911 Contrasting “Carrasco” and Forest Ant Communities in the Chapada Diamantina, Bahia, Brazil by Henrique Jesus de Souza¹ & Jacques Hubert Charles Delabie²* AbStrACt This study contrasts the structure of ant communities of the “carrasco” (deciduous forest) and forest (semideciduous forest) remnants in the buf- fer zone of the Chapada Diamantina National Park, bahia, brazil. Our aim was to compare the richness, composition, vertical partitioning and guild organization at the level of leaf-litter, ground-surface and lower vegetation between the habitat types. Ants were sampled at six sites within each habitat by manual extraction in the leaf-litter, pitfall traps at ground level and baited pitfall traps on tree-trunks and shrubs 1.5 m above the soil surface. A total of 132 ant species was collected belonging to 34 genera and seven subfamilies. According to the analyses performed, the habitat types had equivalent species richness and a distinct species and functional composition at the vertical strata level. Furthermore, a greater vertical partitioning was observed in the forest ant community than in the carrasco ant community. The results indicated a distinct biogeographical association between the ant fauna of the two habitat types and suggested that ant communities in carrasco and forest remnants in the buffer zone of the Chapada Diamantina National Park are especially interesting for conservation and ecological research. Key Words: ant communities, “carrasco” vegetation, Chapada Diaman- tina. INtrODUCtION In the central region of the Chapada Diamantina, bahia, brazil, “carrasco” (deciduous forest) and forest (semideciduous forest) remnants occur as isolated fragments surrounded by other vegetation formations, especially those associ- ¹ Programa de Pós-Graduação em Ecologia e biomonitoramento, Instituto de biologia, Universidade Federal da bahia. rua barão de Geremoabo, s/n. Ondina, CEP. 40170-000, Salvador, bahia, brasil. ² Centro de Pesquisas do Cacau, Laboratório de Mirmecologia, Convênio UESC/CEPLAC, CEP. 45600-970, Itabuna, bahia, brasil. *Corresponding author. E-mail: jacques.delabie@gmail.com 912 Sociobiolog y Vol. 59, No. 3, 2012 ated with “Caatinga” (tropical dry forest) and “Cerrado” (tropical savanna) biomes, and by an anthropogenic matrix, mainly dominated by agricultural activities (MMA 2007). They are of high biogeographic and conservation interest because their peculiar physiognomy and floristic composition are so dramatically different from that in other vegetation formations in brazilian semi-arid domains (Funch et al. 2005, Queiroz et al. 2005). Apparently carrasco and forest support a highly distinctive ant community because the habitat structure differs markedly between them, while in the carrasco the aspect of more open vegetation should create an arid microcli- mate for the ant assemblages of different strata, favouring the persistence of tolerant species, similar to those found in warm and open habitats (Andersen 2000). In the forest, the greater vegetation stratification, with well-defined canopy and understory, should soften the microclimate on the ant foraging surface, favouring the existence of an ant community more sensitive to the higher insolation levels of the habitat. Furthermore, considering the fact that ant species in local communities may differ markedly in their preferences with regard to food and nesting and foraging substrates (Yanoviak & Kaspari 2000), it is expected that differences in ecological properties of the vertical strata between the carrasco and forest range constrain the types of ants and guilds that may occur in them. This study contrasts the structure of ant communities of the carrasco and forest remnants in the buffer zone of the Chapada Diamantina National Park, bahia, brazil. Our aim was to compare the richness, composition, vertical partitioning and guild organization at the level of leaf-litter, ground-surface and lower vegetation between the habitat types. The following hypotheses were tested: 1) the species richness of ants in the vertical strata differs between the carrasco and the forest; 2) the vertical strata of the habitat types have a distinct species composition of ants; 3) in the forest, due to higher habitat structural heterogeneity, there is greater vertical partitioning in ant commu- nity; 4) the guild composition of ants in the vertical strata differs between the habitat types. MEtHODS Study area The study was carried out at 12 sites located in Cascavél, county of Ibicoara, in the buffer zone of the Chapada Diamantina National Park, State of bahia, 913 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil brazil (13°10’-13°16’S 41°21’-41°23’W). The sites studied currently comprise the legal reserves of two farms in the region (Caraíbas farm and Londres farm), which are characterized by the occurrence of two vegetation types: carrasco (deciduous forest) and forest (semi-deciduous forest), both at an interface between the Caatinga (tropical dry forest) and the Cerrado (tropical savanna) biomes. The climate of the region is tropical semi-arid, ranging from sub- humid to dry and is marked by two defined seasons: dry (May-September) and rainy (October-April). The mean annual rainfall is approximately 1.062 mm and the average annual temperature is 20°C (MMA 2007). Carrasco is a shrub-dominated vegetation type of the brazilian semi-arid domain, which is derived from a natural process of ecological succession slowly adjusted to the xeric environment, in line with climate changes, possibly from the Pliocene (Fernandes 1998). It