UNIVERSIDADE FEDERAL DE SÃO CARLOS CENTRO DE CIÊNCIAS BIOLÓGICAS E SAÚDE DEPARTAMENTO DE BOTÂNICA PRISCILLA DE PAULA LOIOLA DISTRIBUIÇÃO ESPACIAL DE ESPÉCIES ARBÓREAS DE CERRADO: FILOGENIA E TRAÇOS DE DEFESA CONTRA HERBIVORIA SÃO CARLOS 2010 UNIVERSIDADE FEDERAL DE SÃO CARLOS CENTRO DE CIÊNCIAS BIOLÓGICAS E SAÚDE DEPARTAMENTO DE BOTÂNICA PRISCILLA DE PAULA LOIOLA DISTRIBUIÇÃO ESPACIAL DE ESPÉCIES ARBÓREAS DE CERRADO: FILOGENIA E TRAÇOS DE DEFESA CONTRA HERBIVORIA Dissertação apresentada ao Programa de Pós Graduação de Ecologia e Recursos Naturais para obtenção do título de mestre em Ecologia e Recursos Naturais Orientação: Prof. Dr. Marco Antônio Batalha SÃO CARLOS 2010 Ficha catalográfica elaborada pelo DePT da Biblioteca Comunitária da UFSCar L834de Loiola, Priscilla de Paula. Distribuição espacial de espécies arbóreas de cerrado : filogenia e traços de defesa contra herbivoria / Priscilla de Paula Loiola. -- São Carlos : UFSCar, 2010. 27 f. Dissertação (Mestrado) -- Universidade Federal de São Carlos, 2010. 1. Ecologia de comunidades. 2. Cerrado. 3. Traços funcionais. 4. Herbivoria. 5. Filogenia. 6. Distribuição espacial. I. Título. CDD: 574.5247 (20a) Dedico esse trabalho à minha mãe, Lourdes, pelo seu amor, paciência, apoio e esperança incondicionais. AGRADECIMENTOS - Ao Prof. Dr. Marco Antônio Batalha, pela orientação e amizade, sempre presentes. - Ao Danilo, pelos nossos nove meses de coleta, pela amizade e ótima compania. - Ao Vinícius, pela inestimável amizade, pelo companheirismo em todas as situações, além da grande ajuda no trabalho. - Ao Igor, por ter sido meu tutor científico por muitos anos, além de um grande amigo. - Ao Marcão, por nunca se cansar das nossas discussões científicas, pela paciência e amizade - Ao Gustavo, pela amizade e presteza em ajudar. - À professora Maria Inês e à Maristela, pelo gentil auxílio com as análises químicas. - À FAPESP, por conceder a bolsa e o financiamento para a realização do trabalho. - À toda equipe do laboratório, pela compania e discussões científicas imprescindíveis para minha formação. - Ao Felipe, por dividir comigo o amor, a cumplicidade e a vida, além da incrível paciência durante os períodos difíceis. - À minha família, que me deu suporte para que eu tomasse decisões, confiança, dedicação e muito amor. RESUMO A herbivoria deve promover um padrão de distribuição de traços de defesa em espécies de plantas coocorrentes, já que plantas similares são mais vulneráveis a herbívoros especialistas. O conservantismo dos traços é geralmente observado em linhagens de espécies de plantas, e assim, as relações filogenéticas entre as espécies também devem estar relacionadas à distribuição espacial. Dessa forma, nós esperamos que os traços de defesa em espécies coocorrentes sejam mais diferentes do que o esperado. Esperamos ainda que os traços contra a herbivoria estejam conservados, e que as distâncias filogenéticas de espécies coocorrentes sejam maiores do que o esperado pelo acaso. Em um fragmento de cerrado do estado de São Paulo, nós analisamos 100 parcelas, com 25 m 2 cada, e amostramos todos os indivíduos arbóreos. Para cada espécie, nós medimos alguns traços de defesa contra herbivoria, e respondemos se o grau de coocorrência das espécies está relacionado com as diferenças funcionais e com as distâncias filogenéticas. Nós também testamos se os traços de defesa são filogeneticamente conservados. Nós não encontramos correlação entre a coocorrência das espécies com os traços de defesa, nem com as distâncias filogenéticas. No entanto, nós encontramos sinal filogenético para quarto dos nove traços estudados. A ausência de correlações pode ser devido a (1) herbivoria não ser um processo forte na vegetação de cerrado como esperado ou (2) a presença de filtros ambientais, como a falta de água e solos pobres, que devem promover agrupamento filogenético, agindo em conjunto com a dispersão filogenética gerada pela herbivoria. Palavras-chave: Cerrado. Distribuição espacial. Filogenia. Herbivoria. Traços de defesa ABSTRACT Herbivory is expected to promote an overdispersed distribution of traits in co-occurring plant species, since similar plants are more vulnerable to specialised herbivores. As long as conservatism of traits is usually observed in lineages of plant species, phylogenetic relatedness may also be related to spatial distribution. Thus, we expected that defense traits against herbivory were conserved and that phylogenetic distances of co-occurring species were higher than expected by chance. In a cerrado site in southeastern Brazil, we analysed 100 quadrats, with 25 m 2 each, and sampled all woody individuals. For each species, we measured defense traits against herbivory and answered whether the degree of co-occurrence of species was correlated with both functional differences and phylogenetic distances. We also tested whether the defense traits were phylogenetically conserved. On the one hand, we did not find significant correlation between species co-occurrence and neither defense traits nor phylogenetic distances. On