B–7I). The proportion of 31.6% (974 specimens) of the total abundance was accounted for by 34 species categorized as “common,” while 29 species were considered as “rare” (6.13% of the total abundance). Only 21 species were doubletons, and 17 were singletons. Trichaltica scabricula (Crotch, 1873) is recorded for the first time in Mexico (Figure 7A).

No clear patterns of species richness were found with elevation. The greatest number of species (50) was recorded at the site of highest elevation (1080 masl), but this value was not significantly different from values observed at 960 (47 species) and 550 masl (38 species). The smallest number of species (27) was registered at a high elevation site (820 masl); however, it was not significantly different from sites at 730 (36 species) and 550 masl (Table 2). Completeness exceeded 70% in all sites, with a maximum value of 97% at Site 2 (760 masl), although lower values were obtained at both sites from the Carricitos y Tinajas locality (730 and 820 masl). The slope of the Clench model was less than 0.1 for each of the sites (Table 4).

Regarding seasonal analysis, the species richness increased progressively and significantly from early dry season (40 species) to early rainy season (78 species). The value decreased to 76 species during late rainy season, although this change was not significant. Seasonal values of estimated species richness through nonparametric indexes and the Clench model followed the same pattern as that of sites, because all values were above 70% of completeness, with slopes under 0.1 for all the seasons (Table 4).

We found significant variations in abundance and diversity of Chrysomelidae between sites of differing elevation (Kruskal Wallis Test, H=100.7, p<0.0001). However, these parameters did not show a specific trend with the increase or decrease in elevation. For example, the highest abundance (665 individuals) was present at the lowest site (500 masl), whereas the lowest value (173 individuals) was obtained at the second lowest elevation (550 masl). Also, differences in abundance obtained between the site of lowest elevation (665 individuals) and site at 960 masl (561 individuals) were not significant, while the number of specimens (440) at the highest site was not statistically different from values observed at lower sites (960, 820 and 760 masl). The lowest abundances were present at 550 (173 specimens) and 730 masl (231 specimens), and these values were significantly different from other elevational sites (Tables 5, 6).

As observed with abundance, no clear patterns of diversity were found with elevation, although all sites were significantly different from each other. The highest dominance (D=0.406) and lowest entropy (H’=1.716) were obtained in the site of lowest elevation (500 m), indicating that the lowest diversity was found at this site (1/D=2.46; e^H’=5.56). However, the second highest value of diversity in the study area (1/D=12.66; and e^H’=21.47) was obtained at the second lowest elevation site (550 meters). Diversity decreased progressively from the third (730 meters) to the fifth elevational site (820 m), and then increased from 960 masl to the highest elevation site (1080 masl), where the lowest dominance (D=0.0508) and highest values of entropy (H’=3.328) and diversity (1/D=19.66; e^H’=27.88) were obtained (Table 6).

According to the PERMANOVA analysis, the leaf beetle composition between sites was statistically different (SS_total=33.16; SS_within-group=23.22; F=5.492, p=0.0001), and almost all pairwise comparisons were significantly different, except for Site 2 (760 masl) and Site 3 (960 masl) (F=1.595, p=0.1151) (Table 7). The Bray-Curtis index also showed that Site 2 and Site 3 were the most similar in the study area, with 59% faunistic similarity. Values below 50% were obtained for all other comparisons, with the lowest percentages associated with Site 7 when compared to all others (Table 8). Cluster analysis suggested that three groups were generated on the basis of the difference in leaf beetle composition between elevation sites: Site 5 and Site 6 (Group 1); Site 2, Site 3 and Site 4 (Group 2); Site 1 and Site 7 (Group 3) (Figure 8).

Figure 8. 

Cluster analysis of chrysomelid composition by elevational site in Sierra de San Carlos, Mexico. Delimitation of groups is indicated by red dotted line.

Abundance and diversity values showed significant variation between seasons (H=92.29, p<0.0001). The lowest number of specimens (433) was recorded at early dry season, and abundance increased significantly at the end of the season (888 specimens). Abundance decreased at early rainy season (690 specimens), although this reduction was not significant; then, the value increased significantly at the late rainy season, which recorded the highest abundance (1,070 specimens) in this study (Tables 9, 10). Similarly, the highest dominance (D=0.146) and the lowest values of entropy (H’=2.736) and diversity (1/D=6.84; H’=15.42) were obtained in the early dry season. These values changed significantly at late dry season (1/D=9.90; H’=17.01) and also during the early rainy season, when the highest diversity was obtained (1/D =18.68; H’=33.38) (Table 10). Monthly temperature and precipitation data (Figure 9) showed a significant positive correlation (p<0.05) with the number of species recorded by month in Sierra de San Carlos, whereas abundance was not correlated with environmental parameters (Table 11).

Figure 9. 

Historical monthly data of precipitation and temperature within Sierra de San Carlos, Mexico.