can be characterized as a deciduous, closed, and unistratified shrubland intermingled with lianas, with an irregular canopy and sparse emergent trees rarely reaching 6-8 m (Araújo & Martins 1999, Queiroz et al. 2005). The trees Eremanthus capitatus (Spring.) Macleish, Bowdichia virgilioides Kunth, Jacaranda irwinii A.H. Gentry, Himatanthus lancifolius (Müll. Arg.) and Pouteria ramiflora (Mart.) radlk. are characteristic of this vegetation formation in the region. The forest studied is part of a group of remnants of semi-deciduous for- est associated with the brazilian Atlantic Forest biome that occurs in this region, especially at altitudes above 800 m, where the moist and sandy soils favour the occurrence of canopy trees generally larger than in other vegetation formations of the Chapada Diamantina (10-20 m) (Funch et al. 2005). In general, these remnants have diverse floristic composition and most species have wide geographical distribution, occurring in various formations, from northern South America to southern brazil (e.g. Aspidosperma discolor A.DC., Copaifera langsdorffii Desf., Tapirira guianensis Aubl., Schoepfia obliquifolia turcz. and Terminalia brasiliensis Cambess.). Sampling Ants were collected in April 2010, at the end of the rainy season in the region. Within each habitat type, six sites were sampled at least 500 m distant from each other to ensure the independence of the data collected. In addi- tion, 10 sampling points were established for each site at 50 m intervals to 914 Sociobiolog y Vol. 59, No. 3, 2012 facilitate the sampling of individuals from different colonies on the leaf-litter, ground-surface and lower vegetation. For the sampling of the leaf-litter ants, 0.5 m² samples were manually extracted from the soil surface. The collected material was placed in plastic bags for the manual processing of each sample. The screening was performed using sieves of 55 mm mesh attached to white plastic containers to facilitate the detection of organisms captured that were later removed with forceps and brushes (Sarmiento 2003). The non-particulate material retained in the sieves (e.g. leaves, branches and twigs) was transferred to other containers and inspected in detail in order to enhance the effectiveness of the screening. Ground-foraging ants were sampled using pitfall traps installed at ground level. The traps consisted of standard volume and diameter plastic containers partially filled with ethylene glycol to preserve the organisms collected. The containers were covered with plastic dishes supported by toothpicks to prevent flooding by rain and were left in the field for 24 h (Andersen 1990). Arboreal-foraging ants were sampled using baited pitfall traps on tree- trunks and shrubs 1.5 m above the soil surface. Smaller plastic containers were attached inside the pitfall traps containing small quantities of honey and sardine as an ant attractive (bestelmeyer et al. 2000). The larger compart- ment of the containers also was partly filled with ethylene glycol to preserve the specimens collected. The baited pitfall traps operated in parallel with the pitfall traps installed at ground level. The ants were identified to specific level according bolton (1995, 2003). Species that could not be confidently named were identified to morphospe- cific level. Voucher specimens were deposited at the Hymenoptera collection of the Museum of Zoolog y, Federal University of bahia (MZUFbA) and at the Myrmecolog y Laboratory of the Cocoa research Center (CPDC) in brazil. Data analysis The observed species richness in the carrasco and forest was determined by the rarefaction curve (Mao tau) (Colwell et al. 2004), and expected spe- cies richness was estimated using the 1st order Jackknife estimator (Heltshe & Forrester 1983), calculated with Program EstimateS version 7.5.2 (50 randomizations) (Colwell 2006). The effects of habitat type on the mean 915 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil species richness of ants in the vertical strata were examined using the Student t-test in the bioEstat version 5.0 (Ayres et al. 2007). The effects of habitat type on the ant species composition in the vertical strata were examined using Analysis of Similarity (ANOSIM) in PrIMEr version 5.0 (Clarke & Gorley 2001). The Similarity Percentage procedure (SIMPEr) was then used to determine which species were good difference discriminators among the sample groups analyzed (Clarke & Warwick 2001). The presence/absence data of ant species in the vertical strata of each site were submitted to analysis of Non-metric Multidimensional Scaling (nMDS) using the bray-Curtis index as a measure of association to order the ant assemblages of habitat types in a two-dimensional space. The relative abundance of ant species in the vertical strata of each habitat type (leaf-litter: n = 60, ground-surface, n = 60, lower vegetation, n = 60) was measured from the species occurrence frequency in the sample sets. The number of individuals was not used as a measure of abundance because this estimate is greatly influenced by the patterns of nesting and foraging strategies of ant species (bestelmeyer et al. 2000). The vertical partitioning of ant communities was characterized by obtaining the proportion of species that occupied an only one stratum and the propor- tion of species that occupied more than one vertical stratum in the set of sites for the same habitat type. ANOSIM was used to examine the similarity in ant species composition among samples from different vertical strata. In addition, a cluster analysis was carried