the other hand, we found phylogenetic signal for four out of nine defense traits. The absence of correlations may be due to (1) herbivory not being as strong as we expected in cerrado vegetation or (2) the presence of environmental filters, such as drought and nutrient-poor soil, promoting phylogenetic clustering, counteracting phylogenetic overdispersion by herbivory. Key-words: Cerrado. Defense traits. Herbivory. Phylogeny. Spatial distribution. LISTA DE TABELAS TABELA 1. Correlation between the co-occurrence of species in a southern cerrado site and the defense traits against herbivory………………………………………………………… 25 TABELA 2. Phylogenetic signal to defense traits, measured from the variance of the phylogenetic independent contrast in tree species of a southern cerrado site……………… 24 LISTA DE FIGURAS FIGURA 1. Phylogenetic tree of species sampled in a southern cerrado site. The relationship among species was based on Davies et al. (2004)................................................................. 27 Sumário 1 Introduction 10 2 Methods 12 2.1 Study site 12 2.2 Defense trait data 12 2.3 Phylogenetic data 14 3 Results 15 4 Discussion 16 Acknowledgements 18 References 19 Tables 25 Figure 27 10 1. INTRODUCTION The coevolution of herbivores and plants has been proposed as a major factor promoting the diversity of defense traits against herbivory in plants (Becerra 1997, 2007). Related insects often feed on plants that share common traits to which they are adapted (Berembaum 1983; Becerra 1997). Consequently, plant-herbivore interactions may contribute to the interactions among plants (Carson & Root 2000; Fine et al. 2006), changing their abundance and spatial distribution (Olff & Ritchie 1998; Maron & Crone 2006). Studies on plant-herbivore coevolution and its impact on plant defenses have focused primarily on interactions that involve a small number of species or populations (Agrawal & Fishbein 2006; Becerra 2007). Here we focused on the question of whether coevolution creates spatial patterns that structure plant community at fine scale. Herbivory could limit the coexistence of plants that share common defense traits, since selection would favor trait divergence in co-occurring similar plants (Becerra 2007; Kursar et al. 2009). The effect of the herbivory on diversification should be stronger for narrowly coevolved systems involving few interacting species, which tend to develop specific adaptations to the features of their counterparts (Van Zandt & Agrawal 2004; Becerra 2007). Consequently, a plant similar to its neighbour should be more vulnerable to a specialised herbivore able to handle many of its defense traits (Becerra 2007; Kursar et al. 2009). Community spatial structuring is more likely to be observed at fine spatial scales, within a uniform habitat, than at large spatial scales, which encompass different habitats (Cavender- Bares et al. 2004, 2006). High specialisation in defense strategies and high differences in the trait values of co-occurring species (that is, trait overdispersion) occur at smaller scales, probably due to stronger selective pressure for trait divergence in the locally present plant species (Becerra 2007). Thus, as coevolutionary specialisation increases, and spatial scale decreases, co-occurring species tend to be more dissimilar (Becerra 2007). Therefore, we expected to find overdispersion of defense traits against herbivory at fine spatial scale. Strategies against herbivory might include nutritional quality, such as proteins and anti- proteins; physical characteristics, such as spines, trichomes, and leaf toughness; toxicity, such as alkaloids; phenology; regrowth capacity; and indirect defenses, like volatiles and branching architecture (Agrawal & Fishbein 2006). Synergistic interactions among multiple traits may potentially provide a greater level of defense than would be possible if the traits were present independently (Berenbaum et al. 1991; Stapley 1998). Therefore, traits acting together may 11 provide a more efficient defense strategy to the individual, and the herbivore pressure may act in the syndrome of traits to divergence in space. Recently, we are becoming increasingly aware that the evolutionary processes, particularly the way that traits evolve within lineages, influence community assembling (Webb et al. 2002; Chazdon et al. 2003; Cavander-Bares et al. 2004). Evolution of plant defenses against herbivory may be central to understand tropical biodiversity (Kursar et al. 2009). In general, there is a great conservatism of functional traits in plant lineages (Ackerly 2003; Reich et al. 2003), including chemical and morphological characteristics that constrain the use by herbivores (Ward & Spalding 1993; Futuyma & Mitter 1996). As a consequence, closely related plants generally present closely related herbivores and pathogens (Becerra 2007; Gilbert & Webb 2007). As herbivores present a selective pressure for trait divergence on co- occurring plants (Becerra 2007; Kursar et al. 2009), we expected to find co-occurring species that