Seasonal species composition and faunistic similarity were analyzed using monthly values. Significant differences in species composition between months were found (SS_total=33.16; SS_within-group=27.33; F=1.395, p=0.0009). Pairwise comparisons showed that differences were obtained principally between early (January, February, March, April and May) and later months (July, September, October, November and December) (Table 12). The highest similarity occurred between September-October (89.4%), January-February (72.1%), November-December (69.8%), April-May (69%), and February-March (63.6%), whereas other comparisons were less than or close to 50% (Table 13). Three clusters were formed according to monthly leaf beetle composition: January to June (Group 1), July to October (Group 2), and November-December (Group 3) (Figure 10).

Figure 10. 

Cluster analysis of chrysomelid composition by month in Sierra de San Carlos, Mexico. Delimitation of groups is indicated by red dotted line.

The Correspondence analysis showed no significant associations in the number of species obtained by site for each month (Total Inertia=0.08796, Chi²=74.147, df=66, p=0.23014). However, the association of abundance between sites and months was significant (Total Inertia=0.32237, Chi²=993.24, df=66, p=0.0000). The most clear associations were observed between abundance obtained at lower elevations (500 and 550 masl) and the period comprised by September to December, while the number of specimens in Site 3 (960 masl) were highly related to August. January and February were principally associated with Site 5 (730 masl) and Site 2 (760 masl). The March-May period was associated with Site 6 (820 masl), and the abundance found at the highest elevation site (1080 masl) was predominantly related with June (Figure 11).

Figure 11. 

Correspondence analysis of chrysomelid abundance obtained per month at each elevational site in Sierra de San Carlos, Mexico.

The 113 species of Chrysomelidae recorded in this study document that the Sierra de San Carlos represents a proportion close to 50% of the total leaf beetle species richness presently reported from Tamaulipas (Niño-Maldonado et al. 2014a) and 5% of the total reported Mexican leaf beetle fauna (Ordóñez-Reséndiz et al. 2014). Also, Trichaltica scabricula represents a new species record for the chrysomelid fauna of Mexico, since this species was previously known only from numerous states in the United States, including Texas (Riley et al. 2003). Additionally, the high proportion of individuals identified as unnamed morphospecies in this and other studies suggests that the actual species richness of Chrysomelidae from Tamaulipas is still greater. The number of species recorded here from Sierra de San Carlos is lower when compared to other similar studies conducted with similar methods in the region. The most related study was done in Peregrina Canyon, approximately 100 km to the southwest, near Ciudad Victoria also in the state of Tamaulipas, from which 240 total samples, 2,228 specimens and 157 species were obtained (Sánchez-Reyes et al. 2014). In Sierra de San Carlos, 1,008 samples and 3,200 specimens were obtained, but only 109 species were recorded. Still, even with the low number of species obtained, Galerucinae was the most dominant subfamily, which is consistent with the observed patterns in various other studies (Bouzan et al. 2015), including those conducted in other areas of Tamaulipas and in other states in Mexico (Niño-Maldonado et al. 2005, Ordóñez-Reséndiz and López-Pérez 2009, Sánchez-Reyes et al. 2014).

Although the sampled area for both studies was almost the same (33,600 m² in Sierra de San Carlos vs. 37,500 m² in Peregrina Canyon), the total and site-season inventory completeness in Peregrina Canyon was close to 70% (Sánchez-Reyes et al. 2014), suggesting a higher proportion of species to be added to that inventory (Jiménez-Valverde and Hortal 2003), while values for this study were close to or above 90%. Moreover, seven different vegetational communities were sampled in this study, while only three sites were studied in Peregrina Canyon. These data indicate that the number of chrysomelid species present in Sierra de San Carlos was considerably less than the sites within the Sierra Madre Oriental. Because leaf beetles are dependent of their associated vegetational communities, we expected to find a higher species richness at our study area. Our data suggest that the insular nature of the geographical location of Sierra de San Carlos (Arriaga et al. 2000) results in lower species richness of Chrysomelidae when compared with habitats connected to the Sierra Madre Oriental, as is the Peregrina Canyon. Also, the lesser richness of plants in Sierra de San Carlos (441 compared to 676 species, Martínez 1998, Briones-Villarreal 1991) is probably directly correlated with the smaller number of leaf beetles, when compared with Peregrina Canyon in Altas Cumbres Natural Protected Area, where at least 1,164 species of vascular plants have been documented (García-Morales et al. 2014). Contrary to species richness, the higher diversity values obtained in Sierra de San Carlos possibly reflect a lower degree of anthropogenic disturbance and a higher quality of natural resources and vegetational communities (Arriaga et al. 2000). These factors may favor a more balanced ecological process, leading to higher evenness in abundance of species and thus the higher diversity values (Magurran 2004).