out using the bray Curtis index as a measure of association to produce a similarity dendrogram. The functional composition of ant communities was compared by assign- ing species to guilds based on the model proposed by brandão et al. (2009). In this model, species that share the same food resources and use the same strategies to occupy their niches are grouped in the same guild based on a series of ecological, morphological and behavioral attributes adopted by the authors in range habitat types of the Neotropics (e.g. body size, diet, foraging activity and substrate nesting ). The adoption of this classification approach has allowed a greater understanding of the structure of ant communities and the factors that determine their organization in habitats of different biogeographic regions (Andersen 1995, brandão et al. 2009, Silvestre et al. 2003). The ant guild organization was characterized by calculating the percent- age amplitude (the proportion of the total number of species) in the vertical 916 Sociobiolog y Vol. 59, No. 3, 2012 strata of each habitat type. The functional similarity was assessed using the Functional Similarity Index, which considers how many guilds are in each location and how many are common between them, the number of species recorded by location and number of shared species within the guilds, by observing the minimum number of species in each guild in each locality (Silvestre et al. 2003). rESULtS A total of 132 ant species belonging to 34 genera and seven subfamilies was collected in the 12 sites sampled, including two exotic species, Tetramorium simillimum and Paratrechina longicornis (Appendix 1). The most species-rich genus was Pheidole with 26 species (19.7% of the total) and Camponotus with 20 (15.2%). In the carrasco, 83 species of 30 genera were observed against 81 species of 25 genera in the forest. Despite the equivalence in the ant species richness between the carrasco and the forest, only 32 species were common Fig. 1. Percentage of ant species for subfamilies in the carrasco and forest in the buffer zone of the Chapada Diamantina National Park, bahia, brazil. Stacked bars show cumulative proportions of total species records from different subfamilies of ants. 917 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil to both habitats. It means that 75.7% of the total species collected in this study occurred in a single habitat type. The carrasco and forest showed little variation in the percentage distribu- tion of ant species by subfamilies. Myrmicinae was the dominant subfamily with approximately 50% of the species collected, while Ecitoninae had the lowest percentage of species, occurring only in the forest (Fig. 1). Despite considerable sampling effort in this study, the species accumulation curves for the two habitat types showed no indication of reaching a plateau, and therefore many more species are likely to be found in them (Fig. 2). The observed and expected ant species richness was similar between the carrasco and the forest. Furthermore, with the exception of lower vegetation, where a statistically significant difference was observed, the mean species richness of Fig. 2. Observed (Mao tau) and expected ( Jack1) richness curves of ant species in the carrasco and forest in the buffer zone of the Chapada Diamantina National Park, bahia, brazil. A total of 180 samples (60 samples of 0.5 m² of leaf-litter, 60 ground pitfall traps and 60 baited pitfall traps) were collected in each habitat type, April 2010. 918 Sociobiolog y Vol. 59, No. 3, 2012 ants in the vertical strata did not differ among the habitat types (leaf-litter: t = -0.72, df = 10, p = 0.49; ground-surface: t = 0.59, df = 10, p = 0.57; lower vegetation: t = 2.51, df = 10, p = 0.03; and all strata combined: t = 0.91, df = 10 , p = 0.38) (Fig. 3 a-d). Multivariate analysis according to the ant presence/absence data showed a clear separation between the carrasco and forest sites. Vertical strata of habitat types showed a distinct species composition of ants, particularly at the lower vegetation level (ANOSIM, leaf-litter: Global r = 0.48, p = 0.002; ground-surface: Global r = 0.41, p = 0.002; lower vegetation: Global r = Fig. 3. Mean (± SD) species richness of ants in the vertical strata (a: leaf-litter, b: ground-surface, c: lower vegetation, d: all strata combined) per site in the carrasco (n = 6, in each case) and forest (n = 6, in each case) in the buffer zone of the Chapada Diamantina National Park, bahia, brazil, April 2010. 919 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil 0.73, p = 0.002; and all strata combined: Global r = 0.56, p = 0.001). The ordinations produced by nMDS showed that ant assemblages sampled in sites of the same habitat type are more similar to each other than those from other habitat types (Fig. 4 a-d). SIMPEr analysis indicated that the ant species that contributed most to the dissimilarity in the assemblage compositions among habitat types were: Odontomachus chelifer, Crematogaster distans and Tetramorium simillimum at the leaf-litter level; Pheidole sp.18 gp. Flavens, Solenopsis sp.3 and Odon- tomachus chelifer at the ground-surface level; and Pheidole sp.18 gp. Flavens, Linepithema pulex and Ectatomma muticum at the lower vegetation level (table 1). Within the two habitats, more than half the species were found in only one stratum (carrasco 67.4% and forest 70.5%), and the number of species exclusive to the ground-surface level was greater than in the lower vegetation Fig. 4. Non-metric multidimensional scaling (nMDS) showing differences in the species composition of ants of the vertical strata sampled (a: leaf-litter, b: ground-surface, c: lower vegetation, d: all strata combined) in the carrasco sites and forest sites in the buffer zone of the Chapada Diamantina National Park, bahia, brazil, April 2010. 