are also phylogenetically distant. The cerrado is characterised by marked rainfall seasonality and experiences a pronounced dry season (Gottsberger & Silberbauer-Gottsberger 2006). The nutrient poor, well drained, acid soils are additional environmental constraints for plant growth in this vegetation types (Gottsberger & Silberbauer-Gottsberger 2006). In communities where resources are not abundant, such as the cerrado (Gottsberger & Silberbauer-Gottsberger 2006), plants tend to concentrate their investments in defenses against herbivory (Coley & Barone 1996; Fine et al. 2006), since they cannot replace damaged tissues as fast as in communities with abundant resources (Janzen 1974; Coley et al. 1985). So, herbivory is expected to be strong in cerrado plant communities, selecting species with a great investment in defense traits and determining their spatial 13 distribution. Additionally, the cerrado vegetation seems to present a high specificity between herbivores and host plants (Marquis et al. 2002; Dyer et al. 2007). For example, Diniz and Morais (1997) found that faunal similarities among host plant genera, and even within genus, were low in cerrado vegetation. In communities with such a specialisation between herbivores and host plants, the effects of herbivory on the diversity of defense traits tend to be high, and the co- occurring species are expected to present different defense traits (Coley & Barone 1996; Becerra 2007). This divergence is expected to occur at fine spatial scales, since insect herbivores play an important divergence pressure, especially in distances smaller than 10 meters (Coley & Barone 1996; Becerra 2007). Thus, we expected to find an overdispersion among defense traits of co-occurring cerrado plant species at fine spatial scale. In a nutshell, we addressed whether the levels of species co-occurrences were correlated 12 with defense traits against herbivory and with phylogenetic distances. We answered three questions: (1) are species that share similar traits against herbivory spatially more dispersed than expected by chance?; (2) are defense traits conserved on the phylogeny of the sampled plant species?; and (3) are closely related species spatially more dispersed than expected by chance? 2. METHODS 2.1 Study site We surveyed a woodland cerrado site (21°58’05.3”S, 47°52‘10.1”W), in São Carlos, São Paulo State, southeastern Brazil. The soil is a dystrophic Oxisoil, and the mean altitude is 850 m (Santos et al. 1999). Regional climate is mesothermic, subtropical, with rainy summer and not severely dry winter (Cwa; Köppen 1931). According to data collected on the meteorological station of São Carlos, from 1969 to 1998, the annual mean temperature is 21.3ºC and the monthly precipitation is 131.3 mm. 2.2 Defense trait data In the rainy season of 2008, we placed a grid with contiguous 100 quadrats (each one with 25 m 2 ), in which we sampled all woody individuals with stem diameter at soil level equal to or larger than 3 cm (SMA 1997). We identified the sampled species with an identification key (Batalha & Mantovani 1999) and comparing the collected material to vouchers lodged at Federal University of São Carlos and Brazilian Institute of Geography and Statistics herbaria. For species with more than ten individuals inside the quadrats, we randomly picked ten individuals for trait measurement (Cornelissen et al. 2003). For the species with less than ten individuals inside the quadrats, we made an extra effort, looking for other individuals nearby the quadrats and trying to reach 10 individuals per species. 13 We collected expanded leaves, without symptoms of herbivore and pathogen attack, and measured the following defense traits: specific leaf area, water content, trichomes, latex content, toughness, chemical defenses, and nutritional quality (Agrawal & Fishbein 2006). Specific leaf area is positively related to mass-based maximum photosynthetic rate, or its potential relative growth (Cornelissen et al. 2003). Lower values of specific leaf area tend to correspond with relatively high investments in leaf defenses, particularly structural ones (Cornelissen et al. 2003), indicating fast growth and low palatability (Agrawal & Fishbein 2006). Low values of specific leaf area are correlated to low herbivory rate (Neves et al. 2010). Water content is also related to palatability, and a leaf with low water content might resist herbivory (Agrawal & Fishbein 2006). We collected two leaves per individual to assess specific leaf area and water content. We reserved the leaves in a fresh recipient and weighted them still fresh. We digitalised the leaves to determine leaf area with ImageJ 1.33 software (Rasband 2004). After that, we oven-dried each leaf sample at 80°C for 72 h, and then weighted the dry mass to obtain the specific leaf area (Cornelissen et al. 2003). We sampled another leaf of the individuals and collected latex, an important physical defense against herbivory (Agrawal & Fishbein 2006; Agrawal & Konno 2009). We measured latex from 10 replicates