Elevation is one of the most important factors driving ecological communities, because the abiotic factors and biotic variables together modify species richness and composition of assemblages. Recent evidence suggests that the most common elevational pattern is the increase of diversity and species richness at intermediate elevations (Rahbek 2005, McCain and Grytnes 2010, Sanders and Rahbek 2012, Guo et al. 2013), which has been documented for various groups of Coleoptera (Escobar et al. 2005, Fernández et al. 2010), including Chrysomelidae (Furth 2009, Sánchez-Reyes et al. 2014). Other studies have shown a decrease of species richness with increasing elevation for various groups of insects (Wolda 1987, McCoy 1990, Sánchez-Ramos et al. 1993, Araújo and Fernandes 2003, Maveety et al. 2011, Jones et al. 2012). Indeed, it has been observed that species richness, abundance and diversity of Lepidoptera decreases with increasing elevation in Cerro El Diente (Meléndez-Jaramillo et al. 2015), which is one of the sampling localities in our study. However, the leaf beetle communities here analyzed did not show any consistent correlation with the elevation, such as those observed in other studies of Chrysomelidae along elevational gradients (Furth 2009, Sánchez-Reyes et al. 2014, Bouzan et al. 2015). It has been determined that small elevational ranges, even with more intensive sampling (as was done in Sierra de San Carlos), could exhibit non-unimodal patterns, due to a limited geographical range over which such patterns could be expressed (Guo et al. 2013). For example, the locality of Cerro El Diente includes an elevation range from 400 to 1200 masl, but in our study only four sites were sampled at this area. It is possible that a more stepped sampling design might result in consistent patterns of species richness, abundance and diversity, as observed in other studies conducted exclusively at this locality (Meléndez-Jaramillo et al. 2015, Sánchez-Reyes et al. 2015a). However, we suggest that the insular nature of Sierra de San Carlos, its geomorphic properties, and the lack of connectivity between elevational sampling sites in different localities, could be the main drivers of results here obtained, because these factors have been recognized as key determinants of biodiversity patterns within elevational gradients (Bertuzzo et al. 2016).

Although a consistent elevational pattern was not found, the high proportion of inventory completeness through all methods employed indicates that the faunistic composition obtained at each site is representative; so, the values of abundance and diversity were reliable (Jiménez-Valverde and Hortal 2003). On this basis, we affirm that 1) highest values of diversity from Sierra de San Carlos were present at the highest and lowest elevations at the Cerro El Diente locality (Site 4, Cloud forest at 1080 masl; and Site 1, Submountain scrub at 550 masl), and also 2) the highest species richness was recorded in the highest elevation site from the same locality (Site 4, 1080 masl). This could be due to the vegetational composition and the characteristics at that area, since the contrasting more humid areas in the Cloud forest in the highest elevation site must be favoring the higher values obtained of diversity and species richness. Besides, we found higher values of species richness and diversity at sites within Cerro El Diente, compared with other sites from the Ejido Carricitos y Tinajas and San Nicolás localities. Land area is a determining factor in shaping communities within elevation gradients (Körner 2007). However, this factor might not affect species richness and diversity in our study, as fewer species were collected from more extensive sampling areas, such as Site 7 and both sites from the Ejido Carricitos y Tinajas locality; contrarily, intermediate vegetational communities from Cerro El Diente, even with sampling plots distributed randomly within corresponding elevational intervals, covered a smaller area (Figures 2, 3) but presented higher values of species richness and diversity. Considering that chrysomelid communities are directly influenced by plant-associated variables (Bach 1981, Řehounek 2002, Aslan and Ayvaz 2009, Sen and Gök 2009), we attribute these results to the higher quality of the vegetational communities at the Cerro El Diente locality (Arriaga et al. 2000, CONABIO 2007). These findings highlight the significance of this locality within Sierra de San Carlos, and constitute support for its designation as an area with extreme priority for conservation. Also, the presence of Cloud forest in Sierra de San Carlos and the associated communities of leaf beetles are remarkable, since this ecosystem has a relict distribution in Mexico and is considered as an important center for high levels of biodiversity and endemism (Gual-Díaz and Rendón-Correa 2014); thus, our results contribute to the categorization of Sierra de San Carlos as a sky island and emphasize its protection urgency.

When analyzing species composition and beta diversity between sites, we observed that Sites 2 and 3 were the only sites with the same composition and a high faunistic similarity, while almost all other comparisons were different, which is contrary to other findings that show a high similarity of Lepidoptera between elevation and seasons at Cerro El Diente (Meléndez-Jaramillo et al. 2015). It has been determined that habitat heterogeneity and diverse characteristics of conservation areas can promote high beta diversity or low similarity between sites, regardless of the distance (Linzmeier and Ribeiro-Costa 2009). This was observed with the lowest site of Cerro El Diente and the other three sites within the same locality, as they formed different faunistic groups, although they are geographically close to each other. Moreover, we found that the lowest elevational sites from Sierra de San Carlos (Site 1 and Site 7) formed a faunistic group even when these were from distant localities. Undoubtedly, this is due to similar vegetational communities, which lead to similar leaf beetle faunas. Conversely, both sites were very different in terms of abundance of species, thus reflecting the specific responses of each species to abiotic and biotic characteristics at each site. This evidence suggests that differences in microhabitats result in very different assemblages of species, owing to an almost entirely different plant composition and also to the climatic or abiotic variation (Ødegaard 2006).