920 Sociobiolog y Vol. 59, No. 3, 2012 and leaf-litter, respectively (table 2). ANOSIM based on presence/absence data showed a greater similarity in the species composition between the vertical strata of the carrasco (Global r = 0.38, p = 0.001) than of the forest (Global r = 0.51, p = 0.001). Furthermore, the cluster analysis showed that table 1. Species contributing most to percentage dissimilarity in the species composition of the vertical strata between the carrasco and forest in the buffer zone of the Chapada Diamantina National Park, bahia, brazil. Vertical strata Species Percentage contribution to dissimilarity Leaf-litter Odontomachus chelifer 8.37 Crematogaster distans 4.76 Tetramorium simillimum 4.20 Solenopsis sp.1 4.12 Pheidole sp.8 gp. Fallax 4.12 Ground-surface Pheidole sp.18 gp. Flavens 3.13 Solenopsis sp.3 2.51 Odontomachus chelifer 2.20 Linepithema neotropicum 2.19 Wasmannia auropunctata 2.09 Lower vegetation Pheidole sp.18 gp. Flavens 4.35 Linepithema pulex 4.35 Ectatomma muticum 4.35 Crematogaster victima 3.78 Wasmannia auropunctata 2.98 table 2. Number and percentage of stratum-specialists species of ants in the vertical strata of the carrasco and forest at the buffer zone of the Chapada Diamantina National Park, bahia, brazil. Vertical strata Carrasco Forest Leaf-litter 5 (6.0%) 5 (6.4%) Ground-surface 30 (36.1%) 35 (44.9%) Lower vegetation 21 (25.3%) 15 (19.2%) total 56 (67.4%) 55 (70.5%) 921 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil Fig. 5. Similarity dendrogram (bray Curtis index) showed the similarity relations in the species composition of ants of the vertical strata in the carrasco and forest in the buffer zone of the Chapada Diamantina National Park, bahia, brazil, April 2010. Fig. 6. Percentage of species for ant guilds in the vertical strata of the carrasco and forest in the buffer zone of the Chapada Diamantina National Park, bahia, brazil, April 2010. Stacked bars show cumulative proportions of total species records from different ant guilds: (AA) Army ants, (AFP) Arboreal Feeding on Pollen, (AP) Arboreal Predators, (DA) Dominant Arboreal, (DP) Dacetini Predators, (EGP) Epigaeic Generalist Predators, (FC) Fungus Collectors, (GE) Generalists and (HGP) Hypogaeic Generalist Predators. 922 Sociobiolog y Vol. 59, No. 3, 2012 the ant species composition on the ground-surface and lower vegetation was more similar in the carrasco than in the forest (Fig. 5). On the other hand, the ant species composition in the lower vegetation and leaf-litter was more similar in the forest than in the carrasco (Fig. 5). In both habitat types, the vertical strata varied markedly in guild composi- tion (Fig. 6). The most frequently recorded guilds were Generalists (mostly species of Pheidole, Dorymyrmex and Linepithema) in the leaf-litter and ground-surface, and Dominant Arboreal (mostly species of Camponotus and Crematogaster) on the lower vegetation. Furthermore, the ant community functional organization in terms of vertical strata differed markedly between the carrasco and forest. The Functional Similarity Index for the leaf-litter ant assemblages was 21.4%, while for the ground-surface ant assemblages it was 35.3% and 34.7% for the lower vegetation ant assemblages. DISCUSSION Although the sampling methods used in this study did not provide exhaus- tive information on ant communities in the habitats studied, they were enough to perform comparative analysis. The adoption of complementary collecting methodologies (bestelmeyer et al. 2000, Sarmiento 2003) and planning of a larger sampling period (Delabie et al. 2000) are necessary to obtain a more complete and exhaustive picture of the ant communities from the carrasco and forest remnants in the buffer zone of the Chapada Diamantina National Park, bahia, brazil. The Myrmicinae taxonomic dominance can be attributed to the fact that this subfamily has the greatest generic and specific diversity between subfamilies of ants in the Neotropics (brown 2000). Pheidole, for example, comprises one of the ant genera with the greatest number of species around the world, a phenomenon known as hyperdiversity (Wilson 2003). Some of the more well-founded explanations for the existence of this phenomenon are related to a successful combination of small size, population factors and an appropriate set of adaptations to occupy new niches and exclude competi- tors (Wilson 2003). The occurrence of exotic species in the habitats studied alerts to the danger of biological invasions, which have produced significant impacts on the eco- 923 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil logical integrity of many natural habitats (Mikissa et al. 2008). Tetramorium simillimum and Paratrechina longicornis are species from the Old World that are widely distributed in savannas and forests in brazil (Delabie et al. 2007, Silvestre et al. 2003). Paratrechina longicornis also been recorded in grassland in Mexico (Gove et al. 2009) and in Australian savannas (Andersen et al. 2007). Although these taxa have shown a low relative abundance in the collection sites, it is of fundamental importance to carry out an effective ecological monitoring to assess the expansion of these species in the region. The results of this study showed that ant communities inhabiting the carrasco and forest remnants