from each species by cutting the tip of an intact leaf in the field and collecting the exuding latex onto a 1 cm disc filter paper. When latex stopped flowing and was absorbed on the filter paper, we covered it with another dry filter paper. We oven-dried the discs at 75°C for 24 h, and weighted them (Agrawal & Fishbein 2006). We measured leaf toughness with a force gauge penetrometer (dinamometer DFE 010, Chatillon). We used the conical tip to penetrate the surface in each side of the mid-rib. For statistical analyses, we used the mean of these two measures. Leaf toughness is related to nutritional and defense components (Agrawal & Fishbein 2006) and is considered the main defense trait against herbivore activity (Coley & Barone 1996). Trichomes are also important physical defenses against herbivory. We used five replicates for each species and counted trichomes on upper and lower sides of 28 mm² leaf discs with a dissecting microscope (Agrawal & Fishbein 2006). We determined presence of chemical compounds on leaves following procedures described by Falkenberg et al. (2003). We determined presence of alkaloids, terpenoids, and tannins – chemicals frequently found in Brazilian plants that could work as defense against herbivores (Lima 2000). We used a series of three essays, Mayer, Dragendorff, and Wagner reactions to determine presence of alkaloids, considering positive the samples that reacted to at least two essays (Falkenberg et al. 2003). We used Liebermann-Burchard and Salkowisk reactions to 14 test the presence of terpenoids and a ferric chloride reaction to determine presence of tannins (Falkenberg et al. 2003). Nutritional quality may also influence the attack to vegetal tissue, since higher C:N ratios might decrease the nitrogen acquisition by the herbivore (Agrawal & Fishbein 2006), which may lead to decreased herbivore attack due to the difficulty to assess the nitrogen (Coley & Barone 1996; Agrawal & Fishbein 2006). Total leaf carbon (C) and nitrogen (N) concentrations were measured from five replicates from each species to assess the C:N ratio. This measure is considered as an indicator of plant nutritional quality (Agrawal & Fishbein 2006). The C and N concentrations samples were determined by the Laborarory of Stable Isotopes of the University of São Paulo. Our defense traits are related to chewing insects, since, in tropical plant communities, folivorous insects are the most important consumers, and chewing insects contribute with 75% or more of the annual leaf consumption (Coley & Barone 1996). We used the mean of each trait to calculate the pairwise Euclidean distances of species, standardising the data to zero mean and unit variance, with the Vegan package (Oksanen et al. 2009) in R environment (R Development Core Team 2009). We calculated the pairwise co- occurrence (C) based on proportional similarity (Schoener 1970) as: Cih= 1 -0.5 ∑│pij -phj│, where Cih is the co-occurrence of species i and h, pij is the proportion of occurrences of the ith or hth species in the jth quadrat. With this index, we obtained the co-occurrence degree of species in the space: the closer to 1 the C value, the more co-occurring the pair (Schoener 1970). As the 100 quadrats were close to each other, we used 50 randomly picked quadrats to calculate the co-occurrence index. So, we decreased the possible bias due to the proximity of the individuals within two quadrats. We obtained the co-occurrence matrix, using the Picante package (Kembel et al. 2009) in R environment (R Development Core Team 2009). Then, we compared the correlation coefficient between C values and the Euclidean distances of all species pairs to a null model generated from 1,000 randomisations of the trait matrix (Mantel test; Manly 2000), also using the Vegan package (Oksanen et al. 2009). 2.3 Phylogenetic data We examined the spatial phylogenetic structure of co-occurring woody species in cerrado by comparing the degree of co-occurrence of species pairs in the quadrats to the phylogenetic 15 distance between them. We initially constructed a phylogenetic tree using Phylomatic software, a phylogenetic database and a toolkit to assembly phylogenetic trees (Webb & Donoghue 2005). The generated tree was based on references of molecular phylogenies (APG III 2009). We assigned branch lengths to the phylogenetic tree using the Branch Length Adjustment averaging algorithm of the Phylocom 4.1 software (Webb et al. 2009). The branch length was based on minimum ages of nodes determined for genera, families, and higher clades according to Davies et al. (2004), by spacing undated nodes evenly between nodes in the trees. Finally, we compared the correlation coefficient between C values and phylogenetic distances of all species pairs to a null model, in which the phylogenetic relationships among species were randomised (Mantel test; Manly 2000). We investigated whether the functional traits were phylogenetically conserved or convergent in the phylogeny of local tree assemblage, using the analysis of traits module implemented