Regarding seasonal analysis, patterns observed at Sierra de San Carlos were different from those recorded in the most related study in Peregrina Canyon, Tamaulipas (Sánchez-Reyes et al. 2014), as our highest values of abundance, species richness and diversity were recorded in the rainy season. These findings are supported by other studies, because the dominance and increased abundance of adult leaf beetles in rainy or wet seasons is the most common result in studies of seasonal variation of this taxon and other insects (Ødegaard 2006, Furth 2009, Bouzan et al. 2015), which is highly related to the increase in plant density during this period. Besides, considering that higher seasonal peaks of abundance are associated with more marked seasons (Wolda et al. 1998), we suggest that the more seasonally dry environmental conditions at Sierra de San Carlos could be a primary factor in the species of Chrysomelidae being absent from samples or less active in the dry season, due to reduction in quality and availability of host plants, and to abiotic conditions (Medeiros and Vasconcellos-Neto 1994, Awmack and Leather 2002, Ishihara and Ohgushi 2006). Furthermore, evidence has shown that phytophagous insects locate temporary refugia when environmental conditions are less suitable (Janzen 1973), thus not being sampled in sweep catches and resulting in low values of abundance, species richness and diversity during dry seasons in non-refugia sites.

According to cluster analysis, three groups were formed, based on faunistic similarity between months: November and December, January to June, and July to October. This inconsistency between the four climate seasons (dry/rainy) and the clustering of months by species compositions is possibly due to differential species responses to seasonal variations, as their temporal niche requirements are very distinct (Linzmeier and Ribeiro-Costa 2013). Similarly, seasonal variations, monthly composition and similarity patterns in leaf beetles reflect different ecological and biological strategies of each species, since reproduction and adult feeding phases may be different at certain times of the year. In Chrysomelidae, this has been attributed to abiotic variation during seasons, such the photoperiod, temperature, relative humidity, and to the quality and availability of host plants (Linzmeier and Ribeiro-Costa 2008, 2013). Also, seasonal and monthly patterns could reflect an existing sequence between Chrysomelidae and other phytophagous species to avoid competition and allow optimum resource exploitation (Pérez-Barroeta and Gurrea-Sanz 1994).

In addition to the influence of the geographical location of Sierra de San Carlos, the responses of the chrysomelid community in this study, being elevational, seasonal or both, are suggested to be driven by host plants and vegetational associated variables (Bach 1981, Erelli et al. 1998, Řehounek 2002, Aslan and Ayvaz 2009, Sen and Gök 2009). So, although not investigated here, the study of relationships between Chrysomelidae and their associated vegetational communities is very important to understand the distribution patterns of this taxon. However, it has been noted that the interplay between the biotic and abiotic environment shapes consumer diversity along elevational gradients (Sundqvist et al. 2013). Therefore, changes in abundance, species richness and diversity within each locality, site and season/month could be driven by other factors related to elevation change, such as climate variables or abiotic environment (temperature, humidity, Körner 2007), which could be influencing the activity patterns of Chrysomelidae and the capacity of each species to obtain resources (Flinte et al. 2011, Bouzan et al. 2015). This influence of climate has been demonstrated in this study, since precipitation and temperature were significantly correlated with species richness; other studies on Chrysomelidae have reported similar effects (Linzmeier and Ribeiro-Costa 2008, Sánchez-Reyes et al. 2014).

Besides, the results obtained through Correspondence analysis confirm the same tendency of an interaction between abiotic and biotic factors on distribution of Chrysomelidae, because they show that abundance of the leaf beetle community at each site is associated with specific months. This was observed, for example, with the significant association of the lowest sites and abundance obtained at the rainy months, or that observed between Site 3 (960 masl) and August. Consequently, these associations suggest unique and specific temporal-site conditions for the communities of Chrysomelidae, which surely are the result of monthly changes in both environmental (abiotic) conditions and plant variables along the elevational gradient. Since elevational and temporal responses of biological communities arise from the effects (direct or indirect) of these gradients on each species (Hodkinson 2005, Sundqvist et al. 2013, Guo et al. 2013), future research on Chrysomelidae in Mexico needs to be targeted at the specific species patterns and their relationships with environmental variation, as well as at specific interactions between leaf beetle species and other ecosystem components.