have equivalent species richness, contradicting the results of the literature from different biogeographic regions, which have suggested that the species richness for these taxa differs markedly between vegetation formations with deciduous and semi-deciduous features (Vasconcelos & Vilhena 2006, Andersen et al. 2007). In spite of this equivalence, the greater number of ant species found in the lower vegetation of the carrasco can be attributed to greater availability of the food supply and nesting sites for ants in this stratum, due to the high density of shrub and emergent trees (ribas et al. 2003). In fact, the carrasco has a dense shrubland capable of supporting a greater number of ant species in the lower vegetation than the forest, where the greater stratification of the vegetation, mainly in the upper strata, usually maintains an arboreal ant fauna more associated with the canopy (brühl et al. 1998, Delabie et al. 2007). Despite the differences in the sampling effort and collecting methods em- ployed, the diversity of ants observed in the forest sites was lower than that found by Martins et al. (2006) in four remnants of semi-deciduous forest in the same region, and by surveys carried out in the Atlantic Forest in southern bahia, where the ant fauna inhabiting especially the leaf-litter is much greater (Alves et al. 2008, Delabie et al. 2007, Delabie & Majer 1999). On the other hand, the ant diversity observed in the carrasco sites was greater than that documented by other authors in open habitats, such as in vegetation types related to the Caatinga biome in the brazilian semi-arid region (Leal 2003, Santos et al. 1999, Soares et al. 2003). In this study, several species presented a strong association with a particular habitat. In fact, the carrasco and forest shared only about one-quarter of the ant species. This result, associated with the discontinuity in species composi- 924 Sociobiolog y Vol. 59, No. 3, 2012 tion observed in the multivariate analysis, indicated a distinct biogeographi- cal association between the ant fauna of the two habitat types. While in the carrasco there was a higher prevalence of open habitat taxa of ants related with the semi-arid zone (e.g. Ectatomma, Dorymyrmex, Pseudomyrmex and Solenopsis), in the forest, a considerable prevalence of forest-associated fauna related with remnants of brazilian Atlantic Forest was observed (e.g. Stru- migenys, Pachycondyla, Labidus and Crematogaster). The greater overlap of species between vertical strata within the carrasco was due to the presence of a group of species with generalist and tolerant habits. For example, Ectatomma brunneum, Odontomachus chelifer, Camponotus cingulatus, Pheidole sp.18 gp. Flavens and Linepithema pulex had favored their distributions probably due to better competitive strategies in obtaining food resources and occuping nesting substrates (brandão et al. 2009, Fernández 2003, Lattke 2003, Silvestre et al. 2003). Thus, the plasticity of foraging and nesting of these species contributed to homogenizing the composition of ant assemblages between the vertical strata. The data obtained by Vasconcelos & Vilhena (2006) and Campos et al. (2008) in savannas reinforce the results of this study that in simplified structure habitats the ant species composition varies less between the vertical strata, especially between the ground-surface and lower vegetation. In contrast to this pattern, in the forest, the ant community was characterized by a greater vertical partitioning, with more than 70% of species associated with only one habitat stratum. This pattern was similar to that reported by brühl et al. (1998) in a primary forest in Malaysia, suggesting that greater vegetation stratification in forests provides a broad spectrum of permanent habitats for ant species allowing a greater vertical partitioning. Analysis performed applying the guild model indicated that the functional composition of ant communities in terms of vertical strata was significantly different between the carrasco and forest remnants in the same region. This result suggested that possible differences in the intrinsic properties of the strata of the habitat types may determine the amplitude and species com- position of ants within the guilds. For example, the Dominant Arboreal ant guild presented high percentage amplitude in the forest lower vegetation that might limit the ant species distribution in this stratum of other guilds. This 925 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil pattern is well documented in tropical forests, where ant colonies with high numbers of individuals can become dominant, because of the potential to raise territory defense (Dejean et al. 2007). On the other hand, in the carrasco lower vegetation, the persistence of other ant guilds typical of vegetation (e.g. Arboreal Predators and Arboreal Feeding on Pollen) together with the expansion of ant guilds typical of the ground-surface (Epigaeic Generalist Predators) in this stratum, seem to decrease the percentage amplitude of the Dominant Arboreal ant guild. Correlation tests between different ant guild traits and ecological properties associated with habitat vertical strata are encouraged in future studies to assess the real contribution of these properties in determining species patterns and functional composition in ant communities. For example, supply and distri- bution of extrafloral nectaries in the habitat and the Dominant Arboreal ant guild amplitude; availability and variety of prey and the Epigaeic Generalist Predator, Specialists and Dacetini