in Phylocom software (Webb et al. 2009). In this method, we analysed the variance of the phylogenetically independent contrasts (PICs) to test which traits presented phylogenetic signal. Phylogenetic signal may be defined as the tendency of closely related species to resemble each other (Webb et al. 2009). If the evolution of some trait is conserved in the phylogeny, the divergences will be small and related species will be similar to each other (Webb et al. 2009). To test the significance of the phylogenetic signal, we compared the observed variance of the PICs with pseudo-variances generated by randomisation of the trait values in the phylogenetic tree (Blomberg et al. 2003). We also used the Picante package (Kembel et al. 2009) in R environment (R Development Core Team 2009). 3. RESULTS We sampled 2,062 individuals, comprising 61 species and 29 families (Fig. 1, Supporting information). We obtained trait defense data to specific leaf area, water content, trichomes, latex, toughness, alkaloids, terpenoids, tannins, and nutritional quality (Supporting information). We did not find significant correlation between the co-occurrence of species and neither defense traits against herbivory (P > 0.05, Table 1) nor phylogenetic distances (R = 0.10; P = 0.09). However, we found significant phylogenetic signal for specific leaf area, water content, latex, and nutritional quality (Table 2). 16 4. DISCUSSION We investigated the influence of defense traits against herbivory and phylogeny on patterns of co-occurrence of tree species in cerrado. We expected to find co-occurring defense traits more different than expected by chance, due to the decrease of conspecific individuals caused by herbivory at fine spatial scale (Coley & Barone 1996; Becerra 2007). However, our results indicated that trait similarity did not limit species co-occurrence at fine spatial scales of woody cerrado plant species. Evidence of correlation between the defense traits and species co-occurrence would indicate a determinant role of herbivory in the cerrado. So, the herbivory in cerrado may not be strong enough to determine the spatial distribution of the woody species at fine spatial scale. In addition, environmental filters and herbivory may act in opposite directions, generating random co-occurrences relative to functional differences and phylogenetic distances. Thus, the herbivory may not decrease the functional similarity and the phylogenetic relatedness of the co-occurring plant species in cerrado. The strength of the herbivory in plants is associated mainly to nutritional quality of the soil that may result in leaves with low value for herbivores (Fine et al. 2006). More nutrients available in the soils are related to species less defended against herbivory (Neves et al. 2010). In cerrado, the soils present low nutritional quality, especially low availability of nitrogen and phosphorus, and might present low herbivory rate in relation to other vegetation types (Gottsberger & Silberbauer-Gottsberger 2006; Neves et al. 2010). For example, leaves of forest trees, where soils were richer in nutrients and organic matter, suffered consumption rates three times higher than those recorded for cerrado leaves (Neves et al. 2010). Although leafcutter ants remove about 13-17 percent of the annual leaf production in savannas (Costa et al. 2008), quantification of the impact of herbivores in savannas are very rare (Marquis et al. 2002; Costa et al. 2008). Therefore, herbivory may not be strong enough to determine the spatial distribution of the woody species. The several environmental filters in cerrado, such as drought, fire, and nutrient-poor soils (Gottsberger & Silberbauer-Gottsberger 2006), also affect the co-occurrence of plant species at fine spatial scale (Silva & Batalha 2009; Silva & Batalha 2010). Environmental filters tend to select co-occurring species with similar niches, that is, species with similar functional traits, because only those species with similar tolerances to the environmental constraints may survive (Chase 2003). In this case, one would find two opposing forces: herbivory leading to 17 trait overdispersion and environmental filters leading to trait clustering. As selection, by definition, is the relationship between a trait and fitness (Strauss & Irvin 2004), there would be a trade-off between overcoming herbivory and overcoming water and nutrient stress. For instance, in African savannas, bottom-up soil nutritional factors and top-down herbivory have both been suggested to control distribution patterns (Stock et al. 2010). In our study, most of functional traits may also be selected by environmental filters, such as specific leaf area, water content, toughness, and nutritional quality (Cornelissen et al. 2003; Pais & Varanda 2003; Neves et al. 2010). Thus, environmental filters and herbivory may act in opposite directions, with similar strength (Cavender-Bares et al. 2006), creating random co- occurrences relative to functional differences and phylogenetic distances. The phylogenetic