Predator ant guild amplitude. Although the lack of information about the insect diversity of the brazilian semi-arid region has not yet attracted the attention of most researchers (bravo & Aguiar 2005), the findings of this study suggest that ant communities in carrasco and forest remnants in the buffer zone of the Chapada Diamantina Na- tional Park have great potential for conservation and ecological research. ACKNOWLEDGMENtS The authors acknowledge the administrators of the Caraíbas and Londres farms for facilities for the field study; thanks to tatiana Cabral, Alexandre Camanho and José rodrigues for help collecting the ant data, to Adrienne Hurley toledo for improving the English manuscript; they also acknowledge their grants from CAPES and CNPq. rEFErENCES Alves, H.S.r., J.H.C. Delabie, I.C. Nascimento, J.r.M. Santos, r.M. Oliveira & M.S. Moreau. 2008. Uso de formigas para identificação de mini-corredores ecológicos na APA Itacaré/ Serra Grande, bahia, brasil. revista Sitientibus Série Ciências biológicas 8:71-79. Andersen, A.N. 1990. The use of ant communities to evaluate change in Australian terrestrial ecosystems: a review and a recipe. Proceedings of the Ecological Society of Australia 16:347–357. 926 Sociobiolog y Vol. 59, No. 3, 2012 Andersen, A.N. 1995. A classification of Australian ant communities based on functional groups which parallel plant life-forms in relation to stress and disturbance. Journal of biogeography 22:15-29. Andersen, A.N. 2000. A global ecolog y of rain forest ants: functional groups in relation to stress and disturbance. In: Agosti, D., J.D. Majer, L.E. Alonso & t.r. Schultz (eds.). Ants: standard methods for measuring and monitoring biodiversity. Smithsonian Institution Press, Washington 25-34. Andersen, A.N., L.t.V. Ingen & r.I. Campos. 2007. Contrasting rainforest and savanna ant faunas in monsoonal northern Australia: a rainforest patch in a tropical savanna landscape. Australian Journal of Zoolog y 55:363–369. Araújo, F.S. & F.r . Martins. 1999. Fisionomia e organização da vegetação do carrasco no planalto da Ibiapaba, estado do Ceará. Acta botanica brasiliensis 13:1-13. Ayres, M., M. Ayres Jr., D.L. Ayres & A.A. Santos. 2007. bioEstat versão 5.0: aplicações estatísticas nas áreas das ciências biológicas e médicas. Sociedade Civil Mamirauá, MCt, belém, Pará, brasil. bestelmeyer, b.t., D. Agosti, L.E. Alonso, C.r .F. brandão, brown, J.H.C. Delabie & r. Silvestre. 2000. Field techniques for the study of ground-dwelling ants: an overview, description, and evaluation. In: Agosti, D., J.D. Majer, L.E. Alonso & t.r. Schultz (eds.). Ants: Standard Methods for Measuring and Monitoring biodiversity. Smithsonian Institution Press, Washington 122-144. bolton, b. 1995. A new general catalogue of the ants of the World. Harvard University Press, Cambridge. 504 pp. bolton, b. 2003. Synopsis and classification of Formicidae. American Entomological Institute, Gainesville. 370 pp. brandão, C.r.F., Silva, r.r. & J.H.C. Delabie. 2009. Formigas (Hymenoptera). In: Panizzi, A.r. & J.r.P. Parra (eds.) bioecologia e nutrição de insetos: base para o manejo integrado de pragas. Embrapa Informação tecnológica, brasília 323-369. bravo, F. & C.M.L. Aguiar. 2006. Inventories and Situation of Insect Diversity in the brazilian Semi-arid. In: Queiroz, L.P., A. rapini, A.M. Giuliett (org.). towards greater knowledge of the brazilian semi-arid biodiversity. Ministério da Ciência e tecnologia, brasília 87-89. brown, J.W.L. 2000. Diversity of ants. In: Agosti, D., J.D. Majer, L.E. Alonso & t.r. Schultz (eds.). Ants: Standard Methods for Measuring and Monitoring biodiversity. Smithsonian Institution Press, Washington 45-79. brühl, A.C., G. Gunsalam & E.K. Linsenmair. 1998. Stratification of ants (Hymenoptera, Formicidae) in a primary rain forest in Sabah, borneo. Journal of tropical Ecolog y 14:285-297. Campos, I.r., C.t. Lopes, W.C.S. Magalhães & HL. Vasconcelos. 2008. Estratificação vertical de formigas em Cerrado strictu sensu no Parque Estadual da Serra de Caldas Novas, Goiás, brasil. Iheringia. Série Zoologia 98:311-316. Clarke, K.r. & r.N. Gorley. 2001. Primer v5 users manual/tutorial. PrIMEr-E., Plymouth, UK, 91 pp. 927 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil Clarke, K.r. & r.M. Warwick. 2001. Change in marine communities: An approach to statistical analysis and Interpretation (2nd edition). PrIMEr-E, Plymouth, UK, 144 pp. Colwell, r.K. 2006. EstimateS: Statistical estimation of species richness and shared species from samples. Versão 8.0. http://purl.oclc.org/estimates. Colwell, r.K., C.X. Mao & J. Chang. 2004. Interpolating, extrapolating, and comparing incidence-based species accumulation curves. Ecolog y 85:2717-2727. Dejean, A., b. Corbara, J. Orivel & M. Leponce. 2007. rainforest canopy ants: the implications of territoriality and predatory behavior. Functional Ecosystems and Communities 1:105-120. Delabie, J.H.C., b.L. Fisher, J.D. Majer & I.W. Wright. 2000. Sampling effort and choice of methods. In: Agosti, D., J.D. Majer, L.E. Alonso & t.r. Schultz (eds.). Standard Methods for Measuring and Monitoring biodiversity. Smithsonian Institution Press. Washington 145-154. Delabie, J.H.C., Jb. Jahyny, I.C. Nascimento, C.S.F. Mariano, S. Lacau, S. Campiolo, S.M. Philpott & M. Leponce. 2007. Contribution of cocoa plantations to the conservation of native ants (Insecta: Hymenoptera: Formicidae) with a special emphasis on the Atlantic Forest fauna of southern bahia, brazil. biodiversity Conservation 16:2359–2384. Fernandes, A. 1998. Fitogeografia brasileira. Multigraf, Fortaleza. 