pattern of co-occurring species depends on the evolutionary history of species traits, the interaction between herbivores and plants, and the environmental filters (Webb et al. 2002; Becerra 2007). However, we found neither clustered nor overdispersed phylogenetic pattern at fine spatial scale. In some genera of tropical forests, phylogeny and defense traits may not be correlated due to the rapid evolution of antiherbivore defenses in one genus of tropical forest (Kursar et al. 2009). However, in our community-level approach, this absence of correlation may be partly attributed to the random distribution of most defense traits in the phylogeny of cerrado plants. Wherever defense traits are not conserved in the phylogeny, neither herbivory nor environmental filters may determine the phylogenetic pattern of the spatial distribution of plants (Becerra 2007; Silva & Batalha 2009). Although trait conservatism is very common in lineages of plants (Prinzing et al. 2001; Ackerly 2003; Reich et al. 2003), we may expect a random distribution of traits in rich communities, such as the Brazilian cerrado (Castro et al. 1999). In these rich communities, trait evolution tends to be a complex mixture of conservatism and convergence (Webb et al. 2002). As a consequence, overall trait conservatism in the phylogeny of cerrado woody species may be hard to detect (Silva & Batalha 2009). Nevertheless, the phylogenetic signals observed in some defense traits reinforced the idea that environmental filters and herbivory may act in opposite direction, generating random species co-occurrences. As a matter of fact, most of the conserved traits we found (specific leaf area, leaf water content, and nutritional quality) are traits also related to the environmental filters (Cornelissen et al. 2003). For instance, nutrient-poor soils are positively correlated to low specific leaf area (Ordonéz et al. 2009). Comparisons between the same genus of cerrado and forest species showed that specific leaf area is lower in the cerrado, where availability of water and nutrients in the soil are low (Hoffmann & Franco 2008). 18 Cerrado physiognomies with intense drought and poor soils are expected to assemble species with low specific leaf area, leaf water content, and nutritional quality (Pais & Varanda 2003). Thus, conserved traits that act as defense traits against herbivory may be selected by herbivores towards an overdispersed pattern (Kursar et al. 2009) and, at the same time, be selected by environmental filters towards a clustered pattern (Webb et al. 2002), generating a random co-occurrences of species. Our results also supported that the associations of herbivores and plants may not be as tight as predicted by literature. More diffuse associations may have more generalist adaptations, because selection pressures of defense traits conflict with multiple herbivores (Strauss & Irwin 2004). In communities where the species present such interactions, community-level overdispersion in defense traits is less likely to occur (Becerra 2007). Although there is evidence of high degree of specialisation between host plants and herbivores in cerrado (Diniz & Morais 1997; Marquis et al. 2002), few species of plants were studied, preventing to a certain extent the generalisation of this statement. Here we studied the whole tree community and the defense traits associated to chewing insects, the most important consumers in tropical plant communities (Coley & Barone 1996). In this community context, random spatial distribution of defense traits against herbivory may also be explained by a diffuse interaction between plants and herbivores. Plant-herbivore coevolution and its impact on defense traits have focused on a small number of species and populations (Agrawal & Fishbein 2006; Becerra 2007; Kursar et al. 2009). However, our approach focused on the community, which may be structured by a combination of processes acting simultaneously and in opposite directions (Kraft et al. 2007), such as environmental filtering and herbivory. Consequently, a clear pattern of functional and phylogenetic overdispersion may not emerge in plant defense traits in cerrado. Moreover, it is important to include all possible species that may contribute with information about environmental filtering and species interactions. Accordingly, future studies in cerrado should also consider the herbaceous species, which represents twice the number of tree species (Castro et al. 1999). Acknowledgements We are grateful to Fapesp, for financial support and scholarship granted to the first author, 19 to D.M. Silva, A.R. Nascimento, and V.L. Dantas for helping us in field work. References Ackerly D. D. (2003) Community assembly, niche conservatism, and adaptive evolution in changing environments. Int. J. Plant Sci. 164, 165-84. Agrawal A. A. & Fishbein M. (2006) Plant defense syndromes. Ecology 87, 132-49. Agrawal A. A. & Konno K. (2009) Latex: a model for understanding mechanisms, ecology, and evolution of plant defense against herbivory. Annu. Rev. Ecol. Evol. Syst. 40, 311-31. Angiosperm Phylogeny Group (2009) An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Bot. J. Linn. Soc. 161, 105-21. Batalha M. A. & Mantovani W. (1999) Chave de identificação das espécies vegetais vasculares baseada em caracteres vegetativos para a ARIE Cerrado Pé-de-Gigante (Santa Rita do Passa Quatro, SP). Rev. I. Florestal 11, 137-58. Becerra J. X. (1997) Insects on plants: macroevolutionary chemical trends in host use. Science 276, 253-6. Becerra J. X. (2007) The impact of herbivore-plant coevolution on plant community structure. Proc. Nat. Acad. Sci. 104, 7483-8. Berembaum M. R. (1983) Coumarins and Caterpillars: a case for coevolution. BioScience 33, 194-5. Berenbaum M. R., Nitao J. K. & Zangerl A. R. (1991) Adaptative significance of furanocoumarin diversity in Pastinaca sativa (Apiaceae). J. Chem. Ecol. 17, 207-215. Blomberg S. P., Garland, T. & Ives, A. R. (2003) Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution 57, 717–45. 20 Carson W. & Root R. (2000) Herbivory and plant species coexistence: community regulation by an outbreaking phytophagous insect. Ecol. Monogr. 70, 73-99. Castro A. A. J. F., Martins F. R., Tamashiro J. Y. & Shepherd G. J. (1999) How rich is the flora of the Brazilian cerrados? Ann. Mo. Bot. Gard. 86, 192-224. Cavender-Bares J., Ackerly D. D., Baum D. A. & Bazzaz F. A. (2004) Phylogenetic overdispersion in Floridian oak communities. Am. Nat. 163, 823–43. Cavender-Bares J., Keen A. & Miles B. (2006) Phylogenetic structure of Floridian plant communities depends on taxonomic and spatial scale. Ecology 87, 109-22. Chase J. M. (2003) Community assembly: when should history matter? Oecol. 136, 489-98. Chazdon R. L., Careaga S., Webb C. & Vargas O. (2003) Community and phylogenetic structure of reproductive traits of woody species in wet tropical forests. Ecol. Monogr. 73, 331-48. Coley P. D., Bryant J. P. & Chapin F. S. (1985) Resource availability and plant antiherbivore defense. Science 230, 895-99. Coley P. D. & Barone J. A. (1996) Herbivory and plant defenses in tropical forests. Annu. Rev. Ecol. Evol. Syst. 27, 305-35. Cornelissen J. H. C., Lavorel S., Garniel E. et al. (2003) A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Aust. J. Bot. 51, 335- 80. Costa A. N., Vasconcelos H. L., Vieira-Neto E. H. M. & Bruna E. M. (2008) Do herbivores exert top-down effects in Neotropical savannas? Estimates of biomass consumption by leaf- cutter ants. J. Veg. Sci. 19, 849-54. Davies T. J., Barraclough T. G., Chase M. W. et al. (2004) Darwin`s abominable mystery: insights from a supertree of the angiosperms. Proc. Nat. Acad. Sci. 101, 1904-9. Diniz I. R. & Morais H. C. (1997) Lepidopteran caterpillar fauna of cerrado host plants. 21 Biodivers. Conserv. 6, 817-836. Dyer L. A., Singer M. S., Lill J. T. et al. (2007) Host specificity of Lepitoptera in tropical and temperate forests. Nature 448, 696-700. Falkenberg M. B., Santos R. I. & Simões C. M. O. (2003) Introdução à análise fitoquímica. In: Farmacognosia: da planta ao medicamento. (eds C. M. O. Simões, E. P. Schenkel, G. Gomann et al.) UFRGS, Porto Alegre. Fine P. V. A., Miller Z. J., Mesones I. et al. (2006) The growth-defense trade-off and habitat specialization by plants in Amazonian forests. Ecology 87, 150-62. Futuyma D. J. & Mitter C. (1996) Insect–plant interactions: the evolution of component communities. Philos. T. Roy. Soc. B. 351, 1361-6. Gilbert G. S. & Webb C. O. (2007) Phylogenetic signal in plant pathogen-host range. Proc. Nat. Acad. Sci. 104, 4979-83. Gottsberger G. & Silberbauer-Gottsberger I. (2006) Life in the cerrado: a South American tropical seasonal vegetation, Vol. 1. Origin, structure. dynamics and plant use, Reta Verlag, Ulm Hoffmann W. A. & Franco A. C. (2008) The importance of evolutionary history in studies of plant physiological ecology: examples from cerrados and forests of central Brazil. Braz. J. Plant Physiol. 20, 247-56. Janzen D. H. (1974) Tropical blackwater rivers, animals and mast fruiting by the Dipterocarpaceae. Biotropica 6, 69–103. Kembel S., Ackerly D., Blomberg S. et al. (2009) Picante: R tools for integrating phylogenies and ecology. R package version 0.6-0. Available from URL: http://www.R-project.org. Köppen W. (1931) Grundriss der Klimakunde. Gruyter, Berlin. Kraft N. J. B, Cornwell W. K., Webb C. O. & Ackerly D. D. (2007) Trait evolution, community assembly, and the phylogenetic structure of ecological communities. Am. Nat. 22 170, 271–83. Kursar T. A., Dexter K. G., Lovkam J. et al. (2009) The evolution of antiherbivore defenses and their contribution to species coexistence in the tropical tree genus Inga. Proc. Nat. Am. Soc. 106, 18073-8. Lima M. I. S. (2000) Substâncias do metabolismo secundário de algumas espécies nativas e introduzidas no Brasil. In: Ecofisiologia vegetal. (ed W. Larcher) Rima, São Carlos. Manly B. F. J. (2000) Multivariate statistical methods. Chapman & Hall, New York. Maron J. R. & Crone E. (2006) Herbivory: effects on plant abundance, distribution and population growth. Proc. R. Soc. Lond. 273, 