340 pp. Fernández, F. 2003. Subfamilia Myrmicinae. En: Fernández, F. (ed.). Introducción a las hormigas de la región Neotropical. Instituto de Investigación de recursos biológicos Alexander von Humboldt, bogotá 113-148. Funch, L. S., r.r. Funch, r. Harley, A.M. Giulietti, L.P. Queiroz, F. França, E. Melo, C.N. Gonçalves & t. Santos. 2005. Florestas estacionais semideciduais. In: Juncá, F.A.; r.r Funch & W. rocha (orgs.) biodiversidade e conservação da Chapada Diamantina. Ministério do Meio Ambiente, brasília 181-193. Gove, A.D., J.D. Majer & V. rico-Gray. 2009. Ant assemblages in isolated trees are more sensitive to species loss and replacement than their woodland counterparts. basic and Applied Ecolog y 10:187–195. Heltshe, J.F. & N.E. Forrester. 1983. Estimating diversity using quadrat sampling. biometrics 39:1073-1076. Lattke, J.E. Subfamilia Ponerinae. En: Fernandéz, F. (ed.). Introducción a las hormigas de la región Neotropical. Instituto de Investigación de recursos biológicos Alexander von Humboldt, bogotá 261-276. Leal, I.r. 2003. Diversidade de formigas em diferentes unidades da paisagem da Caatinga. In: Leal, I.r., M. tabarelli & J.M. Silva (eds.). Ecologia e conservação da Caatinga. Editora da Universidade Federal de Pernambuco, recife 435-460. Majer, J.D. & J.H.C. Delabie. 1999. Impact of tree isolation on arboreal and ground ant communities in cleared pasture in the Atlantic rain forest region of bahia, brazil. Insectes Sociaux 46:281-290. 928 Sociobiolog y Vol. 59, No. 3, 2012 Martins, L.C.b., J.r. Santos, I.C. Nascimento, N.S. Lopes & J.H.C. Delabie. 2006. Assembléias de Formicidae epigéas no entorno do Parque Nacional da Chapada Diamantina, bahia, brasil. revista Sitientibus Série Ciências biológicas 6:306-316. Mikissa, J.b., J.H.C. Delabie, J.L. Mercier & D. Fresneau. 2008. Preliminary assessment on the interactions of Wasmannia auropunctata in native ant communities (Hymenoptera: Formicidae) of a mosaic gallery forest/savannah in Lope National Park, Gabon. Sociobiolog y 51:207-218. MMA, Ministério do Meio Ambiente. 2007. Plano de manejo para o Parque Nacional da Chapada Diamantina. Ministério do Meio Ambiente, brasília. 248 pp. Queiroz, L.P., F. França, A.M. Giulietti, E. Melo, C.N. Gonçalves, L.S. Funch, r.M. Harley, r.r. Funch & t.S. Silva. 2005. Caatinga. In: Juncá, F.A.; r.r. Funch & W. rocha (orgs.) biodiversidade e conservação da Chapada Diamantina. Ministério do Meio Ambiente, brasília 95-120. ribas, C.r., J.H. Schoereder, M. Pic & S.M. Soares. 2003. tree heterogeneity, resource availability, and larger scale processes regulating arboreal ant species richness. Austral Ecolog y 28:305–314. Santos, G.M.M., J.H.C. Delabie & J.J. resende. 1999. Caracterização da mirmecofauna (Hymenoptera: Formicidae) associada à vegetação periférica de inselbergs (Caatinga – Arbórea – Estacional – Semi-decídua) em Itatim, bahia, brasil. revista Sitientibus Série Ciências biológicas 20:33-43. Sarmiento, C.E. 2003. Metodologías de captura y estúdio de las hormigas. En: Fernandéz, F. (ed.). Introducción a las hormigas de la región Neotropical. Instituto de Investigación de recursos biológicos Alexander von Humboldt, bogotá 201-210. Silvestre, r., C.r.F. brandão & r.r. Silva. 2003. Grupos funcionales de hormigas: el caso de los grêmios del Cerrado. In: Fernandéz, F. (ed.). Introducción a las hormigas de la región Neotropical. Instituto de Investigación de recursos biológicos Alexander von Humboldt, bogotá 113-148. Soares, I.M.F., A.A. Santos, D. Gomes, J.H.C. Delabie & I.F. Castro. 2003. Comunidades de formigas (Hymenoptera: Formicidae) em uma “ilha” de Floresta Ombrófila Serrana em região de caatinga (ba, brasil). Acta biologica Leopoldensia 25: 197-204. Vasconcelos, H.L. & J.M.S. Vilhena, 2006. Species turnover and vertical partitioning of ant assemblages in the brazilian Amazon: A comparison of forests and savannas. biotropica 38:100-106. Wilson, E.O. 2003. La hiperdiversidad como fenómeno real: el caso de Pheidole. En: Fernandéz, F. (ed.). Introducción a las hormigas de la región Neotropical. Instituto de Investigación de recursos biológicos Alexander von Humboldt, bogotá 363-370. Yanoviak, S.P. & M. Kaspari. 2000. Community structure and the habitat templet: ants in the tropical forest canopy and litter. Oikos 89:259–266. 929 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil APPENDIX Appendix 1. Frequency of occurrence of ant species in the vertical strata of the carrasco and forest in the buffer zone of the Chapada Diamantina National Park, bahia, brazil. A total of 180 samples (60 samples of 0.5 m² of leaf-litter, 60 ground-surface pitfall traps and 60 baited pitfall traps in the lower vegetation) were collected in each habitat type. The guild of each species is also given: (AA) Army Ants, (AFP) Arboreal Feeding on Pollen, (AP) Arboreal Predators, (DA) Dominant Arboreal, (DP) Dacetini Predators, (EGP) Epigaeic Generalist Predators, (FC) Fungus Collectors, (GE) Generalists, and (HGP) Hypogaeic Generalist Predators. Species Carrasco Forest Guild Leaf litter Ground surface Lower vegetation Leaf litter Ground surface Lower vegetation Dolichoderinae Azteca sp. DA 0 0 3 0 0 0 Dorymyrmex thoracicus Gallardo, 1916 GE 0 0 0 0 3 0 Dorymyrmex sp.1 GE 0 0 0 0 2 0 Dorymyrmex sp.2 GE 0 0 0 0 1 0 Dorymyrmex sp.3 GE 0 1 0 0 0 0 Dorymyrmex sp.4 GE 0 1 1 0 0 0 Dorymyrmex sp.5 GE 0 0 0 0 1 0 Linepithema aztecoides Wild, 2007 GE 0 7 1 0 0 0 Linepithema cerradense Wild, 2007 GE 0 1 1 0 0 0 Linepithema pulex Wild, 2007 GE 0 15 0 0 33 2 Linepithema neotropicum Wild, 2007 GE 0 0 0 0 4 0 Ecitoninae Labidus praedator (Fr. Smith, 1858) AA 0 0 0 0 2 0 Labidus coecus (Latreille, 1802) AA 0 0 0 0 2 0 Ectatomminae Ectatomma brunneum Smith, 1858 EGP 1 19 13 8 1 0 930 Sociobiolog y Vol. 59, No. 3, 2012 Ectatomma muticum Mayr, 1870 EGP 0 36 7 9 0 0 Ectatomma opaciventris roger, 1861 EGP 0 18 1 0 0 0 Ectatomma tuberculatum (Olivier, 1792) AP 0 0 1 0 0 0 Gnamptogenys moelleri (Forel, 1912) EGP 0 0 0 1 0 0 Gnamptogenys regularis Mayr, 1870 EGP 0 1 0 0 0 0 Gnamptogenys sulcata (Fr. Smith, 1858) EGP 0 0 1 0 0 0 Formicinae Brachymyrmex heeri Forel, 1874 GE 3 0 0 0 0 0 Brachymyrmex patagonicus Mayr, 1868 GE 0 0 0 0 0 1 Brachymyrmex sp.1 GE 0 0 0 0 0 0 Brachymyrmex sp.2 GE 0 0 0 0 1 0 Brachymyrmex sp.3 GE 0 0 0 1 1 0 Camponotus agra (F. Smith, 1858) DA 0 0 0 0 0 3 Camponotus arboreus (F. Smith, 1858) DA 0 0 1 0 0 1 Camponotus blandus (F. Smith, 1858) DA 0 1 0 0 0 0 Camponotus cingulatus Mayr, 1862 DA 0 3 15 0 0 1 Camponotus crassus Mayr, 1862 DA 0 0 3 2 2 3 Camponotus fastigatus roger, 1863 DA 0 2 6 0 0 0 Camponotus leydigi Forel, 1886 DA 0 1 3 0 1 0 Camponotus melanoticus Emery, 1894 DA 0 2 7 0 0 0 Camponotus novogranadensis Mayr, 1870 DA 0 6 2 0 0 0 Camponotus punctulatus andigenus Mayr, 1868 DA 0 1 0 0 0 0 931 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil Camponotus renggeri Emery, 1894 DA 0 2 8 0 2 10 Camponotus rufipes (Fabricius, 1775) DA 0 0 2 0 0 0 Camponotus (Myrmaphaenus) sp.1 DA 0 0 1 0 0 0 Camponotus (Tanaemyrmex) sp.2 DA 0 0 3 0 0 1 Camponotus (Myrmaphaenus) sp.3 DA 0 0 1 0 0 0 Camponotus (Myrnothrix) sp.4 DA 0 0 0 1 2 1 Camponotus sp.5 DA 1 0 0 0 0 0 Camponotus sp.6 DA 0 0 0 0 0 1 Camponotus sp.7 DA 0 0 0 0 0 3 Camponotus sp.8 DA 0 0 0 0 0 3 Myrmelachista sp.1 GE 0 0 1 0 0 0 Myrmelachista sp.2 GE 0 0 0 0 8 1 Myrmelachista sp.3 GE 0 0 0 1 0 1 Nylanderia fulva (Mayr, 1862) GE 0 2 0 0 0 0 Nylanderia guatemalensis (Forel, 1885) GE 0 0 0 0 1 0 Nylanderia sp.1 GE 0 0 0 0 2 0 Nylanderia sp.2 GE 0 0 0 0 1 0 Nylanderia sp.3 GE 0 0 0 0 2 0 Paratrechina longicornis Latreille, 1802 GE 2 0 0 0 0 0 Myrmicinae Acromyrmex balzani Emery, 1890 FC 0 1 0 0 0 0 Acromyrmex subterraneus molestans Santschi, 1925 FC 0 2 0 0 0 0 Atta sexdens (Linnaeus, 1758) FC 0 1 0 0 0 0 Cephalotes atratus (Linnaeus, 1758) AFP 0 0 0 0 0 4 Cephalotes minutus (Fabricius, 1804) AFP 0 0 1 0 0 0 932 Sociobiolog y Vol. 59, No. 3, 2012 Cephalotes pallidus Andrade, 1999 AFP 0 0 0 0 1 0 Cephalotes pavonii (Latreille, 1809) AFP 0 0 4 0 0 0 Cephalotes pusillus (Klug, 1824) AFP 1 3 22 0 2 3 Cephalotes sp. AFP 0 0 0 1 0 0 Crematogaster curvispinosa Mayr, 1862 AD 1 0 1 1 0 1 Crematogaster distans Mayr, 1870 DA 4 2 2 4 0 1 Crematogaster victima Smith, 1958 DA 1 1 9 0 0 0 Crematogaster sp.1 DA 0 0 0 0 0 3 Crematogaster sp.2 DA 0 0 0 1 0 0 Crematogaster sp.3 DA 0 0 0 2 0 0 Cyphomyrmex salvine Forel, 1899 FC 2 0 0 0 0 0 Cyphomyrmex sp.1 gp. Strigatus FC 0 3 0 0 0 0 Cyphomyrmex sp.2 FC 0 0 0 0 3 0 Hylomyrma balzani (Emery, 1894) FC 0 0 0 1 3 0 Mycetophylax sp. FC 0 0 0 0 1 0 Nesomyrmex sp. GE 0 0 1 0 0 0 Oxyepoecus sp.1 GE 0 2 0 0 0 0 Oxyepoecus sp.2 GE 0 3 0 0 0 0 Pheidole sp.1 gp. Diligens GE 0 1 0 0 1 0 Pheidole sp.2 gp. Diligens GE 0 0 0 0 0 1 Pheidole sp.3 gp. Diligens GE 0 8 0 0 13 1 Pheidole sp.4 gp. Diligens GE 0 18 3 0 9 11 Pheidole sp.5 gp. Diligens GE 0 4 10 0 3 0 Pheidole sp.6 gp. Diligens GE 1 1 0 0 0 0 Pheidole sp.7 gp. Flavens GE 0 7 0 1 2 0 Pheidole sp.8 gp. Fallax GE 0 11 0 2 6 1 Pheidole sp.9 gp. Fallax GE 0 1 0 0 0 0 Pheidole sp.10 gp. Diligens GE 0 0 1 0 0 0 933 de Souza, H.J. & J.H.C. Delabie — Contrasting Ant Communities in brazil Pheidole sp.11 gp. Diligens GE 0 0 0 0 1 0 Pheidole sp.12 gp. tristis GE 0 6 0 0 0 0 Pheidole sp.13 gp. Fallax GE 0 0 0 0 1 0 Pheidole sp.14 gp. Diligens GE 0 1 0 0 0 1 Pheidole sp.15 gp. Diligens GE 0 2 0 0 0 0 Pheidole sp.16 gp. Fallax GE 0 1 0 0 2 0 Pheidole sp.17 gp. Diligens GE 0 5 0 0 1 0 Pheidole sp.18 gp. Flavens GE 1 17 2 1 2 0 Pheidole sp.19 gp. tristis GE 0 4 1 1 9 0 Pheidole sp.20 gp. Flavens GE 0 2 0 0 5 0 Pheidole sp.21 GE 0 0 0 0 1 0 Pheidole sp.22 GE 0 0 0 0 0 1 Pheidole sp.23 gp. Diligens GE 0 0 0 0 3 0 Pheidole sp.24 gp. Diligens GE 0 0 0 1 4 0 Pheidole sp.25 gp. Diligens GE 0 0 0 0 1 0 Pheidole sp.26 GE 0 0 0 0 1 0 Pogonomyrmex naegelii (Forel, 1878) GE 0 1 0 0 2 0 Procryptocerus sp. DA 0 0 0 0 0 1 Sericomyrmex sp. FC 0 2 0 0 0 0 Solenopsis globularia (Smith, 1858) GE 0 1 0 0 0 0 Solenopsis sp.1 GE 0 2 1 3 2 9 Solenopsis sp.2 GE 0 0 0 3 0 1 Solenopsis sp.3 GE 0 0 0 0 5 1 Solenopsis sp.4 GE 0 0 0 0 1 0 Strumigenys appretiatus (borgmeier, 1954) DP 0 1 0 0 0 0 Strumigenys denticulata (Mayr, 1887) DP 0 0 0 0 1 0 934 Sociobiolog y Vol. 59, No. 3, 2012 Tetramorium simillimum (F. Smith, 1851) GE 3 0 0 2 0 0 Trachymyrmex fuscus Emery, 1834 FC 0 10 0 0 0 0 Wasmannia auropunctata roger, 1863 GE 3 7 2 0 2 0 Wasmannia rochai Forel, 1912 GE 2 4 0 4 1 0 Ponerinae Anochetus oriens Kempf, 1964 EGP 0 1 0 0 0 0 Hypoponera sp.1 HGP 2 0 0 0 0 0 Hypoponera sp.2 HGP 0 0 0 1 0 0 Hypoponera sp.3 HGP 0 0 0 0 1 0 Odontomachus chelifer (Latreille, 1802) EGP 0 4 2 4 39 0 Odontomachus meinerti Forel, 1905 EGP 3 1 0 0 0 0 Pachycondyla bucki (borgmeier, 1927) EGP 0 0 0 0 2 0 Pachycondyla harpax (Fabricius, 1804) EGP 0 3 0 0 1 0 Pseudomyrmecinae Pseudomyrmex gracilis (Fabricius, 1804) AP 0 0 7 0 0 0 Pseudomyrmex oculatus (F. Smith, 1855) AP 0 0 9 0 0 0 Pseudomyrmex schuppi (Forel, 1901) AP 0 0 1 0 0 0 Pseudomyrmex tenuis (Fabricius, 1804) AP 0 3 0 0 0 0 Pseudomyrmex termitarius (F. Smith, 1855) AP 0 0 0 1 1 0 Pseudomyrmex sp.1 gp. Pallidus AP 0 0 1 0 0 0 Pseudomyrmex sp.2 gp. Pallidus AP 0 0 2 0 0 1