2575-84. Marquis R. J., Morais H. C. & Diniz I. R. (2002) Interactions among cerrado plants and their herbivores: Unique or typical? In: The cerrados of Brazil. (eds P. S. Oliveira & R. J. Marquis) Columbia University, New York. Neves F. S., Araújo L. S., Espírito-Santo M. M. et al. (2010) Canopy herbivory and insect herbivore diversity in a dry forest-savanna transition in Brazil. Biotropica 42, 112-8. Oksanen J., Kindt R., Legendre P. et al. (2009) Vegan: community ecology package. R package version 1.15-1. Available from URL: http://www.R-project.org. Olff H. & Ritchie M. E. (1998) Effects of herbivores on grassland plant diversity. Trends Ecol. Evol. 13, 261-5. Ordoñez J. C., van Bodegom P. M., Witte J. P. M. et al. (2009) A global study of relationships between leaf traits, climate and soil measures of nutrient fertility. Global Ecol. Biogeogr. 18, 137-49. Pais M. P. & Varanda E. M. (2003) Variation in plant defenses of Didymopanax vinosum (Cham. & Schltdl.) Seem. (Apiaceae) across a vegetation gradient in a Brazilian cerrado. Acta Bot. Bras. 17, 395-403. Prinzing A., Durka W., Klotz S. & Brandl R. (2001) The niche of higher plants: evidence for 23 phylogenetic conservatism. Proc. R. Soc. Lond. 268, 2383-9. R Development Core Team (2009) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. Available from URL: http://www.R- project.org. Rasband W. (2004) ImageJ: Image process and analysis in Java. National Institutes of Health, Bethesda. Reich P., Wright I., Cavender-Bares J. et al. (2003) The evolution of plant functional variation: traits, spectra and strategies. Int. J. Plant Sci. 164, S143-64. Santos J. E., Paese A. & Pires J. S. R. (1999) Unidades da paisagem (biótopos) do campus da Ufscar. Ufscar, São Carlos. Schoener T. W. (1970) Nonsynchronous spatial overlap of lizards in patchy habitats. Ecology 51, 408-18. Silva I. A. & Batalha M. A. (2009) Co-occurence of tree species at fine spatial scale in a woodland cerrado, southeastern Brazil. Plant Ecol. 200, 277-86. Silva I. A. & Batalha M. A. (2010) Woody plant species co-occurrence in Brazilian savannas under different fire frequencies. Acta Oecol. 36, 85-91. SMA. Secretaria do Estado do Meio Ambiente (1997) Cerrado: bases para conservação e uso sustentável das áreas de cerrado do Estado de São Paulo. SMA, São Paulo. Stapley L. (1998) The interaction of thorns and symbiotic ants of an effective defense mechanism of swollen-thorns acacias. Oecol. 115, 401-5. Stock W. D., Bond W. J. & van de Vijver C. A. D. M. (2010) Herbivore and nutrient control of lawn and bunch grass distribution in a southers African savanna. Plant Ecol. 206, 15-27. Strauss S. Y. & Irwin R. E. (2004) Ecological and evolutionary consequences of multispecies plant-animal interactions. Annu. Rev. Ecol. Evol. Syst. 35, 435-66. Ward L. & Spalding D. F. (1993) Phytophagous British insects and mites and their food-plant 24 families: total numbers and polyphagy. Biol. J. Linn. Soc. 49, 257-76. Webb C. O., Ackerly D. D. & Kembel S. W. (2009) Phylocom: software for the analysis of phylogenetic community structure and character evolution. Bioinformatics 24, 2098-100. Webb C. O., Ackerly D. D., McPeek M. A. & Donoghue M. J. (2002) Phylogenies and community ecology. Annu. Rev. Ecol. Syst. 33, 475-505. Webb C. O. & Donoghue M. J. (2005) Phylomatic: tree assembly for applied phylogenetics. Mol. Ecol. 5, 181-3. Van Zandt P. A. & Agrawal A. A. (2004) Specificity of induced plant responses to specialist herbivores of the common milkweed Asclepias syriaca. Oikos 104, 401-9. 25 Table 1. Correlation between the co-occurrence of species in a southern cerrado site (21º58’05.3”S, 47º52‘10.1”W) and the defense traits against herbivory: specific leaf area, water content, toughness, trichomes, latex, nutritional quality, alkaloids, terpenoids, and tannins. No correlation was significant. Trait R P Specific leaf area 0.08 0.19 Water -0.05 0.70 Toughness 0.14 0.09 Trichomes 0.01 0.13 Latex -0.01 0.29 C:N ratio -0.04 0.72 Alkaloids -0.02 0.44 Terpenoids 0.01 0.35 Taninns 0.11 0.16 26 Table 2. Phylogenetic signal to defense traits, measured from the variance of the phylogenetically independent contrasts in tree species of a southern cerrado site (21º58’05.3”S, 47º52‘10.1”W). Test for phylogenetic signal in each trait sampled: specific leaf area (cm2 g-1); leaf water content (mg cm-2); leaf toughness (N); trichome density (cm-2); latex content (mg); C:N (carbon: nitrogen ratio); and presence of alkaloids, terpenoids, and tannins determined following Falkerberg et al. (2003). Obseved PICs Random PICs P Specific leaf area 0.016 0.023 0.033* Water 0.017 0.023 0.044* Toughness 0.021 0.024 0.26 Trichomes 0.036 0.024 0.919 Latex 0.011 0.024 0.029* C:N 0.012 0.024 0.001*** Alkaloids 0.016 0.024 0.265 Terpenoids 0.021 0.022 0.303 Tannins 0.017 0.019 0.426 27 Fig. 1. Phylogenetic tree of species sampled in a southern cerrado site (21º58’05.3”S, 47º52‘10.1”W). The relationship among species was based on Davies et al. (2004). dissertação_priscilla[1].pdf dissertação_priscilla[2].pdf